Document mbvRnX1wRrV6Kpr9O9oMKKO5B

PREGNANCY OUTCOMES IN SANTA CLARA COUNTY 1980-1985 GEORGE DEUKMEJIAN GOVERNOR STATE OF CALIFORNIA SL 033206 PREGNANCY OUTCOMES IN SANTA CLARA COUNTY 1980*1985 Epidemiological Studies and Surveillance Section California Department of Health Services 2151 Berkeley Way Berkeley, CA 94704 May 1988 033207 SL Acknowledgements These studies were conducted by the California Department of Health Services (CDHS) In collaboration with Its contractor, the California Public Health F undation. The research was carried out under the leadership of Shanna Swan, Ph.D, and Raymond Neutra, HD, DrPH. The principal authors of this report were: Shanna Swan, Ph.D (Section 1), Margaret Vrensch, Ph.D (Section 2), Hargaret Deane, MPH (Section 3), Gary Shaw, Ph.D (Section 4), and Kathleen Claxton, MA (Section 5). The following CDHS and contract staff contributed In the areas Indicated: Data collection, analysis, and management: Luana Acton, John Andrew, Julie Brown, Martha Duncan, David Epstein, Laura Fenster, Peggy Hartsell, Alexandra Hunter, Warren Kennell, Robin Levinson, Pam Lindsey, Jane Lipscomb, Jorge Lizarraga, Lorraine Malcoe, Shelly Mandel, Carol Mattison, Carol McElheney, Kimberly Newton, Juan Perez, Csaba Polony, Theresa Ramirez, Dennis Shusterman, MD, Sharon Trammel, Kitty Tsui, Lome Wallace, Soora Wi. Administrative support: Jill Touloukian. Editorial support: David Epstein. Clerical support: Dolce Brookman, Susan Freeman, Julie Lammers. Technical Consultants to the Santa Clara County Pregnancy Outcome Studies were: D lia Alvarez, Stephen Coray, MD, Louise Detwiler, David Erickson, MD, Brenda Eskenazi, Ph.D, Bernice Giansiracusa, MD, Nancy Hicks, Robert Morgan, MD, Henry Ongerth, Beverly Paigen, Ph.D, Diana Petitti, MD, MPH, Arthur Reingold, MD, Kenneth Rothman, DDS, DrPH, Steve Samuels, Ph.D, Nicholas Toth, MD, Donald Wharton, MD. Assistance and data were also provided by the Santa Clara County Health D pertinent, the Great Oaks Water Company, the Fairchild Camera Company, and by International Business Machines Corporation. Without the assistance of the individuals, companies, and agencies mentioned, and the cooperation of the families that participated, their physicians, and the medical records staffs of Santa Clara County hospitals, these studies would not have been possible. SL 033208 TABLE OF CONTENTS E&ge 1. Overview Across Studies.................................................................................................1.1 2. Pregnancy Outcome Follow-up Study: Pregnancy Outcomes in Four Census Tracts of Santa Clara County 1980-1985....................................2.1 3. Comparison of Original and Follow-up Fairchild Study Populations..............................................................................................................3.1 4. Cardiac Study: Follow-upCluster Investigation................................................ 4.1 5. Material Mass Balance Analysis................................................................................. 5.1 SI- 033209 Table Titles Pregnancy Outcome Follow-up Study. Pregnancy Outcomes In Four Census Tracts of Santa Clara County 1980-1985 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 2.11 2.12 2.13 2.14 2.15 2.16 2.17 2.18 2.19 2.20 2.21 2.22 2.23 2.24 2.25 2.26 2.27 2.28 1980 U.S. Cenaus characteristics used to select two control census tracts for two census tracts served by GOWC Percent distribution of responses to household enumeration of four census tracts In Santa Clara County Response to interviews among women for whom pregnancy was- reported in enumeration Number of eligible pregnancy outcomes reported by interviewed women Validation of reported spontaneous abortions Validated definite spontaneous abortions among all pregnancies and among a single randomly chosen eligible pregnancy Number of definite spontaneous abortions by year and census tract Number of live births by year and census tract Spontaneous abortion rates by year and census tract Odds ratios and 95% confidence intervals for spontaneous abortions in exposed vs. unexposed areas by year Compariaons of ascertainment of live births and spontaneous abortions conceived during 1980 and 1981 and identified during a 1983 interview study and a 1986 interview study Spontaneous abortion rates by selected variables Gestational weeks of spontaneous abortions by census tracts and time periods Multiple logistic model of risk of spontaneous abortion Validation of reported birth defects Congenital anomalies reported by respondent end validated by medical record or physician's report Number of reportable malformations among singleton live births by census tract and year Number of singleton live births by census tract and year Reportable malformation rates by census tract and year Odds ratios and 95* confidence intervals for reportable malformations in exposed vs. unexposed areas by time period Congenital anomaly rates for live births by condition: residents of Santa Clara County study areas (1980-1985) Listing of congenital malformations occurring in more than two children and maternal risk factors Number of low birth weight babies among all births reported on birth certificates Number of singleton live births among all births reported on birth certificates Low birth weight rates among all births reported on birth certificates Odds ratios and 95% confidence intervals for low birth weight in exposed vs. unexposed areas by time period among all births reported on birth certificates Number of low birth weight babies by census tract and year among interviewed women Number of singleton live births with known birth weight by census tract and year among interviewed women SL 033210 2.29 Low birth weight rates by census tract and year among Interviewed v men 2.30 Odds ratios and 95% confidence Intervals for low birth weight In exposed vs. unexposed areas by time period among Interviewed women 2.31A Comparisons of mean estimated cumulative TCA (ppb) exposure among women with pregnancies ending in live birth and those ending In spontaneous'abortion 2.31B Percent distribution of spontaneous abortions and live births by quartlles of estimated TCA (ppb) exposure 2.32 Comparisons of mean estimated probability of exposure to water from well #13 among women with pregnancies ending In live birth and those ending In spontaneous abortion 2.33 Comparisons of mean estimated cumulative probability of no exposure to water from well #13 among women with pregnancies ending in live birth and those ending In spontaneous ebortlon (SAB) 2.34 Comparisons of mean cumulative TCA (ppb) exposure to women with pregnancies resulting In normal live births or In live births with malformation 2.C.1 Observed and theoretical cumulative distributions 2.0. 1 Number of definite spontaneous abortions among all eligible pregnancies 2.0. 2 Total singleton live births among all eligible pregnancies 2.D.3 Definite spontaneous abortion rates using all pregnancies by year and census tract 2.0. 4 Odds ratios and 95% confidence intervals for spontaneous abortions in exposed and unexposed areas by year using all eligible pregnancies 2.0:5 Stillbirths among all eligible pregnancies Comparison Of Original And Follow-uo Fairchild Study Populations 3.1 Total reported spontaneous abortions by population at risk and census tract 3.2 Multiple logistic regression result comparing women who stayed with women who left 3.3 Percent distribution of population subgroups by materal risk factors Cardiac Study: Follow-up Cluster Investigation 4.1 Cases and controls born inside and outside the study area 4.2 Cases and controls by maternal residence during first trimester relative to study area 4.3 Maternal first trimester residence, employment, or frequent visits t the study area compared to SCC 4.4 Odds ratios for cardiac anomalies by home, tap water consumption inside versus outside study area after adjustment for potential confounders 4.5 Estimated TCA score for cases and controls bom in the study area 4.6 Estimated TCA score for cases and controls where maternal drinking water exposure was obtained by interview Material Mass Balance Analysis 5.1 Material balance report analysis - data sources and contacts 5.2 Data items required for materials balance analysis 5.3 DHS tabulations of solvent issues and purchases by month and year SL 033211 5.4 Fairchild waste disposal records 5.5 Problems with waste manifests 5.6 MBR/DHS differences in interpretation of manifests 5.7 Comparison between MBR average monthly solvent purchases for 1981 and wastewater discharge permit average monthly solvent usages for 1982 5.8 Quarterly comparison of materials balance report with DHS calculations SL 033212 Figure Titles Overview Across Studies 1.1 Map of scudy areas 1.2 Four census tracts In Santa Clara County 1980-1985 1.3 Odds ratios and 95% confidence Intervals for spontaneous abortions in exposed vs. unexposed areas for 1980-1981 and 1982-1985 1.4 Odds ratios and 95% confidence intervals for congenital malformations in exposed vs. unexposed areas for 1980-1981 and 1982-1985 1.5 Odds ratios and 95% confidence intervals for low birth weight in exposed vs. unexposed areas for 1980-1981 and 1982-1985 1.6 Estimated TCA exposure during 1st month of pregnancy by pregnancy outcome original and new exposed areas combined (1981) 1.7 Estimated percent of water from well #13 in 1981 for census tracts in the Great Oaks Water Company 033213 SL 1. Overview Across Studies Introduction This report presents the results of two epidemiological studies of pregnancy utcomes and two related exposure assessments. These were conducted by- the Epidemiological Studies and Surveillance Section of the California Department of Health Services (CDHS) with the cooperation of its contractor, the California Public Health Foundation, and the Santa Clara County Health Department. These studies were conducted to follow-up on two studies released by CDHS in January, 1985 (1). In the previously released studies the Department reported a doubled rate of sp ntaneous abortions and tripled rate of birth defects in women who resided in a census tract (Los Paseos) which received drinking water from a well which had been contaminated by leakage of solvents from an underground storage tank operated by the Fairchild Semiconductor Company in South San Jose. At the same time, the Department documented a 2.6-fold increased rate of major cardiac malformations among the children of women residing in the s ven-census tract area through which the contaminated water was distributed. The first of these studies ruled out confounding as the cause f these findings, and for both the role of the leak in causing these excesses was uncertain. Therefore, the Department conducted two pldemiological studies and two exposure assessments to followup on these results. The first study in this report, the Pregnancy Outcome Follow-uo Study (POPS), is similar in design to the first interview study. However, in POFS, women in a census tract adjacent to Los Paseos, which was thought to also have been exposed to water from the contaminated well, were interviewed about their pregnancies in 1980-81, in addition to re-interviewing the women in Los Paseos. Furthermore, this study looked at pregnancies in 1982-85 as well, so that the effect of closing the contaminated well could be analyzed. In a companion report on water exposure and pregnancy outcome released concurrently by the Department, this study is referred to es Fairchild II. 1.1 SL 033214 The second pare of the POFS study, Comparison_of Original and Follow-up Fairchild Study Populations, compares women Interviewed In both the first Fairchild study and In POFS with those who were not reinterviewed. The 1985 Cardiac study was based on hospital data only and no interviews wer conducted. In the Cardiac Study: Follow-up Cluster Investigation, cases and normal controls were interviewed in order to obtain more detailed information on exposure to water, to rule out confounding as an explanation for the observed association, and to look at the prevalence of cardiac def cts In more recent births. Th 1985 studies contained limited Information about the level and extent of the water contamination. In order to obtain more precise exposure information, hydrogeologic modelling of water flow within the distribution system of the water company which had operated the contaminated well was c nducted. This is the subject of The Great Oaks Water Distribution Study, presented separately (2). Foil wing the discovery of the leak, the Fairchild company prepared a mat rials balance report at the request of the Regional Water Quality C ntrol Board. This report Indicated no appreciable leakage from the failed tank before late May 1980. Because of the importance of the timing of the leak in interpreting the excesses documented in the epidemiological studies, the Department conducted a reanalysis of the Fairchild report. This Material Mass Balance Analysis is contained in this report. These studies reflect an unusually complete effort on the part of the Department to investigate the causal relationship between the clusters of adverse pregnancy outcomes documented in 1985 and the contamination episode. They represent three years of work, cooperation of 10,000 households and interviews with 1,500 women and have been extensively reviewed by nationally recognized scientists with expertise in these areas. The Department feels that these studies, In conjunction with the studies released in January 1985, address the "Fairchild question" as completely as is scientifically possible. This Section summarizes each of the component stud! s. 1.2 SL 033215 PREGNANCY OUTCOME FOLLOW-UP STUDY Background and Methods In lace 1981, residence of Che Los Paseos neighborhood In Souch San Jose nociced a possible excess of spontaneous abortions and birch defects In cheir neighborhood. Report of this concern was received by the Santa Clara County Health Department following the announcement of the leak of toxic chemicals from an underground waste solvent storage tank operated by the Fairchild Semiconductor Company. Contaminants from this leak were measured in the drinking water of a well (Well #13) serving this community. A number f chemicals were measured; a high level of the solvent 1,1,1-trlchloroethane (TCA) was of particular concern. Despite the fact that no human or animal studies conducted up until that time suggested adverse reproductive effects f the chemicals which had been identified, the California Department of Health Services, in cooperation with the Santa Clara County Health Department, decided to conduct two epidemiological studies to address this pr blem. One f these, an interview study of women who had been pregnant in 1980-81 while living in either the concerned community (Los Paseos) or a demographieally similar control tract, had the following objectives: (1) to determine whether a cluster of spontaneous abortions and congenital malf nations had actually occurred, (2) to detenlne whether any maternal risk factors could account for this cluster and (3) to detenlne whether the cluster was related to the leak of chemicals into Well #13. This study found that (1) spontaneous abortions and malformations had occurred above the expected rates, (2) maternal risk factors could not explain these excesses and (3) the evidence was insufficient to detenlne whether the leak of chemicals into Well #13 caused the excess. The Pregnancy Outcome Follow-up Study was then undertaken to address this last, unanswered question. The Pregnancy Outcome Follow-up Study (POFS) had the following objectives: (1) to detenlne whether the high rates seen in Los Paseos in 1980-81 persisted in 1982-85, after the well with known contamination was closed and (2) to study rates in a tract adjacent t Los Fas s, in which Well #13 was 1.3 SL 033216 located, to saa whether rates vara high there in 1980-81. For this purpose a second unexposed tract was selected for comparison. Rates in the two new tracts were also compared for pregnancies in 1982-85. A map showing the study areas is included (Figure 1.1). A schematic representation of the study design is shown in Figure 1.2. This study also utilized results of hydr geologic modelling (Great Oaks Vater Company Distribution Study) to estimate the exposure to water from Veil #13 and to TCA for each study pregnancy. A household enumeration was conducted in the four study areas in order to Identify women who had been pregnant in 1980-85. Subsequently, women Identified through the enumeration as having been pregnant were interviewed ab ut each eligible pregnancy, those for which the woman resided in the study area throughout the pregnancy. These Interviews Included a complete pregnancy history and detailed information on exposure to vater as well as dem graphics and potential confounders such as parent's occupation and industry, chemicals, medications, smoking, alcohol and caffeine. All spontaneous abortions Included in the analysis were validated through medical record review or physician interview. Spontaneous abortion analyses wer based on one pregnancy per woman. Birth defects included in the analysis were categorized as "reportable" by the California Birth Defects M nitorlng Program (CBDMP) and validated by medical chart review. Birth weight analyses were also conducted, using birth weight as recorded on the birth certificate. Raavlta Enumeration was completed in 10,063 households (91.6%) and identified 2,713 w men who had been pregnant, of whom 95% were interviewed. Of these, 1376 pregnancies were eligible for further study. A total of 181 spontaneous abortions were validated. After selecting one pregnancy per woman, 103 spontaneous abortions were available for analysis. (This selection decreased the spontaneous abortion rate from 13.2% to 9.9% but did not affect the study's conclusions with respect to year or census tract.) SL 033217 1.4 The race f ap ncaneous abort! n in Che new exposed area In 1980*81 was 4.7%, considerably lower chan Che overall rate. Thus, there was no evidence that this area was ac higher risk chan expected before Che contaminated well was closed. It is interesting to note that the odds ratio for comparing Los Paseos to its control tract in 1980-81 (OR-3.5) was similar to that seen in the original study (OR-2.4). No elevation in rates was seen after 1982 for either the exposed areas (10.8%) or the control areas (10.6%). Odds ratios c sparing exposed and control tracts are shown in Figure 1.3. A large number of potential risk factors were analyzed, and a number of risk fact rs for spontaneous abortion were identified including: maternal age Less than 20 or over 35, multiple prior fecal loss, and alcohol use of four or m re drinks per week. All of these risk factors have been previously rep rted in the literature. Adjusting for these and other variables did not alt r the results on study area or year. A t tal of 37 validated birth defects were considered reportable by CBDMP. As in the prior study, an unusually high prevalence of birth defects (10.4%) was seen in the original exposed tract for 1980-81, but the prevalence in the new exposed tract was unusually low during the same time period (1.9%). After 1982, the frequency of birth defects in the exposed and control areas were similar (2.9% and 1.9%) and consistent with the prevalence reported by CBDMP for the five Bay Area counties (2.5%). Odds ratios for reportable birth defects are shown in Figure 1.4 by study year. The frequency of low birth weight was also examined. The original study had rep rted a significant absence of low birth weight babies in the exposed area. In this study there was little difference seen between exposed and c ntrol areas in 1980-81 (2.4% and 2.6%) or during 1982-85 (3.8% and 4.0%). These results on low birth weight are summarized in Figure 1.5. Hydrogeologic modelling was conducted to estimate the flow of water from Veil #13 throughout the Great Oaks Vater Company distribution system and the levels of TCA women were likely to have received during their pregnancy based on their address and dates. Modelling was not carried out for 1980 pregnancies because the engineer who conducted this analysis concluded that 1.5 SL 033218 it was unlikely that any homes received any significant exp sur t TCA before January 1981. This study finds that the new exposed tract had a considerably higher likelihood of exposure to water from Veil #13 during 1981 (60%-100%) than the Los Paseos area (20%-60%) (Figure 1.7). Thus, the low spontaneous abortion rate in the new exposed tract in 1981 is inc nsistent with the assumption that water from Well #13 was associated with high rates. Using the address reported on interview and the estimated TCA exposure at that address during the first month of pregnancy, it was possible to c nstruet an exposure measure for each pregnancy. If the contaminated water from Well #13 had been responsible for the cluster of adverse outcomes in 1980-81, one would expect to see a higher TCA score for women with spontaneous abortions and birth defects than for those with normal live births. This was not the case. Mean estimated levels of TCA In these three gr ups are similar with the highest level of TCA associated with normal live births. This is shown in Figure 1.6. In summary, there Is no evidence from these data that rates of adverse pregnancy outcomes were elevated in the new exposed tract during 1980-81. If anything, there was an absence of adverse outcomes. Since hydrogeologic modelling shows that this new tract was more likely to have received c ntamination from Well #13 than the Los Paseos area, which was originally studied, the leak is a very unlikely explanation for the cluster of adverse outcomes seen in Los Paseos in 1980-81. This study also demonstrates that recent rates in both exposed and control areas agree with expected rates for adverse pregnancy outcomes. COMPARISON OF ORIGINAL AND FOLLOW-UP FAIRCHILD STUDY POPULATIONS Two of the census tracts in POFS had been examined three years earlier during the first interview study. This provided the opportunity to compare women Interviewed in both studies to women who were in the first study, but not in the second because presumably they had moved away. In both the exposed and contr 1 census tracts, higher rates of SAB were found am ng 1.6 SL 033219 women wh had apparently moved away than among women who still lived in the tw areas (exposed tract: 24.4% vs. 19.3%, control tract: 15.7% vs. 7.9%). However, the odds ratios comparing exposed and unexposed tracts were similar among these two groups. CARDIAC STUDY: FOLLOW-UP CLUSTER INVESTIGATION Severe cardiac defects were among the adverse reproductive outcomes reported by residents of the Los Paseos area. These defects occur Infrequently (about 3 In 1000 live births) and the interview study did not have enough births to determine whether an unusual number of these rare outcomes had occurred. Therefore, the Department conducted a second study in which hospital records thr ughout Santa Clara County were reviewed in order to identify cardiac defect cases among births in 1981 and 1982. Analysis was then carried out for the time period during which births were potentially exposed (January 1981 though September 1982). This study (1), released in 1985, documented that there were two cases in Los Paseos, which is more than expected. It als examined defects among births to women living in the seven census tracts receiving water from the Great Oaks Water Company which operated Well #13. This study found ten cardiac anomalies among women living in the seven census tracts when only four would be expected based on the prevalence in the remainder of the county (2.9 per 1000). This original study was based on hospital records and addresses at birth nly and no Interviews were conducted. Therefore, women who drank only b ttled water, or who moved to the seven census tract area shortly before giving birth would have been mistakenly counted as exposed. Furthermore, there were insufficient data to determine whether the prevalence of these defects returned to normal after Well #13 was closed. Therefore, the Cardiac Follow-up Cluster Investigation was undertaken. A decision was made to interview all cardiac cases bom in 1981-83, as well as a random sample of normal live births, in order to study the time period after the well was closed and to obtain more detailed information about exp sure and possible confounders. 1.7 SL 033220 The California Birth Def cts M nit ring Program (CBDMF) ascertained cases of cong nital cardiac anomalies through a review of medical records at each of the eleven hospitals in Santa Clara County, as well as at the California Children's Service and one out-of-county referral hospital. Cases used for analysis were those for which a major cardiac defect had been diagnosed by ne year of age among births in 1981-83. The entire seven census tract area served by the Great Oaks ffater Company was considered potentially exposed to contaminated water. Hydrogeologic m delling was later used to Identify areas most likely to receive this water. The onset of the exposure period remained uncertain. However, the end f the exposure period is known; Veil #13 was removed from service on D camber 7, 1981. Allowing one week for residual contamination in the distribution system and considering the first seven weeks after conception t be the period during which exposures are most likely to affect the fetal h art, it is estimated that births after September 6, 1982 are unlikely to have had exposures relevant to these defects. Therefore, births which ccurred in January 1, 1981-September 6, 1982 are considered potentially exp sad if the mother lived in the seven-census tract area. The analysis was conducted in two parts. Children bom between September 1, 1982 and December 31, 1983 were considered unexposed and analyzed separately. Prevalence rates were calculated for this time period but mothers of these children were not interviewed. On the other hand an attempt was made to locate and interview the mothers of all cases born in the study p rlod throughout Santa Clara County. An equal number of live born controls w re selected at random from birth certificates throughout the county and interviewed as well. Mothers were interviewed concerning their water exp sure through consumption and showering and bathing, at home, work and other places regularly visited. The interview also included information on alcohol consumption, education, employment, smoking, diabetes, epilepsy and family history of cardiac disease. Using the hydrogeologic modelling presented in the Great Oaks Water Distribution Study, each case and control with first trimester residence in the Great Oaks Water Company service area during the study period was 1.8 SL 033221 assigned a score, reflecting exposure to TCA via Veil #13 vater during the first trimester. Ss.gyl.ts During the comparison time period (9/1/82*12/31/83) there were three eases b m in the study area, which was slightly less than expected. In this area th prevalence was 1.7 per 1000 compared to 3.2 in the remainder of the c unty. These data have been presented separately (3). Overall, 106 mothers of cases bom in the study period (78.3%) and 114 c ntrol mothers (84.2%) were interviewed. After correcting for address (when the first trimester address assigned a different exposure than the birth address) or water use (when woman drank no tap water), the odds ratio was very similar to.that seen in the original study; the original study had pr sented a relative risk of 2.2, while this follow-up study found an odds rati of 2.3 after these adjustments for misclassification. The estimate fr m the first cardiac study did reach traditional levels of statistical significance (95% Cl, 1.2, 3.9) although the 95% confidence interval for this follow-up study Included 1.0 (95% Cl, 0.61, 8.4). Adjustment for p tential confounders had little effect on these estimates. Of the interviewed cases and controls, only six cases and four controls c uld be assigned an exposure score. This is because all other study subjects had first trimester residence before January 1981 or outside the study area. Of the six eases for whom scores could be assigned, only two lived in the study area, the remaining four visited frequently. For c ntrols, one visited the area and three lived there. Therefore, numbers were too small to determine the relative exposure of cases and controls and an additional random sample of normal controls was selected. Exposure scores were based on census tract of residence at birth, since the additional controls were not interviewed. The mean score for cases suggested a somewhat higher likelihood of exposure than among controls, but the difference did not approach statistical significance. 1.9 SL 033222 This study sis examined the temp ral distribution f cases throughout the study peri d. From the beginning of 1982 until the end of the study period (September 1982), seven cases would be expected, if the excess prevalence seen in 1981 In Che study area continued at the same level until the well was closed. However, only two cases were seen. These cases were bom in April, 1982, and none were bom In the study area after that time until after September 1982. This deficit of cases, compared to the expected excess in January-September 1982 is statistically significant (p-,03). Dissuasion Alth ugh the prevalence of cardiac defects returned to normal after the study period, this drop did not coincide with the closing of Veil #13. There was a significant deficit of cardiac cases in the study area during 1982, when seven would have been expected, had the elevated prevalence seen in 1981 continued. This evidence makes the leak an unlikely explanation for the cluster in 1981. Vhile interview data provided information with which to reclassify a number of subjects with respect to exposure, the effect measure changed little after these adjustments. The exposure scores which could be assigned to the few cases in the study area and study period were slightly higher on average than that for controls, but of these cases, only two lived in the study area. Therefore, though this information suggests slightly higher exposure of cases that controls within the study area, it Is based on extremely small numbers. On balance, it appears unlikely that the excess of cardiac defects seen in 1981 was due to waterborne exposure to TCA from Veil #13, although it probably will never be possible to determine conclusively what role the leak played in this observed cluster. However, CBDMP is conducting ongoing surveillance of these and other birth defects and will continue to investigate clusters related to water or other exposures. 1.10 SL 033223 GREAT OAKS WATER DISTRIBUTION STUDY Background and Methods The 1985 Pregnancy Outcome Study made the assumption that all women residing in the Los Paseos area in 1980*81 were equally likely to have been exposed to contaminated water from the Fairchild leak. This assumption was made for simplicity and because no other data were available. Two pieces of information were missing. The first concerns the timing and extent of the leak. When did it start and how rapidly did the concentration of chemicals Increase in the contaminated well? These are difficult questions to answer, because no water testing for these chemicals was done in that area prior to December 1981. The second question concerns the distribution of water within the service area of the Great Oaks Water Company. The Great Oaks Wat r Distribution Study was conducted in an attempt to answer these questions. The Great Oaks Water Distribution Study uses computerized hydrogeologic methods to model the distribution of water flow within the system. Input to this model are: the Great Oaks Water Company pipe network, well elevations, pumping records for each well and records from meters to establish user demand. Using these data the proportion of water from Well #13 at each of 124 nodes, or small areas, within the system, was estimated. These estimates are Independent of any assumptions about the amount of TGA in the system. The timing of the Initial leak, and the rate at which TCA increased at Well #13, was estimated using several alternative models. RsguUg The Water Distribution Study shoved that the distribution of water from Well #13 was very non-uniform across the seven census tracts that make up the Great Oaks Water Company service area. The distribution patterns changed som what from month to month, but not a great deal. Similarly there was little variation across nodes within a census tract. Ranges of the mean proportion of water from Well #13 received in each census tract are shown In 1.11 SL 03322^ Figure 1.7. As can ba se n, tha census tract riginally considered exposed to Well #13 (Los Paseos) was less exposed than the adjacent tract which was considered as the new exposed area in POFS. The results on the timing of the 1 ak are less certain, but none of the scenarios considered would predict that the contamination reached Well #13 in any appreciable amounts before January of 1981. MATERIAL MASS BALANCE ANALYSIS Backgrounds and Methods A materials balance analysis is commonly used in engineering studies to assess the Inflow and outflow of materials within a system. Such an analysis can be used to estimate the emount of material lost, for example, thr ugh a leak. Such an analysis was conducted by the Fairchild Corporation in the spring of 1982 at the request of the Regional Vater Quality Control Board. This report indicated no appreciable leakage before May 1980. Because the timing of the leak was important for interpreting the epid mlological studies, the Department decided to reconstruct the Fairchild analysis. Vith the cooperation of the Fairchild company and local agencies, a large number of documents were reviewed for this purpose. Due to limitations in the available data and in resources available for this task, the Department was only able to partially reconstruct the original analysis. Results The Materials Balance Analysis did not identify sufficient documentation to provide an accurate estimate of the onset of the underground tank leak at Fairchild. Nevertheless, the review did not identify any data which would appreciably alter the estimate of May 1980 which was contained in the original Materials Balance Report. An Initial leak at this time would be consistent with the estimates obtained in the Great Oaks Vater Distribution Study for when contamination reached Veil #13. The Department found that the components used to solve the imbalance equation, such as solvents 1.12 purchased, evaporated or hauled, must be considered unreliable and the extent of solvents lost remains uncertain. DISCUSSION These studies represent an unprecedented effort on the part of the CDHS to conduct a cluster investigation. Together with the original Fairchild studies released in January 1985, they have required six years and over one million dollars to conduct. They have also required the patience and participation of the citizens of four census tracts who have cooperated during these years. The weight of the evidence is against the leak from the Fairchild plant as a cause of the original cluster. Nevertheless, there is little doubt that a cluster did occur, and its cause will probably never be kn wn with certainty. Continued work by the Department on prenatal exp sures may shed light on causes 6f this and other such frequently ccurrlng clusters in the future. 1.13 SL 033226 References 1. Epidemiological Studies Section, California Department of Health Services: Fregancy Outcomes in Santa Clara County, 1980*1982. Reports of Two Epidemiological Studies, California State Publications Section. Reports 7540-958-1301-5, 1985. 2. Murphy, P: The Great Oaks Vater Distribution Study, 1988. 3. Swan SH, Shaw G, Harris J, Neutra RR: Congenital cardiac anomalies in relation to water contamination. Santa Clara County, California 1981-83; submitted for publication to Amer J Epid, 1988. 1.14 SL 033227 1.15 SL 033228 Pregnancy Outcome F II w-Up Study Figure 1.2 Four Census Tracts in Santa Clara County 1980 - 1985 Census Tracts Original Study Araa 5120.12 Naw Study Araa 5120.11 Original Study Araa 5120.08 Naw Study Araa 5053.0204 1980 - 1981 sed* ; ..v. .v.v.v.v.y.y. . -.V v . . .'.v.v.vlv.v v.v "T" ! '.v/.v.v.v.v.v*-* v!v..\y.\\ Dac. 1981 C ntaminatad Wall Closad Unexp osed 1982- 1985 U n exp 0 s e d a Exposure to watar contaminatad by a leaking chemical waste tank at the Fairchild Camara Company. 1.16 SL 033229 Pregnancy Outcome Follow-Up Study Figure 1.3 Odds Ratios and 95% Confidence Intervals for Spontaneous Abortions In Exposed vs. Unexposed Areas For 1980- 1981 and 1982- 1985 1980-1981 i- Original Areas (- ) New Areas 1.17 SL 033230 1982-1985 Original Areas e-- ^-------- ) New Areas iiiiiilili i i i i i__ i i i i 1/11 1/10 1/9 1/8 1/7 1/S 1/5 1/4 1/3 1/2 1.0 2 10 Odds Ratio Pregnancy Outcome Follow-Up Study Figure 1.4 Odds Ratios and 95% Confidence Intervals for Congenital Malformations In Exposed vs. Unexposed Areas For 1980- 1981 and 1982- 1985 (OO l/l 1980- 1981 <- (r 1982- 1985 (r Original Areas A/--) -) New Areas -) Original Areas ---------- ) New Areas 033231 L.J.. I__l_J__I I I I I J__1 I I I I 1/11 1/10 1/9 1/8 1/7 1/6 1/5 1/4 1/3 1/2 1.0 2 3 4 5 6 7 Odds Ratio /y I 15 Pregnancy Outcome Follow-Up Study Figure 1.5 Odds Ratios and 95% Confidence Intervals for Low Birth Weight In Exposed vs. Unexposed Areas * For 1980 - 1981 and 1982- 1985 1980- 1981 (-----------------#] --) Original Areas (-- --0---------- :--------------- ) New Areas SL 033232 1982- 1985 <----e- ) Original Areas (- -0-) New Areas 1--1--J ,, 1 1 1 1 1 _i_ i i i i i i i i i 1/10 1/9 1/S 1/7 1/S 1/S 1/4 1/3 1/2 1.0 2 3 4 5 S 7 8 9 10 * All certificated births Odds Ratio Pregnancy Outc ma F II w-Up Study Flgura 1.6 Estimated TCA Exposure During 1st Month Of Prognancy by Pragnancy Outcoma Original and Naw Exposad Araas Combinad (1981) (25) (98) (6) Pregnancy Outcome (n) SL 033233 1.20 Figure 1.7 Estimated Percent of Water from Well *13 In 1981 for Census Tracts In the Great Oaks Water Company ] 0% - 20% ] 10% -40% 20% - 60% ] 40% - 80% 60% - 100% SL 033234 2. Pregnancy Outcome Follow-up Study: Pregnancy Outcomes In Four Census Tracts of Santa Clara County 1980-1985 Background Two studies released in 1985 by the California Department of Health Services (CDHS) in cooperation with Santa Clara County (SCC) Health Department documented a significant excess of spontaneous abortions and birth defects in an area of the county served by the Great Oaks Water Company (GOWC) (1). These investigations were initially conducted because the community was concerned about a possible association between adverse reproductive outcomes and a leak of toxic waste into the community's water supply from the Fairchild Camera Company. In December of 1981, l,l,l*trichloroethane (TCA) was measured in one GOWC well (Well #13) at 1,700 parts per billion, a level 8.5 times higher than the current California state action level. TCA also was detected at GOWC Well #8, but at levels below the standard. Other rganic solvents, primarily 1,1-dichloroethylene (DCE), were detected at low levels in both wells. This study is an extension of one of the previous investigations, the interview study of pregnancy outcomes In relation to water contamination in 1980*81 in San Jose, California (1). The original study examined adverse pregnancy outcomes including spontaneous abortions (SABs), birth defects, and low birth weight occurring among 1980 and 1981 pregnancies in two census 2.1 traces. One tract (5120.12) was served by the GOVC and presumably received water from Veil #13 that was contaminated with TCA and DCE. This tract Is also known as the Los Paseos area. The comparison census tract (5120.08) was demographleally similar to Los Paseos and was not served by GOVC. In 1982 the drinking water of the comparison tract was found to be free of contamination. The original study shoved: (1) that spontaneous abortions and congenital anomalies in the Los Paseos area occurred more frequently than In the control area; (2) the excess could not be attributed to differences in known maternal risk factors between the two areas; and (3) the Indirect evidence regarding the extent and timing of exposure to contaminated water within the Los Paseos area was insufficient to determine whether the contaminated water of Well #13 of the GOVC caused the excess. Rationale This current study is designed to clarify and resolve questions raised by the rlginal study. In particular, adverse pregnancy outcomes In the Los Pase s area (5120.12) and control area (5120.08) are compared for the period 1982 to 1985. Because the contaminated well was closed in December of 1981 and because only trace levels of contaminants were detected In other GOVC wells between 1982 and 1985, persistence of the excess of spontaneous abortions and birth defects into 1982 to 1985 would suggest that it was unlikely that the excess identified in the first study was due to exposure to the contaminated water. The current study also compares adverse pregnancy outcomes occurring in 1980-85 in another census tract served by GOVC (5120.11) to a comparable area not served by GOVC. Prior to the 2.2 SL 033236 hydrogeologic exposure analysis reported in the Great Oaks Vater Distribution Study (GOWDS) (2), we assumed that residents of tract 5120.11 and 5120.12 were equally likely to have been exposed to water from Uell #13 and to toxic chemicals leaked from Fairchild. The hydrogeologle analysis (2) indicated that residents of 5120.11 were actually more likely to have been exposed to water from Well #13 than residents of 5120.12. If the leak was responsible for the the adverse pregnancy outcomes in 5120.12, we would expect to see a similar or greater rate of adverse pregnancy outcomes in 5120.11 during 1980*1981 than in the Los Paseos area. Similarly this area (5120.11) should show elevated relative risk estimates compared to an unexposed but otherwise comparable tract. STUDY OBJECTIVES The present study had the following objectives: 1. To determine if the elevated rates of spontaneous abortions and birth defects in 1980-81 observed in the previously studied census tract (5120.12) persisted during 1982-85, after the well with known contamination was closed. 2. To determine the rates of adverse pregnancy outcomes (spontaneous abortions, birth defects and low birth weight babies) in a census tract (5120.11) which had a potential for exposure similar to the previously studied tract (5120.12) during the period covered by the original study, 1980-81, and the follow-up period, 1982-85, and compare these rates to an additional unexposed area. 2.3 SL 033237 3. To determine if exposure to various levels of contaminated well water were associated with differences in rates of adverse pregnancy outcomes. Classification of environmental levels of water contamination in the study areas are described in the GOUDS (2). Methods, results and conclusions for the third objective are presented in a separate section following the methods, results and discussion for the first two objectives. METHODS Definition of Adverse Pregnancy Outcomes This study focused on the following three classes of adverse pregnancy outcomes: (1) Spontaneous abortions: fetal losses occurring at less than 20 weeks gestation. (2) Congenital anomalies: codes 740-759.9 of the International Classification of Disease, 9th Revision, Clinical Modification, diagnosed at birth or within the first six months of life. The term "congenital anomalies" refers to a combination of "congenital SL 033238 2.4 malformations," due to poor formations, and "c ngenital deformations,"due to mechanical constraint f rces. (3) Low birth weight: Newborns weighing less than 2500 grams at birth. For completeness, data also was collected on stillbirths. However, since these are rare events, the study did not have adequate statistical power to detect even large differences in the occurrence of stillbirths between xposed end unexposed areas. Selection -af StMfly Amg The original exposed tract (5120.12) was chosen because it was presumed to have received water from the contaminated well and because of concerns of residents in that area. Census tract 5120.08 was chosen as the comparison census tract for 5120.12 because the two tracts had similar demographic characteristics, were near to each other geographically and the water supply for tract 5120.08 was shown to be free of contamination by TCA and DCE in 1982. Similar criteria were used to select a comparison area for tract 5120.11. Census tract 5120.11 was chosen because the limited exposure Information available at the beginning of this study suggested that this tract had exposure to Veil #13 water which was comparable to tract 5120.12. There are essentially three sources of home tap water in Santa Clara County: ground water, surface water and mixed ground and surface water. GOVC used ground water. The only other water company in central SCC which primarily provides ground water and whose water showed no contamination with TCA or 2.5 SL other volatile organics in recent sampling was the City of Santa Clara Water Company. For this reason, we limited our search for a comparison area for tract 5120.11 to the city of Santa Clara. In this area, two adjacent tracts, 5053.02 and 5053.04, were demographically the most similar to 5120.11. Both were used to achieve a sample size comparable to 5120.11. They were treated as a single entity in the study and for convenience are referred to as 5053.0204. F r simplicity in what follows, the original census tracts, 5120.12 and 5120.08, will be called the "original study areas" and the census tracts added for this study, 5120.11 and 5053.0204, will be called the "new study areas." Figure 1.1 shows the location of these areas in Santa Clara County. Table 2.1 presents demographic characteristics and fertility of women residing in the four areas obtained from the US Census (1980). Enumeration of Pregnant Women In the Study Areas In 1986 CDHS staff obtained addresses of all households in the four study areas from the Center for Urban Analysis in San Jose. We then enumerated members of each household living in the study areas by mail, phone, or home visit in order to locate women who had been pregnant in 1980 to 1985. The informant, preferably the female head of household, was asked to list all household members on a standard form and Indicate any pregnancies occurring in the study period. Copies of the letters and enumeration form are presented in Appendix A. SL 033240 2.6 Some men filled out the census form ev n though we requested that women provide the Information. To assess whether male respondents reported all pregnancies, we reenumerated a 10% sample of the forms completed by men which indicated that a woman of reproductive age lived in the household but which did not report any pregnancies for the woman. We contacted the oldest female between the ages of 18 and 45 in each household and asked if there were any pregnancies during the. study period among women in the household. Interview Methods to Ascertain Information on Pregnancies Information from the enumeration forms was used to Identify women aged 15-49 who were pregnant anytime in 1980 through 1985. These women were then sent a letter explaining the study and requesting their participation in a telephone interview. The letter was followed by a telephone call from a trained CDHS interviewer. Vomen who could not be reached by phone were contacted in person. Consenting women were interviewed in detail about their pregnancies that occurred between 1980 and 1985. Pregnancy eligibility was determined by Interviewers using the criteria below. Outcomes of eligible pregnancies and maternal and paternal factors which might be related to the pregnancy outcomes were assessed via questionnaire. Eligibility criteria for each reported pregnancy were: 1. Live birth, stillbirth, or spontaneous abortion. 2. Estimated date of conception on or after January 1, 1980. 2.7 SL 033241 3. Pregnancy terminated on or bef re December 31, 1985. 4, Woman lived in the study area the entire time of her pregnancy. Copies of the questionnaire and letter are included in Appendix A. Validation of Spontaneous Abortions Rep rts of spontaneous abortions were validated as soon as possible after the interview. Validation was achieved through the woman's attending physician or through additional interviews conducted by a CDHS physician or nurse. The validation forms are presented in Appendix A. Spontaneous abortions were then categorized as either definite, probable, possible, not a spontaneous abortion or not validatable. . Only definite spontaneous abortions were included in the analysis. A definite spontaneous ab rtlon was one confirmed by a MD or medical record which indicated a dilation and curettage (D&C), or a positive pregnancy test with an appropriate clinical course. A probable spontaneous abortion was a pregnancy identified by a positive home pregnancy test, but having an atypical history or a pregnancy of at least six weeks after the last menstrual period (LHP) with positive symptoms of pregnancy and a typical history, late, painful menses (less than six weeks post LHP) with symptoms of pregnancy or a late, painful menses without symptoms of pregnancy (at least six weeks post LMF) was categorized as a possible spontaneous abortion. A th rapeutic abortion, ectopic pregnancy or a reported 2.8 SL 033242 spontaneous abortion with a t tally atypical history vas categorized as "not a spontaneous abortion". If a women reported a spontaneous abortion during the initial telephone interview but no MD report or medical records were available and the women herself was unavailable for an interview by CDHS medical staff, the spontaneous abortion was categorized as not validatable. In the previous interview study (1>; Kaiser Health Plan was sent the names of all women in the study areas who had not been enumerated as pregnant in 1980*81. Approximately 21%, or 732, of these women were Kaiser members. Their medical records for 1980 and 1981 were then reviewed. Of these 732 w men, 95% had, in fact, no record of having been pregnant in 1980 or 1981. Of the remainder, most pregnancies resulted in therapeutic abortions. Only one spontaneous abortion was Identified in the review. As a result of this effort we are reassured about the nearly complete ascertainment of spontaneous abortions which occur in pregnancies for which medical care has been sought. Therefore, this type of validation was not repeated in this current study. Validation of Birth Defects Rep rted birth defects were first reviewed by a CDHS physician. Suspected defects were then validated through physician report or medical or hospital r cord review. A physician from the California Birth Defects Monitoring Program (CBDMP) reviewed all suspected reportable defects and classified them according to criteria used by CBDMP as: 1) a CBDMP reportable malformation, 2) a CBDMP reportable deformation, 3) a CBDMP non-reportable defect, or 4) not a defect (3). 2.9 SL 033243 In che previous study (1) we reviewed the medical records from the hospital of birth for a random sample of 300 children whose mothers were interviewed and reported no congenital anomalies. Only one of these children had a birth defect noted on the medical record, indicating no serious under reporting. We did not repeat this validation of negative responses In this study. Also in the previous study, birth certificates were used to identify suspected congenital anomalies in children born to parents living in any of the study areas at the time of birth, but whose parents moved, after the child's birth and therefore were not enumerated at the time of census. In this way, four congenital anomalies among 287 live births were identified. Because there is little information on maternal risk factors for congenital anomalies on birth certificates, data on congenital anomalies obtained from birth certificates are of limited value. Thus, we did not repeat the review of birth certificates to identify congenital anomalies in children born to parents who were not identified in the enumeration. Identification of Low Birth Weight Babies The occurrence of low birth weight babies was ascertained differently than spontaneous abortions or congenital anomalies. Birth weight as reported on birth certificates in the vital statistics file for SCC were obtained for all live births to women who were residents in the study areas at the time of delivery and had an estimated date of conception between January 1, 1980 and December 31, 1984. Separate analyses were conducted for births to all women who were residents in the study area at the time of delivery whether or not they were interviewed and for births to those women identified in the interview study who resided in the study area for their entire pregnancy. 2.10 SL 033244 Analytic and Statistical Methods Spontaneous Abortions: Numerators for spontaneous abortion rates were those spontaneous abortions classified as "definite" after validation. Denominators were definite spontaneous abortions plus live births. Rates were computed for each study area and year and for the two time periods 1980-81 and 1982-85. Only pregnancies with estimated dates of conception prior to April 1985 were Included in the analyses of spontaneous abortions because the final eligibility date of December 31, 1985 precluded full term births with later c nception dates. For the original and new study areas, unadjusted odds ratios and 95% confidence intervals comparing the exposed to unexposed areas were computed for each year and grouped time period. Sp ntaneous abortion rates and 95% confidence intervals were also computed f r numerous potential covariates. Definite malformations were subtracted from the live births for these comparisons to eliminate the possibility of diluting any associations. Multiple logistic analyses were used to assess effects of study areas and time periods while controlling for potential c variates. A variable was included in the logistic analyses if its unadjusted relative risk for high to low risk category was statistically significant at the .05 level or was greater than or equal to 2.0 regardless of significance level. Vomen who reported IUD use (n - 17) at the time of pregnancy were excluded from the logistic analyses to restrict analysis to idiopathic spontaneous abortions. Vomen reporting use of water from a 2.11 SL 033245 private veil (n - 5) were also excluded because their water source was not on of those under study. Because of the lack of independence of multiple pregnancies occurring in the same woman and resulting bias in spontaneous abortion rates which are based on ell pregnancies, analyses of spontaneous abortions were based on a single randomly chosen eligible pregnancy per woman. For purposes of comparison, rates end unadjusted comparisons between areas and time periods were also computed for all reported pregnancies. Birth Defects: Numerators for birth defect rates were birth defects classified as CBDMP reportable malformations after validation. Denominators were all reported live singleton births. Rates were computed for each study area and year and for the two time periods 1980-81 and 1982-85. For the original and new study areas, unadjusted odds ratios and 95% confidence Intervals comparing the exposed to unexposed areas were computed for each year and grouped time period. Birth Weight: Low birth weight rates were calculated for 1980-85 in the four study areas; numerators were numbers of babies with birth weight less than 2500 grams; denominators were all live singleton births. Computations were made for all births with certificates and for those births to interviewed women. Births to women reporting diabetes or rubella during pregnancy were excluded. 2.12 Sl> Analysis of birchs to interview d women were based on a single randomly chosen eligible pregnancy per woman. Unadjusted odds ratios and 95% confidence intervals comparing exposed to unexposed areas for the years 1980-81 and 1982-85 were computed separately for the original and new study areas. Variable classification criteria: Methods for classifying most variables were based on precoded interview questions. Methods for coding and classifying occupation and Industry, chemicals and medications were more complicated and are presented here. Occupations and Industries: All mothers' and fathers' job histories were c ded according to the 1980 Bureau of Census Occupation and Industry three digit codes by a trained research assistant unaware of pregnancy outcome. Occupational exposures were analyzed for first trimester employment only. Codes were grouped into three Industries of particular Interest (electronics, health care and child care), three related occupational groups, (electronics workers, health care workers and child care workers), solvent exposed occupations, and sedentary versus non-sedentary occupations. Th electronics industry and occupations were of interest because of p tential exposure to organic solvents, while the health care industry and occupations were of interest because of potential exposure to biologic and pharmaceutic agents. Census codes for the electronics Industry included electronic components, computers and communication equipment while occupational codes included operator, printer, metal plater and assembler. The codes for health care industry included hospitals, doctor's offices and 2.13 SL 033247 laboratories. Child care providers f rmed a separate industry category. Health care occupations included physicians, nurses and allied health care professionals. Again, child care providers formed a separate occupation category. Classification of solvent exposed occupations was much broader and Included a list compiled by CDHS industrial hygienists and occupational physicians of 35 probably exposed and 37 possibly exposed occupations (see Appendix B). For the comparison of sedentary and non-sedentary occupations, secretaries, typists and general office clerks were classified as sedentary while teachers and waitresses were classified as non-sedentary. Chemicals: Responses to an open-ended question (with specific prompting for s lvent exposure) regarding exposure to chemicals or solvents at home or at work, or in some other activity during the first trimester, were categorized as solvent, gasoline, or pesticide exposure. Gasoline exposure referred to exposure to gasoline at self service stations. An insufficient number of w men reported lead exposure or specific solvent exposures to allow analysis by these exposures. Medications: Based on frequency of response, responses to an open-ended question regarding first trimester use of prescription medications were classified into the following 10 categories: tetracycline, other antibiotics, antihistamines, estrogens, corticosteroids, narcotics, fertility medications, thyroid medications, pain medications, and prenatal vitamins. 2.14 033248 si* Enumeration: zssms. A total of 10,063 households were enumerated between April and October, 1986. Distributions of response rates by method of enumeration for each study area are shown in Table 2.2. Refusal rates were very low in the four areas but were somewhat lower in the exposed areas (2.5% and 2.6%) than in the unexposed areas (3.7% and 4.2%). Across all areas the refusal rate was 3.3%, and the vacancy rate 3.4%. Another 1.2% of households did not appear vacant, but no response was obtained after seven door-to-door visits and 0.5% households could not be enumerated because of language problems. This gave an overall response rate of 91.6%. Approximately 360 forms were completed by men who reported no pregnancies but indicated that a woman of reproductive age lived in the household. A t tal of 114 households were randomly selected from these to be reinterviewed. Vomen in 28 (25%) of these households could not be reached by phone for further questioning about their pregnancy history. Of 110 w men listed in the remaining 86 households, only two were reported to have pregnancies upon telephone interview with women in the household. If this group of households were representative of the other households in which men reported no pregnancies in women of reproductive age, only about eigjit women with potentially eligible pregnancies would have been missed in our study. 2.15 SL 0332^ Numbers f Women Interviewed-end Pregnancies Rep rted: According to the household enumeration a total of 2,713 women were reported t be pregnant sometime during the study period. Of these, 2,581 (95%) were Interviewed; two percent refused and we were unable to locate or had language difficulties with three percent. Of interviewed women, 1,476 rep rted no eligible pregnancies and 1,105 women reported one or more eligible pregnancies (table 2.3). Numbers of pregnancy outcomes reported p r woman and the distribution of outcomes for eligible pregnancies are given in Table 2.4. Sp ntaneous Abortions; Results of validation of 260 reported spontaneous abortions are shown in Table 2.5; 211 (81.2%) were considered definite. Of these 211, 181 were conceived prior to April, 1985 and therefore are included in subsequent analyses. The crude spontaneous abortion rate among all reported pregnancies conceived January 1, 1980 * April 1, 1985 was 13.2%. Multiple pregnancies per woman are not Independent events and number of subsequent pregnancies depends on pregnancy outcome. Thus, the 13.2% is an overestimate of the spontaneous abortion rate. The rate based on a single randomly chosen pregnancy per woman was 9.9% (Table 2.6). The numbers of definite spontaneous abortions and live births by study areas, years, and grouped time peri ds are given in Tables 2.7 and 2.8. The 2.16 SL 033250 spontaneous abort! n rates are shown in Table 2.9. The rates range from a low of 0% in 1980 in the original control area (5120.08) to 28.6% in 1985 in the new control area (5053.0204). There are no apparent trends with year, grouped time period, or exposure status. Appendix C fits the spontaneous abortion rates by census tract and area to the normal distribution expected if all the rates came from a single binomial distribution. Even though the rates appear to be quite variable they are compatible with those expected by chance. Odds ratios and 95% confidence Intervals for spontaneous abortions in exposed versus unexposed areas for each year and grouped time periods for the original and new study areas also show no trends. Thus, these data are not consistent with the hypothesis that the leak was associated with the increased rate of spontaneous abortions observed in 5120.12 in 1980-81 (Table 2.10). As in the first interview study (1), the odds ratio for the original exposed versus the original unexposed area for 1980 and 1981 was significantly gr ater than one (OR " 3.52; 95% Cl: 1.21, 10.31). However, the odds ratio for the new study areas for 1980 and 1981 was less than one (OR - 0.32; 95% Cl: 0.09, 1.14). Combining the original and new study area for 1980 and 1981 yields an odds ratio for exposed versus unexposed areas of 1.4 (95% Cl: 0.9, 2.1). For 1982-85, the post-exposure time period, the odds ratios for the original study areas is 1.00 and for the new study areas is 1.31; combining the original and new study areas yields an odds ratio of 1.1, (95% Cl: 0.8, 1.3). 2.17 SL 033251 Table 2.11 compares Che spontan us ab rtion results from this study with those found In the original study (1) conducted In 1983. As shown, the estimated odds ratios for exposed to unexposed area in 1980-81 were very similar from the two studies: this study gave slightly higher relative risks estimates than the original study. Tabulations of spontaneous abortions rates and odds ratios by census tracts and years using all eligible pregnancies are contained in Appendix D. Tabulation of stillbirths by year and area are also In Appendix D. Spontaneous abortion rates and 95% CIs for numerous potential risk factors and other covariates are presented In Table 2.12. Variables for which there was at least a 2-fold relative risk between highest and lowest risk category (and for which there were more than 10 pregnancies for highest and lowest risk category) or for which the relative risk was significantly greater chan 1 included: woman's age at pregnancy, number of previous spontaneous fetal 1 sses, month prenatal care began, IUD use at time of pregnancy, first trimester bleeding, nausea, nausea medication use, prenatal vitamins, antibiotics, alcohol use, probable solvent exposure at work, father's occupation of electronics worker or health care worker, used water from a private well, water filter use around the time pregnancy began, mother's and father's ethnicity, first trimester cold tap water consumption, and bottled water use anytime during pregnancy. In addicion, average gestational weeks differed among time periods and study areas (Table 2.13). 2.18 SL 033252 To decide if any of these factors might alter th odds ratios for area and time period, several mutiple logistic models were considered. Using the established criteria, a preliminary model was fit that included study area (four census tracts), time periods (1980*81 and 1982-85), an interaction between census tracts and time periods, any vs. no home cold tap water, maternal age ( 20, 21-34, 35+), number of previous miscarriages or stillbirths (0, 1, 2 or more), ethnicity (White or Hispanic versus others), ale hoi consumption (< 4 drinks/week, 4+ drinks per week), probable versus n t probable solvent exposure on the job, father's occupation as electronics worker, father's occupation as health care worker, and water filter use around time pregnancy began. Despite its lack of association to sp ntaneous abortion, smoking (none, > 1/2 pack per day, or about a pack a day r more) was included in the model because of the reported association b tween smoking and SAB in other studies (4). Use of prenatal vitamins or antibiotics was not included in the model (in spite of the observed ass elation with spontaneous abortion) because their use varied with pregnancy length and month prenatal care began. Furthermore, specific drugs were not solicited and may be over or underreported among women with different outcomes. Bleeding was not included in the model because it may be the first symptom of spontaneous abortion. Bottled water was not Included in the model because we did not ask first trimester use; women with live births would have five to seven months more time to have drunk bottled water during their pregnancy than women with SABs. Results for the preliminary model vising father's occupational variables are not presented b cause these variables were not found to significantly contribute to the model or to alter the other parameter estimates when included or excluded from the model and missing data for these variables reduced the number of 2.19 SL 33253 pregnane1 s chat c uld b analyzed (number with missing Inf rmatl n - 38). Results for seven other models are presented In Appendix E. Model 1 uses all the variables listed above except father's occupational variables; model 2 used only White and Hispanic women; model 3 used only women reporting first trimester nausea; model 4 used only women reporting no nausea during the first trimester; model 5 used only pathology confirmed SAfis; model 6 used only early SABs (those with estimated gestational week of nine weeks or less); and model 7 excluded pregnancies conceived in 1980 because of the uncertain contamination status of the water in 1980. In order to provide confidence limits for the odds ratios associated with study areas and time periods and with water filter use among tap water and non* tap water drinkers a final model (model 8) was fit which entered the eight census tract and time period combinations as a single variable and four water filter use and cold tap water use combinations (Table 2.14). The parameter estimates of the other covariates in this model are very similar to those found in model 1. To summarize the results for model 8, women under 20 or over 34 were more likely to have an SAB (p-.OOl), as were women with prior fetal loss (p-.lO). Whites and Hlspanics had higher rates of SAB than other ethnicities (p - .01) and women classified as having jobs that probably Involved solvent exposure had elevated odds ratio compared to other worn n (p-.008). Women reporting cold home tap water use without filters were 5.5 (95% Cl 2.1, 14.0) times as likely to have had SABs as those using no tap water while women consuming tap water but using filters were only slightly more likely to have had SABs chan non-tap water drinkers OR - 1.6, (95% Cl; 0.3, 8.6). A separate CDHS document (5) further explores this association. Neither alcohol consumption nor sm king were significantly 2.20 SL 033254 related to occurrence of SAB (p-.60 and .94 respectively). All of the ORs for study areas and time periods were less than the original exposed area during 1980-81, but only the new exposed area for 1980-81 and the original c ntrol area for 1980-81 had odds ratios significantly less than the rlglnal exposed areas. The odds ratios in Table 2.14 are also rearranged to nable easy comparison to the unadjusted odds in Table 2.10, and are quite similar to the unadjusted ORs. N ne of the models presented in Appendix altered the basic impression that the odds ratio of spontaneous abortion was only elevated in the original exp sed study area in 1980-81 and that the odds ratio was equally elevated relative to both the original control area and the new exposed area during that time. Omitting pregnancies conceived in 1980 (model 7) had very little effect on the estimated odds ratios. S me other results are of interest in comparing models 1 through 6. The del restricted to Whites and Hispanics (model 2) gave very similar results t the model that Included ethnicity as a parameter (model 1). Women reporting first trimester nausea (model 3) appeared to have different patterns of odd ratios for some factors than women reporting no nausea (m del 4). Of Interest, women reporting nausea compared to non-nauseous women had higher odds ratios for tap water consumption, alcohol use, and probable solvent exposure. Among pathology confirmed spontaneous abortions (model 5) compared to all definite spontaneous abortions (model 1), odds ratios were somewhat lower (cl ser to unity) for tap water use and two or more prior fetal losses. The 2.21 SL 3325s odds rstio for n n-White or n n-Hispanic ethnicity were even lower and the odds ratio for probable solvent exposure even higher among the pathology confirmed cases. No pathology confirmed SABs occurred among women reporting filter use. Odds ratios for prior spontaneous fetal loss were higher among early e sp ntaneous abortions (model 6) than all definite spontaneous abortions (m del 1). None of the women classified as probably solvent exposed had early spontaneous abortions yielding a very low estimated odds ratio for this factor. Birth Defects: Table 2.15 presents results of the validation of all reported birth defects. One hundred sixteen women reported having a child bom with a birth defect during the study period. Of these, 36 (32t) validated birth defects met the CBDMF classification of a reportable malformation and one validated case met the CBDMP classification of a reportable deformation. Diagnoses of cases of c ngenltal anomaly by study area, month and year of birth are presented in Table 2.16. The anomalies are grouped into malformations and deformations. Only one of the 37 cases of a reportable defect was an Isolated deformation. Therefore, the remainder of the analysis included only the 36 malformations. Numbers of reportable malformations, numbers of singleton births, and malformation rates by census tracts and years are presented in Tables 2.17-2.19. Table 2.20 presents odds ratios and 95% confidence Intervals for all reportable malformations In exposed versus unexposed areas by time period. A four-fold excess of birth defects is seen in the original exposed census tract for the period 1980-81. However, the new exposed census tract demonstrates a deficit for this time period. During the 1982-85 study period a slight, non-significant excess in birth defects among exposed c mpared with unexposed census tracts was found. Malformation rates and odds ratios were computed for the following potential covariates: parental ethnicity , education and occupation, mother's work sch dule, chemical exposure, home pest control application, diabetes, hypertension, prescription medications, IUD use, amniocentesis, chorionic villus sampling, x-ray exposure, fever, nausea, alcohol consumption and cigarette smoking. For those exposures which might not have been constant thr ughout the pregnancy, first trimester exposure was examined separately. C variates associated with a 2-fold or greater increase in risk of malformations among those exposed compared to those unexposed to the factor are as follows: narcotic use, chest x-ray first trimester, a work schedule involving rotating between two work shifts and father's occupation in childcare. For each of these covariates only one malformation occurred among women exposed to the factor. First trimester alcohol consumption was associated with a 2-fold deficit In the malformation rate among exposed. Although the number of cases of any specific defect was very small in this study, rates for specific defects among study participants were calculated and are presented in Table 2.21. This table contains a list of all categories of congenital anomalies for which published CBDMP rates are 2.23 Si 33257 available. In det raining races, children vich aulCiple defects ar included in aore Chan one category. Races froa Che Centers for Disease Control prograa in the Atlanta Metropolitan Area are given as an additional coaparison rata. No condition occurred aore than four tiaes in the study p pulation with the aajority of conditions appearing only once. Because of th small number of overall eases (37) the rates for individual conditions are very unstable and therefore, of limited value. This table also dea nstraces the numerous conditions which are equally prevalent in the general population but did not occur among the study area population. Of the conditions which occurred among children born in the four study areas but which are not in the CBDMP published list, only four conditions occurred a re than once: cardiac malformations, (ICD9-CM-745.4,.5) (n-2), undescended testicle (ICD9-CM-752.5) (n-2), duplicated ureter,(ICD9-CM-753.4) (n-2), pectus excavatum (ICD9-CM-754.81) (n-2), for a rate of 1.71/1000 live births for each condition. Table 2.22 lists covariates of interest for the four categories of c ngenital malformations occurring in aore than two children: pyloric stenosis, anomalies of the genitourinary system, hypospadias and syndactyly. No pattern of risk factors are common to a specific malformation. 2.24 033258 SL Birth Weight: Numbers of low birth weight babies, numbers of singleton births, and low birth weight rates from all certified births by census tract and years are shown in Tables 2.23 to 2.25. Rates in the grouped time periods ranged from a low of 1.5 in the new unexposed (5053.0204) area in 1980*81 to a high of 4.8 in the new exposed area (5120.11) in 1982*85; there were no apparent trends in rates among study areas or time periods. In the original study area, low birth weight occurred somewhat less frequently in the exposed compared to the unexposed areas in both time periods OR (1980-81)- 0.7; OR (1982-85) - 0.6 (Table 2.26). The occurrence of low birth weight was s mewhat higher in the exposed versus unexposed new study areas OR (1980-81) - 1.7; OR (1982-85) -- 1.2. Combining the original and new study areas, the odds ratios for low birth weight in exposed versus unexposed areas were 1.07 for 1980-81 and .98 for 1982-85. In the analysis of low birth weight among all certificated births the following risk factors of low birth weight were evaluated as potential c nfounders; mother's age, race, pregnancy complications (pre-eclampsia, eclampsia), presence of and time of prenatal care, previous spontaneous fetal loss, prematurity (< 36 weeks gestational age), parity (nulliparity or more than four prior births versus 1-3 prior births), and infant's sex. Adjustment for these factors individually using the Mantel-Haenszel method, or simultaneously using logistic regression did not alter the odds ratios associated with time periods and study areas. 2.25 Result* f c births to interview <1 women wer similar to results for all certificated births, (Tables 2.27 to 2.30). The estimated odds ratio for 1980-81 births was lower in the original study areas (0.3 for interviewed women compared to 0.7 for all certificated births) but was higher in the new study areas (4.4 for interviewed women versus 1.7 for all certificated births). All of the 95% CIs Included unity. In addition to factors examined in relation to low birth weighc in all certificated births, first or second trimester bleeding or nausea, maternal sm king and drinking habits, mother's and father's education and employment status, mother engaged in shiftwork or worked more than 40 hours per week, and mother's ponderal index were also examined In interviewed women. Only variables found to be significantly associated with low birth weight were included as potential confounders in multiple logistic analyses. These variables were: any vs. no cigarette smoking, premature delivery (<36 we ks), low weight gain 20 lbs) (models were examined which included and excluded these variables), previous spontaneous fetal loss, first trimester ale hoi consumption (none, 1-3 drinks/week, 4 or more drinks/week), nausea during second or third trimester, and job in the electronics industry (yes: % low birth weight - 14.8%, n - 27; no: % low birth weight - 3.0%; n - 871; p - .001). The observed lack of relationship of low birth weight to study areas and time periods persisted after adjustment for these factors. 2.26 SL 033260 DISCUSSIQE Sppntanms Atortlani Overall rates of definite spontaneous abortion of -13.2% (among all eligible pregnancies per woman) and 9.9% (among a single eligible pregnancy per woman) are compatible with published rates of spontaneous abortion among rec gnized pregnancies. The most often cited rates range from 10*20% (4, 6) but rates as low as 7% have been reported (7). This combined with enumeration and interview response rates exceeding 90% and with the evidence fr m the original study (1) that there was very little underreporting of pregnancies among enumerated Kaiser Health ?lan members suggests that we had satisfactory ascertainment of pregnancies. This conclusion may not apply to Asians who had a significantly lower spontaneous abortion rate (4.2%). Average gestational age of five spontaneous abortions to Asian women (mean age - 14.20 weeks) were somewhat greater (p-.08) than those to 98 Uhite or Hispanic women (mean age - 11.0 weeks) suggesting that Asian women with early SABs may have been less likely to have been enumerated as pregnant. Ass clations observed between spontaneous abortion and woman's age at pregnancy, prior spontaneous fetal loss, and nausea are consistent with other studies (4, 8). Although not statistically significant, there was a tv -fold relative risk for spontaneous abortion associated with consuming four or more alcoholic drinks per week' in univariate analysis; the odds ratio was reduced to 1.4 after adjustment for other factors but was 2.3 in the model restricted to women reporting nausea. Associations of alcohol 2.27 SL 033261 consumption and SAB In other studies heve not b en totally consistent but suggest e deleterious effect of alcohol consumption (4, 9). The unadjusted odds for spontaneous abortion and cigarette smoking suggested a a dest association in this population (OR-1.9). However, the adjusted odds for smoking one or more packs a day were very close to unity.in all models except that restricted to early spontaneous abortions (OR-2.0). Other studies have reported significant associations of smoking with sp ntaneous abortions, but the relative risks were well below two among light smokers In one study (9) and did not hold after adjustment for maternal age, social class, and oral contraceptive use in another study (reviewed in 4), The elevated, unadjusted odds ratio for spontaneous abortion among those classified as probably occupationally exposed to solvents persisted after adjustment in all models except the model restricted to early spontaneous abortions. As with alcohol exposure, the estimated odds ratio for ccupational solvent exposure was highest in the model restricted to women rep rting first trimester nausea. Although recall bias is a possible explanation for these findings, the initial categorization of solvent exposed occupation was made on occupational code alone. Only after a job was categorized on that basis were the job activities and reports of solvent exposure used in making the classification of probable solvent exposure. Solvent exposure has been Implicated in spontaneous abortion risk but not consistently so (10). A nearly completed CDHS study will more thoroughly address the possible association between parental solvent exposure and spontaneous abortion. 2.28 033262 A separate report discusses at length the finding that c nsumption of unfiltered tap water is significantly associated with spontaneous abortion independent of study area or time period (5). To summarize, our findings are consistent with other studies with respect to observed rates and risk factors for spontaneous abortion. Furthermore, our results for the original study areas and the original time period (1980*81) are very consistent with those reported for the original study. This suggests that potential differential migration out of the exposed area either did not occur or was of insufficient magnitude to seriously alter the observed odds ratio for exposure in the original areas and time period. It seems unlikely, therefore, that we would fail to detect an elevated odds ratio for exposure among 1980-81 pregnancies In the new study areas if it had occurred. In fact, we observed a non-significant odds ratio of less than ne for exposure in the new study areas in 1980-81 and an overall odds rati of very close to one after combining the original and new study areas. Adjusting for covariates did not alter these findings. Given that the new exp sed area (5120.11) was more likely to receive water from Veil #13 [see Great Oaks Water Distribution Study (2)], it seems very unlikely that the bs rved excess of spontaneous abortions in the original study area (5120.12) would be due to exposure to water from Veil #13. A subsequent section presents more detailed exposure analysis of 1981 pregnancies in areas 5120.11 and 5120.12. 2.29 SL 033263 Birth Defects and Birth Weight As with spontaneous abortions, our overall rates of birth defects and low birth weight are similar to published rates from other studies (See Table 2.21 and Reference 11). The findings for original study areas and time periods are quite similar to those reported in the original study (1); there was an excess of malformations and a deficit of low birth weight babies in the riginal exposed area In 1980-81. However, there was a deficit of malformations and an excess of low birth weights in the new exposed area in 1980-81 so odds ratios of exposure in 1980-81 for either malformations or low birth weight were close to unity when the original and new areas were combined. 2.30 0^ ANALYSES OF 1981 PREGNANCY OUTCOMES IN RELATION TO HTORPSTOLPCTO EXPOSURE ESTIMATES Methods Exposure estimates: The Great Oaks Vater Distribution Study (GOWDS) provided estimates of the probability of receiving water from Veil #13 of the GOVC for each month in 1981. The estimates for each node (junctions of water distribution pipes) apply to a group of addresses around that node. The study further provided estimates of the proportion of each month that Well #13 was operational and f th concentration of TCA in water from Veil #13 each month. The author f GOWDS calculated these estimates with no knowledge of the distribution of pregnancy outcomes in the census tracts. The limitations of these estimates are presented in the report (2). Ve assigned a node number to the addresses of each reported eligible pregnancy conceived in 1981 in census tracts 5120.12 and 5120.11. The estimated exposure to water from Veil #13 for the 1st, 2nd and 3rd months of each pregnancy were then obtained for the appropriate node and months from Table XU in the GOWDS (2) and multiplied by the percent of each month that the V 11 #13 was pumping. Each probability was then multiplied by the estimated concentration of TCA in Veil #13 water for the appropriate month to give an estimated TCA exposure for the month. Cumulative exposure for the first two or the first three months of pregnancy were obtained by adding the estimates for the first two months and for the first three months of pregnancy. It was assumed that the probability of exposure was zero after December 31, 1981. The maximum of the two probabilities of exposure to Veil 2.31 033265 SL #13 water for the first two m nths f pregnancy and f the three probabilities of exposure for the first three months of pregnancy were also analyzed, as were the cumulative probabilities of no exposure to Veil #13 water during the first, the first two and the first three months of pregnancy. Statistical Methods Comparisons of estimated TCA exposures and probabilities of exposure to Veil #13 were made with analysis of variance (to compare means) and Vlleoxon rank sum tests (to compare medians). Raavlci Spontaneous Abortions Estimated TCA exposure for the first month of pregnancy and estimated cumulative TCA exposure for the first two months and first three months of pregnancy were very similar among pregnancies ending In live births versus those ending in SAB (Table 2.31A). This was true using either all reported pregnancies per woman or using a single random eligible pregnancy per woman. Similar results were obtained when exposure was assumed to be zero for women wh reported either no home tap water consumption during the first trimester or use of a water filter beginning around the time of conception. Using all reported SABs or only those considered "definite" SABs upon validation did not alter the results. Futhermore, percent distributions of spontaneous abortions and live births (excluding chose born with defects) by quartlles of estimated TCA exposure during the first month of pregnancy did not indicate increasing risk of spontaneous abortion with increasing estimated TCA exposure in either census tract 5120.11 or 5120.12 (Table 2.31B). 2.32 & Similarly, no differences were detected in the probabilities of exp sure to water from Well #13 between pregnancies ending in live births and those ending in spontaneous abortion (Tables 2.32 and 2.33). Birth Defects The six live births classified as CBDMP reportable malformations occuring in 1981 had significantly less cumulative exposure to TCA, lower maximum pr bability of exposure to water from Well #13 and higher cumulative probability of no exposure to water from Well #13 during their first trimester than live births without reportable malformations (Table 2.34). ggnclaaion These results support the finding that exposure to contaminated drinking w*ter from Well #13 was not related to risk- of spontaneous abortion or birth defects In.pregnancies conceived in 1981. 2.33 0332&7 Si* References 1. Epidemiological Studies Section, California Department of Health Services: Pregnancy Outcomes in Santa Clara County 1980-1982: Reports of Two Epidemiological Studies. California State Publications Section. Report 7540.958-1301-5, 1985. 2. Murphy P: The Great Oaks Water Distribution Study, 1988. 3. California Birth Defects Monitoring Program: Procedure Manual Vols I and II California Department of Health Services, Berkeley 1984. 4. Kline J, and Stein Z: Spontaneous abortion in Perinatal Epidemiology, M Bracken, (ed.) Oxford Unlv. Press, London, 1984; 23-51. 5. California Department of Health Services: Water Exposure and Pregnancy Outcome, 1988. 6. Warburton D: Effects of common environmental exposures on spontaneous abortion*of defined karyotype. In: Prevention of Physical and Mental Congenital Defects, Part C: Basic and Medical Science, Education, and Future Strategies 1985; 31-36. 7. Hemmlnki K, Niemi ML, Saloniemi I, Vainio H, Hemmlnki, E: Spontaneous abortions by occupation and social class in Finland. Int. J. Epidemiology 1980; 9:149-53. 8. Petltti, DB: Nausea and pregnancy outcome. Birth 1986; 13(4): 223-226. 9. Kline J, Levin B, Stein Z, Susser M, Wharburton D: Epidemiologic detection of low dose effects on the developing fetus. Environmental Health Perspectives 1981; 42:119-126. 10. Lindbohm ML, Tasklnen H, Hemmlnki K: Reproductive health of working women: Spontaneous abortions and congenital malformations. Public Health Rev. 1985; 13:55-87. 11. Kleinman JC, Kessel SS: Racial differences in low birth weight. N Engl. J Med 1987; 317:749-53. 2.34 6* 9V Table 2.1 1980 U.S, Census Characteristics Used to Select Two Control Census Tracts for Two Census Tracts Served By GOWC GOWC Tracts 5120.12 5120.11 Median Housing Cost $104,600 $101,800 Median Persons Per Unit 3.2 3.6 Median Age Females Race White Black Aslan Facific/Islander Other 26.6 84.5% 3.6% 6.8% 5.2% Spanish Origin1* 11.2% Fertility0 82.0% 25.1 81.0% 4.0% 4.9% 10.1% 16.3% 62.2% Control Tracts 5120.08 5053.0204 $113,800 $ 94,800a 3.2 2.9* 26.5 30.4* 84.0% 2.9% 6.9% * 6.1% 12.5% 60.7% 80.1% 1.8% 10.0% 8.1% 16.7% 59.1% Numbers are the average of tracts 5053.02 and 5053.04. Persons of Spanish origin may be of any race, c Defined as the number of live births divided by the estimated female population, age 15-44, multiplied by 1,000. 1599.tbl/13 2.35 033*69 Table 2.2 PercenC Distribution of Responses to Household Enumeration of Four Census Tracts in Santa Clara County EXPffgfid Unexposed fteSP9PS0 Enumerated: 1st mailing 2nd mailing Do r to door canv ssing Combinations of mailings and do r to door 37.1 17.2 39.0 0.5 5120.11(H3 5120.08(0') % 5Q53.0204(N) 35.3 16.0 41.4 31.2 16.4 42.3 31.9 18.5 41.3 0.3 0.4 0.1 a12IAi % 32.1 17.6 41.6 0.3 All Methods 93.8 Wot Enumerated: Refusal Vacancies Unenumerable Language Problem 2.6 2.5 1.0 0.1 93.0 2.5 3.3 0.9 0.3 90.3 3.7 3.6 1.7 0.5 91.8 4.2 2.4 0.5 1.1 91.6 3.3 3.4 1.2 0.5 All Reasons 6.2 7.0 9.5 8.2 8.4 TOTAL 2079 2637 3402 2373 10977 *Tocal is greater chan Che sum of the four areas because 486 additional eligible hous holds were discovered after the first mailing. These households were enumerated with the second mailing and door to door canvassing. The refusal rate was 3.1 % and the vacancy rate was 9.5% in these households. 0-Original study area N-New study area 2.36 gk 033270 Table 2.3 Response to Interviews Among Women For Whom Pregnancy was Reported In Enumeration Respondents: Women with 1 or more eligible pregnancies Women with no eligible pregnancies Non*respondents: Refused Unable to locate and language problems TOTAL --H-- 1105 1476 45 87 2713 41 54 2 3 100 2.37 SL 033271 Table 2.4 Number of Eligible Pregnancy Outcomes Reported by Interviewed Women Ineligible pregnancies') 1 2 3 4 6 TOTAL N(women) 802 245 48 8 2 1105 V Distribution 72.6 22.2 4.3 0.7 0.2 100.0 Distribution of Outcomes Among Reported Eligible Pregnancies Singleton Live Births Twin Births (5 twin pairs) Spontaneous Abortion Stillbirth 1199 10 260 11 1599-5.12 2.38 SL 033272 Table 2.5 Validation of Reported Spontaneous Abortions Validation Categories: NfSABI Included in analyses: Definite 211 Not Included in analyses: Probable** Possible0 Not a Spontaneous Abortion Not Validatablee 12 13 11 13 % Distribution 81.2 4.6 5.0 4.2 5.0 TOTAL 260 100.0 Confined by MD or medical record, D & C or positive pregnancy test with appropriate subsequent clinical course. Positive hone pregnancy test with an atypical history or a pregnancy of 6 weeks post LMP, with positive symptoms of pregnancy and a typical history. Late painful menses with symptoms of pregnancy if < 6 weeks post LMP, or without symptoms if > 6 weeks post LHP. Therapeutic abortion, ectopic pregnancy or totally atypical history. Neither woman nor MD nor medical record available for validation and did not fit into above categories. 1599-6.12 2.39 Si* 033273 Table 2.6 Validated Definite Spontaneous Abortions Among All Pregnancies and Among a Single Randomly Chosen Eligible Pregnancy8 Total File Random Eligible Pregnancy NfSAB^ 181 103 N(SAB + UY9 birth) 1376 1038 8 Conceived prior to April 1, 1985 Crude SAB rate* 13.2 9.9 1599-7.12 2.40 SL 033274 Exposed: 1980 1981 (1980-81) Unexposed: 1982 1983 1984 1985 (1982-85) TOTAL Table 2 Number of Definite Spontaneous Abortions9 by Year and Census Tract CENSUS TRACT EXPOSED UNEXPOSED __ JUV.H1WJ __ygiftn/__________ ivi juj j.uiwtrtj -lPIMHJ___ iiriAL. 4 9 (13) 1 3 ( 4) ( 5) (12) (17) 0 5 ( 5) 1 5 ( 6) ( 1) (10) (ID 6 22 (28) 2 5 12 3 (22) 35 3 7 8 2 (20) 24 ( 5) (12) (20) ( 5) (42) (59) 7 2 11 3 (23) 28 2 2 2 4 (10) 16 ( 9) ( 4) (13) ( 7) (33) (44) 14 16 33 12 (75) 103 2-4l SL 033275 a Number of spontaneous abortions among a single random eligible pregnanacy per woman. 0 - Original study area N - New study area 1583.1-11 Table 2. TIKE FEEIQP Exposed: 1980 1981 1980-81 Number of Live Births by Year and Census Tract CENSUS TRACT EXPOSED UNEXPOSED S12Q.12(Q)_____5120.11(N) ____ (BOTH)_______ 5120.08(0)____ 5051^0204(N) 44 32 ( 76) 34 48 ( 82) ( 78) ( 80) (158) 52 51 (103) 16 22 ( 38) (BOTH) ( 68) ( 73) (141) TOTAL 146 153 (299) Unexposed: 1982 1983 1984 1985 1982-85 43 44 52 16 (155) 54 48 77 14 (193) ( 97) ( 92) (129) ( 30) (348) 32 57 55 18 (162) 31 38 47 10 (126) ( 63) 160 ( 95) 187 (102) 231 ( 28) 58 (288) (636) SL 033276 TOTAL 231 275 (506) 265 164 3 Number of live births ,among a single random eligible pregnancy per woman. 0 - Original study area N - New study areaa (429) 935 Table 2.9 Spontaneous Abortion Ratesa by Year and Census Tract TIME PERIOD Exposed: 1980 1981 (1980-81) EXPOSED 5120.12(0) 5120.11(N> 8.3 22.0 (14.6) 2.9 5.9 ( 4.7) CENSUS TRACT UNEXPOSED (BOTH) 5120.08(0) 5053.0204(H) (BOTH) ( 6.0) (13.0) ( 9.7) 0.0 8.9 ( 4.6) 5.9 18.5 (13.6) ( 1.5) (12.1) ( 7.2) Unexposed: a 1982 1983 1984 1985 (1982-85) 4.4 10.2 18.8 15.8 (12.4) 5.3 12.7 9.4 12.5 ( 9.4) ( 4.9) (11.5) (13.4) (14.3) (10.8) 18.0 3.4 16.7 14.3 (12.4) 6.1 5.0 4.1 28.6 ( 7.4) (12.5) ( 4.0) (113) (20.0) (10.3) TOTAL 3.9 12.6 ( 8.6) 8.1 7.9 12.5 17.1 (10.6) SL 033277 TOTAL 13.2 8.0 (10.4) Among a single random eligible pregnancy per woman. 0 - Original study area N - New study area 1583-3.11 9.6 8.9 ( 9.3) 9.9 ro -*c** Ul XT' O w u> to o CO Table 2.10 Odds Ratios and 95% Confidence Intervals for Spontaneous Abortions In Exposed vs. Unexposed Areas by Year Tine period 1980 1981 (1980-81) 1982 1983 1984 1985 (1982-85) ORIGINAL STUDY AREAS (5120.12/5120.08) OR ____________ 95% Cl_______ NEV STUDY AREAS (5120.11/5053.0204) OR 95% Cl 10.62a 2.78 (3.52) 0.25 2.84a 1.15 1.12a (1.00) (0.56, 202.66) 0.48a (0.86, 9.03) 0.30* (1.21. 10.31) (0.32a) (0.06, 1.12) 0.81 (0.61, 13.33) 2.38 (0.47. 2.84) 2.08 (0.22, (0.54, 5.69) 0.40 1.87) (1.31) (0.05, 4.96) (0.07, 1.23) (0.09, 1.14) (0.15, 4.35) (0.54, 10.59) (0.49, 8.93) (0.07, 2.29) (0.59, 2.88) Snail sample adjustment Table 2.11 Comparisons of Ascertainment of Live Births and Spontaneous Abortions Conceived During 1980 or 1981 and Identified During a 1983 Interview Study and a 1986 Interview Study Census Tract 5120.12 Census Tract 5120.08 1980-1961 Live births Definite SAB All Reported SABs l\J in 1983 Study___________ 195 35 (19.4%) 40 (21.6%) 1986 {>tutJy 97 22 (18.5%) 26 (21.1%) 1983 Studv 182 21 (10.3%) 22 (10.8%) 1983 Study 1986 Study 1986 Study 127 10 (7.3%) 13 (9.3%) OR for definite SABs 5120.12 vs. 5120.08 2.1 29 OR for all reported SABs 5120.12 vs. 5120.08 2.3 2.6 SL 033279 Table 2.12 Spontaneous Abortion Rates by Selected Variables Variable Woman's age at pregnancy SAB Rate (*) N (Pregnancies^* 95% Confidence Interval < 20 21-34 . 35+ 17.4 69 ( 8.5, 26.4) 8.3 836 ( 6.4, 10.2) 21.0 105 (13.2, 28.8) Number of prior pregnancies 0 10.1 298 ( 6.7, 13.5) 1 8.7 355 ( 5.8, 11.6) 2 8.5 200 ( 4.6, 12.4) 3 or more 16.0 156 (10.3, 21.8) Number of previous spontaneous abortions r stillbirths 0 9.3 829 ( 7.3, 11.3) 1 12.3 138 ( 6.8, 17.8) 2 or more 21.0 43 ( 8.8, 33.2) a # spontaneous abortions + # live births - # birth defects among a single random eligible pregnancy per woman (overall SAB rate - 10.2%). NC confidence intervals not calculated if rates based on fewer than five pregnancies. 2.45 SL 033280 Table 2.12 (Continued) Month prenatal care began 1 2 3 4+ None Number of prior live births 0 1 2 3+ IUD use at time f pregnancy Yes No First trimester: Diabetes Yes No SAB Rfrt? (%) 14.4 7.5 4.1 0.0 100.0 10.1 9.7 10.0 13.8 29.4 9.9 0.0 10.3 N (Pregnancies)8 313 464 172 41 15 368 402 160 80 17 986 8 998 95% Confidence Interval (10.5, 18.3) ( 5.1, 9.9) ( 1.1, 7.1) NA NA ( 7.0, 13.2) ( 6.8, 12.6) ( 5.4, 14.7) ( 6.2. 21.4) (10.5, 54.4) ( 8.0, 11.8) ( 0.0, 28.2) ( 8,4, 12.2) 2.47 SL 033281 Table 2.12 (C ntinued) Variable Hypertension Yes No SAB Rate 7.1 10.3 Fever Yes No 9.1 10.4 Bleeding Yes No Naus a Yes No 31.8 8.6 7.3 14.4 Nausea medication use among women who reported first trimester nausea Yes No 1.3 8.1 Smoking first trimester None < 1/2 pk/day 1+ pk/day 9.4 12.9 16.0 1591.1 N (Pregnancies't* 28 978 33 . 942 85 924 607 396 95% Confidence Interval < 0.9, 23.3) ( 8.4, 12.2) ( 0.0, 18.9) ( 8.5, 12.4) (21.9, 41.7) ( 6.8, 10.4) ( 5.2, 9.4) (10.9, 17.9) 75 ( 0.0, 3.9) 530 ( 5.8, 10.4) 862 ( 7.5, 11.4) 93 ( 6.1, 19.7) 50 ( 5.8, 26.2) 2.48 SL 033282 Table 2.12 (Continued) Variable First trimester alcohol use SAB Rate f%) None <4 drinks/vk 4+ drinks/wk 9.9 9.9 19.4 First trimester prescription medication use None Prenatal vitamins Antibiotics Antihistamines Estrogens Corticosteroids Anti-nausea Thyr id medication Non-narcotic pain medication Narcotics Diabetes medication Hypertension medication 11.8 3.1 3.7 9.1 0.0 50.0 0.0 7.1 11.1 11.1 0.0 0.0 N (Pregnancies)3 669 305 31 720 32 27 11 7 2 35 14 9 9 4 3 95% Confidence lnsrvgl ( 7.6. 12.2) ( 6.6. 13.3) ( 5.5. 33.3) ( 9.4. 14.2) ( 0.0. 9.1) ( 0.1, 18.9) ( 0.2, 40.5) ( 0.0, 32.2) NC ( 0.0, 8.1) ( 0.2, 31.8) ( 0.3, 39.8) ( 0.3, 39.8) NC NC 1591.1 2.49 SL 033283 Table 2.12 (Continued) Harlablt sab Rate (%> Any employment during first trimester Yes No First trimester job involved: 10.3 10.2 Probable solvent exposure Possible solvent exp sure Non*sedentary work Sedentary work 25.0 6.3 11.9 13.5 Industry Electronics 13.1 Hospital 11.1 Doctor's Office/Clinic 15.0 Child Care 11.1 Occupation Electronics Hospital Child Care 11.1 14.5 15.6 N (Pregnancies^8 638 363 20 16 42 96 175 63 20 45 27 62 32 95% Confidence Interval < 7.9, 12.7) ( 7.1, 13.3) < 6.8, 40.7) ( 0.2, 30.0) ( 2.1, 21.7) ( 6.7, 20.3) ( 8.1, 18.1) ( 3.3, 18.9) ( 3.2, 37.9) ( 1.9, 20.3) ( 2.4, 28.9) ( 5.7, 23.3) ( 3.0, 28.2) 1591.1 2.50 Si 33284 Table 2.12 (Continued) Variable SAB Rate (%) Any first trimester shift work Yes No shiftwork 14.8 9.4 Type of first trimester shift work Evenings Nights Rotate two or three shifts Other 14.8 10.0 15.8 15.2 Father worked three months bef re pregnancy Yes 9.9 No 17.6 Father's occupation Electronics worker Health care worker 20.0 21.4 N (Preenancies)a 115 523 61 10 19 33 971 34 15 14 95% Confidence Interval ( 8.3, 21.3) ( 6.9, 11.9) ( 5.9, 23.7) ( 0.3, 44.5) ( 5.5, 60.9) ( 3.0, 27.5) ( 8.0, 11.8) ( 4.8, 30.4) ( 4.3, 48.1) ( 4.8, 46.6) 2.51 SL 033285 Table 2.12 (Continued) y.ftriafrls SAB Rate (%> First trimester chemical exposure at work or home or in other activity None reported Solvents Gasoline Pesticides 10.3 17.1 5.6 13.0 Pest control work done at residence during first trimester Yes 10.0 No 10.1 Exposure during three months before pregnancy began or first trimester to: Video display terminals Yes No Chest x-ray Yes No 11.3 10.0 15.8 10.0 N (Pregnancies)* 747 76 71 23 40 943 204 803 19 974 95% Confidence Internal ( 8.1. 12.5) ( 8.6. 25.6) ( 0.3. 11.0) ( 2.8. 33.0) ( 0.7. 19.3) ( 8.2, 12.0) ( 7.0. 15.6) < 7.9, 12.1) ( 3.4. 38.0) ( 8.1, 11.9) 2.52 SL 033286 Table 2.12 (Continued) Variable Pelvic x-ray Yes No SAB Rate (*) 5.6 10.3 Other x-ray Yes No Used water from a private well 6.2 10.3 Yes 20.0 No 10.1 Water filter use around time pregnancy began Yes 4.5 No 10.7 Mother's ethnicity White Hispanic Asian Black American Indian Other 11.5 10.9 4.2 0.0 0.0 0.0 N Pregnancies )* 18 984 81 915 5 994 67 926 755 101 118 20 4 9 95% Confidence Interval ( 0.1, 26.7) ( 8.4, 12.2) ( 1.0, 11.5) ( 8.3, 12.3) < 0.5, 71.6) ( 8.2, 12.0) ( o.o, 9.5) ( 8.7, 12.7) ( 9.2, 13.8) ( 4.8, 17.0) ( 0.6, 7.8) ( 0.0, 13.9) NC ( 0.0, 25.0) 1591.1 2.53 SL 033287 Table 2,12 (Continued) Satiable SAB Rate r%) Mother's education 0-8 yra 9-11 yrs 12 (High School grad) 13-15 16+ (college grad or more) 20.0 10.6 9.1 9.5 11.8 Father's ethnicity Vhite Hispanic Asian Black Am rican Indian Other 11.0 10.6 4.5 4.2 0.0 17.7 Father's education 0-8 yrs 9-11 yrs 12 (High School grad) 13-15 16+ (college grad or more) 16.7 10.6 11.3 10.0 9.6 N (Pregnancies 10 47 309 347 289 737 113 111 24 3 17 12 47 212 321 405 95% Confidence Interval ( 2.5, 55.6) ( 1.8, 19.4) ( 5.9, 12.3) ( 6.4, 12.6) ( 8.1, 15.5) ( 8.7, 13.3) ( 4.9, 16.3) ( 0.6, 8.4) ( 0.1, 20.7) NC ( 3.8, 42.4) ( 2.1, 45.9) ( 1.8, 19.4) ( 7.0, 15.6) ( 6.7, 13.3) ( 6.7, 12.5) 1591.1 2,54 SL 033288 Table 2.12 (Continued) Variable First trimester cold home tap water consumption SAB Rate (*) Yes 12.0 No 3.3 First trimester cold tap water consumption home or work Yes 11.4 No 4.1 Bottled water use at home anytime during pregnancy Y s 4.8 N 14.2 Bottled water use at home or work anytime during pregnancy Yes 5.6 No 14.6 N (Pregnancies')* 775 213 837 147 418 586 480 520 95% Confidence Interval ( 9.7, 14.3) ( 0.9, 5.7) ( 9.3, 13.6) ( 0.9, 7.3) ( 2.8, 6.9) (11.4, 17.0) ( 3.5, 7.7) (11.6, 17.6) 2.55 SV Table 2.13 a Gestational Weeks of Spontaneous Abortions by Census Tracts and Tine Periods CENSUS TRACT EXPOSED UNEXPOSED TU1E.FERIQB------- ________________ 513Lllffl)_____yBgyXUr5--1,2.0...08(01 1980 - 1981 5053.0204(N) BOTH ----TIVOTMALL mean + std err (N) 9.8 9 (13) 9.8 1.7 ( 4) 9.8 12.6 12.8 12.7 10.9 + 2.8 2.2 ( 5) ( 6) 1982-1985 mean + std err (N) 12.1 -9 (22) 12.3 8 (20) 12.2 9.8 9 (23) 10.4 9 (10) 10.0 11.2 SL 033290 MEANS 11.2 11.9 10.3 11.3 Among a single randon eligible pregnancy per woman. 0 - Original study area N - New study area P values for comparisons of average gestational weeks are 1) SABs conceived in 1980 or 1981 compared to 1982 - 1985 (p-.80); 2) SABs conceived in 5120.12 versus 5120.11 versus 5120.08 versus 5053.0204 (p-.60); 3) SABs conceived in 1980 or 1981 in 5120.12 or 5120.11 versus 1980 or 1981 in 5120.08 or 5053.0204 versus 1982-85 in 5120.12 or 5120.11 versus 1982-1985 in 5120.08 or 5053.0204 (p-.02) Table 2.14 Multiple Logistic Model of Risk of Spontaneous Abortion Mother's Age <20 20 - 34 35+ Prior spontaneous fetal loss 0 1 2+ Ethnicity White or Hispanic Other Ale h 1 consumption < 4 drink/week 4 + drinks/week Smoking None 1/2 pack/day 1 or more packs/day Pr bably solvent exposed job No Yes H me cold tap water and water filter use N tap water, no filter No tap water, filter used Tap water, no filter Tap water, filter CB 2.6 1.0 2.6 1.0 1.6 2.2 1.0 0.3 1.0 1.4 1.0 1.1 1.2 1.0 4.8 1.0 4.2 5.5 1.6 S5% Cl (1.3, 5.6) (1.0, 6.3) (0.9, 3.0) (0.9, 5.3) (0.1, 0.8) (0.4, 4.4) (0.5, 2.2) (0.4, 3.1) 1 0.001 0.10 0.01 0.60 0.94 (1.5, 15.5) 0.008 (0.4, 44.3) (2.1, 14.0) (0.3, 8.6) 0.002 2.57 SL 033291 Table 2.14 (Continued) Census tract & time period Original exposed 1980-81 Original exposed 1982-85 New exposed 1980-81 New exposed 1982-85 Original control 1980-81 Original control 1982-85 New control 1980-81 New control 1982-85 Original areas Exposed/unexposed 1980-81 Exp sed/unexposed 1982-85 02. 1.0 0.9 0.2 0.5 0.2 0.7 0.9 0.4 5.0 1.2 95% Cl (0.4, 1.9) (0.04, 0.6) (0.2. 1.2) (0.06, 0.7) (0.3, 1.6) (0.3, 2.7) (0.1. 1.0) New areas Exposed/unexposed 1980-81 Exposed/unexposed 1982-85 0.2 1.4 2.58 033292 SL Table 2.15 Validation of Reported Birth Defects Validation Categories: California Birth Defects Monitoring Program (CBDMP) Reportable Defect (Malformation) (Deformation) CBDMP Non-Reportable Defect Not a defect Not validatable NfDefects) % Distribution 37 (36) ( 1) 33 45 1 32 28 39 1 TOTAL 116 100 1598-1.2 2.59 St 33 2g3 TabU 2.16 Congenital Anomalies Reported by Respondent and Validated by Medical Record or Physician's Report Old Unexposed Census Tract (5120.08) Malformation Month and Year of Birth One set of toes shorter, toes aren't separated on that foot 01/80 Coloboma of right eye 10/80 Hypospadias 10/80 Low set nasal bridge, tricuspid regurgitation pat nt foramen ovale 03/82 Pyloric stenosis 07/83 Persistent hypoplastic vitrins with overt microphalmic of right eye 07/83 Pyloric stenosis 04/84 2.60 033^ Table 2.16 (Continued) New Exp sed Census Tract (5120.11) Malformation No palpable testes Accessory left nipple Pectus excavatum D uble collecting system of left kidney Synostosis Pyl ric stenosis C ngenital abscense of fibula Axial congenital lens opacity at anterior subcapsular region Branchial cleft cyst Bony abnormality of medial side of right foot Deformation Facial asymetry Month and Year of Birth 03/80 09/80 07/82 09/82 01/83 03/83 08/83 10/84 11/84 02/85 Month and Year of Birth 02/85 2.61 Si 33295 Table 2.16 (Conclnued) Old Exposed Census Trace (5120.12) Malformation Month and Year of Birch Midline fistula 05/80 Tw toes on both feet webbed 07/80 Cerebal palsy with hypotonia, significantdevelopmental delay, persistent microcephaly, omphalocele, congenital club foot with deformity of right side 07/80 Talipls equinovous appearance, ' asymetrical chest wall, elongated ears, pectus carlnatum 08/80 Tricuspid atresia, atrial septal defect with right to left shunt and CHF, ventricular septal defect, asymetrical simian crease on left hand 02/81 Bilateral hypoplasia of lung, hypoplasia of left kidney, hypoplasia of left testis, polycystic kidney, absent anus 03/81 Duplication of ureters 09/81 Cleft lip 09/81 Undescended testicle 12/81 Cleft palate, Robin anomaly 12/81 Congenital anal agenesis 11/83 Hypospadias 02/84 Pyloric stenosis 05/84 Two sets of webbed toes, an extra toe on left foot 2.62 03/85 0336 si* Table 2.16 (Continued) Malformation Month and Year of Birth Total anomalous pulmonary venus connection to right atrium 02/85 Deformation Month and Year of Birth Clubbed foot 05/80 New TJnexposed Census Tract (5053.0204) Malformation Month and Year of Birth Pectus excavatum 03/81 Hypospadias (primary) 09/82 Cleft lip 05/83 Left hydronephroses, left hydroureter,ureterovesico junction obstruct, left ureteral achalasia 05/84 Deformation Month and Year of Birth C ng nltal dislocation of right hip 10/82 2.63 SL 033297 Table 2.17 Number of Reportable Malformations Among Singleton Live Births by Census Tract and Year CENSUS TRACT EXPOSED UNEXPOSED IME-IERIQP ______________ 5120.12(0) 5120.ll(N) Exp sed: (BOTH) 5120.08(0) 5053.0204(H) BOTH TOTAL 1980 1981 (1980-81) 4 6 (10) 2 0 ( 2) ( 6) ( 6) (12) 3 0 ( 3) 0 1 ( 1) ( 3) ( 1) ( 4) 9 7 (16) Unexposed: 1982 1983 1984 198S (1982-85) 0 1 2 2 ( 5) 2 3 2 1 ( 8) ( 2) ( 4) ( 4) ( 3) (13) 1 2 1 0 ( 4) 1 1 1 0 ( 3) ( 2) ( 3) ( 2) ( 0) ( 7) 4 7 6 3 (20) SL 033298 TOTAL 15 10 (25) 7 4 (11) 36 0 -- Original study area N - New study area 6Z TIME PERIOD Exposed: 1960 1981 (1980-81) Table 2.18 Number of Singleton Live Births by Census Tract and Year CENSUS TRACT EXPOSED UNEXPOSED ________5120,12*01. 5120.11(H) (BOTH) 5120.0810) 5053.0204INI BOTH 52 44 ( 96) 45 60 (105) ( 97) (104) (201) 62 63 (125) 25 28 ( 53) ( 87) ( 91) (178) Unexposed: 1982 1983 1984 1985 (1982-85) 59 56 62 25 (202) 69 62 97 17 (241) (124) (118) (159) ( 42) (443) 42 71 70 26 (209) 45 48 57 14 (164) ( 87) (119) (127) ( 40) (373) TOTAL 298 346 (644) 334 217 . (551) O - Original study area N -- New study area TOTAL 184 195 (379) 211 237 286 82 (816) 1195 033300 Table 2.19 Reportable Halformation Rates by Census Tract and Year CENSUS TRACT EXPOSED UNEXPOSED I1HE PEBIQD------------------------- 5l2fi.l2fO) 5120.11(H)(BOTH) Exposed: 5120.08(0) 5053.0204(N) (BOTH) 1980 1981 (1980-81) 7.7 13.6 (10.4) 4.4 0.0 (19) (6.2) (5.8) (6.0) 4.8 0.0 (2.4) 0.0 3.6 (1.9) (3.4) (1.1) (2.2) Unexposed: 1982 1983 1984 1985 (1982-85) 0.0 1.8 3.2 8.0 (2.5) 3.1 4.8 2.1 5.9 (3.3) (16) (3.4) (2.5) 7.1 (2.9) 2.4 2.8 1.4 0.0 (19) 2.2 2.1 1.8 0.0 (1.8) (2.3) (2.5) (1.6) (0.0) (1.9) TOTAL 5.0 2.9 (3.9) 2.1 1.8 (2.0) 0 - Original study area N - New study area TOTAL 4.9 3.6 (4.2) 1.9 3.0 2.1 3.7 (2.5) 3.0 Table 2.20 g Odds Ratios and 951 Confidence Intervals for Reportable Malformations in Exposed vs. Unexposed Areas by Tine Period ORIGINAL NEU STUDY AREAS STUDY AREAS (5120.12/5120.08) (5120.11/5053.0204) _______________________fiR_____________ 2Al_I_______________ Qg_____________ 95% 1 Tine period 1960-81 4.3 (1.2, 14.7) 0.9 (0.1, 6.6) 1982-85 1.2 (0.4, 4.5) 1.7 (0.5, 5.9) g Snail sanple adjustment irz Aow Table 2.21 Congenital Anomaly Rates for Live Births by Condition: Residents of Santa Clara County Study Areas (1980-1985) (Rates per 1,000 Live Births) (ICD-9-CM) Condition(Code)__ Number Anencephaly (no brain) (740) Spina bifida (741) Hicrocephalus (small head) (742.1) Hydrocephalus (742.3) Anophthalmia (743,0) 0 0 1 0 0 Microphthalmia (small eyes) (743.1) 1 Glaucoma (743.2) 0 ngenital cataract (743.30*34) 1 Ear defects v/deafness (744.0, 744.23) 0 (includes microtia) Singl ventricle (heart) (745.3) 0 Endocardial cushion defect (745.6) 0 Pulmonic valve atresia (746.0) 0 Pulmonic valve stenosis /insufficiency(746.01,746.02) Aortic valve stenosis /insufficiency(746.3,746.4) Mitral valve stenosis /insufficiency (746.5,746.4) 0 0 0 0 0 Common truncus (745.0) 0 Transposition of great vessels (745.1) 0 S.F. Study Area Bay Area Rate^ 0.2 -- 0.4 0.8 0.9 -- 0.4 -- 0.1 0.8 0.2 -- 0.1 0.8 0.2 -- 0.3 -- 0.1 -- 0.4 -- 0.3 --- * * - 0.2 -*"*- 0.1 -- 0.1 -- 0.2 0.7 Atlanta RateC 0.5 0.7 0.6 1.2 0.1 0.1 0.1 0.3 0.2 0.1 0.4 0.1 0.2 0.3 0.2 0.1 0.4 si. 33302 2.68 Table 2.21 (Continued) - (ICD-9-CM) Condition(Code)___________ Kuraber Tetralogy of fallot (745.2) 0 Hyp plastic left heart (746.7) 0 Congenital heart block (746.86) 0 Coarctation of the aorta (747.1) 0 Anomalous pulmonary venous return (747.41,747.42) Ch anal atresia (748.0) 1 0 Anomalies of larynx, trachea, bronchus, (748.2,748.3) 0 Absent,small,cystic lung (748.4,748.5) 1 Cleft palate alone (749.0) 0 Cleft lip alone (749.1) 2 Cleft lip and cleft palate (749.2) 0 T-E Fistula, atresia, stenosis (750.3) 0 Pyloric stenosis (750.5) 4 Atresia/stenosls-small intestine (751.1) 0 Atresia/stenosislarge intestine, anus (751.2) 2 Hirschsprung's disease (751.3) 1 Biliary atresia (751.61) 0 Study Area Rate3 S.F. Bay Area Rate*5 0.4 -- 0.4 -- 0.1 -- 0.5 0.8 0.2 -- 0.1 - * 0.1 0.8 0.5 -- 0.8 1.7 0.5 -- 0.5 -- 0.3 3.3 2.3 . 0.1 1.7 0.4 0.8 0.1 * . 0.1 Atlanta Rate 0.4 0.3 <0.1 0.4 0.1 0.1 0.1 0.7 0.5 0.4 0.7 0.3 1.2 0.3 0.4 0.2 0.1 824.tbx/3 2.69 SL 033303 Table 2.21 (Continued) (ICD-9-CM) C9ndlt?afcode) Number Hypospadias/epispadias (752.6) 3 Indeterminate sex (752.7) 0 Renal agenesis/dysgenesis (753.0) 1 Cystic kidney disease (753.1) 1 Exstrophy of urinary bladder (753.5) 0 Sc liosis, lorsosis of spine (754.2) 0 Congenital dislocation of hip (754.30, 754.31, 754.35) 1 Clubfoot (754.50, 754.51, 754.53, 754.60, 754.73) 2 Extra thumb, great toe-preaxlal Polydactyly (755.01, 755.03) 0 Syndactyly (webbing digits) (755.1) 2 Ch ndrodystrophies (756.4) 0 Osteodystrophies (756.5) 1 Diaphragmatic hernia (756.6) 0 Omphalocele/gastrosehisis (756.70,756.71) 1 S.F. Study Area Bay Area Rate* Rate0 2.5 3.0 -- 0.2 0.8 0.4 0.8 0.3 -- 0.1 -- 0.6 0.8 1.2 1.7 1.7 -- 0.3 1.7 0.8 - * - 0.1 0.8 0.2 -- 0.3 0.8 0.2 Atlanta RateC 2.9 0.2 0.4 0.3 <0.1 0.1 0.7 2.3 0.2 0.9 0.1 0.1 0.3 0.6 824.tbx/3 2.70 OSS304 Table 2.21 (Continued) (ICD-9-CM) Conditionlode) Down's syndrome (758.0) Other autosomal anomalies (758.1, 758.2, 758.3, 758.5) Sex chromosome anomalies (758.5, 758.7, 758.8) Congenital viral infection (771.0, 771.1, 771.2) Number 0 0 0 0 Study Area Rate* S.F. Bay Area Ratfck 0.9 . - * 0.3 . - * 0.1 0.3 Atlanta RateC 1.1 0.4 0.1 0.1 *Total number of live births in study area b Five Sen Francisco Bay Area counties January 1, 1983-June 30, 1983. 1196. C1979-1983 Source: Report of the California Birth Defects Monitoring Program, Congenital Malformations Among Live Births, January 1, 1983, - June 30, 1983, Berkeley. California. 824.3 2.71 SL 033305 Congenital HaUpraatlpn (ICD-9-CH) Pyloric stenosis (750.5) case #1 case #2 case #3 case #4 Table 2.22 Listing of Congenital Malformations OccurIng In More Than Two Children and Maternal Risk Factors Exposed census tracts Age of other at birth < 20 __or > 35 Chemical* exposure at work or home Hone pestcontrol Use of a tap water filter VDTa U no no no yes no no no no yes no yes yes yes no no yes no no no no yes no yes no Cigarette SB9llRS yes yes no no Anomalies of genitourinary system (751.2, 751.3) case #1 yes no no yes no no no case #2 yes no no no yes no no case #3 no no no no no yes yes Hypospadias (752.6) case #1 case #2 case #3 no yes no no no no no no no no no no no yes no no no no no yes no SL 033306 Syndactyly and other malformations of feet (754.7) case #1 no no no yes no no yes case #2 yes yes no no no no no case #3 yes no no no no no no case #4 yes yes no no no no no None of these women reported the following possible risk factors: diabetes, hypertension, eclampsia, seizures, fever, exposure to X-rays the three months prior to or during pregnancy, or a private well as their source of drinking water. /,0t 0 U Z` Congenital Halfaimatlgn (ICD-9-CH) Alcohol* sapawaptien Pyloric stenosis (750.5) case #1 yes case #2 no case #3 no case #4 no Prescription* medication other than prenatal vitamins no no yes no Anomalies of genitourinary system (751.2, 751.3) cas #1 yes case #2 no 1 case #3 yes Hypospadias (752.6) case #1 case #2 case #3 no no yes Syndactyly and other malformations of feet (754.7) case #1 no case #2 no case #3 no case #4 no First trimester exposure no yes no no no no yes no no no Table 2.22 (Continued) Nausea' no yes yes yes Amniocentesis no no yes no yes no no no no no no no yes no no no yes no no no no no yes no Sliding no no no no no no yes no no no no no no no C-sectlon no no yes yes Family history of congenital malformations no no no no yes no no no yes no no no no no no no no no no yes yes no no no Table 2.23 Number of Low Birth Weight Babies among All Births Reported on Birth Certificates CENSUS TRACT EXPOSED UNEXPOSED IIMLmiQB----- 5JL2Q. 12fQ)------aMLlUM)____ LfiQIii)512Q.QB(0) Exposed: 5053.0204(H) (BOTH) TOTAL 1980 1981 (1980-81) 4 2 (6) 5 ( 9) 3 ( 5) (8) (14) 4 8 (12) 2 ( 6) 15 1 ( 9) 14 (3) (15) (29) Unexposed: 1982 r\j 1983 v4*i 1984 1985 (1982-85) 3 4 3 2 (12) 8 9 4 9 (30) (ID <U) ( 7) (11) (42) 7 6 6 8 (27) 6 6 4 2 (18) (13) (12) (10) (10) (45) 24 25 17 21 (87) TOTAL 18 38 (56) 39 21 (60) 116 SL 033308 0 - Original study area N - New study area Table 2.24 Number of Singleton Live Birth Among All Births Reported on Birth Certificates CENSUS TRACT EXPOSED UNEXPOSED IIHE PERIPP-------SLEfl-UfO)5120.11(N1 Exposed: (BOTH)5120.08(01 5053.0204(N1 (BOTH! TOTAL 1980 1981 (1980-81) 146 133 (281) 164 130 (294) (312) (263) (575) 188 184 (372) 115 89 (204) (303) 615 (273) 536 (576) (1151) Unexposed: N 1982 m 1983 1984 1985 (1982-85) 107 131 121 117 (476) 179 130 151 168 (628) (286) (261) (272) (285) (1104) 169 172 149 174 (664) 101 119 122 119 (461) (270) 556 (291) 552 (271) 543 (293) 578 (1125) (2229) TOTAL 757 922 (1679) 1036 665 (1701) 3380 0 -- Original study area N - New study area SL 033309 1597.tb3/9 TIME PERIOD Exposed: 1980 1981 (1980-81) Table 2.25 Low Birth Height Rates Aaong All Births Reported on Birth Certificates CENSUS TRACT EXPOSED UNEXPOSED 5120.12(0) 5120.11(N) (BOTH) 5120.08(0) 5053.0204(H) (BOTH) TOTAL 2.7 1.5 (2.1) 3.1 2.3 <2.7) (2.9) (1.9) (2.4) 2.1 4.4 (3.2) 1.7 1.1 (1.5) (2.0) 2.4 (3.3) 2.6 (2.6) (2.5) Unexposed: 1982 1983 1984 1985 (1982-85) 2.8 3.1 2.5 1.7 (2.5) 4.5 6.9 2.7 5.4 (4.8) (3.8) (4.9) (2.6) (3.9) (3.8) 4.1 3.5 4.0 4.6 (4.1) 5.9 5.0 3.3 1.7 (3.9) (4.8) 4.3 (4.1) 4.5 (3.7) 3.1 (3.4) 2.4 (4.0) (3.9) TOTAL 2.4 4.1 (3.3) 3.8 3.2 (3.5) 3.4 SL 033310 0 - Original study area N - New study area 1597.tbl/9 Table 2.26 Odd* Ratios and 95% Confidence Intervals for Low Birth Weight in Exposed vs. Unexposed Areas by Tine Period Among All Births Reported on Birth Certificates Tine period ORIGINAL STUDY AREAS (5120.12/5120.08) _________Q&__________ S51 Cl________ NEW STUDY AREAS (5120.11/5053.0204) OR 95% Cl 1980-81 1982-85 0.7 <0.2, 1.8) 0.6 (0.3, 1.2) 1.7* 1.2 (0.5, 6.0) (0.7, 2.2) te t aSnail sanple adjustment Table 2.27 Number of Low Birth Weight Babies by Census Tract and Year Among Interviewed Women9 CENSUS TRACT EXPOSED UNEXPOSED HUE EERIQP____ 5120.12101____ 5120.llfN)_____(BOTH)_________ 5120.08(0) Exposed: 5053.02Q4(N> (BOTH) TOTAL 1980 1981 (1980-81) 0 1 (1) l <1> 3 (4) (4) (5) 3 0 (3) 4 3 0 (3) 7 (6) (0) (6) (ID Unexposed: 1982 1983 1984 1985 (1982-85) 1 1 1 0 (3) 1 (2) 1 (2) 4 (5) l <1> (7) (10) 0 2 ( 2) 4 2 2 ( 4) 6 3 1 < 4) 9 0 0 ( 0) 1 (5) (5) (10) (20) TOTAL 4 11 (15) 11 5 (16) 31 SL 033312 Based on a single random eligible pregnancy per woman. 0 - Original study area N - New study area 1597.tb7/9 uno rMuuti Exposed: 1980 1981 (1980-81) Table 2.28 Number of Singleton Live Births With Known Birth Height by Census Tract and Year Among Interviewed Women* CENSUS TRACT EXPOSED UNEXPOSED ____ 214V. !_____tgmm ____ ?uv.wm ____ 5P53.020411!)__ (BOTH) -TOTAL 42 32 ( 74) 33 48 ( 81) ( 75) ( 80) (155) 52 51 (103) 16 21 ( 37) ( 68) ( 72) (140) 143 152 (295) Unexposed: * 1982 43 VO 1983 42 1984 52 1985 16 (1982-85) (153) 54 47 77 14 (192) ( 97) ( 89) (129) ( 30) (345) 32 57 55 18 (162) 30 37 47 10 (124) ( 62) 159 ( 94) 183 (102) 231 ( 28) 58 (286) (631) TOTAL 227 273 (500) 265 161 (426) 926 SL 0333X3 Based on a single random eligible pregnancy per woman. 0 - Original study area N - New study area Table 2.29 TIME PERIOD Exposed: 19B0 1981 (1980-81) Low Birth Height Rates by Census Tract and Year Among Interviewed Women8 CENSUS TRACT EXPOSED UNEXPOSED 5120.12(0) 5120.1KN) __ (BOTH)____ ^12(LM(0) 5053.0204(N) (BOTH) TOTAL 0.0 3.1 (1.4) 3.0 6.3 (4.9) (1.3) (5.0) (3.2) 5.8 5.9 (5.8) 0.0 0.0 (0.0) (4.4) (4.2) (4.3) 2.8 4.6 (3.7) Unexposed: 1982 1983 1984 1985 (1982-85) 2.3 2.4 1.9 0.0 (2.0) 1.9 2.1 5.2 7.1 (3.7) (2.1) (2.3) (3.9) (3.3) (2.9) 0.0 3.5 5.5 0.0 (3.1) 6.7 5.4 2.1 0.0 (4.0) (3.2) 2.5 (4.3) 3.3 (3.9) 3.9 (0.0) 1.7 (3.5) (3.2) TOTAL 1.8 4.0 (3.0) 4.2 3.1 (3.8) 3.3 SL 033314 Based on a single random eligible pregnancy per women. 0 - Original study area N - New study area Table 2.30 Odds Ratios* and 95% Confidence Intervals for Low Birth Weight in Exposed vs. Unexposed Areas by Time Period Among Interviewed Women Time period ORIGINAL STUDY AREAS (5120.12/5120.08) _________OR__________ 95% Cl________ NEW STUDY AREAS (5120.11/5053.0206) OR___ 95% Cl 1980-81 0.3 (0.05, 1.9) 6.6 (0.2, 83.0) 1982-85 0.7 (0.2, 2.6) 0.9 (0.3, 2.9) - t/i Source: SCC12 t4 0 Small sample adjustment 5t t 0 1597.tb5/9 Table 2.31A Comparisons of Mean Estimated Cumulative TCA (ppb) Exposure Among Women with Pregnanciesa Ending in Live Birth and Those Ending in Spontaneous Abortion All PregnaneUg Per VPfflflll Random Eligible ?reenancvb 1st month of 1st, 2nd 1st, 2nd & 11 BifiEMncy month 3cd gnth 1st month of 1st, 2nd H Biggnancy month Live Births All Reported SABs (mean square error) 104 25 1052 1057 (1296212) 2119 1960 (4904497) 3202 2749 (9862612) 80 904 16 1002 (1428763) 1809 2063 (5580761) p(ANOVA) p(UlLCOXON) .98 .75 .52 .60 .56 .43 .76 .70 .23 .32 D finite SAB's (mean square error) p(ANOVA) p(WILCOXON) 20 824 (1196151) .39 .27 1681 (4784766) .41 .36 2469 (9660986) .33 .35 12 928 (1467716) .95 .47 1996 (5742787) .80 .48 ro a m Includes pregnancies with estimated dates of conception in 1981 in census tracts S120.ll and 5120.12. 1st, 2nd & 2Ed wonth 2797 3098 (11791667) .75 .33 3044 (12100605) .82 .48 Zero exposure for women reporting water filter use or no cold home tap water consumption. SL 033316 1595.tbl/3 Table 2.31B Percent Distributions of Spontaneous Abortions3 And Live Births^ by Quartiles of Estimated TCA (ppb) Exposure CT 5120.11 TCA (ppb) 1st month 0- 199 200- 899 900-2199 2206+ TOTAL Definite SABs No. (%) 1 (6) 3 (18) 1 (6) 1 (6) 6 Live births No. (%) 16 (94) 14 (82) 15 (94) 15 (94) 60 All SABs No. (%) 2 (ID 3 (18) 1 (6) 4 (21) 10 Live births No. 16 (89) 14 (82) 15 (94) 15 (79) 60 CT 5120.12 TCA (ppb) 1st month Definite SABs No. _ (1) 0-42.99 5 (42) 43-399.99 3 (23) 400-1259.99 2 (17) 1260+ 4 (27) TOTAL 14 Live births No. f> 7 (58) 10 (77) 10 (83) 11 (73) 38 All SABs No. (%> 5 (42) 4 (28) 2 (16) 4 (27) 15 *Uses all eligible pregnancies per woman in 1981. b Live births minus six Infants with birth defects. Live births No. 7 (58) 10 (71) 10 (83) 11 (73) 38 2.83 SL 03331V Table 2.32 Comparisons of Mean Estimated Probability of Exposure to tfater from Hell #13 Aaong Uomen with Pregnancies* Ending in Live Birth and Those Ending In Spontaneous Abortion (SAB) u Live Births All R ported SABs (mean square error) p(ANOVA) p(WILCOXON) 104 25 vs. Definite SAB's (mean square error) p(ANOVA) p(UILCOXON) 20 All Pregnancies Per Woman 1st Month of Preanancv Maximum of 1st, 2nd Month Maximum of 1st, 2nd & 3rd Month .40 .42 (06) .67 .83 .38 (.06) .81 .63 .45 .46 (06) .84 .83 .44 (.06) .77 .69 .48 .47 (.06) .90 .98 .45 (.06) .59 .54 H 80 16 12 Random Eligible Pregnancy*1 1st month of frswancy Maximum of 1st, 2nd Month Maximum of 1st, 2nd & 3rd Manlh .37 .42 (08) .52 .38 .40 (.09) .72 .59 .41 .47 (.08) .51 .43 .46 (.09) .60 .49 .45 .49 (.09) .59 .70 .49 (.09) .64 .70 a*t*e o Includes pregnancies with estlnated dates of conception in 1981 In census tracts 5120.11 and 5120.12. Zero exposure for wonen reporting water filter use or no cold hone tap water. Table 2.33 CoaparIsons of Hean Estivated Cumulative Probability of Ho Exposure to Water froa Well #13 Aeong Wooten with Pregnancies Ending in' Live Birth and Those Ending in Spontaneous Abortion (SAB)a Ail fKCTfflClgg Per Woven Random Eligible Pregnancy1* Live Births All reported SABs (vean square error), p (ANOVA) p (WILCOXON) H 104 25 1st vonth of pregnancy .80 .58 (.06) .67 .84 1st, 2nd vonth .43 1st, 2nd 6 3rd vonth .33 .44 (.07) .87 .76 .36 (08) .66 .50 1st vonth of 1st, 2nd 1st, 2nd & H pregnancy Bsnsh 3rd vonth 80 .63 .50 .41 16 .58 (.08) .52 .38 .42 (ID .39 .47 .31 (.14) .29 .59 SL 033319 e u* Definite SABs 20 .60 .45 .36 12 .60 .43 (vean square error) p (ANOVA) p (WILCOXON) (.06) .81 .63 (.07) .68 .46 (08) .67 .44 (-09) 72 .59 . (11) .53 .56 g Includes pregnancies with estivated dates of conception in 1981 in census tracts 3120.11 and 5120.12. Zero exposure for woven reporting water filter use or no cold hove tap water consumption. .31 (14) .39 .61 Table 2.34 Comparisons of Mean Cumulative TCA (ppb) Exposure to Vomen with Pregnancies* Resulting in Normal Live Births or in Live Births with Malformation Live Births with No Defects 1st 1st, 2nd 1st, 2nd & 3rd H month month month 98 1091 2204 3334 Live Births with CBDMPb Reportable Malformations 6 419 758 1063 (mean square error) p(ANOVA) p(WILCOXON) (1206735) .15 .22 (4855557) .11 .09 (10138658) .09 .04 Average Estimated Maximum Probability of Exposure to Vater from Veil #13 Live Births No Defects Live Births with CBDMpk Reportable Malf rmations (mean square error) p(ANOVA) p(WILCOXON) 98 6 1st month .41 .17 (.06) .01 .01 1st & 2nd month .47 1st, 2nd & 3rd month .49 .23 (.05) .02 .04 .28 (.06) .03 .07 Average Estimated Cumulative Probability of No Exposure to Vater from Veil #13 1st month 1st & 2nd month 1st, 2nd & 3rd Live Births No Defects 98 .59 .41 .31 Live Births with CBDMpk Reportable Malformations 6 .83 .71 .62 (mean square error) p(ANOVA) p(WILCOXON) (.06) .01 .01 (.07) .008 .03 (.08) .009 .02 Includes all reported live births conceived in 1981 in census tracts 5120.11 and 5120.12. California Birth D fects Monitoring Program 1595.tb2/3 2.86 SL 033320 Appendix A. Letters, Forms, and Questionnaire Enumeration Cover Letter *1*11 Of C*tQ*Nt*--m*|tM AHO Wtt>*w OIMO DEPARTMENT OF HEALTH SERVICES *111 tfMfltv ** mcniT. c* utv* OlOtOf CfU(Mf;i*N. Onw Dear Resident: We are conducting an Important health study In Santa Clara County. In 1981, a leak of potentially dangerous chemicals into a well of the Great Oaks Water Company was discovered. The contaminated well was promptly closed. In response to community concern about possible health effects of the leak, state and county researchers studied birth defects and miscarriages in an area of San Jose that may have received the contaminated water and in a similar neighborhood with a different source of water. More birth defects and miscarriages occurred in 1980 and 1981 in the area that may have received the contaminated water. Because we do not know when the leak began, we could not conclude that the contaminated water was responsible for the problems. We do know the con* taminated well was closed in December of 1981. Therefore, we are expanding the original study to see if the higher rates of birth defects and miscar riages continued after the contamination stopped. This study also will include two new neighborhoods in addition to the two previously studied areas. WE NEED YOUR HELP TO CONDUCT THE NEW STUDY. We need a list of all the In dividuals residing at this address. We also need to know whether any of the women in the household were pregnant at any time during the six year period from January 1, 1980 through Oecember 31, 1985. We are Interested in all pregnancies regardless of whether they ended with a live birth, stillbirth, miscarriage or abortion. We will contact women who have been pregnant during the study period and ask them to participate in an interview. Please have an adult woman living at this address fill out the enclosed form. It does not matter whether she has ever been pregnant. An adult man may complete the form if there are'no adult women in the household. We need to receive a form from your household WHETHER OR NOT THERE ARE ANY WOMEN In the household. We also need this Information now even though you may have filled out a form like this in the past. All Information on these forms will be kept confidential as required by the State Health and Safety Code and will be used only for this study. No In dividuals or families will be Identifiable In any analysis or report. If you have any questions, please call Kathleen Claxton or Margaret Wrensch at (41S) 540-2669. Thank you for reading this letter and for taking the time to return the form to us. A pre-addressed, stamped envelope is enclosed. Sincerely, Margaret Wrensch, Ph.D. Epidemiologist Department of Health Services State of California Chief of Health Protection Santa Clara County Health Oeparment 408/299-5858 2.87 033321 SL Enumeration Form Please have an adult woman who lives at this address complete this form. If there are no adult women in the household, then an adult man may complete the form. Please list each person In your household. LIST YOURSELF FIRST then each member of your household, Includ ing Infants. Start with the oldest person, then list the next oldest, etc. INCLUDE ANYONE WHO lives here or is staying here and has no other home, whether related to you or not; has a home elsewhere but who stays here most of the time; lives here but is away temporarily. DO NOT INCLUDE people currently away In the armed forces; college stu dents currently home only weekends or less; or people currently away from home in institutions. First Name Your name: Niddle Initial Household members: .Pit, Print. Birthdate Sex Last Name MO DAY YR M/F For each female born before 1970, was she ever pregnant anytime in 1980-85? (Ue are interested in |JJ. pregnancies Including live births, still births, miscarriages and abortions.). Please Check Yes or No. YES NO (Continue on back if necessary.) In case we need more information, what is your: home phone number _ hat Is the best time to reach you? a.m. p.m. work phone number When did you move to this address? ___ month year Please give us the name, address and telephone number of a relative, friend, or organiza tion who would be apt to know how we can reach you, if you move away. Name Address THANK YOU 2.88 Phone Number SL 033322 Interview Cover Letter |U1f 0 ANO WHfA*f ACtNCV department of health services 11)1 tlHttCf HiBltlV, CA 4704 415/540-2669 OIO*Gl OfUCMfjtAN V* recently received the form from your household in response to the first phase of our health study. Thank you for your cooperation In returning that information to us. The form Indicated that you were pregnant sometime between 1980 and 1985. Ve now would like to interview you concerning your pregnancies. The interview is usually conducted by telephone. Your participation in the Interview Is entirely voluntary, and you may terminate the Interview at any time. All Information received during the interview will be kept confidential as required by the State Health end Safety Code. It will be used only in statistical summaries and no individuals will be identified in an/ analysis or report. An interviewer will telephone you within the next few weekt. to answer any questions you may have and to schedule an interview. As we explained in the previous lecter, this study is an extension of a recent study of miscarriages and birth defects in Santa Clara County. The original study shoved that in 1980 end 1981 more birth defects and miscarriages occurred in an area of San Jose served by the Croat Oaks Water Company then In a similar nelghborhood with a different source of water. This new study will cover 1980 to 1985 and will add two neighborhoods to those originally studied. This study will help to clarify the reasons for the previously observed excess of miscarriages end birth defects end will tell us whether this excess still persists. To obtain useful information, a high proportion of the women living in all the study areas who were pregnant anytime between 1960 and 19B5 must participate. Wo hope you will participate. If you have any questions please write or call us. Ve greatly appreciate your help in this study. Sincerely, Epidemiologist Epidemiological Studies and Surveillance Section 1512.5 Chief of Health Protection Santa Clara Co. Health Dept. 408/299-5858 2.89 033323 Cover Letter Requesting Permission to Contact Physician About Reported Spontaneous Abortion STAff O* *HO AGtNCV DEPARTMENT OF HEALTH SERVICES till MMIltT WAV ttrativ. ca tint 415/540-2669 010*01 OCUVMCjiam Gw Dear In a racant interview you reported having a laiacarriagt In Aa the interviewer nantlonad, we would Ilka youc peculation to contact your phyalelan for aora nodical information. Plaaaa chack tha aneloaad conaant fora to aaa that tha data and doctor'* name and addrasa arc correct. Plaaaa fill in any slating information; than aign the and raturn It in tha aneloaad atanped, aalf-addraaaad anvalopa. Thank you again for your halp in thia atudy. Sincaraly, Dannlt Shu*Canaan, KO, HPH Hadical Epideniologlat Epldenlological Studiaa and Surveillance Section 1513.5 2.90 SL 033324 Permission Form to Contact Physician About Reported Spontaneous Abortion STUDY ID DATE OF MISCARRIAGE (SPONTANEOUS ABORTION): NAME OF DOCTOR: ADDRESS: Tha California Stata Dapartaant of Haalth Sarvieaa haa ay paraisaion to obtain inforaation concavnlng ay aiscarriaga (spontanaou* abortion) occurring on tha data givan abova. SIGNED: 1313/rlt.J 2.91 SL 033325 Spontaneous Abortion Validation Form t*H Of -WHIM tHO *IHI Gt*CT DEPARTMENT OF HEALTH SERVICES mi imiin witnrf. c* *> 415/540-2669 CitOtGf 0IUIJI1N 0.w<w Dear During our study of pregnancy outcomes tn Sants Clara County, statsd that sha had had a miscarriage in Enelosad Is an authorisation forn Cor you to rolsass Information concerning this miscarriage. To verify tha diagnosis and complete our study va would Ilka tha following Information: Vas tha raportad miscarriage a spontanaous abortion (lass than 20 waaks gatcation)7 I 1 YES [ 1 NO If "yaa." what was tha causa of, or tha eltcuiastancas surrounding, tha raportad miscarriage? If *no,* ptaasa describe what happened. If It was an Induced abortion, was it performed because of known or suspected fatal abnormalities? Please return this to us as soon as possible In the enclosed stamped self- . addressed envelope. Your cooperation Is graatly appreciated. Sincerely, Dennis Shustermen, H.D., M.P.H. Epidemiological Studies Section 1514.5 2.92 SL 033326 Cover Letter Requesting Permission to Contact Physician About Reported Birth Defect SIATt or CAIWOSMIA--WtAITM AMO AGfWCY DEPARTMENT OF HEALTH SERVICES mi RIMIII WA num. ca fA70A 415/540-2669 610*01 DCUCMIllAH CamTM** i*: Dear In a recent Interview you reported that your ehtld wee diagnosed as having a birth defect. As the Interviewer Mentioned, we would like your P*mi**ion to contact the physician who wade the diagnosis for more nodical Information. Flaasa check the enclosed consent fora to see that the child's name, and the doctor's name and address are correct. Please fill in any alsslng Information; than sign the form and return It In the enclosed scamped, salf-addressed envelope. Thank you again for your help In this study. Sincerely, Dennis Shusterman, HD, MPH Medical Epidemiologist Epidemiological Studies and Surveillance Section 1515.5 2.93 SL 033327 Permission Form to Contact Physician About Reported Birth Defect STUDY ID ___ ______________ CHILD*S NAME: ____________________ APPROXIMATE DATE OF DIAGNOSIS: BIRTH DEFECT DIAGNOSED: __________________ NAME OF DOCTOR: ADDRESS: ______ Tha California SCaca Dapartment of Health Services has ay permission to obtain Information about cha birch defaces ay ehlld has. SIGNED: lSlS/rl.5 2.94 SL 033328 Birth Defect Validation Form sun Of C(l0*HI*--MfAlTH AMO WflUlf ACfNtT DEPARTMENT OF HEALTH SERVICES 111) KMIUi w* CA NMl *15/5*0-2669 OfOSGI OtU>MtJiAN G--r *v Daar During our study of pregnancy outcomes In Santa Clara County, atacad that you had dlagnoaad birth dafsct(s) In bar child Enclosed la an authorisation fora for you to ralaaaa Information concerning the birth daface(s). To verify the diagnosis and complete our study please reeord the conditions you have diagnosed In this ehlld. * Pill lit Plaingltfl Condition Ploaso return this to us as soon aa possible In the enclosed stamped self* addressed envelope. Your cooperation Is greatly appreciated. Sincerely Dannie Shusteraan, M.D., H.f.H. Epidamlologleal Studlas Saetlon 1516.S 2.95 0333*9 si- Questionnaire TIME INTERVIEW STARTED _______________________ TOTAL TIME 7/ INTERVIEWER'S INITIALS \q/ ID ___ ______________________ CT HH* W* DATE _____________________ 1/ jj/ PREGNANCY OUTCOME QUESTIONNAIRE Hallo. May I speak to . I*ii from the California Department of Haalch Sarvlcaa. I' interviewing for tho haalth study of woman in Sapca Clara County who havo boon pregnant. Did you racoivo a lctcor from us concornlng this study? PAUSE. (IP NO. READ LETTER OR SEND LETTER IF REQUESTED AND SET A TIME TO CALL BACK.) * I'd llko to interview you now if It's convenient. All information you give mo is confidential and will bo used only in statistical summaries. PAUSE. IF NOT CONVENIENT TIME FOR INTERVIEW. SET A TIME TO CALL BACK. IF R HESITATES.TRY TO FIND OUT WHY. TO ANSWER ANY OBJECTIONS AND TO PERSUADE THE R TO ALLOW HERSELF TO BE INTERVIEWED. R HAY CONFIRM THE LEGITIMACY OF THE STUDY BY PHONING DAVID EPSTEIN COLLECT AT (415) 540-2934. IF R REFUSES DON'T PUSH, LEAVE THE DOOR OPEN FOR SECOND TRY. 1. First of all. what is your blrthdaco? MONTH DAY YEAR .19/ [ ] DK IF DON'T KNOW: How old aro you? 25/ 2. Havo you over boon pregnant. ovon if onlv for a short time? Plaaso Include any pregnancy regardless of whether it endec1 In a live bitth, stillbirth, Miscarriage, abortion or was an ectopic pregnancy. K 1 VES 2t 1 NO 91 1 DR IF NO OR DON'T KNOW, TERMINATE INTERVIEW. SAY: Thank you very much for your help in this study. We are only interviewing women who havo been pregnant. 3. When did your most recent pregnancy end? 27/ MONTH YEAR IF HOST RECENT PREGNANCY ENDED BEFORE JANUARY 1. 1980, TERMINATE INTERVIEW. SAY: We're primarily interested in pregnancies that occurred during or after 1980, so I won't need to ask you anymore questions at this time. Thank you very much for your help in this study. If MOST RECENT PREGNANCY ENDED SINCE JAN 1. 1980 CO TO QUESTION 4. 28/ SL 033330 4, How I would like Co ask you some questions to summarize your entire pregnancy history Let's start with your very first pregnancy. ASK a-E AS APPROPRIATE. Did you have another pregnancy after this one? IF YES, ASK A-E: IF NO ASK: ASK 4-F AB CHECK HERE IF TWIN OR HVLTIPLE BIRTH. Did this pregnancy end with a live birth, stillbirth, miscarriage, elective abortion or was it an ectopic pregnancy? PREGNANCIES AFTER THE 1ST: How did this pregnancy end? IF LIVE BIRTH: Did the baby have any abnormalities or birth defects noticed at birth or with in the first 6 months of life? outcome LIVE MISCARRIACE STILLBIRTH ECTOPIC ELECTIVE CURRENTLY BIRTH BEFORE 20 WK AFTER 20 WK TUBAL____________ ABORTION PREGNANT YES NO ll ) it t 21 J 31 I 41 ) St ) St J ll J 21 1 ll ) it ) 2l J 3t 1 ft ) Jt I Si 1 ll ) 2l ) ll ) It ) zl ) ll 1 it 1 zi 1 11 1 21 ) 21 1 11 ) 21 1 21 1 jl I ft 1 5t 1 St ) jl 1 it 1 5l 1 ft 1 Jl J ft ) 51 1 ft I Jl 1 ft 1 Si ) Si ) If I ll J ll ] ll ] 2( 1 2( 1 21 ] It 1 ll 1 ll ) 2( I Jl 1 ft 1 Si 1 fl 1 ll 1 2 l 1 ll ) 21 1 2l I Jl 1 ft ) St ) ft 1 ll 1 2 l 1 ll ) 21 ] 2t J Jl 1 ft ] Si 1 St 1 2( 1 21 1 21 1 ll ) 2 ( J Jl ) ft ] Si 1 fl ) ll 1 St 1 4F. AFTER LAST, COUNT PRECNANC1ES: I count that you've had a total of ____ () pregnancies. Old you have any others chat we might have missed? Any which were very short or ended in abortion? IF SO, ADD TO ABOVE. J22/ 46, WHEN HISTORY COMPLETE: Excuse me a moment. I'm going to mark a calendar with the months you were pregnant between 1980 and 198S and check whether you were living in our study areas during that pregnancy. I'll then ask you some more detailed questions about those pregnancies. HARK CALENDAR. -2- 033331 CD I How long did Cha pregnancy last? IF LIVE: UhaC was cha baby's birthday? OTHERWISE: Whan did cha pregnancy end? IF BECAN ON/AFTER 1/1/80 AND ENDED ON/BEFORE 12/31/85 Ac what address did you live during this pregnancy? Did you live anywhere else during Chat cine7 MONTHS WEEKS MONTH DAY YEAR ADDRESS When did you ova In? KN YR ------- ------- ------- ------- -- --..... ------- ------- -------- --------- ------- ------- ------- ------- ------- ------- ----- U- ------ -- ------- ---- ------------ _ --- _ ------ ---- ------- ---- ------------ ---- ----------- ---- ------------ ---- ----------- ---- ----- ---- ---- --- --- ---- __ ---- -- --- ---- CHECK BOX IF ELIGIBLE ELIGIBLE [ ] Jl I ] f! t 1 9? I 1 21? ( 1 145 ( ! 177 l 1 2oe I 1 233 I 1 254 l 1 293 ELIGIBLE IF: 1. Live birch, sclllblrch, olscarriaga 2. Began on/afcer 1/1/80 3. Endad on/before 12/31/85 4. Rasldanc In scudy area entire tlma of pregnancy -J 1306.6 033^Z IF NO PREGNANCIES BECAN ON OR AFTER JAN l, 1980 OR ENDED ON OR BEFORE DEC 31. 1985 TERMINATE INTERVIEW AND SAY: Ue're primarily Interested in pregnancies chat ended In a live birch, miscarriage or stillbirth and that occurred In the study areas between 1980 and 1985, so I don't need to ask you any more questions at this time. Thank you very much for your help In this study. OTHERWISE. RESUME THE 1NTERVIFW WITH THE FIRST ELIGIBLE PREGNANCY. I've marked a calendar with the months you were pregnant for each pregnancy between 1980 and 1985. I'm going Co ask you some questions about each pregnancy that ended with a (live birth, stillbirth or miscarriage) during this time. CHECK CALENDAR AND CO TO FIRST ELIGIBLE PRECNAHCY. * --1 of _. ,, _ ZZ4/ Ufa begin with the pregnancy that ended of and that began around J*5/ MONTH YEAR -- of : and that you said resulted In a MONTH YEAR OUTCOME that correct? (IF NO. CORRECT DATES AND OUTCOME] : Is IF THE PREGNANCY ENDED IN A: TWIN BIRTH, USE TWIN FORM FOR SECOND BABY. OTHER MULTIPLE BIRTH: USE TWIN FORM FOR OTHER BABIES AND MARK IT 'TRIPLETS*, QUADS', ETC. MISCARRIAGE CONTINUE TO NEXT QUESTION STILLBIRTH SKIP TO QUESTION & ON PACE 5 LIVE BIRTH SKIP TO QUESTION 2 ON PAGE 6 -4- SL 033333 5. FDR MISCARRIAGES ONLY a. Did you have a pregnancy ease Cor this pregnancy? 2 ( ] YES: 1. Waa cha ease dona: J ( J ae homo? 2 { j ae a doccor'a offlea? j ae i hospleal or clinic? 4 1 ] or ae soae oeliar placa? (DESCRIBE) 2 ( J NO a [ 1 dk 2. tfaa eha outcome of ehla case postclva? J I I YES 2 [ 1 NO a l ) DK b. Did you sa# a physician Cor ehla alacarriage? 1 [ ] YES: 1. WhaC waa eha physician's or clinic's nama? I t 1 NO PHYSICIAN FIRST NAME LAST OR 2. Did you have a D and C Cor ehla pregnancy? J t J YES 2 t I 0 CLINIC <. FOR STILLBIRTHS ONLY SKIP TO QUESTION S ON PACE 8 a. Did you sac a physician or go eo a hospleal Cor ehis seillblreh? J [ ] YES: l. Uhae wars cha physician's and hospleal's naaas? 2 l ] NO PHYSICIAN'S FIRST NAME HOSPITAL SKIP TO QUESTION 8 ON PACE 8 LAST CITY 333/ 334/ 335/ 336/ 33T/ 33a/ 333/ 340/ 341/ *5* Si 033334 7. FOR LIVE BIRTHS ONLY a. Vas eha baby a boy or a girl? 1 ( ) BOY 2f ] GIRL b. What nan* appear* on eh* baby'* birth certificate? FIRST MIDDLE LAST 344/ e. Was th* baby born in a hospital,clinic or at hone? J( ] HOSPITAL OR CLINIC: What was th* nan* of th* hospital or clinic and vhat city vas it in? 345/ ^-- MAME OF HOSPITAL OR CLINIC CITY 346/ 2( ] HOME -6* SL 033335 d. Lie aa ravtaw this baby's health again: Did ths baby hava any abnornalltlas or birch dafaces notlcad ae birth or within cha first 6 aonths of Ufa? j( ) YES: 1. Vhae did tha doctor say tha baby's birch dafaet or abnoraaliey was? PROBE FOR SPECIFIC CONDITION. 349/ 250/ 2. Was this diagnosis-aada ae a hospital or a doctor's offlea? t[ J HOSPITAL: Uhac was cha naaa of Cha hospital and what city was le In? _____________________________ HOSPITAL _____________________________ 351/ CITY Vhae was tha doctor's naaa? FIRST. 2[ ) DOCTOR'S OFFICE: Vhae was cha doctor's naaa? LAST FIRST LAST What was <hls/har) addrass? 355/ STREET CITY 3. Doss any blood rolaetva of tha child hava this condition? It J YES: Uhac ralatlonshlp Is tha parson to tha child? ft ) NO l J DR - ...... 21 ) HO 9{ ] NOT APPLICABLE ------- ---- a. Is this child aliva today? J[ 1 YES: SKIP TO QUESTION 8 PAGE 8 2( J NO: 1. Whan did tha child dla? ______ MONTH ____ DAY _____ YEAR 2. Vhat was cha eausa of daach? te*/ 254/ 356/ lsa/. ' 365/ 3. In what city, county, and stata did tha daach occur? CITY COUNTY STATE -7- 366, 336 0** 5^ ASK ALL RESPONDENTS: . When did you first see a doctor for this oratnancv? MONTH OF PRECNANCY 9. Were you using on ItJD whan you became pregnant? I l 1 YES 2 l ] NO g [ ] DK 367/ 370/ 10. Now, I'm going to ask you about soma ehings that may hava occurred during this pregnancy or in the three months before, that being from of to of _______________________________ MONTH YEAR MONTH YEAR 3 At anytime during this pregnancy or in the three months before did you have any of the following diseases or conditions: YES NO DK Diabetes? 1 ( 1 2 l IP II IF YES: Did you taka any medication for this? 1 ( Ml 1 9 (I MEDICATION Hypertension J[ or high blood pressure? IF YES: Did you take mad* lcation to control It? ] [ ) 21 121 )J M 1 9U MEDICATION c. Rubella or German measlee? I-1 ] 2 l ) 9 | ] A. Did you take any (other) prescription drugs anytime during this pregnancy or the three months before? J [ ] 2 ( ] 9 [ ] IF YES: Was During what this before months of the pregnancy? pregnancy? HI __________ HI __________ III III __________ ______________ III 379/ 383/ 391/ 397/ 403/ 409/ IF YES; Whet was it? 1) 2) 3) 11 1 It 1 It 1 410/ 8* SL 033337 following complications: a. Frs-aelampsla, aclampsla, or toxoaia? YES NO j( 1 2( J DK 9( J IF YES: During what months of prognancy? _________________________ __ 435/ b. SoUuroo? t. Vaginal blooding oxcopt at tho and of tho prognancy? 4. Caesarean soctlon7 IF LB OR SB: 2 l 1 21 ) St ) j ( ] ] g[ ] j[ ] g( ] g( j 440/ 445/ 4sg/ IF YES: What was tho roason for hawing a caosaroan section? _______________ PROBE FOR HOST SPECIFIC INFORMATION KNOWN 45J/ 12. At anytlmo during this prognancy did you havo a favor of 100 or moro7 2( ] YES: Did you hava a favor: during tha 1st3 mos7 jJ J YES g{ J NO g{ J DK ' 2nd 3 nos7 2( J YES *t 1 NO S ( ] DK 3rd 3 most Jl J YES 2t J NO 91 ] DK at tho ond of tho prognancy7 2( 1 NO F( 1 DK 2[ 1 YES 2t 1 NO * g( ] DK 13a. About how much did you walgh at tha baginning of this prognancy? b. How many pounds did you gain during this prognancy? LBS. LBS. 452/ 453/ 454/ 455/ 456/ 460/ -9. SL 033338 14. In th three months before or ae any time during this pregnancy, that would be from of through _____ of .were you employed or MONTH YEAR MONTH YEAR did you do volunteer work? 1 I JYES 2 { ]NO-SKIP TO QUES' ION 15. 453/ 471/ a. What was your occupation _or -lob' . b. What Industry was this Job In? e. What was the name of the company or agency you worked for? d. In what city was that? a. What kind of activities did y,ou do day to day at work? .......... JOB *1 JOB 2 472/ 478/ 480/ 484/ 488/ f. What month and year did you start this Job? MONTH YEAR (. During what months of this pregnancy did you work at this job? h. How many hours/week did you work? 1. Old you have any other Jobs during this pregnancy? J ( j YES - CO TO NEXT COLUMN, ASK 14a. 2( I NO - CO TO QUESTION 14J. MONTH YEAR 508/ 816/ 824/ 11 ) YES * RETURN TO QUESTION 14a AND ATTACH SUPPLEMENT JOB FORM UNTIL ALL JOSS ARE COVERED. 25/ 2 1 ] NO - CO TO QUESTION 14J . M/ -10- SL 03333g 1AJ. Old you have any typo of work schedule ochor chan only daytine houra during ehla pragnaney or In Cha 3 montha before? 3 I J YES 2 t J HO: SKIP TO QUESTION IS S30/ IT YES: Did you work: ft. Evaninga only (3-llpa)? b. Nlghta only (llpa * 7am)7 Co Rotate between 2 ehlfta? d. Rotate between 3 ahlfta? Other SPECIFY Waa Chla bafora pragnaney 7f 1 til T1 1 7f 1 1I 1 During whae aontha of pragnaney? .. . 337/ ***/ s*7/ 848/ 551/ 13. How I'd Ilka to aak you about tone axpoauraa you aay hava had during ehla pragnaney or in tha 3 noncha bafora. O. Vara you ragularly, at laaae onea a waak, exposed to ehaaleala or aolvanca at hooa, at work, or In aona ochar activity? For axampla, paint, laquar, varnlah; organic ehaaleala or xolvanta such aa paint thinnar, gaaollna, karoaana, fuel oil; or to paatlcldaa or Intact apraya. 3 l J YES 2t 1 HO: SKIP TO QUESTION 15b p t J DK: SKIP TO QUESTION 15b *56/ IT YES: What particular ehaaleala or solvents wara you axpoaad to? Waa thla During what nontha CHEMICAL OR bafora cha of tk pragnaney What ware SOLVENT pragnaney? wara you axpoaad? you doing? 1) ______ _____________________ 2) ______ _____________________ 3) ______ _____________________ 4) ______ _____________________ How often? ______ 557/ ______ 555/ ______ 553/ ______ 553/ b. Vara you ragularly axpoaad to a video dlaplay terainal? 3 [ 1 YES: a l 1 no 5 ( J OK ______ _____________________ ______ SOS/ 806/ *11- 0333^0 SL 16. In the three month* before chi* pregnancy or at any time during chi* pregnancy: a. Did you have a chest x-ray? YES NO DK l[]r[)S[) IF YES: Has thl* before the During what month(s) pregnaney? of pregnancy.? 1[ J _________________________ 616/ b. Did you have a pelvic or abdominal x-ray? e. Were you exposed to any other x-rays or other radioactive materials? During this pregnancy: I [ ]2[ ]9 [ ] J[ J __________________________ 621/ Z { |*( ]*f J 2( J __________________________ 627/ d. Old you have amniocentesis? 1 { ] 2( )9l ] . Did you have chorionic villus sampling? j [ ] *[ ) j( J 633/ 638/ -12- 17. Did you smoke cigarettes during ehts pregnancy? 11 1 YES ?l 1 NO SKIP TO QUESTION 18 3t 1 DK SKIP TO QUESTION 18 IT YES: How much did you smoke: 1/2 pack a day or lass a. during the 1st 3 months b. during cha 2nd 3 months e. during the last 3 months M1 It J It 1 about pack a day 21 1 Ifl 2( 1 1-1/2 packs a day .or more S[ 1 t 1 St I not at all U) U1 UI 643/ 644/ 645/ 646/ IS. Did you drink any type of alcoholic beverages during this pregnancy any bear vino, or hard liquor? It 1 YES gf 1 WO SKIP TO QUESTION 19 gf 1 DK - SKIP TO QUESTION A3 *47/ IT YES: abouc ovary day? a. During cha 1st 3 months, abouc ofcan did you drink: 1[ ] On days whan you did drink abouc how many drlnka did you havo? It ) 1 It J 20 Jl ] 4+ b. During cha 2nd `3 months, about how oftan did you drink:;{ ] On daya whan you did drink abouc how many drlnka did you hava? 3l J 1 It ) 20 J[ ] e. During the last 3 months about how oftan did you drink:! On daya whan you did drink abouc how many drinks did you have? M 1 1 *t J 20 3 | J 4* several times a week? 21 1 It 1 St 1 about once a week? several times a month? once a not month at or lass? all? 31 J 3t 1 3t 1 4\ ) St J St I */' 4( 1 649/ S l J fit) 650/ 651/ 4\ 1 5 t 6 t 1 652/ 653/ -13- 03331*2 19. The nxc group of questions is about the liquids you drank during this pregnancy BX. for the three months before this pregnancy. a. What types of liquids did you usually have each day during this time period? b. About how many glasses, cups or ounces of liquids did you drink per day? Pi**** include all sources of liquids; coffee, tea, soft drinks, juices, water, milk, beer, wine, other liquor, any liquid. ______________ or CUSSES OR CUPS (S ounces - approx. 1 cup) I J OK c. During lust the first trimester of this pregnancy was this amount different? 21 ) HO *1 J OK 11 ) YES; How many glasses or cups did you drink per day? CUSSES OR CUPS 654/ 6S7/ 658/ 20a. Ac home, did you use bottled water at all during this pregnancy or for the three months before this pregnancy? 2[ J yes :( J HO: SKIP TO 21 g( 1 DK: SKIP TO 21 b. Did you usually use bottled water at home? 2( ] YES 2( J HO 91 1 DK 661/ 662/ 21. During this pregnancy or In the 3 months before did you use a water filter at your tap? it ] YES*. 1. When did you begin using it? MONTH YEAR 2. How often did you change It? 2t J NO S( J DK 663/ 664/ 668/ -14* SL 033343 22. Did you usually drink water Cron a prlvaCa wall during Chit time? jt J YES: 2( ] HO St] DK Where was tha wall? STREET CITY 23a. AC home, about how many glasses or cups of cold tap water or beverages made of unhaatad tap wacar did you drink each day? This should Include spy beverages ado of unheated wacar such as orange juice. Iced tea, or other beverages made frea eoneantrate. ( ] NONE: SKIP TO 24 ____________________ GLASSES OR CUPS or [ 1 OK 670/ 671/ 672/ 674/ b. And during lust the first trimester of this pregnancy was this amount different? j( 1 YES: H1 It 1 DK How many glasses or cups did you drink par day? GLASSES OR CUPS 575/ 576/ 578/ 2*. ir THE WOMAN DID NOT WORK DURING THIS FRECNANCY OR IN THE 3 MONTHS BEFORE SKIP TO QUESTION 26. a. And at work, did you use bottled water at all during this pregnancy or for the three months before this pregnancy? 1 t 1 YES Jl J NO: SKIP TO 2S | ( ] DK b. Did you usually use bottled water at work? 1 t J TES 2 t J NO 9 t 1 DK 680/ -15* SL 033344 23a. AC work, abouC how many giants or cup* of cold fan water or bavtragts Bade of unheated tap water did you drink each day? This should Include nv beverages Bade of unheated water such as orange juice, iced tea, or other beverages made froa concentrate. [ ] NONE: SKIP TO 26 ___________________ or [ ) DK: CLASSES OR CUPS 81/ 683/ b. And during lust the first trimester of this pregnancy was this amount different? 2 I J YES: How many glasses or cups did you drink per day? AMOUNT 684/ 21 1 NO 2l j DK 685/ 687/ 26. Were you nauseous during this pregnancy? If 1--3TE5ft 1 NO CO TO QUESTION 27688/ IF YES: How often were you nauseous: a. During 1st 3 mo? every day or almost every day *1 ) b. During 2nd 3 mo? 2I 1 at least ones a wk but not every day *t 1 less than once a week ] not at all *t 1 m 3[ 1 4[ ) DK 9t 1 9[ ] c. During Last 3 mo? 1 [ ] 2( 1 3( 1 d. Did your nausea affect your consumption of cold drinks? 4( 1 9I 1 1 ( ) YES: Did your consumption Inc 2 [ J NO, OR IT STAYED ABOUT THE SAME ? 1 t I INCREASE 2 [ ] DECREASE 889/ 690/ 691/ SO 2/ 893/ 9 t ) DON'T KNOW a. Did you take ancl*nausea medication. 2 ( J YES 2 l J NO IF YES: Uhat was that? 9 l ] DK MEDICATION 891/ 895/ -16* SL 033345 27. In the Ytir before this pregnancy or any etna during this pragnancy, waa acruccural pasc control or termite control work dona In your raaldance? 1 YES 1. What chemical waa uaed7 or [. J DK NAME 2. What waa the main pasc balng controlled? 697/ 698/ 699/ 3. What was the name of the company that did tha post control work? 700/ 4. What aonth and year waa tha work dona? 2 l J NO M I DK MONTH YEAR 701/ 21. Row I'd Ilka to ask a few questions about the father of this pregnancy. [INTERVIEWER: CHECK PREGNANCY CALENDAR FOR THE APPROXIMATE MONTH AND YEAR CORRESPONDING TO 3 MONTHS BEFORE THIS PREGNANCY BEGAN.] In lust the 1 months before this pregnancy began, that would be from of MONTH W to _ of 19 . was the father of thts pragnancy ea^loyed or did he do MONTH volunteer work? 705/ 1 t 1 YES 2 1 ] NO: SKIP TO QUESTION 29 p [ J DK: SKIP TO QUESTION 29 713/ In this lob:JOB al__________________ _________________ JOB 2 a. What was his occupation? b. What was the nano of the company or agency - ha_.worked for? c. In what city was that? d. What kind of activities did he do day to day at work? 711/ 718/ 722/ 726/ e. Did he have any 1 [ ]YES: CO TO 28a AND SAY: other Jobs during In the other Job he had: this time period? 2 [ ] NO: CO TO NEXT QUESTION 17* 1 1 J YES: CO TO 2Sa ATTACH FATHER'S SUPPLEMENTAL JOB FORM AND SAY: In tha other job ha had: 2 [ ] NO: CO TO NEXT QUESTION 736/ 737/ SL 033346 mwn. 29. Which of tha following base describes hi* raca or athnlc group: 1[ ] Hispanic 5( ] Asian 2( I Black J Or so*a ochar group? Jl I White Si } Aaeriean Indian Si J DON'T KNOW SPECIFY 30. What*a tha hlghasc grada in school char ha had eontplacad aC Cha tiae of chis pregnaney? 2110-8 2 I J 9-11 J [) 12 (HIGH SCHOOL GRADUATE) 4 t J 13-15 (SOME COLLECT OR TECHNICAL SCHOOL) 5 I J 16 or aora (COLLEGE GRADUATE) 9 I J DK 733/ 739/ ( J ADDITIONAL ELIGIBLE PREGNANCY: FILL OUT AND ATTACH THE "ADDITIONAL PREGNANCIES* FORM. ( J NO MORE ELIGIBLE PREGNANCIES: CO TO NEXT PACE. -IB- 0333^7 I'd Ilka Co conclude with a faw nora questions about yourself. 31. How tall ara you? ___________ FEET INCHES 32. Uhlch of cha following bait daicrlbas your raca or achnlc group: it ] Hispanic S[ ] Aslan 2( ] Black el ] Or soaa ocher group? 21 ] White 91 j DON'T KNOW 41 ) Aaerlcan Indian SPECIFY 33. Vhat Is Cha hlghast grada In school chat you hava completed? 1[ 1 0-8 S[ ] 9*11 J[ J 12 (HIGH SCHOOL CRADUATE) <( ] 13*13 (SOME COLLEGE OR TECHNICAL SCHOOL) 5( ] 16 or Bora (COLLEGE GRADUATE) 91 ) DK If wo naad co gac In Coueh with you again, can you glva a cha naaa and addrass and calaphona number of a ralatlvs, frland, or organisation who would ba apt Co know how wa can roach you If you aova away7 740/ 743/ Thosa ara all ay qua*Cions, do you hava any questions you'd Ilka co ask aa bafora wa finish cha Interview? ANSWER QUESTIONS OR REFER TO A SUPERVISOR. Thank you very nuch for helping us In this study. We may want co review your medical records for more Information about this (these) pregnancy(las). If so, wa will send you a consent form that asks you for permission co do so. TIME INTERVIEW FINISHED COMMENTS_________________________ NOTE IF RESPONDENT HAD DIFFICULTY WITH ANY QUESTIONS OR WAS UPSET OR DISTURBED BY THE INTERVIEW. DHS/ESS/MW/6/3/86 19* 0^ Appendix B. Occupations Used In Classifying Solvent Exposure Probably Solvent Exposed: *48 Chemical engineers 73 Chemists, except biochemists 115 Chemistry teachers, postsecondary 188 Painters, sculptors, craft*artists 203 Clinical laboratory technologists 224 Chemical technicians 505 Auto mechanics except apprentices 506 Auto mechanic apprentices 507 Bus, truck, stationary engine mechanics 508 Aircraft engine mechanics 514 Auto body and related repairers 515 Aircraft mechanics except engine 518 Industrial machinery repairers 519 Machinery maintenance occupations 538 Office machine repairers 539 Mechanical controls and valve repairers 544 Millwrights 565 Tile setters, hard and soft 579 Painters, construction and maintenance 585 Plumbers, pipefitters, steamfitters 587 Plumber, pipefitter, steamfitter apprentices 637 Machinists except apprentices 639 Machinist apprentices 657 Cabinet makers and bench carpenters 658 Furniture and wood finishers 669 Shoe repairers 679 Bookbinders 703 Lathe and turning machine set*up operators 704 Lathe and turning machine operators 705 Milling and planing machine operators 734 Printing machine operators 735 Photoengravers and lithographers 748 Laundering and dry cleaning machine operators 753 Cementing and gluing machine operators 759 Painting and paint spraying machine operators Hand painting, coating, and decorating Possibly Solvent Exposed: 56 Industrial engineers 78 Biological and life scientists 83 Medical scientists 114 Biological science teachers, postsecondary 133 Medical science teachers, postsecondary 189 Photographers 213 Electrical and electronic technicians 223 Biological technicians 828.3 2.96 SL 033349 225 Science technicians, N.E.C. 523 Electronic repairers, communications and industrial equipment 525 Data processing equipment repairers 526 Household appliance and power tool repairers 533 Hise. electric and electronic equipment repairers 535 Camera, watch, musical instrument repairers 536 Locksmiths and safe repairers 547 Specified mechanics and repairers, N.E.C. 549 Not specified mechanics and repairers 634 Tool and.die makers except apprentices 635 Tool and die maker apprentices 636 Precision assemblers, metal 645 Patternmakers and model makers, metal 649 Engravers, metal 654 Sheet metal worker apprentices 655 Miscellaneous precision metal workers 675 Hand molders and shapers except jewelers 678 Dental lab and medical appliance technicians 683 Elect and electronic equipment assembler 684 Miscellaneous precision workers, N.E.C. 695 Power plant operators 696 Stationary engineers 708 Drilling and boring machine operators 737 Misc. printing machine operators 745 Shoe machine operators 774 Photographic process machine operators 783 Welders and cutters 784 Solderers and brazers 785 Assemblers dumber Indicates 1980 U.S. Census Code. 828.3 2.97 05335 App ndix C. la there more variation In spontaneous abortion rates than expected bv chance? Background In the Pregnancy Outcome Follow-up Study a large number of spontaneous abortion rates have been estimated. There are four study areas and six study years each of which provides an independent estimate (Table 2.9). Ve ask: are these 24 estimates more variable than expected? To answer this question it is necessary to model the expected distribution. In this discussion we have assumed that each of these 24 estimates arises from an Independent binomial experiment. Ve have assumed that the probability of "success" (in this case, pregnancy ending in spontaneous abortion) is constant over all 24 trials. Ve then fit a normal distribution to the observed rates. Mtshodg A frequency distribution was constructed using the 24 observed rates. These range from .00 to .286 with a mean of .099. The observed and expected frequency distributions are contained In Table 2.C.I. The numbers of pregnancies on which these rates are based range from 14 to 85. Rates in 1985 are based on only three months of data and sample sizes for this year are particularly small (14 to 21). Including 1985 data, the mean sample size was 43.25. To simplify matters it was decided to attempt to fit a normal distribution using a common sample size for all years and areas. It can be argued that if the data fit a normal distribution reasonably well under this assumption, the fit would be even better using the actual sample sizes. Therefore, the expected frequency distribution was calculated assuming that the 24 data points are independent and Identically distributed according to a binomial distribution with a success probability of 10% and a sample size of 43. The observed and theoretical frequency distributions were then compared using the one-sample Kolmogorov-Smirnov statistic (1). 2.98 Raauita The observed and theoretical distribution# agree fairly well. The observed distribution is somewhat skewed, with a longer right tail. The maximum difference between the observed and theoretical frequency was 22.7% and occurred in the interval 0.04 to 0.06. The significance probability for this difference, using the ne-sample Kolmogorov*Smirnov statistic, is about 0.15. Discussion Overall, the wide variation of spontaneous abortion rates seen in this study is not very different from what would be expected by chance, assuming that the rate for each year*eensus tract combination arose from a single repetition of a binomial trial In which the overall probability of success was 10%. The data from the original Fairchild study were unusual, in that these data included the lowest observed rate (0.0%) among the 24 estimates. This occurred in the control tract for that study in 1980. This study also found one of the highest rates (21.4%), which occurred in the exposed tract in 1981. The latter high rate is not suprising, since it was a cluster in that tract which brought about the study. This high rate was, In a sense, the "index ease". However, the low rate in the control tract for 1980 can only be ascribed to chance, since, across all study years, rates in that tract averaged .096, which is very close to the overall mean. 2.99 033352 S^ Rftfgrgnc? 1. Rohlf FJ, Sokal RR; Statistical Tables: Second Edition. WH Freeman and Co. San Francisco 1981, p. 204. 2.100 SL 033353 Table 2.C.1 Observed and Theoretical Cumulative Distributions Value 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 0.30 Number of observation* ___ less than value 1 3 Observed 0.0417 0.1250 Cumulative Probability Theoretical 0.0385 0.0922 0.0033 0.0328 9 0.3750 0.1480 0.2270* 10 0.4167 0.3291 0.0871 13 0.5417 0.5000 0.0417 14 0.5833 0.6709 -0.0876 16 0.6667 0.8520 -0.1853 18 0.7500 0.9078 -0.1578 20 0.8333 0.9616 -0.1283 22 0.9167 0.9864 -0.0697 23 0.9583 0.9961 -0.0378 24 1.0000 1.0000 0.0000 Maximum difference A-0.2270 Significance probability using one-sample Kolmogorov-Smirnoff statistic approximately 0.15. 2.101 SL 033354 cn f 0 u0s1 CO m Ul 2.102 Appendix D Table 2.D.1 T1HE PERIOD Exposed: 1980 1981 1980-81 Number of Definite Spontaneous Abortions Among All Eligible Pregnancies8 EXPOSED UNEXPOSED 5120.1210) 5120.11IN) BOTH 5120.0810) 5053.0204IN) BOTH 8 14 (22) 4 6 (10) (12) (20) (32) 0 10 (10) 2 8 (10) (2) (18) (20) TOTAL 14 38 (52) Unexposed: 1982 1983 1984 1985 1982-85 12 8 17 3 (A0) 7 12 11 4 (34) (19) (20) (28) ( 7) (74) 9 10 17 4 (40) 6 2 3 4 (15) (15) (12) (20) ( 8) (55) 34 32 48 15 (129) TOTAL 62 44 (106) 50 25 (75) 181 0 - Original study area N - New study area beginning before April, 1985 ro --* COJ CO Co Co CO Or O) Table 2.D.2 Total Singleton Live Births Among All Eligible Pregnancies* EXPOSED UNEXPOSED TIME PERIOD _S1?Q.12(9) ____ i2Q..mm . (BOTH) Exposed: 5120.08(0) 5053.0204(N) (BOTH) TOTAL 1980 1981 (1980-81) 53 44 (97) 45 60 (105) (98) (104) (202) 63 64 (127) 25 28 (53) (88) 186 (92) 196 (180) (382) Unexposed: 1982 1983 1984 1985 (1982-85) 59 56 62 24 (201) 65 62 97 17 (241) (124) (118) (159) ( 41) (442) 42 71 70 23 (206) 45 48 57 15 (165) (87) 211 (119) 237 (127) 286 ( 38) 79 (371) (813) TOTAL 298 346 (644) 333 218 (551) 1195 0 - Original study area N - New study area 3 Beginning before April, 1985. Table 2.D.3 Definite Spontaneous Abortion Rates Using All Pregnancies* by Year and Census Tract IIME.KBISB _ Exposed: 1980 1981 1980-81 EXPOSED 5120.11(N) 13.1 24.1 (18.5) 8.2 9.1 ( 8.7) UNEXPOSED1 BOTH_______ ____ 5120.08(0)____ 5053.0204ff0 BOTH TOTAL (10.9) (16.1) (13.7) 0.0 13.5 ( 7.3) 7.4 22.2 (15.9) ( 2.2) 7.0 (16.4) 16.2 (10.0) (12.0) Unexposed: 1982 1983 1984 1985 1982-85 16.9 12.5 21.5 11.1 (166) 9.7 16.2 10.2 19.1 (12.4) (13.3) (14.5) (15.0) (14.6) (14.3) 17.7 12.4 19.5 14.8 (16.3) 11.8 4.0 5.0 21.1 ( 8.3) (14.7) 13.9 ( 9.2) 11.9 (13.6) 14.4 (17.4) 16.0 (12.9) (13.7) TOTAL 17.2 11.3 (14.1) 13.1 10.3 (12.0) 13.2 O -- Original study area N - New study area * Beginning before April 1985 SL 033357 1594.tb5/ll cn r* o ip ol Cd sore Table 2.D.4 Odds Ratios And 95% Confidence Intervals For Spontaneous Abortions in Exposed and Unexposed Areas by Year Using All Eligible Pregnancies9 Time period 1980 1981 (1980-81) 1982 1983 1984 1985 (1982-85) ORIGINAL STUDY AREAS (5120.12/5120.08) NEW STUDY AREAS (5120.11/5053.0204) m___________ iH-CI_____________ Qfi_____________ 95% Cl 20.18b 2.04 (2.88) 0.95 (1.14, 357.77) (0.83, 5.00) (0.30, 6.37) . (0.37, 2.46) 1.01 (0.38, 2.74) 1.13 (0.53, 2.40) 0.75b (1.03) (0.17, (0.64, 3.37) 1.66) 1.01b 0.35 (0.51) 0.80 3.88b 1.94b 0.89b (155) (0.20, 5.10) (0.11. 1.11) (0.20, 1.29) (0.26. 2.56) (0.95, 15.87) (0.56, 6.70) (0.20, 3.87) (0.82, 2.94) 9 Beginning before April 1985 Small sample adjustment 1594.tb6/ll Exposed: 1980 1981 (1980-81) Table 2.D.5 Stillbirths Aaong All Eligible Pregnancies* EXPOSED UNEXPOSED -CIWTHi taizo.QBto) 5U53.0204(N) --lMTHl .__ TOTAL 1 1 (2) 0 2 (2) (1) (3) (A) 1 0 (1) 3 0 1 (1) 3 (1) (1) (2) (6) Unexposed: 1982 1983 1984 1985 (1982-85) 0 0 0 0 (0) 1 (1) 0 (0) 1 (1) 0 (0) (2) (2) 0 0 (0) 1 0 0 (0) 0 1 0 (i) 2 1 0 (1) 1 (2) (0) (2) (A) TOTAL 1 5 (6) 3 1 (A) 10 0 - Original study area N - New study area Beginning before April 1985 SL 033359 1594.tbl/ll SL 033360 VARIABLE Any vs. no home cold tap water consumption Appendix E. Multiple Logistic Models of Risk of Spontaneous Abortion Model --1------------------ 2--------------3----------------- &5i_ 4.6 (.0005) 5.5 (.0005) 11.0 (.02) 4.2 (.009) 2.8 (.05) 3.7 (03) Mother's age: <20 20 - 34 35 + 2.7 1.0 2.9 (.001) 2.6 1.0 2.2 (.008) 4.3 1.0 1.8 (.008) 1.8 1.0 2.8 (08) 2.8 1.0 1.6 (08) 2.6 1.0 1.7 (17) Prior spontaneous fetal loss: 0 1 2+ 1.0 1.6 2.1 (-10) 1.0 1.6 2.4 (09) 1.0 1.8 3.0 (17) 1.0 1.9 2.2 (.20) 1.0 1.6 1.4 (.51) 1.0 3.8 4.8 (.0003) Ethnicity: White or Hispanic Other Alcohol consumption <4 drlnks/wk 4 or more drinks/wk 1.0 0.3 (.01) 1.0 1.5 (.46) NA 1.0 1.4 (.55) 1.0 0.4 (20) 1.0 2.3 (32) 1.0 0.1 <0l) 1.0 0.9 (.88) 1.0 0,1 (05) 1.0 1.3 (73) 1.0 0.3 (09) 1.0 0.6 (60) Any vs no probable solvent exposure at work 4.8 (.008) 3.2 (08) 6.7 (03) 2.2 (.40) 5.7 (.01) .001 (0.86) 4.6 (.0005) 2.6 1.0 2.6 (.003) 1.0 1.7 1.4 (24) 1.0 0.3 (.02) 1.0 1.7 (.39) 5.5 (.005) Smoking: none < 1/2 pack/day 1 or more packs/day 1 1.0 1.1 1.1 (.96) 1 1.0 1.1 1.2 (.92) Model 1-------------------h____________ 5____________ ____________ L 1.0 0.4 0.9 (.49) 1.0 1.9 1.0 (4) 1.0 0.8 1.1 (93) 1.0 0.9 2.0 (51) 1.0 1.1 1.0 (.94) Any vs no water filter used beginning around tim pregnancy began 0.4 (0.17) 0.3 (0.10) 0.6 0.53) 0.2 (0.17) 0.0008 (0.71) 0.4 (0.31) 0.4 (0.18) X9 0 rvj to-* 00 tfi Time period and census tract: 1980-81: original exposed (0E) new exposed (NE) original control (0C) new control (NC) 1982-85: original exposed (0E) new exposed (HE) original control (0C) new control (NC) 1.0 .17 .20 .90 .87 .51 .71 .38 (.03) 1.0 .17 .15 .55 .92 .SO .72 .40 (.09) 1.0 .36 .14 .78 .80 .61 1.08 .54 (25) (Rearrangement to compare exposed to unexposed areas) 1980-81 0E/0C NE/NC 5.0 6.7 7.1 0.2 0.3 0.5 1.0 .06 .12 .79 .79 .34 .38 .25 (10) 8.3 0.1 1.0 .23 .20 .56 .77 .30 .91 .29 (26) 5.0 0.4 1.0 .22 .31 .84 .57 .25 .94 .30 (.13) 3.2 0.3 1.0 .12 .25 .83 .52 .30 .41 .22 (.05) 4.0 0.1 1982-85 OE/OC NE/NC 1.2 1.3 0.7 2.1 0.8 0.6 1.3 1.3 1.3 1.1 1.4 1.0 0.8 1.4 Model 1----------------2--------------- 2--------------- k5I (N) SABs (N) Live births 92 88 39 52 44 40 87 858 723 535 322 858 858 734 Goodness of fit p-valve Hosmer-Lemeshow .63 .38 .73 .40 .94 .85 . .60 C.C. Brown .83 .95 .59 .46 .75 .92 .66 ap-valve NA - not applicable Model 1 Uses a single random eligible pregnancy per women; excludes women reporting IUD use at time of conception and women reporting using water from a private well. 2 Whites and Hlspanlcs only. 3 Only women reporting nausea during first trimester. 4 Only women reporting no nausea during first trimester. 5 Only pathology confirmed SABs. 6 Only SABs of less than 10 weeks estimated gestational age. 7 Excludes pregnancies conceived in 1980. SL 033362 1599.t2/12 3. Comparison of Original and Follow-up Fairchild Study Populations Background In 1983 interviewing was conducted for a study of pregnancy outcomes starting in 1980*1981 among women living in a residential area that had received drinking water contaminated wlch trlchloroethane (TCA). The contamination resulted from a leak in an underground storage tank at a nearby industrial plant. The exposed study area consisted of one of the two census tracts considered most likely to have received the contaminated water based on their proximity to the well in which TCA was detected. A control census tract was selected based on demographic similarity to the exposed tract. The events leading to the discovery of the contamination, the concentrations of TCA found in the contaminated well, and the selection of the study areas has been covered in previous reports (1). A comparison of pregnancy outcomes In the two areas showed that the exposed area had twice the rate of spontaneous abortion as the control area (21.5% compared to 11.0%). Subsequent analyses adjusting for potential maternal risk factors resulted in little change in the odds ratio determined from the unadjusted data. Zn 1986, a household enumeration was carried out for a follow-up study to determine whether the spontaneous abortion rate difference between the two census tracts had changed following removal of the contaminated well from service, and whether the second presumptively exposed tract had experienced a similar excess. Current residents of both exposed tracts and two control tracts were interviewed concerning pregnancies that occurred during the period covered by the first study (1980*1981) as well as during the subsequent time period (1982-1985). This design offered the opportunity to make several interesting comparisons, as follows: 1. Comparison of interview responses and pregnancy outcomes in women who were interviewed in both studies, using data from both studies. 3.1 033363 SL 2. Comparison of interview responses and pregnancy outcomes In women who were In both studies with responses and outcomes of new residents, using data from the new study only. 3. Comparison of Interview responses and pregnancy outcomes in women who were in both studies with responses of women in the first study who could not be located at the same address during the second study, using data from the first study only. HftEhgdt This report presents some results of the third comparison, raanalysis of the first Fairchild study results separately for women who had changed their place f residence and those who had remained at the same address. For ease of comparison with the original Fairchild results on the entire group, the same variables were used, and the population at risk was defined in the same way as In Fairchild I. Thus, eligible women had lived within their respective study cress for the entire pregnancy, the pregnancy began between January 1, 1980 and December 31, 1981, and ectopic pregnancies and pregnancies resulting in twins were excluded. Spontaneous abortions reported by the mother but not medically confirmed were Included. Of the 401 Fairchild Z pregnancies, 236 pregnancies occurred among women who had remained at the same street address and 165 among w men who had moved away or could not be reached for other reasons (refusals, unable to locate) and thus were not included In Fairchild II. Results Results of analysis show that the crude odds ratios comparing the spontaneous abortion rates in the two census tracts differ in the two groups of women, 2.79 (95% confidence interval 1.99, 3.59) for women who resided at the same address three years later, 1.74 (95% confidence interval 0.97, 2.51) for women not located at the same address three years later, (Table 3.1). However, the 95% confidence intervals overlapped, as shown above, indicating that the difference between the odds ratios was not significant at the 5% level. The spontaneous abortion rates were lower among women who had remained in both the exposed and the control areas (19.3 and 7.9 respectively) than among women who had left 3.2 Si. o33364 (24.4 and 15.7 respectively). The results were similar when the pregnancies at risk were redefined to exclude stillbirths, therapeutic abortions, and unconfirmed spontaneous abortions. The odds ratios and regression coefficients of the multiple logistic regression analysis are given in Table 3.2 along with the 95% confidence intervals for census tract. The distributions of each of the variables in the control and exposed census tracts are shown in Table 3.3. The main features of these tables are as follows: The adjusted odds ratio comparing census tracts was greater than 2.00 regardless of whether the women could be located at the same address three years later; however, it was somewhat lower and not statistically significant among the women who had apparently moved from the area. The significant effect of tap water and alcohol consumption on the spontaneous abortion rate that was found in the original study occurred principally among the women who later moved, and adjustment for these factors might be at least partially responsible for the lower adjusted odds ratio for census tract In this group of women. The finding of greatest interest was that in both census tracts the rate of spontaneous abortion was considerably lower among women located at the same address after three years compared to those women who had apparently moved away. 3.3 0^3 36^ l. Epidemiological Studies Section, California Department of Health Services: Pregnancy Outcomes in Santa Clara County 1980-1982: Reports of Two Epidemiological Studies. California State Publications Section Report 7540-958-1301-5. 3.4 Table 3.1 Total Reported Spontaneous Abortions by Population at Risk and Cehsus Tract Fairchild I Total Same residence three years later Not located three years later U> w Fairchild I Total Same residence three years later Not located three years later CONTROL NN (Pregnancies) ISAB) 210 23 127 10 83 13 203 21 125 10 78 11 _1_ 11.0 7.9 15.7 10.3 8.0 14.1 EXPOSED N (Pregnancies) N (SAB) 191 41 109 21 82 20 180 35 106 19 74 16 _1_ 21.5 19.3 24.4 19.4 17.9 21.6 SL 033367 Excluding therapeutic abortions, unconfirmed spontaneous abortions, and stillbirths. Odd Ratios 2.22 2.79 1.74 2.09 2.51 1.68 Table 3.2 Multiple Logistic Regress! n Results Comparing Vomen Who Stayed With Vomen Who Left Number of pregnancies Number with missing data Number in analysis Number of spontaneous abortions Rate per 100 pregnancies Census tract OR (exposed vs control) 95% confidence limits Lower Upper Risk factor Cold tap water** Previous pregnancies** Previous spontaneous abortions and stillbirths" Age of mother < 20 Age f mother > 34 Smoking (first trimester)** Alcohol (first trimester)** Non-Vhl^s Occupational exposure (first trimester) Month pregnancy identified** IQXAL 401 8 393 64 16.0 Same address Not located 3 vrs later 3 YJ8 latsr 236 165 53 231 162 31 13.1 33 20.0 2.35* 1.30 4.23 2.13* 0.88 2.49* 1.08 5.72 Coifficignta 1.20 0.83 2.05 0.83 5.04 4.10* 1.02 1.37 1.58 1.89 0.97 1.48* 0.70 1.03 0.69 1.71 0.00 0.64 0.36 1.37 0.55 1.66 0.67 0.78 1.36 4.47 1.20 1.80* 0.75 0.44 0.75 * Significant at tha 5% laval. b Logistic coefficients representing tha increase In relative risk for moving from one category of the variable to the next highest. Odds ratios for other variables. Notes: Cold tap water categories are the same as on Table 3.3. Previous pregnancies, previous spontaneous abortions and stillbirths, and month pregnancy identified are actual counts. Smoking and alcohol are as originally coded on the questionnaire. Occupational exposure is exposure to at least one of the following substances at least once a week during the first trimester of pregnancy: paint, lacquer or varnish; organic chemicals or solvents; gasoline, kerosene, fuel oil or other petroleum products; pesticides or herbicides: xrays or radioactive material. 3.6 033368 SL Table 3.3 Percent Distribution of Population Subgroups by Maternal Risk Factors Total number of pregnancies Home tap water consumption None 1*3 glasses/day 4 or more glasses/day Number of previous pregnancies 0 1 2 3+ Number of previous spontaneous abortions or stillbirths 0 1 2+ Age of mother Less than 20 20-34 35 or greater Amount of smoking None 1/2 pack/day or less One pack/day or more Alcohol consumption Never Once a month or less Several times/month About onee/week or more Ethnicity Non-White White Same address 3 Yrs-latcr Control Exposed 127 109 Not located 3 vra later Control Exposed 83 82 14.2 41.7 44.1 10.3 39.3 50.5 9.8 39.0 51.2 15.0 23.8 61.3 20.5 33.9 25.2 20.5 20.2 35.8 18.4 25.7 28.9 34.9 14.5 21.7 30.5 31.7 23.2 14.6 73.0 19.0 7.9 0.0 90.6 9.5 85.0 8.7 6.3 69.0 21.4 4.8 4.8 18.1 81.9 73.4 16.5 10.1 0.9 96.3 2.8 93.6 3.7 2.8 69.7 16.5 1.8 11.9 18.4 81.7 72.3 18.1 9.6 8.4 83.1 8.4 78.3 13.3 8.4 69.9 18.1 4.8 7.2 19.3 80.7 74.4 20.7 4.9 7.3 86.6 6.1 84.2 8.5 7.3 72.0 17.1 4.9 6.1 26.8 73.2 3.7 SL 033369 Table 3.3 (Continued) Month decided pregnent Firet Second Third or leter Oecupetionel exposure (first triaester) Same address 3 vrs Later Control Exposed N t located Control Exposed 31.5 55.9 12.6 33.3 60.2 6.5 41.0 53.0 6.0 28.4 58.0 13.6 3.9 3.7 6.0 3.7 Note: Percents ere based on totals minus missing values. 3.8 SL 033370 4. Cardiac Study: Follow-up Cluster Investigation BACKGROUND On November 17, 1981, it was discovered char a leak of toxic chemicals had occurred from a storage tank located about 2,000 feet from a well (Well *13) which supplied drinking water to nearby residences and industries. When Well #13 was removed from service on December 7, 1981, 1,1,1-trichloroethane (TCA) was found in this well at 1700 parts per billion (ppb). An epidemiologic investigation was conducted by the California Department of Health Services (CDHS) which found a 2.2-fold excess prevalence of severe cardiac anomalies in the area potentially served by the contaminated water supply for births occurring in the period January 1981 through August 1982 compared with births in the remainder of SCC (1,2). That study was not able to identify any sources of bias which could explain the excess prevalence seen. However, the relation between the leak and the prevalence of cardiac anomalies remained uncertain due to the temporal and geographic pattern of the eases relative to the leak (1,2). No excess was found for the succeeding time period, September 1982 through December 1983 (2). On the basis of these findings, the CDHS concluded that a follow-up study be conducted of this excess prevalence. This study was limited in time to the period, January 1, 1981 through September 6, 1982, and in space to children 4.1 SL 033371 bom in the seven census craec area served by Che vacer syscem in which contamination had been found. The objecCives of Che follow-up study reported here were: (i) to determine whether potential confounding factors other chan those examined in the earlier investigation (1,2) explained some the excess of cardiac anomalies; and (li) to determine whether the effect estimate for cardiac anomalies among children, based on a crude measure of exposure, Increased or decreased when exposure to the contaminated water supply was more accurately classified. In the earlier scudy, exposure was defined by place of birch; exposed cases were chose whose residence at time of birth was within the seven census tract area potentially served by che contaminated water supply, hereafter referred to as the study area. If exposure to che waCer supply had increased the risk, using more accurate exposure Information should result in an increased estimate of relative risk. In contrast, if the effect estimate was significantly decreased after correcting for misclassification, it suggests that che original measure of exposure was not indicative of exposure to the water supply or chat exposure was not related to risk. There are four ways misclassification with respect to exposure could have occurred: (i) cases called exposed were not actually exposed, e.g., mother did not live or spend time in the study area during the critical time period for cardiac development or mother did live in the study area but did not drink tap water; (ii) cases called unexposed were actually exposed, e.g., mother lived in the study area during the critical period of gestation but moved before giving birth or mother exposed regularly to the contaminated 4.2 water supply but lived outside the study area; (lii) non-cases called exposed were not actually exposed (same reasons as [1] above); and (iv) noneases called unexposed were actually exposed (same reasons as [ii] above). Another potential source of misclassification is the possibility that homes of ease and control mothers within the study area did not receive TCA contaminated water from Veil #13 during the relevant time period of gestation. This report compares maternal exposures from the study area water supply with maternal water exposures from the remainder of SCC for children born in the period January 1, 1981 through September 6, 1982 with and without severe cardiac anomalies. METHODS The study period corresponded to births occurring between January 1, 1981 through September 6, 1982. This time period was determined based on the critical period during gestation for cardiac development and the date of closure of Well #13. Cardiac development occurs from approximately the third to the seventh week post-conception. In the absence of knowledge of the metabolic characteristics of a particular exposure it is reasonable to assume that any exposure from conception through the seventh week of gestation is relevant. The well was closed on December 7, 1981, and it was assumed that exposure to the contaminated water would still be possible for a subsequent week, i.e., through December 14, 1981, to allow thorough transport of water through the system. If a woman consumed water and also 4.3 concived on December 14, 1981, her baby would be the last birth at risk for exposure. Her expected delivery date would be September 6, 1982 based on a 38 week gestation. Thus, September 6, 1982 represents the last date a child would be born that would have had potential in-utero exposure to the contaminated water supply. The follow-up study was case-control in design. Children with severe congenital cardiac disease (cases) diagnosed within the first year of life were ascertained by the California Birth Defects Monitoring Program among live births in SCC for the study period. A cardiac anomaly included a diagnosis of cardiac malformation (745-747.9, International Classification of Diseases-9) made by echocardiogram, catheterization, surgery, or autopsy. Excluded as cases were; septal defeats, patent foramen ovale, patent ductus arteriosus in a premature infant (less than 38 weeks), pulmonic stenosis, arrhythmia, or functional murmur, unless accompanied by congestive heart failure or any other cardiac anomaly. Septal defects were included if their diagnosis was made by invasive diagnostic procedures (surgery or catheterization). In the earlier study performed by the CDHS, all live births in the study area (N-2151) and in the remainder of SCC (N-36,592) were included to obtain prevalence measures of cardiac anomalies in presumptively exposed and unexposed areas. The ratio of these two measures yielded a 2.2 (95% CI-1.2, 3.9) fold excess of severe cardiac anomalies in the study area compared to the rest of SCC. For the follow-up study, a representative sample of all live births in the county was selected to serve as controls. The number of control births was approximately equal to the number of cases and was 4.4 SL 033374 obtained through random selection of vital records for all live births in SCC for the same time period as the cases. The study Identified 106 children born in the study period with severe cardiac anomalies. Mothers were found and interviewed for 83 (78.3%) cases. Of the 114 controls identified from the vital records files, 96 (84.2%) mothers were found and Interviewed. The primary reason that either a case mother or a control mother was not Interviewed was the inability to locate her after extensive tracing efforts. A new control mother was not substituted by resampling the vital statistics files when a control mother was not located. Resampling was not conducted because it was thought that such a strategy would bias the control group towards mothers who were more likely to have remained at one address. * Interviews were conducted over the telephone or ln*person when necessary. Zn an effort to obtain information on all maternal water exposures, the mother of each case or control was queried on her residence address(es), location of her place of employment, and places she frequently visited during the first trimester of her pregnancy with the index child. She was asked about her consumption of both tap and bottled water for the same time period. She was also asked about the frequency and duration of showering and bathing both at home and at places she frequently visited. In addition, each mother was asked about numerous other potential risk factors for congenital cardiac anomalies such as alcohol consumption, education, employment, age, nausea, race, cigarette smoking, diabetes, epilepsy, and family history of cardiac disease. 4.5 31* sv This report presents effect estimates for cardiac anomalies and first trimester maternal exposures to water (drinking, bathing, or showering) in the study area relative to the rest of SCC. Analyses of maternal water exposures to the study area and to the rest of SCC were also performed with adjustment for a number of potential confounders. The measure of association used was the odds ratio (OR). Confidence intervals (95 per cent) were calculated using Epistat software (3). When an observed value of an individual cell was less than five, the small sample size adjustment procedure according to Cart (4) was used. The summary procedure according to Woolf (5) was used to adjust for potential confounders. Also presented in this chapter is a comparison of cases and controls within the study area with respect to estimated TCA exposure via water. TCA exposure for cases and controls in the study area was estimated using hydrogeologic modelling methods (6). These data were obtained for 1981 from the Great Oaks Water Distribution Study conducted as part of the CDHS effort to follow-up on the industrial leak of 1981. A detailed discussion of that study can be found elsewhere (6). Briefly, the water distribution study provided estimates of the probability of receiving water from Well #13 for each month in 1981 by small areas within the water distribution system and the estimated concentration of TCA in water from Well #13 each month. With this information, a cumulative maternal TCA exposure score for the first trimester was obtained for cases and controls whose birth address was located within the study area. The seven census tracts in the study area were ranked on the estimated exposure to Well #13 for each month. A census tract was ranked 1 if it received the largest proportion of Well #13 water 4.6 033 3l6 S*> and 7 if it received the smallest. Tied rankings were split. Data from the Great Oaks Water Distribution Study Report (6) were extracted to estimate the TCA concentration in Well #13 for each month of 1981. By multiplying the rank by the estimated TCA concentration in Well #13 for the appropriate month it was possible to derive a monthly estimated TCA exposure score for each census tract. To obtain a TCA exposure score for each case and control in the study, a census tract was assigned to each birth address and was assumed to reflect the first trimester address. The sum of the estimates for the first three months of pregnancy for the census tract corresponding to the child's birth address represented the cumulative TCA exposure score for the first trimester. Because the number of control mothers who had first trimester residence in the study area during the study period was small (N-6), an additional 19 live births were randomly selected from the SCC vital records file to serve aS controls. This allowed for a case to control ratio of approximately one to five. RESULTS Interview Data The OR for cases bom inside the study area versus those bom outside the study area was 2.0 (0.47, 5.2), as is shown in Table 4.1. This effect estimate is based on all study subjects whether they were interviewed or not. The value of 2.0 is comparable to the relative risk of 2.2 found for the original investigation which was calculated by taking the ratio of the prevalence of cardiac anomalies among live births in the study area 4.7 SL 033377 (5.6/1000) and cha prevalence of cardiac an aalies among live births in the rest of SCC (2,6/1000). Overall, 78.3 per cent and 84.2 per cent of ease mothers and control mothers, respectively, were Interviewed for this study. Of the mothers of 12 case children born in the study area, nine (75.0%) were located and interviewed. Of the 94 case mothers whose child was born outside the study area, 74 (78.7%) were interviewed, With respect to control mothers, six (85.7%) of the seven controls bom in the study area and 90 (84.1%) mothers of the 107 control children bom outside the study area were Interviewed. Thus, the percentage of mothers found by area and case (case or control) status was similar. An examination of cases and controls by maternal residence during the first trimester (the approximate critical gestational time period for cardiac development) indicated chat address of birth was a fair indicator of first trimester residence in the study area. Using first trimester address instead of birth address, the OR for cardiac anomalies and maternal first trimester residence in the study area was 1.4 (0.47, 3.9). Table 4.2 displays the distribution of cases and controls by location inside or outside the study area based on maternal residence during the first trimester. For the nine case mothers who had children bom in the study area, eight had resided in the same area during the first trimester and one had resided elsewhere during the first trimester. No case mother who had a child born outside the study area was found to reside in the study area during the 4.8 SL 033378 first trimester. For the control mothers, all six mothers who had a child born within the study area also lived in the area during the first trimester. In addition, one control mother who had a child born outside the study area had lived in the study area during the first trimester. Thus, the change in OR from 2.0 for cardiac anomalies and birth address in the study area to 1.4 may be explained by one ease based on birth address being removed from the sttfdy area (misclassification of 11.1%) and one control mother being added to the study area (misclassification of 16.6%). Maternal water exposures were examined to correct for possible misclassification of exposure among women who drank no tap water. Comparing mothers who lived in the study area during their first trimester and who drank at least one glass of tap water per day to mothers who lived outside *> the study area for their first trimester and drank at least one glass of tap water per day revealed an OR of 2.3 (0.61, 8.4). Thus, correcting this source of misclassification increased'the OR from 1.4 to 2.3. These data are shown in Table 4.3. The association between residence inside the study area and consumption of tap water was also examined in the presence of a number of potential confounders, namely maternal alcohol consumption, education, employment, age, nausea, race, cigarette smoking, diabetes, epilepsy, and family history of cardiac anomalies. The adjusted ORs were all generally lower than the unadjusted OR of 2.3 and are shown in Table 4.4. This was particularly true when the OR was adjusted for maternal education level (OR^^-1.5, 95% CI0.44, 5.4). 4.9 SL 033379 A somewhat elevated OR (OR-2.0 95% Cl- 0.58, 6.7) was also found for cardiac anomalies and maternal residence In the study area and showering or bathing more than 50 minutes per week compared to maternal residence outside the study area and showering or bathing for the same length of time (Table 4.3). Among the 18 case and control mothers who reportedly frequented (defined as time spent within SCC at least twice a week or for one week or pore) areas within SCC other than at home or work during the first trimester, four case mothers (none of whom resided in the study area) and one control mother reported visiting the study area, and seven case mothers and six control mothers reported visiting a location outside the study area (Table 4.3). The OR for cardiac anomalies and frequent visits to the study area was 2.6 (0.31, 21.8). The OR for cardiac anomalies and maternal frequent visits to the study area and water consumption (any tap water versus none) was 1.4 (0.15, 13.2). Case and control mothers were also asked whether they regularly showered or bathed at a place in SCC other than at home. No case or control mothers reported showering or bathing regularly in the study area. Similarly, no ease or control mother reported their work location during the first trimester to be in the study area. Water Distribution Study Data As outlined above, the first trimester exposure to TCA in drinking water was estimated by ranking cases and controls based on their census tract of 4.10 033380 birth. This analysis was not limited to the cases and controls whose mothers were interviewed, but rather included all cases and controls identified for the study. Therefore, for this analysis there were 12 cases and 7 controls born in the study area. In addition, 19 births were randomly selected from the SCC vital statistics file of live births to also serve as controls. This provided 12 cases and 26 controls for which to obtain estimated first trimester TCA scores. If the first trimester of a case or a control was in 1980, we were unable to assign a score due to the fact that TCA data were unavailable for 1980 from the Great Oaks Water Distribution Study (6). However, based on the hydrogeologic modelling it is unlikely that there was any TCA exposure in 1-980 (6). Of the 12 cases born in the study area, we were able to assign a score to five. Of the 26 controls, we were able to assign a score to 25. It should be noted that the 19 additional controls were chosen so that their first trimester would have been in 1981 so that a score could be assigned. The mean TCA score for cases was 416.4 and for controls was 586.4. The lower score seen for the cases is an indication that cases were somewhat more likely to have been exposed to TCA. These data are displayed in Table 4.5. Because we observed a mean difference in scores based on birth address, we also looked at TCA scores exclusively for those cases and controls for which interview data were available on residence during first trimester and consumption of tap water in the study area. However, this analysis was limited to only six cases and four controls. For the ease mothers, two lived in the study area and drank tap water and four frequently visited the study area and drank tap water. For the control mothers, three lived in the 4.11 033381 SI* study area and drank tap water and one frequently visited the study area and drank tap water. The mean TCA scores for cases and controls for this analysis were, 271.4 and 473.3, respectively. These data are shown In Table 4.6. Again, the lower average score for cases compared with controls suggests more TCA exposure for case mothers during the first trimester. DISCUSSION In an earlier report (2), It was concluded that the temporal distribution of cardiac cases was some evidence that the leak Into Veil #13 Identified In 1981 was unlikely to have accounted for the entire excess observed In the study area. Vhlle the time of onset of the leak In Uell #13 Is uncertain, solvent levels must have Increased between the time of Initial leak and Its discovery In December 1981 over a period of many months. If this contamination was responsible for the observed excess of cardiac anomalies, one might expect to see the number of cases among full-term births increasing throughout 1981 and into the summer of 1982, l.e., toward the end of the study period. However, the 10 case children born In the study area in 1981 occurred at a frequency of approximately one per month. In addition no children with severe cardiac anomalies (all of whom were full-term births) were born in the study area between May and December 1982 (2). If the leak was the explanation, we would have expected to observe a similar frequency of children with cardiac anomalies In the first nine months (until 9/6/82) of 1982 as was seen in 1981. Since nine cases, who had first trimester exposure to the study area, occurred In the 12 months of 1981, one would have expected to observe approximately seven cases in the first nine 4.12 03^ months of 1982. However, only two cases (both In April) occurred in Che scudy area at this time. The absence of cases betv en January 1, 1982 and September 6, 1982 was statistically significant (p-0.03). Thus, even though the prevalence of cardiac anomalies was higher among 1981 births in the study area than the rest of SCC and vas not higher for 1982 and 1983 births (a pattern which might be argued as consistent with the leak), the drop in the prevalence of cardiac anomalies in the study area began four months before one would have expected to observe a decreased prevalence based on the time of closure of Veil #13. The present study suggests that the findings from the earlier CDHS study of a 2.2 fold excess prevalence wore not due to confounding bias or severe * errors in classification of exposure to the contaminated water supply. Slightly elevated ORs were identified in this study for cardiac anomalies and maternal residence in an area of SCC which potentially received contaminated drinking water in 1980 and 1981. These ORs were found for case mothers who lived or frequently visited the area and drank tap water and for ease mothers who lived in the area and showered or bathed more than SO minutes per week. However, all of the elevated ORs Identified had associated confidence Intervals which Included 1.0, which suggests that sampling variability is a possible explanation of those results. It is Important to emphasize that because these results are based on small numbers, any error in reporting or recording exposure status could severely influence the effect 4.13 SL 033383 estimates found. This is particularly true for those cases and controls whose residence was in the study area. Further, the effect estimates could also be influenced if the mothers of case and control children not located for interview were different from those mothers chat were interviewed with respect to residence in the study area or tap water consumption. There were three case mothers and one control mother who had a child bom in the study area and were not interviewed in this study. If the three case mothers were tap water drinkers and lived in the study area during the first trimester, and the one control mother was not a tap water drinker, then the OR would be 3.2 rather than the OR calculated of 2.3. In contrast, if the one control mother was a drinker and the case mothers were not then the OR would be 1.8. Based on the estimated TCA concentration in the study area water, this study found some evidence to suggest that case mothers were somewhat more likely to have had TCA exposure from their drinking water than control mothers at a time period in gestation critical for cardiac development. The limitations of this estimation process have been described elsewhere (6). 4.14 SL 033384 References 1. Epidemiological Studies Section, California Department of Health Services. Pregnancy outcomes in Santa Clara County, 1980*1962: Reports of two epidemiological studies. California State Publications Section. Report 7540-958-1301-5, 1985. 2. Swan SH, Shaw CM, Harris J, et al: Congenital cardiac anomalies in relation to water contamination: Santa Clara County, California. 198183. Submitted for publication 1988. 3. Gustafson TL: Eplstst statistical package for the IBM personal computer, version 3.0. 1984. 4. Fleiss JL: Statistical methods for rates and proportions. Second Edition. Hew York. 1981. 5. Woolf B: On estimating the relation between blood group and disease. Ann Hun Genet 1955; 19:251-3, 6. Murphy P: The Great Oaks water distribution study. 1988. 4.15 033385 SL Table 4.1 Cases and Controls Bom Inside and Outside the Study Area* Place of Birth Relative to Studv Area Inside Outside *Referent group. Cases 12 94 Controls 7 107 _QR_ 2.0 __ a 95% Cl (0.47, 5.2) 4.16 SL 033386 Table 4.2 Cases and Controls By Haternal Residence During First Trimester Relative to Study Area Residence During First Trimester Relative tff Stydy Area------------------- Inside Cases 8 Outside 75 *Referent group. Controls 7 89 OR -25LSJ , 1,4 (0.47, 3.9) 1.0a 4.17 Si 33387 Table 4.3 Maternal First Trimester Residence, Employment, or Frequent Visits to the Study Area Compared to SCC Residence Inside jWfly Al<?a____ Outside Study Area Maternal Exposure Cases Controls Cases Controls OR 95% Cl Resided & drank tap water Resided & showered/ bathed >50 minutes per week 7 7 Frequently* visited Frequently visited & drank tap water 4 4 3 72 76 2.3 (0.61, 8.4) 4 - 60 71 2.0 (0.58, 6.7) 1 7 6 2.6 (0.31, 21.8) 1 7 3 1.4 (0.15, 13.2) Mothers who frequently visited the study area were not the same mothers who resided and drank tap water in the study area. 4.18 033388 SL Table 4.4 Odds Ratios for Cardiac Anomalies by Home Tap Water Consumption Inside Versus Outside Study Area After Adjustment for Potential Confounders Variable OR __ Woolf Alcohol 2.2 Education 1.5 Employment 2.3 Maternal Age 1.9 Nausea 2.1 Race 1.7 Smoking 2.0 Family history 2.2 Income 2.0 Maternal diabetes 1.7 Maternal epilepsy 2.0 Sex of child 2.0 95% Cl (0.63, 7.5) (0.44, 5.4) (0.64, 8.3) (0.29, 12.4) .(0.58, 7.4) (0.44, 6.3) (0.57, 7.2) (0.62, 7.6) (0.60, 6.5) (0.48, 6.3) (0.55, 6.9) (0.54, 7.3) Hgoemnoe*ity Chi souare 0.25 2.4 0.03 3.5 0.37 2.7 2.0 0.03 0.12 0.02 1.2 1.1 D 2 0.88 3 0.49 1 0.87 2 0.17 1 0.54 2 0.26 2 0.37 1 0.95 2 0.94 1 0.90 1 0.28 1 0.29 Unadjusted OR 2.3 (0.61, 8.4) Any tap water consumed inside study area at home residence versus any tap water consumed outside study area at home residence during the first trimester. 4.19 SL 033389 Table 4.5 Estimated TCA Score For Cases and Controls Born In the Study Area No. No. with scores assigned Sum of score Range Mean score Cases 12 5 2081.8 4.0 - 990.0 416.4 gputrala 26* 25 14661.2 4.0 - 1385.0 586.4 19 additional llva birth* vara salaetad from vital *tatl*tic* fllea to give an approximate 1:5 ratio of cases to controls. 4.20 SL 033390 Table 4.6 Estimated TCA Score for Cases and Controls Where Maternal Drinking Water Exposure* Was Obtained By Interview No. No. with scores assigned Sum of score Range Mean score Cases 11 6 1628.5 4.0 - 990.0 271.4 Santrgls 5 4 1893.0 37.0 - 990.0 473.3 "Mother resided or frequently visited the study area and drank tap water during her first trimester. 4.21 SL 033391 5. Material Maas Balance Analysis Introduction The California Department of Health Services (DHS), in collaboration with its contractor, the California Public Health Foundation, has been conducting several epidemiological and environmental studies in Santa Clara County, California. The epidemiological studies were undertaken as follov-up studies to previous Investigations of the possible reproductive health effects of drinking water potentially contaminated by several organic chemicals, including low levels ( 11 ppb) of freon, 1,1 dlchloroethylene (DCE), and high levels (in the range of 2000 ppb) of 1,1,1 trichloroethane (TCA). For TCA, the California State Action Level and the proposed Maximum Contaminant Level (MCL) for drinking water standards Is 200 ppb. The environmental studies have been performed primarily to provide exposure data for the epidemiological studies. For the latter studies, investigators need to know, as accurately as possible,^who was exposed, to what chemicals, and when the exposure began. This report presents the findings of one of those environmental studies. Background Xn late November 1981, Fairchild Semiconductor Corporation (formerly Fairchild Camera and Instrument Corporation) discovered that one of Its two underground tanks storing solvent wastes at their South San Jose facility had failed, allowing the stored waste to leak into the surrounding soil. The mixture of vastes stored in the tank were regulated according to California and federal hazardous waste management laws. Fairchild notified appropriate local and state regulatory agencies and the Great Oaks Vater Company (GOUC) about the leak. GOVC was the owner and operator of a municipal water supply well (GOVC #13) which was approximately 2000 feet downgradlent from the failed tank. The GOUC distribution system serves groundwater to seven census tracts in Santa Clara County. Fairchild also requested of GOVC to conduct water quality tests on well #13. GOVC agreed and shut off well #13 from the rest of the vater distribution system. Veil #13 was never returned to service within the system. 5.1 SL 033392 Fairchild did not know tha axt nc f cha anvlronaantal contamination resulting from tha laak because tha time that tha leak began and the exact mixture and amount of chemicals which had leaked from tha tank were unknown. However, TCA, acetone, xylene, isopropyl alcohol (IFA), freon and other proprietary mixtures of solvents were used by the company. Fairchild did not know exactly which of these chemicals leaked from the failed solvent waste tank because only a portion of the solvent wastestream was collected in the tank, while other solvent wastes ware collected in drums. The initial water quality testing performed by Fairchild found 1700 to 1800 ppb of TCA in well #13. A few days later samples were taken for water quality testing on all GOVC wells. No other wells shoved contamination except well #8 which had 17 ppb of TCA. The contamination of well #8 was considered to be related to chemical releases from a nearby IBM facility. The concentration of TCA at well #13 had Increased to 5700 ppb. Continuous testing over the next four months of well water showed TCA concentrations ranging from 2800 to 8800 ppb in Veil #13 and from 0.1 to 27 ppb In well #8. Veil #8 was removed from COVC service in March 1982. Freon was detected in well #8 at 25 ppb In April of 1982. Freon and DCG {a degradation byproduct f TCA and a contaminant In the manufacture of TCA) were detected in well #13 from April to July 1982 In the range of 3-11 ppb for each. Residents of the neighborhoods in the vicinity of the contaminated well *13 became very concerned over the possible health effects of drinking water contaminated with these Industrial chemicals. Some residents were particulary concerned about spontaneous abortions and cardiac defects in nearby homes. In response to these concerns, DHS conducted two epidemiological studies of pregnancy outcomes and cardiac defects in the area to address the problem. The results of these studies were reported in January 1985^. The authors used 1980 as the beginning of potential exposure because of the uncertainty of the date of onset of the chemical release. Following the discovery of contamination, the California Bay Area Regional Water Quality Control Board requested Fairchild to determine the extent of Che chemical leakage from the failed tank. Fairchild conducted a study and prepared a report of their findings entitled "Fairchild South San Jose 5.2 SL 033393 2 Material Balance Report May 11, 1982." A copy of the report Is Included as Appendix A of this report. The Introductory paragraph of the report states, "This material balance report for solvent materials at Fairchild's South San Jose facility was prepared at the request of the Regional Water Quality Control Board. . Fairchild has fully cooperated with the Board and other Interested public agencies, and has attempted to determine as precisely as possible the extent of the discharge of solvent materials into the ground at the South San Jose facility following the failure of a solvent waste storage tank."^ The Introduction to the report further states, "The results of this study are based on an exhaustive review of current data, and support the view that solvent balance measurements are, at best. Imperfect and that a 4 variance of up to ten percent (10%) is to be expected." The study report concludes that, "the maximum number of gallons of solvent which may have been lost appears to be 58,394 and "The data, therefore, tends to indicate that the solvent waste tank failed,'at the earliest, after mld-1980."^ DHS regarded the Fairchild Material Balance Report (MBR) as a potential source of information regarding environmental exposure data naeded for the interpretation of their epidemiological studies. However, it was necessary to analyze this report with respect to completeness and accuracy before it could be relied upon. The goal of the present analysis is to qualitatively assess the accuracy of the report's conclusions regarding the date the leak began. It is not intended to be a complete audit of the Fairchild report or a precise estimate of all the chemicals lost. Given these limited goals, DHS allocated $10,000 and nine months to complete the review. The date the leak began is particularly important for interpreting the results of the follow-up epidemiological studies of pregnancy outcomes. It helps to define the period in which residents were potentially exposed to chemicals found in the drinking water. Determining the amount of chemicals spilled is also of interest because It Is an important factor in Fairchild's assessment of when the leak began. 5.3 SL 033394 Determining whet chemicals were spilled is important because the epldeaiological studies are concerned with the reproductive effects of the exposure. Prior to conducting the first epidemiological study, a review of animal data was performed. No study found TCA to be teratogenic in laboratory animals. The toxicological data for DCE was similar. The researchers reviewed the toxicological data for freon, xylene, and IFA as well, and concluded that these chemicals were not causes for concern. However, glycol ethers, which are widely used in the semiconductor industry (including Fairchild), are of concern since they are known reproductive toxins. Unfortunately, no testa to determine the presence of glycol ethers in well water were conducted at the time the other chemical contamination was identified. The reproductive effects of mixtures or other unmeasured chemicals are, similarly, unknown. METHODS Material Balance Analysis A material balance analysis is commonly used in engineering studies to assess the inflow and outflow of substances through systems. Solvents are the substances of concern in the Fairchild analysis because the waste storage tank that failed was specifically intended for solvent waste. Therefore, the balance equation in the case of the Fairchild spill Is: Solvents purchased and used in the system should equal solvents "used up" in processing (primarily evaporation) plus solvent wastes disposed of out of the system. The difference between the estimated inflow and outflow of materials is considered an Imbalance and potentially lost via the failed tank. Chemicals were brought into the Fairchild South San Jose facility in two ways. They were either sent to that facility from the Fairchild Mountain View facility's Chemical Stores Department or they were purchased directly from chemical vendors. An unknown amount was "used up" or evaporated during the manufacturing process. Solvent wastes were collected in either 55 gallon drums or one of two bulk solvent storage tanks until they were shipped for disposal. Review of the Fairchild Report The first step in our analysis of Fairchild's report was t review the report to understand how Fairchild arrived at their conclusions regarding the date the leak began. We needed to know how they calculated the total volume of solvents brought into the facility, the volume of solvents lost through evaporation, and the volume of solvent wastes transported for disposal. the Original Pats Set The next step was to try to review the original data set used for the MBR. While trying to locate this data set, we learned that the MBR was not prepared by Fairchild staff, but was actually prepared by the law firm of Fillsbury, Madison, and Sutro. From what we were able to learn, data for the report was gathered by Fairchild staff and turned over to the attorneys for analysis and the production of the report. We contacted Fillsbury, Madison, and Sutro to request the original set of data they used to produce the report. We were told that all of the data had been returned to Fairchild, and that their office no longer had any data pertinent to our investigation. Review of the Fairchild Documents Since the original data set was not available, we reviewed the documents from which the original data set had been gathered. Fairchild produced hundreds of thousands of documents for litigation purposes. It should be noted that the overall production of documents for litigation took several years, and that potentially more documentation was available to DHS than to the preparers of the MBR. These documents were Indexed, boxed, and held in storage. Fairchild allowed us to review the documents at the San Francisco offices of their attorneys, Landels, Ripley, and Diamond (LKD). We began by reviewing the index of these documents in order to identify and request those documents which we felt would be most useful for our purposes. We then reviewed thousands of pages of documents included under the categories of: plant construction, maintenance and operations, chemical usage, government communications, safety and training, corporate safety, management and organization, security logs, remedial program, testing and environmental 5.5 <,V policy, insurance, and specific Fairchild managers' files. We also reviewed che deposition of a Fairchild employee familiar with operations. From our review of these documents, ve were able to obtain most of the solvent inflow (including the solvent Issues from Mountain View and the solvent purchases from chemical vendors) and the solvent outflow (waste collected and transported for disposal) data. We did not obtain sufficient data to be able to analyze the extent to which chemicals are used up or evaporated during the manufacturing process. We learned that the failed solvent waste tank was equipped with a gauge to measure the tank's fullness. We were Interested In reviewing the logs of these tank gauge readings. The data contained in these logs is an important source of information regarding the rate at which solvent wastes enter the tank. The log is also a record of the amount of solvent which was in the tank at the times it was emptied for waste disposal. Additionally, an analysis of the fill rate of the tank could have shown when the typical fill rate was Interrupted, as in the case of a leak. Fairchild's attorneys provided OHS with one key piece of information regarding glycol ether. Landels, Ripley, and Diamond wrote in a letter to Dwight R. Hoenlg of the DHS Toxic Substances Control Division (TSCD) dated December 13, 1984 that Fairchild used a positive photoresist which was a proprietary blend of 70% cellosolve acetate, 10% N-butyl acetate, and 5% xylene. The cellosolve acetate is primarily glycol ether. We hoped to document the quantity of positive photoresist used at Fairchild in order to estimate how much glycol ether might have passed through the failed waste solvent tank. There were ocher Fairchild documents and depositions made available to us that could have been useful which we could not review because of time and financial constraints. Overall, however, we feel that we were able to review most of the relevant available data. 5.6 033397 Review of Data from Other Sources Ve pursued Cher s urces of data to validate and/or augment the data gathered from the Fairchild documents. Ve considered validating Fairchild's chemical purchases data by contacting their chemical vendors, however, this task would have been too time consuming for the limited scope of our study. Ve attempted to validate waste manifests in two ways. Ve made a request to the TSCD in Sacramento for their copies of all manifests for the Fairchild South San Jose facility. Ve also contacted the waste transport companies for copies of the same manifests. Ve contacted the California Bay Area Air Quality Management District for any information that they had pertaining to the evaporation of solvents used by Fairchild. Ve also pursued another approach to help determine the time that the leak began. Learning more about the mechanism of the tank failure might lead us to some clues as to when it failed. Ve looked at the failed tank which is held in storage in San Jose, and found that many square cuttings had been removed from the tank. In talking to various people knowledgeable about the incident and involved in the litigation that stemmed from it, we learned that some (if not all) of the parties to the litigation had had an analysis done of the tank. A physical and chemical analysis of the failed tank could possibly provide the most important data on the timing of Its failure and leakage. Therefore we considered several options to find out more about the tank construction specifications and what caused it to fall. Ve attempted to get voluntary cooperation from the tank manufacturer, Owens Coming Fiberglas (OCF), for their evaluation of the mechanism and the timing of the failure. Ve asked all the parties to the litigation what they knew about the mechanism of the tank failure. Ve also asked them if they had done an analysis of the tank, and if they would give us the results of any such analysis. Further, we considered performing a physical and chemical analysis of the tank ourselves and contracting a consultant for an interpretation of that analysis. The many agency and private contacts ve made in our data starch are listed in Table 5.1. A summary of the data which we attempted to obtain that which we did obtain is presented in Table 5.2. 5.7 033398 SL RESULTS Review of the Fairchild Material Balance Report (MBR) OHS staff reviewed the Fairchild MBR and found errors, omissions and contradictions which raise questions as to the overall accuracy of this analysis and report. For example, after checking the calculations and figures presented in their tabulations of chemical inflow and waste outflow (Appendices A, B, and C), It is evident that arithmetic errors were made. Soma examples follow: Chemical laflw Esiqri: 1) Using Fairchild's own figures In Appendix A, we calculated the total solvent purchases in 1979 to be 18,148 gallons, 172 gallons more than Fairchild's total of 17,976 gallons. This appears to be due to the omission of 172 gallons for "3,5,1 pre-mix" from the overall total for the year. Similar arithmetic errors exist throughout the report. Despite this, the error in the five-year total of solvent purchases is negligible because the errors "balance out." 2) On page A.2 (solvent summary for 1978), the amount of solvent purchased for November and December are exactly the same for each of the nine solvents listed. Although this is possible, we suspect that it may instead be a clerical error. (See Notes, 1) 3) On page 4, methyl alcohol is listed as a "solvent material...used at Fairchild's South San Jose facility and...accounted for in this material balance analysis." In reality, in the solvent summaries which follow (Appendix A), Fairchild failed to list purchases of methyl alcohol. (See Notes, 2) 4) On page A.5, there is an asterisk attached to the column total for "photoresist" materials. The column has listed as its total 4866 gallons; it actually adds to 3143 gallons. If the asterisk is intended to refer to an explanation for this discrepancy, none is included in the report. 5) Fairchild did not include any data to indicate the amount of solvent in inventory at the time of the MBR. Inclusion of solvent inventory would have served to reduce the solvent "imbalance". 5.8 399 033 Waste Outflow Errors: 6) On page B.2, there are two errors In the addition of the number of drums picked up for disposal. In the 1979 figures, the total number of drums should be 75. The total volume of drummed waste (4,125 gallons) is correct, however. In the 1980 figures, the total number of drums should be 636, and the total volume of drummed waste for that year should be 34,980. 7) On page 2 of the MBR, it is stated that "Small amounts of vapor are inevitably emitted from solvents used in the production process, and these amounts are mixed in with the air. This air from the facility is then scrubbed to purify it before it Is emitted Into the atmosphere." However, the scrubber effluent is not included anywhere in the MBR. (See Rotes, 3) 8) The MBR contains conflicting information concerning the criteria for emptying the tank. Was it regularly emptied, or was it emptied only whan it was full? On this issue, the report contradicts itself on pages 6 and 7. On page 6, the report states that the tank was emptied when it reached capacity, whereas on page 7, it states that solvent in the waste tanks and drums is retained until a "removable volume" is accumulated. Furthermore, there Is no explanation In the report of how the tank's fullness was measured. 9) In mid-1980 (which is the "estimated" time of the start of the leak), there was a sudden change in the method of solvent disposal. Before mid-1980, the majority of solvent wastes hauled for disposal were from the bulk storage tank. After mid-1980, the majority of solvent wastes were disposed in drums. There is no explanation offered in the MBR for this change. 10) More important for this analysis, the amount of waste hauled away appears to have changed dramatically in mid-1980. Before mid-1980, there were bulk disposals of: 5040 gallons (18 December 1978), 3800 gallons (May 1979), 4958 gallons (3 December 1979), and 4500 gallons (4 May 1980). After that date, the amount of solvent disposed of in bulk is an order of magnitude smaller, eg. 500 gallons, 700 gallons, etc., a radical departure from past practice. There is no explanation. Furthermore, there should be separate disposal figures for the large 5.9 0 ("failed") and small solvent tanks after the small tank was Installed in mid*1980. (See Notes, 4) Review of the Fairchild Documents Chemical Inflow Because the original data set used for the MBR was unavailable, we began our analysis by reviewing thousands of Fairchild documents produced for litigation. To conduct this review, we attempted to locate the data items used in producing the MBR. The MBR states: "the Material Mass Balance data for this report has been developed from the following sources: 1. Records of chemical issues to the San Jose Facility from the Fairchild Mountain View Central Stores and South San Jose purchase records. 2. Records of solvent waste removal manifests for shipment to an approved disposal site. 3. Empirical measurements of ingoing and outgoing solvents taken during the first quarter of 1982."6 The MBR states that for the first several years (1977-1980) that the Fairchild South San Jose facility was operating, chemicals were shipped to that facility from Fairchild's Central Chemical Stores Department in Mountain View. This practice was stopped at an unspecified time in 1981. After that time, chemicals were purchased directly from vendors. The chemical shipments from Mountain View, referred to as "chemical issues," were tabulated on a weekly computer-generated report, We found, reviewed, and tabulated solvents from these reports, titled "Stores Weekly Change Reports," for 1978 (except November and December, which we did not locate), 1979, 1980, and January * March 1981. We did not find or review any chemical issues data for 1977 or for later than March 1981. 5.10 St 0331*01 In addition to the chemical lasues, we reviewed hundreds f purchase order releases which documented the order of chemicals from vendors and the receipt of them at the South San Jose facility on Bernal Road. Solvent purchase data were entered directly from the Fairchild documents Into a computerized database management system for editing and analysis. The data Items entered Into the database are included in Appendix B of this report. Duplicate entries were removed after matching on purchase order release number. About 301 of the documents we inputted were duplicates. Ve then used the duplicates to check the quality of our key-entry, and found very few errors. Ve also edited all of the solvent names, removing all of those known to us to be non-solvents which were Incorrectly put in the database. Ve checked the quantity of each solvent entry, and converted all known quantities to gallons. Unknown quantities were left as zero. Ve then grouped the solvents according to the eleven categories in the MBR, but found some solvents which we could not categorize. Upon requesting help in categorizing these remaining solvents, we were advised by Fairchild attorneys that it was unlikely that we would be able to get the necessary information to accurately finish the grouping. This is due to the fact that the Fairchild South San Jose facility has not been operating for several years and the remainder of the Fairchild facilities were sold by their former owners, Schlumberger Technology Corporation, last year. Therefore, former employees most likely to know how solvents were grouped for the MBR are no longer available. For this reason, we have presented our tabulation of the solvent issues and purchases by month and year, without regard to the solvent groupings, in Table 5.3. Ve could not estimate the quantity of glycol ether Fairchild, used because we could not accurately identify all solvents categorized as positive photoresist purchases in 1979-1982. Ve also found issues of ethylene glycol among the 1978 chemical issues. (See Notes, 5) Waste Outflow The waste outflow data we used for our analysis were also found among the Fairchild documents. Fairchild had copies of two types of waste removal 5.11 SL 033402 records. The first type, s Vests Hsuler Record, is the predecessor to the currently used fora, s Hazardous Waste Manifest. Current federal and California state law require the use of the Hazardous Vasts Manifest to document the removal, transport, and disposal of hazardous wasta from the waste's origin to its final disposal site ("cradle to grave"). Ve reviewed thousands of Vasts Hauler Records and Hazardous Vasts Manifests that we found among the Fairchild documents. Ve selected only those records which documented solvent waste removal for our analysis. Ve encountered a number of problems in analyzing the hauler records and manifests. All of the documents which were used to compile this report were photocopies, sometimes having been previously reproduced several times. In many cases this made these documents, or parts of the documents, barely legible or illegible. Further, those which were legible occasionally had problems of being self*contradictory. For example, a components section of a manifest would list only solvents, but the description of the waste might be listed as "acid solution," "contaminated soil and sand," and/or "oil." However, without any further information other than the record of manifest itself, we took the description of the components to be most accurate. Also, six of the manifests listed confusing concentrations of solvents and parcentages of water which leads us to believe that some portion of at least these six shipments was water. However, ve could not interpret the accurate water content of the shipments because of our inadequate understanding of the conventions used by Fairchild and/or the waste transporter staff when they completed the manifests. The water content of the solvent waste transported for disposal artificially inflates the solvent outflow estimates. If the water content of the wastes manifested could be accurately calculated, it would increase the apparent "imbalance" of the MBR. Ve have included an example of a Fairchild manifest in Appendix C of this report. Ve present our findings in three tables. Table 5.4. is a complete listing of all of the manifests included for our analysis, including all those that Fairchild listed in the MBR and those which OHS found and reviewed. Table 5.5 is a listing of the various problems which were discovered in reviewing 5.12 SL 033403 Che records end manifests. Table S.6 lists che differences between what DHS found and what Fairchild presented In the MBR. Evaporation Analysis The MBR states on page 2 that "Fairchild employed Independent laboratories, using EPA approved methods, to conduct tests to determine what portion of the solvent received by the South San Jose Facility Is lost through evaporation. Those tests indicate that evaporation accounts for the dispositon of eleven percent (lit) of all solvent. Such an allowance for evaporative disposition has been included In this study. DHS reviewed February 1982 emmlssions tests reports supplied to us along with the other litigation documents. However, we did not find any corresponding analysis or interpretation of these tests to derive solvent evaporation rates. Tank Gauge Although the MBR does not mention it, the failed tank was equipped with a gauge which measured the stored solvent level in gallons. The 1978 and 1981 logs of meter readings for the failed tank were among the Fairchild documents, tfa requested the 1978 meter readings late In our study, and subsequently did not have time to review them, tfe were informed by Fairchild's attorneys that according to deponents in the litigation, the 1979 tank readings were never found and 1980 tank level readings were accidently destroyed when an acid waste pipe at the South San Jose facility backed up and spilled on them. We reviewed the log of daily meter readings for 1981 which contained readings for "solvent waste treatment." The first entry in this log Is a "0" reading on 1/5/81. The readings average about 500 gallons + 200 gallons for the first four months of 1981. In May, the readings increase to 900 gallons. In June, the readings decrease and drop to zero on 6/23/81. There is no later non*zero entry for solvent waste in the log. In the entries for October of 1981, question marks "?" appear next to the zero entries, but there is no explanation. Fairchild's attorneys Informed us that several Fairchild employees deposed in the South San Jose litigation testified that the solvent tank gauge repeatedly malfunctioned in 5.13 1980 and 1981 and was not rsllad up n as an aeeuracs lndicat r f cha laval of solvanes In tha tank. Va found no axplanaelon of this In tha documents va reviewed, however, as previously stated, we did not review all of the pertinent Fairchild employee depositions. Other Data Reviewed Chanlcal Eaata validation We searched for sources which might have data to validate Fairchild's chemical inflow data. The only source from which we obtained any data was the San Jose/Santa Clara Water Pollution Control Plant in San Jose. They have a file for Fairchild South San Jose, which contains some, although incomplete, information. The file has an undated, though presumably pre1982, Industrial Wastewater Discharge (IVWD) Questionnaire which refers to an attachment list of chemicals stored at the site. This attachment is missing from the file. There is a more complete IVWD Application/Permit dated 10-4-82 which does list quantities of chemicals stored and used per month at the site. The list of chemicals does closely match the MBR, however, the groupings and amounts used per month do not match as closely. Table 5.7 presents a comparison of the data found in the MBR and the SJ/SC WPCP file. It is unfortunate for the purpose of this study that other regulatory agencies did not have similar types of data to be used for validation of Fairchild's chemical usage data. Waste Manifest Validation We attempted to validate the Fairchild Waste Hauler Records and Hazardous Waste Manifests in several ways. We requested the documents from the Program Monitoring and Personnel Section of the TSCD in Sacramento. We learned from them that our request would entail a search of 143 boxes containing hundreds of thousands of manifests. These manifests are not grouped by facility, so virtually every document would need to be reviewed. After a trial review of twelve boxes of 1979 documents did not produce any manifests relevant to our analysis, we decided that we could not continue the search within the time and budget limits of our study. 5.14 033405 Sb We elso contacted International T chnology (IT) Corporation and Chemical Waste Management, Inc., both of vhich hauled wastes from the Fairchild San Jose site. Neither of these companies were able to provide us with the manifests we requested. Evaporation Validation Although we were not able to find or review documentation that we would consider to be full documentation for Fairchild's evaporation analysis, we did review pertinent information from the Ray Area Air Quality Management District (BAAQMD). The BAAQMD regulates Industrial facilities' ambient air pollution emissions through a permitting process. The BAAQMD file for Fairchild's South San Jose facility contained brief process descriptions, engineering evaluation reports, applications, and correspondence regarding permitting for the facility's estimated air pollutant emissions. In a letter dated August 15, 1980 from Fairchild to * the BAAQMD, Fairchild states "records show that for an average month we use approximately 5900 gallons of acids and 2100 gallons of solvents. Experience has shown that 95% of these chemicals stay in liquid form and 5% g are exhausted as fumes." However, a July 1982 Engineering Evaluation by the BAAQMD calculates evaporation rates to be much higher. They calculate evaporation rates based on vapor pressure of each chemical. For solvents typically used at Fairchild, BAAQMD assumed evaporation to be: "Acetone 25%, IPA - 15%, all others (Xylene, HMDS, Photoresist, Resist Strip) - 9 10%." The evaluation report continues to state, "Furthermore, there is an uncertainty regarding the amount of organic solvent that evaporates. The rate has been estimated by the industry to be 10*11%, but there are no firm records to substantiate this. Source test results that have surfaced are not definitive enough to prove this. Records will be required so input and outgo data can be matched to produce a material balance for organic solvents. The BAAQMD prepared a report "Emissions From Semiconductor Manufacturing in the Bay Area, April 1987."^ This report states much higher estimates of 5.15 SL 033406 solvent evaporation. For example, for negative and positive photoresist usage, "90% of organic solvent in photoresist and developer evaporates." 12 For solvent cleaning stations, "30% of all chemicals used evaporate." 13 These varying reports of solvent evaporation, coupled with our inability to find appropriate documentation of Fairchild's evaporation estimates, leads us to conclude that the evaporative loss of 11% of all solvents claimed by Fairchild in their MBR cannot be substantiated. The evaporative loss may be much higher, which would lower the apparent "Imbalance" of the MBR. Comparisons of MBR and DHS Tabulations of Inflow and Outflow In Table 5.8 we have presented a quarterly comparison of MBR and DHS Tabulations of solvent inflow and outflow. The MBR figures are derived from DHS's corrected tabulations of the MBR Appendices A and B. The DHS figures are derived from Tables 5.3 and 5.4. We did not Include 1977 figures in this comparison, because we did not find any data for that year. We did not include evaporation in this comparison because we could not validate Fairchild's estimate of evaporation or calculate an appropriate estimate urselves. The results of this comparison are that DHS tabulated more solvent inflow and lesser outflow than the MBR. DHS tabulations show an increasingly larger imbalance than the MBR for all four years. (See Motes, 6 and 7) Tank Failure Mechanism The failed tank was a resIn*coated fiberglas tank manufactured by Owens Corning Fiberglas (OCF). Its design and engineering specifications were supposed to be compatible with Fairchild's intended use. However, the failure occurred as a degradation from the inside of the tank's resin lining due to the incompatability of the chemicals stored in the tank. An analysis of the tank's specifications and the materials that were stored in the tank should indicate the chemical(s) responsible for its failure. Four days after Fairchild reported the leaking tank and groundwater contamination to the RWQCB, a RWQCB inspector conducted a site visit to 5.16 SL 033407 inspect the tank and th site and to talk with Fairchild about the Incident. The report of that visit, dated 12/9/81 states, "3) (The) Solvent tank was fiberglas, buried directly in the soil (bottom of tank [approximately] 12' below grade.) Preliminary indications are that the acetone dissolved through the tank bottom. (An) Owens Corning rep. was on (the) scene earlier and gave that opinion to Fairchild. This tank was supposedly constructed using a special resin to make it acetone resistant." 14 Because acetone was implicated as the material responsible for degrading the tank, we reviewed the MBR Appendices A.l - A.5 to see what total acetone purchases were for the years 1977-1981. According to the MBR, acetone purchases equalled 43 gallons in 1977, 395 gallons in 1978, 4055 gallons in 1979, 19650 gallons in 1980, and 15349 gallons in 1981. It is obviously even more important to know when acetone or acetone wastes were put into the failed solvent waste storage tank, since that is the only way the acetone could have effected any damage to the tank. However, we could not get any direct measurement or information about when acetone went into the tank. So, lacking more direct information, we analyzed the waste manifests that we had collected to see if and when acetone was listed as a waste component. We were particularly interested in knowing when acetone v*s listed as a waste component on manifests which documented "bulk" waste disposal. "Bulk" differentiates wastes stored in the solvent waste storage tanks from those wastes stored in drums. In order to deduce which of these bulk disposal records most likely pertained to the failed tank, we considered two things: the amount of waste disposed and the date of the disposal. These factors are Important because there were two chemical waste storage tanks used at the Fairchild facility. The first tank, the one that failed, was a large tank (approximately 6000 gallon capacity) installed when the facility was built In 1976. This tank was plumbed to two operating wings of the facility. A second smaller tank (550 gallon capacity) was installed in mid-1980 near the "Chem Mix" building where chemicals were mixed and repackaged for use in the operating wings. This tank was plumbed to drains in the floor of the Chem Mix building. 5.17 Sl> 033^08 Chemical spills v re washed off of the floor and down these drains with water. We found that acetone was listed as a waste component on manifests of bulk disposals over 550 gallons on 6/1/79 (reported as May 1979 in the MBR), 9/30/80, 3/18/81, and 12/4/81 . Acetone was also listed on some manifests of drum disposals from 5/3/79 through 6/15/81. We searched for more Information regarding the OCF representative who had come to Fairchild to inspect the failed tank. We learned that this representative had filed a report regarding his findings. The representative was later deposed and the report was made an exhibit to the deposition. We asked LRD for access to this deposition, but they referred us to OCF's attorneys, Sedgewick, Detert, Moran, and Arnold. We contacted this firm, but were ultimately refused access to the deposition. Next, we contacted OCF directly to request access to the report and deposition. We also requested their expertise In explaining the exact mechanism and possible timing of the failure. OCF did not reply to our request, and they did not give us access to the reports or depositions of any OCF employees. PhYJlcal and Chemical Analysis of the Failed Tank We requested the results of any analysis done on the failed tank from the RWQCB and each of the litigant's attorneys. The RWQCfi had not done a tank analysis. We were told by the attorneys for all the litigants that this was considered attorney work product and subject to attorney-client priviledge. Furthermore, we were informed that there were stipulations in the litigation settlement, which was agreed to by all parties, that information was to remain confidential. Therefore, this Information was not released to us. We considered performing a tank analysis ourselves, however, several knowledgeable contacts held the opinion chat the results of such an analysis, done six years after the tank had been out of use, would be inconclusive. Furthermore, the cost of hiring a consultant to interpret the 5.18 SL 033409 results of such an analysis was outside the limited budget of this Investigation. DISCUSSION The DHS analysis of the Fairchild Material Balance Report did not Identify enough documentation to provide an accurate estimate of the onset of the Fairchild South San Jose underground tank leak. The analysis was hampered by limited time and money, limited data to support the study, and. limited cooperation from the sources who held pertinent data. Ve believe, however, that the analysis does demonstrate that Fairchild also had inadequate data at the time the MBR was prepared to support an accurate calculation of the solvent imbalance at the end of the five year period under study. Furthermore, the MBR conclusion regarding the timing of the leak appears uncertain in light of DHS's Inability to replicate the MBR solvent Imbalance although .no evidence was found that^would appreciably alter that date. Although the material balance methodology may have been an appropriate approach to use in trying to determine Fairchild's solvent imbalance, the data used to fill in the seemingly simple material balance equation for the MBR is complex and portions were likely interpreted inaccurately. Ve restate the essential components of the equation as we see it and discuss the data used to support each component to illustrate this point. The equation is: Solvents purchased and used in the system should equal solvents "used up" in processing (primarily evaporation) plus solvent wastes disposed of out of the system. Solvent Purchase Data The MBR states that ingoing chemical data was compiled from chemical issues from Mountain View and direct purchases by the South San Jose facility. On page 5 of the MBR, it states, "An additional material input to the facility was batch shipments of photoresist materials from two different vendors to the production group during 1980 and 1981. These shipments were accounted for in the incoming mass balance." The compilation of this data was undoubtedly quite complicated. The MBR tables do not present separate 5.19 033^1 SL figures f r each of these components. Furtherm re. the documentation of how each of these components was complied is not available from Fairchild. DHS found documentation for a different total of solvent inflow than reported in the HSR. DHS figures for "issues" and "purchases" data are presented separately In Table 5.3. DHS has included Fairchild's explanation of their interpretation of the differences between the MBR and DHS solvent Inflow tabulations in Notes 6 and 7. However, despite Fairchild's explanation, and even if DHS was to make the adjustments to their solvent Inflow tabulations suggested by Fairchild, the DHS and MBR tabulations would not be completely reconciled. DHS was using Fairchild documents, and yet the Inflow estimates between the two are not equal. Ve can not conclude what the most accurate estimate would be. Evaporation DHS could not validate the 11% evaporation figures used in the MBR. The BAQMD reports which we reviewed did not support that estimate either. The BAQMD data that we presented shows that for many of the solvents used by Fairchild, the evaporation rates are higher than 11%. However, the BAQMD also reports chat more data needs to be collected to determine the accurate amount of solvent which evaporates. Waste Disposal The MBR states on page 4 that the chemical waste disposal records for the early part of the study period may not have been complete. In fact, DHS did find more solvent disposal manifests than Fairchild did for the MBR. However, DHS's tabulation of the waste manifests resulted in a lower calculation of solvent wastes manifested for disposal. In addition, DHS found that six of the manifests listed a significant portion of the waste to be water. If the water content of the waste shipments could be accurately calculated and subtracted from the total solvent waste disposal figure, it would further reduce the amount of solvent disposed. Consequently, the solvent outflow figures as presented in the MBR and the DHS tabulations are both, most likely, inaccurately high. 5.20 Sb The DHS analysis also f und ocher, and possibly more accurate, sources of data which Fairchild could have considered t determine the onset of the tank failure. If the tank gauge readings had been properly kept and analyzed, they would have proven to be an excellent source of data regarding the timing of the tank failure. In fact, the Fairchild spill was the singular event that triggered the development of stricter underground tank monitoring and recordkeeping regulations which were recently mandated by new federal and California state laws. Additionally, more information could have been developed and presented by Fairchild regarding the solvent waste tank's engineering specifications and chemical Incompatibilities which caused it to fail. If acetone did have a significant effect on the degradation of the solvent waste tank's structure, then the degradation could have potentially begun as soon as the tank was put into service. However, there is Inadequate public information available on this issue to answer questions regarding the exact mechanism of the tank failure, or to draw any firm conclusions as to when the tank might have failed. 5.21 SL 033412 References 1. Epidemiological Studies Section, California Department of Health Services: Pregnancy Outcomes in Santa Clara County 1980*1982: Reports of Two Epidemiological Studies. California State Publications Section Report 7540-958-1301-5, 1985. 2. Fairchild South San Jose Material Balance Report, May 11, 1982. 3. Ibid. p.l. 4. Ibid. p.3. 5. Ibid. p.10. 6. Ibid, p.4. 7. Ibid. p.2. 8. Fairchild letter dated August 15, 1980 found in Bay Area Air Quality Management District files. 9. Bay Area Air Quality Management District, "Fairchild Camera and Instrument Corp., MOS Division, Engineering Evaluation", 7-7-82, p.l. 10. Ibid, p.2. 11. Bay Area Air Quality Management District, Emissions from Semiconductor Manufacturing on the Bay Area, April 1987. 12. Ibid, pp.31-32. 13. Ibid, p.32. 14. Regional Water Quality Control Board, Fairchild Semiconductor, San Jose Discharge of Solvents into the ground, 12/9/81 (from 12/8/81 inspection and conference), pp.3-4. 5.22 SL 033413 Notes The Draft Material Mass Balance Analysis was reviewed by over thirty Individuals from regulatory agencies, concerned citizen groups, law firms, consultants, Great Oaks Vater Company and Fairchild, tfe appreciate and thank each of the Individuals who gave us their thoughtful review and time In preparing comments to the draft report. We have taken all comments into consideration In our final report. In particular, Fairchild and their attorneys, LandsIs, Ripley and Diamond, had extensive comments and new information to add to the draft report. The new Information they provided regarding solvent inflow data substantially reduces our confidence in our solvent inflow tabulations as presented in Tables 5.3 end 5.8. However, DHS staff did not have sufficient time or resources before the scheduled release of this report to be able to Independently verify all of Fairchild's new Information. Therefore, we have Included their Information in these notes as opposed to removing our original tabulations. All of the comments with quotation marks in these notes are quoted from written comments of Fairchild as transmitted to us through their attorneys, Landels, Ripley and Diamond (LRD). Other nonquoted notes are explained parenthetically. DHS notes follow the Fairchild comments. 1. "The November/December 1978 solvent purchase Information was evenly divided between the months because available information at the time of the MBR was for the two month period. Therefore, the November/December 1978 numbers are estimates." 2. Methyl alcohol purchases are probably included with the figures for IPA. (Telephone conversation with LRD attorney). 3. "The Draft Report refers to the MBR* s omission of scrubber effluent as a 'waste outflow error.' As was discussed with the Regional Vater Quality Control Board in some detail in 1982 the scrubbers et Fairchild's facility were primarily intended to scrub acids from the alrstream. To 5.23 SL 033414 s me extant, Incidental solvents may also have be n scrubbed from the alrstraaa. The scrubber effluent was discharged to the acid neutralization system at tha Fairchild South San Jose facility. Therefore, there was no relationship between scrubber effluent and the amount of solvent which may have reached the failed solvent waste tank." DHS Included its observation that scrubber effluent is not included anywhere in the MBR for two reasons. First, the MBR includes the statement about emissions scrubbing relevant to solvent evaporation. Secondly, Fairchild submitted solvent emissions calculations documents in 1980 and 1982 to the BAQMD in support of their applications for appropriate permitting to operate their South San Jose facility. In 1982, the recommendation chat the Permits to Operate be issued to Fairchild was qualified by three conditions. Each of these conditions pertained to the use, emissions scrubbing, and disposal of solvents. Clearly, the issues of solvent evaporation, the efficiency of solvent removal by emissions scrubbing, and the measurement of solvent in scrubber effluent is pertinent to a solvent materials balance analysis. In the Fairchild ease, an accurate accounting of solvent in the scrubber effluent would help verify the amount of solvent lost to evaporation and thus further substantiate the estimated amount of solvene which reached and may have leaked from the failed solvent waste tank. "The final 'waste outflow error' noted in the Draft Report is chat the Fairchild MBR does not separately state disposal figures for the large ('failed') and small solvent waste tanks after the small tank was installed in mid*1980. Operations people at Fairchild testified at their depositions that there were several occasions when the waste hauler emptied both the large and small solvent waste tanks into a tanker truck after mid*1980 without differentiating the gallonage between the tanks. Therefore, the data which DOHS believes 'should' be available simply is not available." DHS agrees that if information is simply not available, then that should be clearly stated. The fact that the information was not presented because it was not available is not clearly stated in the MBR. Where 5.24 033^5 informati n is available, we believe it should have been included. The exact date f the installation and peration f the Chem Mix 550 gallon tank should be Included. The fact that the failed 6000 gallon tank was removed and a new temporary 1000 gallon tank was installed on 12/3/81 should be Included. These two pieces of information help define the period when bulk waste disposal figures might have included a combination of the 550 gallon tank and the 6000 gallon tank. "Operational employees from Fairchild repeatedly testified in their depositions that photo resist, in normal operations, would not go down solvent drains to the solvent waste tank but would be disposed of, with stripper materials, by 'super sucker' removal and disposal into drums. Fairchild included photo resist materials in the 1982 MBR because they were part of the incoming and outgoing solvent materials. However, the operational facts concerning the South San Jose facility developed after the MBR was prepared Indicate that the likelihood of any substantial quantity of positive photo resist materials reaching the solvent waste tank was remote. Moreover, as acknowledged in the Draft Report, the Bay Area Air Quality Management District estimates that 90% of solvents in photo resist materials evaporate. Therefore, the likelihood of glycol ethers in the solvent waste tank is reduced even further. For the Epidemiological Studies Unit to raise epidemiological concerns about the 'possibility* of a chemical teratogen in groundwater Is Irresponsible without some substantiation for its concern." DHS has included information about glycol ethers exactly because of the possibility, however remote, that this chemical teratogen could have reached public drinking water. The operational information about the disposal of photo resists containing glycol ethers into drums that Fairchild attorneys have presented here is very relevant to the interpretation of the MBR, and we are pleased to present it with this analysis. However, DHS would like to add that this Information, along with much other relevant information developed after the preparation of the MBR is not aceessable by the public. DHS, in conducting its analysis of the 5.25 SL 033A16 MBR and saklng public this report, has sought to disci sa new information relevant to the KBR and its conclusions regarding which chealcals in what amounts leaked from the failed tank, and what was the onset and duration of the leak. 6. 'DOHS Should Not Include * Developers' and * Chrome Etch' as Solvents in ita-Xgtlll The Draft Report concludes that there were 42,304 gallons of additional 'incoming' solvents to Fairchild's South San Jose facility from 1978 to 1981 which were not reflected in Fairchild's 1982 Material Balance Report ('the Fairchild MBR'). Of those 42,304 gallons, 37,579 gallons were 'positive developers' and 'chrome etch' (categories 10,23 and 32 in DOHS' backup data). Ve are informed by former Fairchild employees that those developers (AZ-351; Micro 351, etc.) were alkaline products, not solvents. Therefore, they were disposed of down the acid neutralization system, not in solvent drums or via the bulk solvent waste tank. Similarly, the chrome etch products were acids, not solvents. That information is supported by the enclosed information from chemical suppliers. As you will see, the manufacturers' information confirms the alkaline nature of the positive developer products and the acidic nature of chrome etch. Recommended disposal is by neutralization. Ve conclude from the above information that 3,5,1 pre-mix (a developer) should not have been included in the Fairchild MBR aa a solvent. This may explain why the only entry for 3,5,1 pre-mix in 1979 (172 gallons) is not added to the solvent totals for 1979 in the Fairchild MBR. In short, 37,579 gallons of the 'additional solvent purchases' referred to in the Draft Report are not solvents." DHS includes this important new information from Fairchild in Note 6 as a probable major correction to our solvent inflow calculations presented in Tables 5.3 and 5.8. DHS did not have the resources to independently corroborate this new information presented to us by Fairchild's attorneys before the scheduled release of this report. 5.26 Si 334i7 DHS staff originally included all three of these ch mieals in solvent inflow tabulations for two reasons. These three chemical groups were listed in Fairchild forms titled "Solvent List and Inventory." Also, the 3,5,1 premix, a developer and another developer, 802, were Included, perhaps erroneously, as Fairchild points out, in the original HBR. DHS presented their backup data to Fairchild attorneys upon their request because Fairchild had serious doubts about DHS tabulations presented in the Draft of this report. The DHS tabulations for the three chemicals separately are as follows: 32,453 gallons of AZ-351 and micro 351 developers (DHS category 10); 4886 gallons of projection developer (DHS category 32); and 240 gallons of chrome etch (DHS category 23). The total for all three groups In 37,579 gallons. Ve have Included a copy of the information from chemical suppliers about these three chemicals which Fairchild enclosed In their comments to us as Note 8. * 7. "* Drop Shipments' of Chemicals to San Jose In addition to including non-solvents in its totals, the Draft Report also appears to 'double count' certain solvent purchases in the years 1978 to 1981. As Indicated in our April 29, 1988 comments, Interviews with former Fairchild employees familiar with the solvent purchase documentation suggest that a number of the purchase orders for solvent materials to South San Jose reviewed by DOHS may have been duplicative of chemicals reflected on the chemical stores reports issued from Mountain View under a master Mountain View purchase order during the years 1978 through mid-1981. Ve are informed that chemicals, Including solvents, were periodically 'drop shipped' by chemical suppliers directly to San Jose under the Mountain View purchase orders. While these 'drop shipments' would create documentation reflecting a direct delivery of chemicals to San Jose, they also would be included in the Mountain View chemical stores reports. Therefore, DOHS may be 'double counting' solvent purchases by recording drop shipments to San Jose both as a 'direct' San Jose purchase and as a 'chemical issue' from Mountain View. 5.27 St 03341s In the shore time since receiving DOHS' backup date, we have not been able Co locate the underlying data which would allow us to specifically docuaent the extent of any double counting reflected in the Draft Report. However, as an example of the phenomenon, it appears that the Mountain*View chemical issues reports for 100% HMDS in 1978 and 1979 would more than account for the * direct purchases' of 100% HMDS Included in the Draft Report backup data. Specifically, the Mountain View chemical issues reflect a total of 145 gallons of 100% HMDS to San Jose In 1978 and 1979. (This total is reflected as part of the 174 gallons of HKDS + 802 in the Fairchild MBR.) The 'direct* purchases of 100% HMDS reflected by DOHS* review of purchase orders total 97 gallons. Therefore, the 'direct* purchases could be merely subsumed within the Mountain View chemical issues reports. While this example is not definitive proof of double counting, it certainly suggests that the Information provided by former employees may be correct." DHS includes this consent from Fairchild as a possible explanation and correction to Tables 5.3 and 5.8. Upon receiving Fairchild's comments on this point, DHS staff re-examined its chemical purchase data, especially with regard to 1978-1980 data. We found 105 purchase order releases for those years, all with the same blanket purchase order agreement number for one vendor. A re-examination of the pertinent 1981 purchase data is more difficult, since Fairchild can not say exactly when in 1981 the method of obtaining chemicals for the South San Jose facility was changed from "chemical Issues" by Mountain View to direct chemical purchases by the South San Jose facility. We were not able to reanalyze the chemical purchase quantities by solvent categories for 1978-1980 before the scheduled release of this report. However, a look at the data reveals that photoresists and 351 developer seem to be the largest volume categories of the chemical purchases for those years. The possible erroneous inclusion of 351 developer has been previously discussed in Note 6. Although Fairchild believes that the 1978-1980 and early 1981 chemical purchase figures may already be accounted for In the chemical issues data 5.28 SL 033419 for th. tiTM port d, they w ,,.t, .. th.y ,t.t. t,, th.tr o TM ,, dflnitlv.ly demon.trot. thi,. Th.r.f.r., DHS c.n ,,.t pr.cl,,d. th.t th, docunnt.d .otv.pt puroh... d.t. tor thl. tl. p.rlod ,, ... .ddtttop.1 ' ch.alc.1 l.suas d.t. d po..lbIy .v.rlo.k.d in th. d.t. tpiUtl.,, phase of th preparation of the MBR. 5.29 SL 033420 `88 93'13 tl i 44 2 488 727 624? Ticvuil F*b In 09 note 8 November, \9M Wide development latitude Compatible with epln/apray develop equipment U Cuatom dilutions available U Compatible with KT11300 series photoresist KTl 3bi Oftvsiopcr m combination with KM 1300 Series Photoresists it designed 10 meet. the demanding mierolithographic and processing requirements of tho semiconductor I'icusiry it is an odorless, aqueous, inorganic, alkaline soljtion, f-ee ot phosphates, which is compatible with batch and in-ime developing processes. Precise manutacture and stringent quality control ensure bateh-to-batch roproducidilify and pioduci quality. KTi 35t Oeveiopet is supplied as a concentrate. Custom dilutions are also available. The standard High Contrast makeup provides optimum resolution and contrast as wan as maximum processing latitude. The Standard High Spood dilution results in very mgh production '.timughput Use of KTl 3!>1 Dnvntnpor msOltg in reduced bath ma*e-op cost. Bath Make-Up To preoaie the standard H gh Contrast and High Speed Ciiutions from the concentrate, mi* KTl 351 Developer anti ce united water by vniu-nq as follows. f.'nvcUOCr Ma<e-U`o K i: 3fi` ocvnloocr Hcnonizea Water Hig-I Cont'OSt H gh Speed 1 0 pan 1 0 part 5 0 pans 3 5 parts *, Ad>*t 13 333 fOO lOrr'BC'a'.urO prior 10 uSS Phyaioal and Chamleal Propertlaa The physical and chemical characteristics are shown m Table V Development Immertlen Immnmo lor approximately 00 seconds m either High Contrast or High Speed KTl 351 Developer maintained at a constant temperature ( * 1*C) within the range 20*-25"C. Use mechanic cal or nitrogon burst (not air) sgitatinn. Rinse immediately in deionised water until resistivity Is within soecifications. Spin dry in air or tore# dry with tittered nitrogon. Fresh drjvnlopKf gives optimum results, it is recommended that the bain solution be replaced at least once each shift. Recirculating Beth Spray Replenish with fresh developer as recommended by the equipment manufacturer. In-Line Spray Control developer temperature at the dispensing head at a conatent temperature (* l*C) with,n the range of 20''25*C Moderate spray pressure is recommended. A typical process will involve spraying either in# High Speed or High Contrast developer on a slowly spinning wafer for 60 seconds, amt overlapping a deionized water rinse with the developing cycle. After a 10-15 second Q.i. rmsc. the wafer is spun ary Table 1 Color Turbidity Sueu itu Gravity. 1U/21/C Normalcy Clear (?5 Pt-Cd cr Inns) None (2 0 NTU's or lt-au 1 0d3 s 0 005 Concentrate 1 39 x 0 Of) N High Speed Make-Up 0 3i - 0.01 N High Contrast Make-Up 0 232 ' 0.005 N F.liration 0 2 Jim apnnnjin 'Other dil,,i,or* *ra aio *va ma. ?0103 SL 033421 4 # <-!* '08 0J/ 1 3 11145 KTI351 Developer S 408 727 6247 Ticu:tl Fb* In 10 November, 1983 Determination oi Normality Reegenta Hydrochloric Actd (mCi) 0.2 N. Standardized. Methyl Rttri IncJir.ainr (0 ?% m methanol) Procedure 1 Pipette 5 ml ot K11351 Developer into a 250 ml Erion- mcyer flask 2 Dilute with approximately 100 ml deionized wator 3. Add 3 drops ot methyl red indicator. 4 Titrate with hydrochiorc acid (0.2N) to a red endpoint. Calculation (ml HCl) X (N HCl) 5 ml KTI 351 N of KTI 351 Normality of a freshly mndo-up batch should be 0.22 to 0.24 n lor the standard High*Contrast make-up of i part developer with 5 parte deionized water and should be 0.30 to 0.32 N for the High Speed mako-up of 1 pan developer with 3Vj parts oeionized water. Safety Precautions Handling HANDLE WiTl- CARE. KT: 35" DCVL.OPLR IS AN ALKALINE SCXU'IION. IT CONTAINS SODIUM MYUHOXIDE. MAYBE HARMFu. IF fiWAI I OWED AVOID CONTACT WITH SKIN AND EYES. AVO.D BREATHING MIS:s. WEAR Cl IEMICAL GOCCLfcS. MUBUrn G..OVES AND PROTECl IVfc CLOl HiNG. ST'll.l ON i.l-AK: I I USH SPILL AREA WITH WATER First Aid IN CASF OF CONTACT, IMMEDIATELY FLUSH EYES A Th LSNTY OF WATER FOR AT LEAST 15 MINUTES CALL A PHYSIC AN 11 USH SK'N Wl'tH WATFN RKM.OVr CON taminated clothing ano wash before re- jse II INI'ALLO. MOVE 101IUSH All) It- Hlti-AMiNG ha:-. S'lO^l'tO. GIVE ARTiFICIA, RESPIRATION |C BREATHING ISDirriCJLT G'VC OxvGLN. Cai 1 A physician IF SWALLOWED, DO NOT INDUCE VOM.T'NG G VC LARGE OUAN'ITICS Qi` WATER. NEVER o'Vw ANYll-iNG BY MOUTH TO AN UNCONSCIOUS PFRSON. CALI. A PHYS ClAN Equipment USE FOLYPROPYLENC. POLYETHYLENE. P-FF 0* FCi. ivA LENT MATERIAL KTI 351 DFVE:.OI'ER iS CCMPAti3. WITH MOST A-_ COMVERC ALLY AVALAQLC WA- !> HANDJNG ECUPMENT Storage S'CRE ONLY IN ORIGINAL CONTAINERS N DRY ARLA A iO' SJO'F (10**32*0). SEA,. CCNTA NCR WHi \ \C M.SF Dlspoeal DISPOSE Ol N ACCORDANCE W'lh I OCAL REG JL-T.0NS. CONTACT YOUR KTI TECHNICAL RE?RESENTA-'VE IF ASSISTANCE IS REQUIRED. Thu information i| basod on our asperienee end It. to ihe beat of Out knOw*ed*. hue end accurate. However, since Ihe condition* (Ot U*V 40it hendung of tho product* are beyond our control, wo mate no guaranty or warranty, expresaad or Imoitod. regarding tnc Information, me use handling, atorege O' poaiaasion of tn# prortuete, or the application ol any proceae deecnbed heroin or the results soughi to be outainoo Nom ng narain inn be construed at a racommenoatiort to ust sny product In violation 01 any patent rights. All sales are subject to Our siansa>0 and conoilionc of salt 11 7Q $onOt n Guort Sun.iyvjiia. CA 94086-5385 H08) 733-3500 2125 w. 7th Street lumna. Arizona 8528 f (802) 96.g?12 Subsidiary nion Carbide Corporation loctronics Division PQ103 i2(X) w. jacaaon Hoad Carrollton. Taxaa 75006 (2141 243-5941 2 Maine* inouatriai Haia Hoaa Wallingford. Conneetieui 00492 (203) 263-0242 SL 033422 11 8 83'' t J IU4J X 48 7i? *J4? F.o, I, 97 MICROPOSIT* 351 DEVELOPER Instructions For Use I. Bath Make-up Dilute MiCROPOSIT 351 DEVELOPER for use as follows: High Spovtf M*fcup (22% Miutlon) High Ratalullon Mckt-up (17% alutlon) MiCROPOSIT i pari by velum# 1 part by volum# DEVELOPER 0oniz#d w*tr 3vk parts by volume 5 parts by volume Mix thoroughly. Proper dilution can bo verified by analysis for normality. See section Vll. Photoresist dissolution rate increases with increas ing developer concentration. Maximum resolution Is obtained at the lower developer concentration where unexposed resist loss is minimized. Shorter exposure times are passible when the higher developer concentration Is used. Production line downtime and potential dilution errors can be avoided with ready to use developers (MICROPOSIT 352. 353, 354. 355 DEVELOPERS). 352 is recommended for high resolution (equivalent to 1.5 make-up above). 354 is recommended for high speed (equivalent to i 3'/a make-up above). II. Temperature Operate MICROPOSIT 351 DEVELOPER between 20* and 25*C, with the temperature controlled ilC. The photoresist dissolution rate -'creases with increasing developer temperature. !n spray equipment, the spray action causes a temperature drop in the developer solution. For th'S reason, developer temperature should be monitored at the substrata surface. tit. Time Immersion: 40-60 seconds. Spin/spray: Varies with equipment Consult your Shipley Technical Sales Representative. Longer development times permit the use ol shorter exposure limes Shorter development times minimize developer attack on the unexposed photo resist. The range recommended is optimum, wc recommend keeping the development time con stant and adjusting the exposure time as neces sary to meet critical dimension requirements IV. Agitation Immersion: Mild, consistent agitation is recommended. Spih/spray: Contact your Shipley Technical Sales Representative. V. Rinse Immersion: Spin/spray: Cascade rinse with deionized water to resistivity specification immedi ately after developing. Overlap deionized water rinse with developer cycle to prevent develops drying on substrate surface. Provide adequate rinsing of buck side of substrates. Vi. Bath Control Immersion: For maximum process control. replace bath with fresh oeveiorjcr solution at least once per snift Kg', bath covered when not in use. Spin/spray: Not applicable. Batch spray: As recommended by equipment manufacturer. E>v '88 83/13 1H43 X 408 72? (247 Ttcticjl F 4c 4 In as Vtt. Determination of Total Alkallna Normality A. Reagents 1. Hydrochloric acid (HCI). 0.1 N. standaidizod 2. Methyl red indicator solution B. Procedure 1. Pipette 5 mis aliquot MlCROPOSIT 351 DEVELOPER bath into a 200 ml Erlenmeyer flask. 2 Add approximately 100 mis deion ized water 3 Add 3 to 5 drops methyl red indicator. 4. Titrate with 0.1 N HCI from yellow to red color change. C. Calculations Normality of ml HCI titrated x N HCI - MlCROPOSIT 351 5 mis DEVELOPER D. Results The normality of freshly made-up MICROPOSH 351 DEVELOPER should be: 1-3V4 make-up 0.31 0.02 N 1:5 make-up 0 23 a 0.02 N Equipment Use polypropylene, polyethylene, polytetrafiuonoethylene. or equivalent materials. Storage Store only in original containers in dry area at 50*-90*F (10*-32*C). Do not store in sunlight. . Close container when not in use. MlCROPOSIT 351 DEVELOPER has a limited shelf life. Disposal MlCROPOSIT 351 DEVELOPER should be dis posed of according to Shipley Waste Treatment Procedure WT 77-1 (pH adjustment with sulfuric acid). Contact your Shipley Technical Sales Representative for details. Properties As Delivered MlCROPOSIT 351 DEVELOPER is manufactured to the highest quality standards and is subjected tc state of the art testing for physical, chemicai and functional properties to assure the user o< maximum lot to lot reproducibility. MlCROPOSIT 351 DEVELOPER is filtered to C.2|i.m absolute directly into clean containers. Certificates of Analysis will be supplied with each shipment upon request. Quality Assurance Materia! Specifications and Analytical Tasting Procedures may be obtained upon request from your Shipley Technical Sales Representative. MlCROPOSIT 351 DEVELOPER, as delivered, will conform to the foliowing specifications: Specific gravity at 20/20*0 1 073-1.093 Color Water white to very pale yellow solution Turbidity Nonturbid Total Alkaline Normality 1.36-1.42 Handling Precautions Warning! Dilute sodium hydroxide solution Harm ful if swallowed. May cause burns. Avoid contact with skin and eyes. Avoid breathing mist or spray Handle with care. Wear chemical goggles, rubber gloves, and protective clothing. Material Safely Data Sheet available upon request. Duo to the nature of MlCROPOSIT 351 DEVELOPER, disposal of it. or residues therefrom, should be made in compliance with federal, state, and local environmental laws. First Aid: If swallowed -- contact physician at once. If eye contact -- flush wilh water lor 15 minutes, contact physician. if skin contact -- flush with water for 15 minutes. If inhaled -- move into fresh air. Spills: Flush with large amounts of water into waste treatment system 1002. m (J.3 A S1300 *Ad Si*00 ;do fiTA.aiu\tmjOvmOvf*.i snSrh. p'NCeYwutokni MT>a' logo yrv utKtomnrk* o' Shitjitry Mic.HOPO'jn osi crvruincfl p9t;i. , '98 93'1 3 IH41 t *98 727 62*7 Tactical Fabl In 94 WASTE File No___ )^77*,]_ A TREATMENT ____________________ _____ __________________________ I j*) Dato L WASTE TREATMENT OF SHIPLEY COMPANY ,'PHOTORE SI ST DEVELOPERS ________________ IN USE WITH MICROELECTRONICS_________________________ J r i i iI m ! ** I. Adjust pH to 5-0 - 5.5 with Sulfuric Add i (HjSO^), I I i 2. Let stir for 30 minutes. 3- Allow pracipitete to settle. k. Decent end fiiter. I 5. Adjust pH to acceptable municipal levels l for discharge. \I t 6.* Sewer effluent I iI A Note: Dispose downstream of other metal bearing waste .as the ehelates If present are still In solution. i i ** Note: For Remover 140 substitute for step 1: Dilute with water 3-1 to precipitate dissolved solids. r> y .!!orn.iCn cofM.uncd hara<n 14 nasad upon data a:ul rmuil* m tm; r>iv;;i.::iyoti of Shipicy Company inc. and is nanavcil tc -<>: . 1 cmctsi *iih Mispact |0 SMpley * SotiW tn t`w nccuiocy oi tnu ..Vt rv.i'.Kin no w.vraplio* .iqm.M or mpl;od. snail de if.-c- 1 in inn.^ni*'0" 0 :n sof,,>,noe mfOfmaiwr no." mail any Ic-yal ilulw* pr Phi-panno* DOnnposart upon Snipiuy Company Inc .i c SL 033425 88 05^13 11134 X 408 72? 8247 Tjctital Fot 1 i- CHZ0M-?CjJ 12-OC // U-S. DEPARTMENT OF labor Occupational Safety and Health Administration MATERIAL SAFETY DATA SHEET Reduirtd undtf USDL Safety end Haartfi Reputations (O' Ship Repairing. Shipbuilding, end ShiMreafcfng Q Cf ISIS, Tt. 117| SECTION 1 Westamerica Chemical Corporation Apontes tSumdtt. Strtt). dry, 5`uc. and Vf Codr) 716-B San Antonio Ave.Palo Alto. CA. 94303 c 4tM.CAU NAMI ANO SYNONYMS throm-Pteh 1200. Chrflm-Pteh 1200-S T KMcnotNCY tele-ons no. <15:493-6517 TAAQC NAMt AND |VN$NVMi * hh53 * "*-`r SECTION II HAZARDOUS INGREDIENTS PAINTS, MtSlftVATIVIS, 0 IOLVCNTE aioment* CRTAlVIT % tlv (Unilil ALLOY! ANO METALLIC COATINGS alloys % tlv * 1 VjtMlCkC solvents additives METALLIC cdatino* AILLtA MfTAL ALLS eOATlNO DA eOAC AHJ* othea* otNena -------------------------M--A-X--A-R--D-O--U-S--M--I-X-T--U-R--E-S---O--f -O--T-M--I-R---U-Q--U--I-O--t--S-O--U--P-S-,--O--N--O-A--S-E--S------------------------1--%-- - tlv (Umtti Ceric Ammonium Nitrate approx. 13; - SECTION III . PHYSICAL DATA SD'LiNa point t*e,i vama aacssuae (mm m.) approx >212 18 SACCIAIC SAAYITY (yO>11 AtACENT, VOLATILC tv volume id VAAOA DENSITY (AlA'll ~Tf SOLUHLITV IN WATER complete pH a*iance n0 oooa orange solution, odor of dilute nitric acid approx 1.14 83'* <1 <1 SECTION IV FIRE AND EXPLOSION HAZARD DATA MINT h*M| jot ipfllliable. >IA FUMMAIkl LIMIT! not aonlfeefaT serttUAL na* aiONTiN* water fpg enoctounca , eif contained breathing apparatus with full facepiece operated in oressure- mano or otner poaitlve pressure mode and full body protective clothing So&y*Mefa<ls \o release hydrogen gas which can font explosive mixtures 033^26 Sb '88 85'U U>53 X *98 727 62*7 Tctie*l ftot In 2* SECTION VIII SPECIAL PROTECTION INFORMATION OSHA aooroved respirator If tlv (2oom1 Is exceeded <s io r as reoulred to maintain exoosure below IwCCmanicai /Ct*tni) OTMta OTteT'yt owavca acid resistant gloves otHtn laoTieTivc tauiiwtNT . . imoervfous clothing and boots. eve laoTfqriOw cnemical splash gogqles and face shield in compliance with OSHA regulations SECTION IX SPECIAL PRECAUTIONS ^*(bUTlSN| TO II TMIM IM HANPUMC AMOItOKlMa -------Stgrg In ft COOl dry-PlACB away from eeeilv ax-Hlyad materials. containers, do not re-use bottle for storing drinking water. "ofutA **tCA'JTlOW tin ..ttt i-ai-il page (2i Ml IS.140 END OF OATA SHEET SIj 033427 form OSHA-20 at., sit 7i Table 5.1 Haterial Balance Report Analysis Data Sources and Contacts (Data requested in parentheses) dfcgncits 1. Department of Health Services, Toxic Substances Control Division, Sacramento and Emeryville (waste manifests) 2. Bay Area Regional Water Quality Control (background of MBR) 3. Bay Area Air Quality Management District (evaporation data) 4. San Jose/Santa Clara Water Pollution Control Plant (chemical storage data) Private Companies 1. Ovens-Corning Fiberglas, the manufacturer of the failed tank (no data provided) * 2. Sedgwick, Detert, Moran, and Arnold, Attorneys for Owens-Corning Fiberglas (no data provided) 3. Landels, Ripley, and Diamond, Attorneys for Fairchild Semiconductor Corporation Schlumberger Technology Corporation (inflow and outflow data) 4. Pillsbury, Madison, and Sutro, Attorneys for Fairchild Semiconductor Corporation who prepared the Material Balance Report, May 11, 1982. (no data provided) 5. The Boccardo Law Firm, Attorneys for citizen plaintiffs in litigation against Fairchild, Ovens Coming Fiberglas, etc. (no data provided) 6. Great Oaks Water Company (background of MBR) 7. IT Corporation, Waste Haulers for Fairchild (waste manifests; not available) 8. Chemical Waste Management, Inc., Waste Haulers for Fairchild (waste manifests; not available) 5.30 238.4 SL 033428 Table 5.2 Data Items Required for Materials Balance Analysis Data Item Chemical Inflow Chemical issues Chemicals purchased Chemicals stored Waste Outflow Waste manifests Evaporation Tank gauge data Auxilary Data Tank analysis Acetone Whv Data _N.eeded estimate inflow estimate Inflow validate inflow estimate outflow validate outflow estimate outflow validate outflow validate outflow estimate time of failure estimate time of failure Source UU> LRD SJ/SC LRD DHS-TSCD LRD BAQMD LRD litigants litigants Percent Obtained 75% 90% 100% (1982) All requested All requested not available - summary only 1981 only not available not available 238.tb2/8 5.31 033^29 Table 5.3 DUS Tabulations of Solvent Issues and Purchases by Month and Year J FMAM J J A S0 N Issues 691 900 763 691 382 636 802 603 1700 1974 NA Purchases 204 48 128 297 199 319 221 280 204 180 300 Total 1978 895 948 891 988 581 955 1023 883 1904 2154 300 Issues Purchases Total 1979 1361 24 1385 1355 1008 2363 1552 710 2262 1312 929 2241 1592 1024 2616 1218 750 1966 416 856% 1272 1050 680 1730 2034 738 2772 2039 . 1113 276 NA 2315 1113 Issues Purchases Total 1980 2129 NA 2129 2466 NA 2466 3844 NA 3844 3031 1310 4341 1172 1036 2208 1602 572 2174 3717 1742 5459 4312 2022 6334 8859 2076 10935 5489 1908 7397 6275 NA 6275 Issues 3217 Purchases 3212 Total 1981 6429 3268 5193 8461 4079 8786 12865 NA 7262 7262 NA 6127 6127 NA 7752 7752 NA 9557 9557 NA 7085 7085 NA 5593 5593 NA 5527 5527 NA 6907 6907 WHOLE D YEAR NA 9142 100 2480 100 11622 1904 NA 1904 16946 6995 23941 6123 112 6235 49019 10778 59797 NA 5372 5372 10564 78373 88937 SL 033430 a See Notes 6 and 7 for more information. XevEEO fiat 05/31/78 12/28/78 04/12/79 05/03/79 06/01/79 08/15/79 11/05/79 11/15/79 12/03/79 02/04/80 02/15/80 02/28/80 03/15/80 03/24/80 05/01/80 05/05/80 06/04/80 06/28/80 07/02/80 08/05/80 Number 1391 2329 272e01307 272e01365 08415a 11629a 10224sJ 10297sj 13875a 10694sj 10762SJ 10822sJ 11287sj 10848sj 11116sJ 15810a 11349sJ 16613a 11533sj 11621sJ Table 5.4 Fairchild Uaste Disposal Records Bulk Volume Gallons 0 5040 0 0 3800 0 400 0 4758 0 0 0 0 0 0 4500 0 500 0 0 No. of. Gallons Drums Per Drum 23 55 00 46 55 19 55 00 30 55 00 14 55 00 48 55 ' 8 55 39 55 83 55 19 55 20 55 00 38 55 00 18 55 17 55 Other Quantity Gallons 0 0 0 0 0 30 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Total Volume 1265 5040 2530 1045 3800 1680 400 770 4758 2640 440 2145 4565 1045 1100 4500 2090 500 990 935 Confirmed y y y y y y y y y y y y y y y y y y y y FrobUw y y y y y y y y y y y KBR Plffereoce y y y y y y Table 5.4 (Continued) Date 09/05/80 09/30/80 09/30/80 10/24/80 11/18/80 12/01/80 01/09/81 01/29/81 02/17/81 03/18/81 03/31/81 05/01/81 06/15/81 09/07/81 11/05/81 12/03/81 12/04/81 12/12/81 12/18/81 Hiafesi H865sj 19129a 12053sJ 12374sj 201142a n/a 088-022005 088-022154 088-022290 008x23533 088-023607 n/a 088-065131 088-085645 008-066547 088-103124 088-2322 088-016789 008-103027 Bulk Voluae GflUgng 0 700 0 0 0 700 0 500 0 639 0 500 0 1595 440 270 580 500 995 Other No. of Gallons Quantity Drums Per Drum Gallons 42 55 0 000 37 55 0 97 55 0 44 55 0 000 22 55 0 000 39 55 0 0 0' 0 39 . 55 0 0 00 43 55 0 000 0 00 0 00 000 000 0 00 Total Volume 2310 700 2035 5335 2420 700 1210 500 2154 639 2145 500 2365 1595 440 270 580 500 995 Confirmed y y y y y n y y y y y n y y y y y y y Problem y y y y y y y y y y y y y y HBR Differenc y y y y y y y y y Confirmed code: y - DHS did find/ review manifest; n - DBS did not flnd/revtew manifest Problem code: see Table 5.5 problems HBR Difference code: y - see Table 5.6 HBR Differences Table 5.5 Problems vich Waste Manifests DATE 05/31/78 WHSER 1391 PROBLEM Components illegible 04/12/79 272-e-01307 Components include an unknown quantity of pump oil as well as solvents 06/01/79 08415a Barely legible volume 11/05/79 10224sj Confusing concentrations; 55-80% water; also, both acids and solvents in components 11/15/79 12/03/79 10297sj 13875a Kerosene listed as a component Volume barely legible 02/04/80 02/28/80 10694sJ 10822sJ Description of components illegible; possibly soil P0C13 listed as a component 03/15/80 11287sJ Illegible record 03/24/80 10848sj Unknown quantity of oil listed as a component 05/01/80 llllfisj Barely legible volume 09/30/80 19129a 40% water 09/30/80 12053sj Oil listed as a component 10/24/80 12374sj Illegible * of drums of HF acid 12/01/80 n/a Date is actually 12/80 (no day specified in MBR) 01/29/81 02/17/81 088-022154 088-022290 Oil listed as a component; 80% water 75%-97% water 03/18/81 088-723533 95% water 05/01/81 n/a Date is actually 5/80 (no day specified in MBR) 06/15/81 088-065131 Components barely legible; illegible quantity of corrosive material 09/07/81 088-085645 Barely legible manifest #, volume, components 11/05/81 088-066547 Barely legible manifest *, volume; components c onfusing/missing 12/03/81 088-103124 Corresponding IT invoice refers to 088-103124 with volume equal to "five drums" 12/04/81 088-102322 12/12/81 088-016789 Components barely legible; pickup after larger, failed tank disconnected Day of the month barely legible; 80% water 5.35 SL 033433 Table 5.6 MBR/DHS Differences in Interpretation of Manifests DATE 08/15/79 11/05/79 11/15/79 03/24/80 05/01/80 08/05/80 09/05/80 09/30/80 12/01/80 01/29/81 03/31/81 05/01/81 NUMBER MBR AMT DH5 AMT EXPLANATION 11629A 0 drums 30 drums MBR does not list this 2 drums record 10224SJ 4 drums 400 gal manifest states 400 gal 10297SJ 25 drums 14 drums 11 drums were empty 10848SJ 22 drums 19 drums 3 drums listed as having contents other than solvents 11116SJ 63 drums 20 drums 43 were HF drums 11621SJ 90 drums 17 drums 73 were HF drums 11865SJ 86 drums 42 drums 44 were HF drums 12053SJ 0 drums 37 drums MBR does not list this record N/A 700 gal 700 gal Unconfirmed; listed in MBR but not found by DHS 088-022154 0 gal 500 gal MBR does not list this manifest 088-023607 8 drums 39 drums manifest states 39 drums N/A 500 gal 500 gal Unconfirmed; listed in MBR but not found by DHS 06/15/81* 088-065132 09/07/81 088-085645 11/05/81 12/04/81 088-066547 088-102322 7 drums 0 drums 0 gal 1595 gal 54 drums 0 gal 440 gal 580 gal Shipment was waste oil--no solvents MBR does not list this manifest manifest states 440 gal MBR does not list this manifest *Not included on Table 1 waste disposals 238.tb6/8 5.36 SL 33434 Solvents Acetone I PA Freon Xylene 1.1,1 TCA HMDS +802 ) Photo-Resist A-30 J-100 3,5,1 Premix M-320 Table 5.7 Comparison Between MBR Average Monthly Solvent Purchases for 1981 and Uastewater Discharge Permit Average Monthly Solvent Usages for 1982 MBR Average Monthly Solvent Purchases (19813 1279 1065 206 477 1131 UVD Permit , Average Monthly Solvent Usage (1982^ 1094 925 67 182 1125 191 217 406 NA 216 228 438 745 0 NA 0 15 238.4 5.37 033'*35 Table 5.8 Quarterly Comparison of Materials Balance Report with DHS Calculations (all figures in gallons)* Solvent Inflow OTR MBR 1978 1 2432 2 1581 3 2777 4 2898 Total 9688 DHS 2734 2524 3810 2554 11622 1979 Total 1 4348 2 4068 3 4578 4 SIZt 18148 6010 6825 5774 5332 23941 Solvent Outflow" MBR 0 1265 0 5040 6305 * 0 6330 0 mi 12883 . DHS 0 1265 0 5040 6305 0 7375 1680 5928 14983 Difference-Imbalance (Inflow -Outflow) MBR 2432 316 2777 Lr21421 3383 . 2734 1259 3810 (-2486) 5317 4348 (-2262) 4578 1.-1399) 5265 6010 (-550) 4094 (-596) 8958 1980 1 8834 2 5875 3 17573 4 18680 Total 50962 8439 8723 22728 19907 59797 11000 10555 11370 8455 41380 10835 8190 6970 8455 34450 (-2166) (-4680) 6203 10225 9582 (-2396) 533 15758 11452 25347 1981 Total Tot. 1 17014 27755 2 18137 21141 3 14951 22235 4 13093 17806 . 63195 - -88937 141993 184297 7 42304* 4434 3250 0 Jt7 35 12419 72987 7 3365 6639 2665 1595 2785 13884 69622 12580 14887 14951 8358 50776 69006 21116 18276 20640 15021 75053 114675 7 45669* Please see Notes, 6 and 7 for more information. Hauled solvent waste only; does not include evaporation. 238.tb8 5.38 sv 0331,36 Appendix A. Fairchild South San Jose Material Balance Report INTRODUCTION This material balance report for solvent materials at Fairchild's South San Jose facility was prepared at the request of the Regional Water Quality Control Board. Fair child has fully cooperated with the Board and other interested public agencies, and has attempted to determine as precisely as possible the extent of the discharge of solvent materials into the ground at the South San Jose facility following the failure of a solvent waste storage tank. In the theoretical ideal, a material balance is based on the law of conservation of matter; i.e. the total mass of all materials entering a fixed system in a given time must equal the total mass of all materials leaving, plus any accumulation that occurs in the system. However, such precision is attain able only under laboratory conditions. In the first instance, industrial realities such as time lags between purchase of materials and the use of those materials in production, stock piling, and inventory accumu lation tend, to exaggerate the volume of materials assumed to be used in production at any particular point in time. More over, as practical matter it is impossible to measure the ! disposition of solvents used in the manufacturing process with one hundred percent (100%) accuracy. For example, some solvent residue remains on surfaces that are cleaned with these -1SL 033437 mat rials; insignificant am unts of water become mixed with solvents from time to time, and small quantities of solvent remain in the "empty" containers in which material was pur chased. Thus, when solvent receipts are balanced against solvent removals, a certain quantity of solvent always appears to have been "used up" in the manufacturing process. No allowance for these inherent variances in solvent balance measurements was used in this study. However, information we have received from consultants and public agencies indicates that measurement variances of between five to ten percent (5-10%) are normal, and can be expected. * Another amount of solvent which is "used up" in the manufacturing process is due to evaporative loss. Fresh air is constantly circulated through the production facility. Small amounts of vapor are inevitably emitted from solvents used in the production process, and these amounts are mixed in with the air. This air from the facility is then "scrubbed" to purify it before it is emitted into the atmosphere. During the first quarter of 1982, Fairchild employed independent laboratories, using E.P.A. approved methods, to conduct tests to determine what portion of the solvent received by the South San Jose facility is lost through evaporation. Those tests indicate that evaporation accounts for the disposition of eleven percent (11%) of all solvent. Such an allowance for evaporative disposition has been included in this study. To evaluate the results of our study of historical data, we attempted to determine the degree of precision which could -2- SL 033438 be attained in a solvent balance study of current data. We carefully measured all incoming and outgoing solvent during the first quarter (January, February and March) of 1962. A balancing of solvents received against solvent waste removed (plus evaporation) during this period accounted for the dis position of only ninety-six percent (96%) of the total solvent for that quarter. The results in this study are based on an exhaustive review of current data, and support the view that solvent balance measurements are, at best, imperfect and that a variance of up to ten percent (10%) is to be expected. METHODOLOGY The present material mass balance focuses on solvents and solvent-based chemicals used in semiconductor manufacturing perations at Fairchild's South San Jose Facility. These materials are used in general cleaning within the facility as well as in the manufacturing processes of wafer cleaning, wafer inspection, wafer coating, wafer etching, and wafer spinning. The following solvent materials are used at Fairchild's South San Jose facility and are accounted for in this material balance analysis: A-30 (photoresist remover) Acetone Isopropyl alcohol (IFA) -3- SL 033439 Freon Hexelmethyldisilane (HKDS) J-100 (photoresist remover) Markham 320 Cleaner Methyl alcohol 3.5.1 Pre-mix Photoresist material 1, 1, 1 Trichioroethane (TCA) Xylene The material mass balance data for this report has been developed from the following sources: 1. Records of chemical issues to the San Jose Facility from the Fairchild Mt. view Central Stores and South San Jose purchase records. 2.- Records of solvent waste removal manifests for shipment to an approved disposal site. 3. Empirical measurements of ingoing and outgoing solvents taken during the first quarter of 1982. The period for which this material mass balance was conducted is January 1977 through December, 1981; i.e. from the opening of the facility until the detection and replace ment of the defective solvent waste tank. The data set forth herein accurately reflects all of Fairchild's records. However, chemical waste disposal for the first part of this period may not be complete. While all of Fairchild's records were thoroughly reviewed, we were unable to complete an equally thorough 4* SL 033A40 review of the files of certain public agencies and haulers of chemical waste. INCOMING CHEMICALS As the first step in establishing the material balance for solvents it was necessary to examine all sources of material inputs to the Fairchild South San Jose facility. In the beginning of the period of study in 1977, chemicals were shipped to the South San Jose facility from Fairchild's Central Chemical stores Department in Bldg. 9, Mt. view. All of these shipments were tabulated on a weekly report. An additional material input to the facility was batch shipments of photoresist materials from two different vendors to the production group during 1980 and early 1981. These shipments were accounted for in the incoming chemical mass balance. In 1981, Fairchild stopped shipping chemical materials to San Jose from its Mt. View facility. Instead, chemicals were purchased by San Jose directly from vendors. San Jose pur chasing records were used to. compile information in this report on chemical shipments from 1981 to the present. Finally, another input taken into consideration was small amounts of materials brought into the facility by manufacturers representatives for demonstration purposes. These quantities were insignificant and, therefore, were not used in determining the overall mass balance. -5033*U SL Included in Appendix A are lists of all solvent received by the South San Jose facility from 1977 through 1981. These receipts are broken down by month and by the type of solvent received. OUTGOING CHEMICALS There are three methods by which waste solvents are col lected for disposal. These methods are: A. Bulk Solvent Tank (Large): The Fairchild building * production areas are equipped with a waste solvent drain col lection system. Certain sinks are reserved for solvent dis posal; no acid waste and no water other than small amounts mixed with solvent waste is placed in them. Solvent waste is then piped from the sinks to a holding tank. The tank is emptied by tanker trucks when it reaches capacity. Until December 1981, when the tank was discovered to have failed, the solvents from wing A and Wing B were collected in a 6000 gallon solvent waste tank. That tank has since been replaced with a temporary 1000 gallon solvent waste tank. This tank is being used pending the installation of a new permanent tank which is to be enclosed in a concrete vault. B. Bulk Solvent Tank (Small): The small bulk solvent tank is located near the Chem Mix building to collect waste solvent from that area. It has a capacity of approximately 550 gallons. Like the larger solvent tank, the small tank is emptied when full by a tanker truck. The Chem Mix Building is -6- SL 033442 equipped with e solvent waste drain which is connected to this tank. C. Drum Solvents:. Certain solvent waste is also col lected at various points in the production process in portable seven-gallon tanks. The seven-gallon tanks are mounted on rolling carts, taken to the Chem nix Building and emptied into 55 gallon drums. These drums are periodically removed by a licensed hauler of chemical waste. Included as Appendix B are tables showing the dates and volumes of solvent materials removed from the South San Jose facility. These tables were prepared from shipping manifests and the records of licensed haulers of chemical waste. CORRELATION OF SOLVENT USAGE TO SOLVENT REMOVAL ' Correllation or balancing of solvent purchase data with waste removal data is difficult for at least two reasons. First, solvent is purchased substantially in advance of the time it is employed in the production process. Second, waste solvent is not removed every month. Solvent accumulations in . the waste tanks and solvent waste drums are retained until a removable volume is accumulated. Thus, removal of any gallon of solvent tends to take place long after that gallon of solvent was purchased. To correct for this disparity, we attempted to determine the time at which solvents were accumulated for disposal. For -7- u3 example, in Nay, 1978, 1265 gallons of waste solvent were removed in drums. (See Appendix C, p. 1.) The next removal of solvent in drums occurred in April, 1979. However, some portion of the 2530 gallons removed in April, 1979 was accumu lated during 1978 (after the May removal of waste solvent in drums). To determine what that portion was, the period between removal is measured (11 months) and the volume of the April, 1979 removal is divided by the number of months during the period (2530 j 11 * 230). Seven months of the solvent waste removed in April, 1979 should be attributed to June - December, 1978. Thus, 1,680 (7 x 230) i added to 1,265 and the total solvent drum waste for 1978 is determined to be 2,875. Similarly, solvent purchases tend to vary markedly from month to month. To determine the representative level of solvent usage for each quarterly period, a three month moving average was calculated. Such an average yields the number of gallons of solvent used in a "normal" month during that quar terly period. (The statistician's worksheet for this calcu lation is included in Appendix C, p. 3.) When the results of these two calculations are compared, there is a strong correlation between quarterly solvent receipts and quarterly solvent removals until mid-1980. Solvent removals consistently increase in relation to increases in solvent receipts through the second quarter of 1980. Thereafter, solvent receipts increase markedly and solvent removals begin to diminish. (A graph illustrating this relationship is -8- SL 033444 presented in Appendix D, p. 2.) Thus, such a comparison tends to indicate that the solvent waste tank failed sometime after mid-1980. The graph presented in Appendix D, p. 3 illustrates the cumulative difference in gallons between total solvents received and total solvent dispositions (removals plus evaporation). Note that the cumulative difference decreases in the first and second quarters of 1980 because the solvents removed during those two quarters exceeded the solvents received (i.e. solvents received in earlier quarters were used and disposed of in the first half of 1980). After mid-1980, the cumulative difference between receipts and dispositions markedly increases. Thus, this comparison further tends to indicate that the solvent waste tank failed sometime after mid-1980. CONCLUSIONS When records of solvent receipts are "balanced" in the manner described above against records of solvent disposals, the end result is that until mid-1980 essentially all of the solvent used at Fairchild's South San Jose facility can be accounted for. From the date the facility opened until mid-1980, 46,371 gallons of solvent were purchased. Of that total, only 1,139 gallons were 1,1,1 TCA, and the majority of that was purchased in the first half of 1980. (See Appendix C.4.) 9- 033^5 Si* From mid-1980 until the discovery and replacement of the defective solvent waste tank in December, 1981, 102,446 gallons of solvent (including 13,956 gallons of 1,1,1 TCA) were received by the South San Jose facility. We are able to account for the disposition of forty-three percent (43%) of this material. Thus, the maximum number of gallons of solvent which may have been lost appears to be 58,394 (102,446 x .57). Thus, within the accuracy parameters already explained, all solvent used at Fairchild's South San Jose facility can be accounted for until mid-1980. After mid-1980, an apparent imbalance occurs. On December# 4, 1981, the failure of the solvent waste tank was discovered. Solvent receipts and disposals returned to a "balanced" status in January 1982. The data, therefore, tends to indicate that the solvent waste tank failed, at the earliest, after mid-1980. 10- u*> APPENDIX A SOLVENT RECEIPTS 1977 - 1981 033^47 SL 1977 SOLVENT SUMMARY (PURCHASES) T sr i s E LiJ U. j UacJJ * *< i 3 <>>, 3i 00 /> 1 . , **-** PHOTO RESIST - -- nf v]i ill va iU n{! ?' 7 aw- 2 iA X 1 'i i_ in H 1 JANUARY :EBRUARY 1ARCH 1 25 2 3? i M 9 1PRIL 1AY JW JULY MJGUST 6 12 10 * <i 5 M 0 10 0 12 V 76 60 IS 8 16 6 n 10 99 76 35 8 12 IEPTEMBER JCTOBER 12? 118 96 8 20 80 28 IOVEMBER 12 50 8 122 12 28 ECEMBER 18 91 1? 70 3? 8 T* M3 362 J21 Jdl -hi OTAL SOLVENTS - 1,M71 GALLONS OTAL 14,1 TCA: 62 GALLONS 20 ML Jii __a. Appendix A.l - 1 PHOTO RESIST 1 A-30 J-100 1978 SOLVENT SUMMARY (PURCHASES) u s u35UJ. w *< 2 mmmmmm%m 1 + ts> I X ? UJ rR\ ^m* i JANUARY 0 171 55 273 18 2 144 0 57 EBRUARY 1ARCH 60 195 155 175 36 12 193 0 64 0 218 55 515 0 4 3? 47 0 \PRIL 0 150 110 54 S 0 8 20 1 20 1AY 18 74 ,0 253 0 1 12 4 0 W -JS. 134 5? 324 0 4 28 7 0 JULY 36 749 110 791 0 0 0 60 0 UGUST 35 142 110 16(1 0 8 0 32 4Q EPTEMBER iCTOBER OVEHBER 53 52? 440 292 p 117 350 9 698 11 33 198 no 378 7 8 0 355 2? 0 28 94 24 5 46 60 21 ECEMBER 33 198 up 32? TOTAL 395 2377 1320 I3RL TOTAL SOLVENTS - 9,688 GALLONS TOTAL 1,1,1 TCA: 79 GALLONS 7 5 4p 60 21 79 57 549 720 250 Appendix A.2 033**9 ACETONE PHOTO RESIST 1979 SOLVENT SUMMARY (PURCHASES) Si s tKsu lx. 1e IXJ --J fc % cn *i S 2 "1 X e tn-*i Kl 8 1 JANUARY 320 640 490 55 2 12 CEBRUARY 1ARCH 288 108 55 648 n490 260 165 4 22 108 11? 10 72 ?? 288 103 58 9 APRIL 1AY JUI_ . JULY AUGUST 330 236 111 569 * 440 433 110 381 108 108 110 368 165 "ft 21 4 108 37? 336 55 5?7 55 330 216 55 3?7 4 108 12 1: 6 1? * 10 SEPTEMBER JCTOBER 612 3?0 523 15? 440 361 660 35? 288 108 144 10 12 I0VEMBER 648 300 110 155 16 22 1C ECEMBER 330 540 110 438 54 30 288 108 f Ti -'L 4055 4470 1926 300? JSL _I2L _A 1220. _M8_ 172 _121 TOTAL - 17976 GALLONS fOTAL 1,1,1 TCA: 327 GALLONS Appendix A.3 033^5 St 1930 SOLVENT SUGARY (PURCHASES) X PHOTO RESIST * ACETONE 2 NO CD uUoScJ. liJ 5 s H* *** %+/% E roo1 ooi-r ID N**1 mK\ a JANUARY :EBRUARY 1ARCH Uii 75R 0 765 no 48 220 199? 400 546 no 0 605 1216 330 114? 491 4 ?52 0 0 ]44 440 0 15 15 0 IPRIL 1UNC JULY 1135 432 220 795 0 648 0 215 550 -52a 2 432 ML im U2L 923 0 0 6 144 324 0 0 22 288 0 0 O 24 292 0 ') 0 283 215 AUGUST 1784 m 4?n 1239 221 0 83 144 0 iEPTEMBER OCTOBER IOVEMBER 4256 2333 _25. 1459 155 2911 2155 130 0 0 3714 1535 55 1155 0 0 81 432 0 0 1'7 288 0 n 40 144 108 lECEMBER w: 113? 155 754 0 0 96 432 108 TPT'L 12252 13492 2177 9430 ioaz. 52 -5fiL 3144 TOTAL SOLVENTS - 50,962 GALLONS TOTAL 1,1,1 TCA: 1, 097 GALLONS Appendix A.4 900 46 ?n 0 20 * 20 36 14 30 42 259 SL 033451 ACETONE P M !________ 1 1PHOTO RESIST 1981 SOLVENT SUfTtARY (PURCHASES) 2 6 to % v B * R _m 2 X 1 in ft i JANUARY FEBRUARY MARCH APRIL BAY_ JE . JULY AUGUST 1?55 55 J65 0 0 144 1006 1083 383 220 605 2C 1?C 576 324 24J? 2509 275 1406 1650 2)2 495 430 108 9?0 935 ?75 685 1320 275 44E 144 0 990 1065 33C 1045 825 12 355 144 216 1815 1630 1?1Q 650 273 NWS 2145 22C in 755 22? 328 28$ 1' I 316 280 288 216 216 1320 880 ??: 715 1870 J96 2(?0 144 648 SEPTEMBER OCTOBER 1265 660 53 1210 248 30S 144 648 m1110 11c no 1100 240 284 144 1440 NOVEMBER 990 825 11: 110 1320 272 292 144 432 DECEMBER 990 550 nn 770 144 48 144 864 "*TAL 53H? 12785 2W5 5721 13570 2m M5E 2532 5255 TOTAL SOLVENTS -64,910 GALLONS TOTAL 1,1.1 TCA: 13,370 GALLONS ' Appendix A. 5 * * * sh 033*52 APPENDIX B SOLVENT WASTE DISPOSALS 1977 - 1981 Date 12/18/78 WASTE BULK SOLVENT SHIPPED FOR DISPOSAL 1970 Volume/Gallons Manifest No. 5,040 2329 May, 1979 12/3/79 Total 1979 3,800 4,958 8,758 (12/9/81 Correspondence from i.T. corpora ation) 13875 5/5/80 6/28/80 9/30/80 D C. 1980 Total 1980* 4,500 500 700 700 6,400 *- j 15810 16613 19129 (12/9/81 Correspondence from I.T. Corpor ation) 3/18/81 May 1981 - 12/03/81 12/15/81 (Date Unclear) 12/18/81 Total 1981 639 500 270 500 995 2,904 088-023533 (12/9/81 Correspondence from I.T. Trans 088-103124 088-016789 088-1030Z7 Appendix B.l SI. SOLVENT DRUM PICKUP 1978 Date 55 Gal Drums _________ Manifest No. 5/31/78 Total 23 23 X 55 1,265 Gal. 1391 1979 1 4/12/79 11/15/79 11/15/79 Total 46 4 25 71 x 55 4,125 Gallons 01307 10224 10297 1980 2/4/80 2/15/80 2/28/80 3/15/80 3/26/80 5/01/80 6/04/80 7/02/80 8/05/80 9/05/80 10/24/80 11/18/80 . Total 48 8 39 83 22 63 38 18 90 86 97 44 673* X 55 37,015 Gallons 10694 10762 10822 11287 10848 11116 11349 11533 11621 11865 12374 201142A 1981 1/09/81 2/17/81 3/31/81 6/15/81 6/15/81 11/5/81 Total 22 39 8 43 7 54 x 55 9,515 Gallons 988-022005 088-022290 088-023607 080-065131 088-065132 088-066547 Includes a small number of empty hydroflouric acid drums. Appendix B.2 33 455 APPENDIX C STATISTICAL WORKSHEETS * SL 033456 t/1 C a 03 9O-' Jan Feb Njr April n.r June July AU| Srpt Oil Nov _cs Jan Frb Mar April Hay June July Auk Srpt Oct Hov Drr Jan Frb Mar April May June July A" A Srpt Oct Hov Dec GAL 3060 3BOO 4950 4500 500 700 700 WASTE bulk __________________ _____________________________________ WASTE DRUMS AVC AVC--------- AVC RUN RATE AVC RUN RATE PEA MONTH PER OUT cut CAL PER MONTH PER CRT AIO 253 410 253 420 420 1260 253 253 420 233 420 1263 253 420 420 2320 230 243 410 y 230 420 230 420 420 3700 230 230 420 230 420 230 420 420 3040 230 230 760 230 760 230 ` 760 760 7320 230 230 760 2330 230 760 220 700 743 9340 220 229 700 220 700 220 700 700 11672 220 220 70S 220 700 1595 220 700 700 13796 1742 733 900 1762 900 3223 1742 900 900 16496 5775 3775 3006 91JO 1733 900 3463 1733 500 767 10796 2090 2090 1052 233 990 990 231 4950 4950 233 233 19496 4730 4730 3557 233 SJ35 5335 213 2620 2420 233 233 20196 605 2707 CUM 759 1494 2104 2074 3564 4251 4935 7134 16392 21960 32619 40900 TOTAL CUM 2019 4014 3964 7914 10004 13799 16607 20930 32000 40744 52113 61176 Appendix C.1 1981 J*n feb tt#r April H.y June July Au* Sept Oct Mo Dec CAL 639 500 1765 WASTE BM AVC AVC mm rate TER HOWTO PER OAT 213 213 213 213 250 250 252 251 252 1 252 252 252 252 252 252 252 CUM 20835 21587 22343 23099 CAL 1210 21*5 440 2750 2970 AVC per norm 605 21*5 <00 917 917 917 59* 59* 59* 59* 594 594 WASTE KWHS AVC RUN RATE TER QRT CUM 1050 917 594 59* 44130 46881 48663 49851 TOTAL CUM 64965 68468 71006 72953 SL 033458 Append' C. 2 T977 19TB 1979 Jen Feb Har April Hay June July Auq Sept Oct Nov Oec Jan 7eb Har April Hay June July Aug Sent Oct How Dec Jan feb Mar Apr! 1 Hay June July Aug Sept Oct Now Dec HtIu *ti. (3) 81 42 9 190 S2 7)0 4S0 78 7)7 IB) 661 900 871 657 167 56? 746 517 1504 I7BZ BOB 'OB 519 1355 1474 1758 1592 I7IB 1494 1050 70)4 1979 1771 1904 44 80 84 157 744 2)6 237 147 358 5B1 811 809 630 527 557 608 972 1101 1198 966 1045 1227 1449 1362 1441 1356 1435 1254 1526 1688 1761 1718 1864 S-- by QmrUr 708 637 742 2201 1714 2631 3209 4038 SOLVENTS _XL|VE0 CW 208 845 1587 3788 5502 8133 11342 15380 I960 1981 1982 Jan Feb * Har April Hay June July Au9 Sept Oct Nov Dec Jan Feb Har April Hay June July Auo Sept Oct Nov Dec Jan Feb Mar - k? HwiM *|. (3) Quarter 2418 2773 4522 3340 1465 1897 4020 4947 9749 580? 7068 6426 3191 4461 9678 5413 5240 3135 414) 654) 4734 4883 474) 3796 5987 5613 7468 2365 3238 3545 3109 2234 2461 3621 6072 6666 7373 6432 5562 .469) 5760 6501 6760 676) 5839 6274 5140 5)87 4787 4474 4842 5132 6)56 5599 9148 7804 16359 19367 16954 1BB62 16801 14103 17087 CIH 38571 46375 62734 82101 . 99055 117917 134718 148821 165908 4232 19612 4468 24080 5343 29423 SL 033459 Appendl* C.3 SOLVENT USED v. SOLVENT DISPOSITIONS (IN GALLONS) Tear 1*77 Solvent* Used* Total Solvent* 1'ied! 11i2 1,587 62 1S7I *,755 7* 1*79 18,081 327 1/BO 6/SO 16,952 671 *7/80 - 12/SO , 35,726 '386 19B1 66,720 13,570 Solveot* NeMoved** Drug* Bulk Evaooratioo Total Out fS 175 175 2.875 5,040 1,073 8,988 4,260 8,757 1,989 15,006 14,816 5,001 1,865 21,680 19,032 1,398 3,930 24,360 9.465 2,904 7,339 19,708 Solvent* Died* 'ear Total Solveot* Used (TCAJ 1*82 19,068 3,616 SUBTOTAL 1977-Mid*1980 66,371 1,13* SUBTOTAL id-60-12/81 102,666 13,956 Solveot* Neooved** Drug* Bulk Evaooratioo 6,750 ' i 9,425 2,097 Total Out 18,272 41 21,949 18,798 5,102 45.849 11 28,497 4,302 11,269 46,068 571 Based an three-oonth Moving average of solvent* racaivad * Based on data solvent stcuoulated for reaoval. APPENDIX C.4 SL 033460 APPENDIX D GRAPHS AND TABLES 3346l 1977 1978 1979 1980 1981 FAIRCHILD'S SOUTH SAW JOSE FACILITY CHRONOLOGY OF OPERATIONS (April) South San Josa Plant begins operation -start-up phase. Only wing A (one-half of present facility) facilitized. Full production commenced in wing A. Further facilitization of plant to prepare for opening of wing B. Chemical inventories built up late in year to prepare for expended operations. (April) Expanded operation begins. Failure of large (bulk) solvent waste tank dis covered in December; appropriate agencies notified, temporary solvent waste tank installed immediately, remediation program begun, materials balance study initiated. APPENDIX D.l SL 033462 r RELATIONSHIP OF SOLVENT USAGE TO SOLVENT DISPOSAL CALLONS 25000 ----------------------------------------------------------------------------------------------------------------------------------------------------------------- KEY: 2COCO 15000 SOLVENT RECEIVED EVAPORATION DRUM DISPOSAL BULK DISPOSAL ("i Is f 10000 \ 5000 v* 2341 1977 l i II11 234123412341234 1 1978 1979 1980 1981 1932 Ap dlx 0.2 ACCUMULATED DIFFERENCES BETWEEN SOLVENTS RECEIVED AND SOLVENTS DISPOSED PLUS EVAPORATION GALLONS 100000 ---------------------------------------------------------------------- ------------------------------------------------ --------- 90000 -- 80000 -- 70000 -- 60000 -- 50000 40000 -- 30000 -- _i '_T' ni:i 4 f ; T1 ' j] t: !; ; 20000 -- ioooo -- [1 f ; r '! : . 1 .i f 12 3 4 1977 234 1978 234 1979 234 1980 1981 1982 SL 033464 Appendix D.3 Appendix B. Database Structure for Fairchild Chemical Purchases Identification Number Chemical Company Bernal Road - Document Type * Blanket P.0. Number - Fuchaae Order Releaae Number Invoice Number - Requisition Number Date Requested Solvent Name Solvent Code Quantity * Container - Unit Unit Cost Box and document number assigned to the document by Landels, Ripley and Diamond. Name of the chemical vendor. Entered to assure that the chemical was purchased for the Fairchild South San Jose facility on Bernal Road. The type of document used to order chemicals. Most documents were purchase order releases. We had very fev purchase requisitions and invoices. The number of the blanket purchase order agreement from which the blanket purchase order releases were drawn. The pre-prlnted number on the purchase order release document. The number on invoice documents. The number on purchase requisition documents. requested. The name of the solvent purchased. The coded grouping in which DHS placed the solvent. The number of units of solvent ordered and denoted as received. The type of container the solvent was dispensed in. Most solvents were sold in one gallon\ bottles, five gallon pails, or 55 gallon drums. The unit of measure by which the solvent was sold. Cost of the solvent per unit. We took this information to help us calculate the total solvents purchased. 5.39 SL 033465 Total Cost Total Gallons Solvent Name, etc. to Total Gallons - Requester Initials Date Received Bill of Lading Illegible Other Problems - Total cost for the quantity of solvent purchased. The total gallons of solvent purchased. This data was taken directly from the chemical purchase document or it was calculated by DHS. These data items are repeated for every solvent listed on a purchase document. There were as many as nine different solvents ordered on a single purchase document. The initials of the Fairchild employee who made the chemical purchase request. The date the chemicals were received. Ve entered a "Y" for yes if there was a bill of lading from the chemical vendor attached to the purchase order. Ve entfcred a code to denote legibility problems. Ve entered a description of the type of problem. 238-aB.8 5.40 SL 033466 ,9 at t ) -2O* . i Tr3o3 . ii a. x :j *l Hazardous Waste M anifests !I a. U1 o ToBJ (ab.