Document JN4QRzYma1MVqEXLaXbryMOd2

B 'E 'P. 1 "5" 1 Mm ortality of Workers Employed II in the Manufacture of Chlordane i Ii Helen H.Wang, M.D., and Brian MacMahon, M.D. I ALretrospective mortality study has been carried out on Central to the suspension proceedings and to the ban workers employed in the manufacture of chlordane and subsequently negotiated was concern that these com- i heptachlor between 1946 and 1976. The study group was comprisedof 1403 white males who worked for more than pounds might be carcinogenic to man The evidence is controversial and derives entirely from long-term feeding I three months at either of the two plants in the United cing these compounds. tnformation on minated employees was obtained from y Administration and supplemented b y I ted by another investigator by individual k>lloH.op. There were 113 deaths observed in the group, 157 expected, giving a standardizedmortality f 72. There was no overall excess of deaths even among workers followed twenty or I more years after entry into the occupation. There was one death from liver cancer An excess of deaths from lung cancer (72 observed, 9.0expected) was not statistically sig- nificantand was not distributed by duration of exposure or I of latency in any pattern suggesting an etiologic role for chlordane-heptachlor exposure. Although diseases of the studies in laboratory animals There are a few crosssectional clinical studies of heavily exposed humans but essentially no data on long-term mortality associated with chronic or heavy exposure in man A study of the mortality of workers engaged in the manufacture of these pesti- cides at the only two plants currently engaged in their production in the United States is reported here While current levels of exposure are reasonably well character- ized, the total exposure experienced by workers in these plants in the past cannot be reconstructed However, it is very likely that a substantial proportion of workers has been exposed to levels well above those prevailing in the general population The study was undertaken at the invitation of the Velsicol Chemical Corporation which owns and operates dmdatory system as a whole showed fewer deaths than the two plants I there was a statistically significant ex- cerebrovascular disease (7 7 observed, Materials and Methods xcess was not related to duration of ex- Velsicol Chemical Corporation began producing chlor- i W u t e or latency and occurred exclusively after termina- dane in its plant at Marshall, Illinois, in 1946 and hepta- \ f h n of employment i - chlor and endrin in the plant at Memphis, Tennessee, in 1952 Personnel records are available for all of the 951 employees at Marshall and the 1425 employees at Memphis who had ever worked in these plants before the spring of 1976, when this study was initiated After exclu- lordane and heptachlor are chlorinated hydrocar- s a n of females, males who worked for three montia or ich were widely used as agricultural and less, and persons with inadequate identifying information, i pesticides until 1975, when the Environmental ion Agency suspended their use for all but subtertermite control and certain other limited applicacomplete ban for agricultural use is now being in. there remained 1403 eligible subjects No information on race was available but all the deaths identified in the course of the study were of whites Among the 973 employees excluded, approximately 7% were excluded because they were female, 64% because they worked for Dcpsnmmt ofEpldcmalogy, Huvatd khDdd Public Health. 677 A=.. Boston, MA 02llS three months or less and 29% because of missing infor- maton ldentifyrng information was submitted to the Social Security Administration (SSA) which provided, for those workers known dea& Mfwmatm on date of death and I k l i a - u p a d hy Pb-at. - DuratkR d F0lkVr-u~ e10 10-19 20-29 30 + Total Manbrt 129 99 193 149 570 Mmpkk' 352 324 1-59 835 TOW Ne. W 481 422' 351' 149 1403 34 3 30 1 250 106 100 0 'Includes one subject who worked in both plants the state in which the beneficiary filed a claim for benefifi. A Social Security Administration search of its records identified 104 deaths in the study cohort through the end of 1975, for which death certificates were sought from the states and registration areas. Contemporaneously, Dr. Sidney Shindell of the Medical College of Wisconsin had been conducting individual follow-up of terminated workers in these plants. He informed us of nine deaths in our cohort which had not been identified by SSA, four of which occurred in 1976 after the termination of the SSA search. These have been added to the known deaths and the closing dates of Dr. Shindell's follow-up - June 30, 1976, for Marshall employees and December 31,1976, for Memphis employ- ees - were used as the termination dates for computing person-years of observation. Dr. Shindell also provided copies of some death certificates which had not been ob- tained by the authors. There remain two deaths for which certificates could not be found. The distribution of deaths by cause was based on death certificate causes coded by one cf the authors (HHW) to b e estimat& using a computer klonson.' The program appties n specific for sex, race and age and calehPfer tim year groups to the comparable during the period of observatio the reference cited, the progra national mortality rates throu males were used for this anal person-years, the beginning d 1,1946, for Marshall and January1,1952, for Memphis the employee was eligible on those dates, or at the end of three months employment if he was first employed later. Standardized mortality ratios (SMR) were estimated by dividing the observed deaths by the expected and multiplying by 100 An iterative method based on mid-p values' was used to compute 95% confidence intervals. An attempt was made to examine the relationship b e tween the intensity of exposure and mortality experience among these workers However, a complete occupational history was not available for each individual Moreover, serum pesticide levels of workers actively employed in the plants in 1975 and 1976' did not correlate with a classification of presumed exposure level based on job category Therefore, analyses according to presumed level of exposure are not presented. Results The distribution of the study group by plant and by duration of follow-up IS given in Table 1. Table 2 gives observed and expected numbers of deaths by cause and estimates of the SMRs. The overall SMR is lower than 100 -Table 2. Observed and Expectad Deaths from Selected Causes. Cause of Death All causes Malignant neoplasms Digestive organs and peritoneum Lung Lymphatic and hematopoielic Other Diseases of circulatory system xhemic heart disease Cerebrovascular disease Other Diseases of respiratory system Diseases of digestive system External causes All other causes ___ _ _ 0bserved Expected 113 157 4 24 29 3 7 a2 12 9 0 1 33 4 aa 63 75 a 37 53 9 17 9 3 9 12 6 3 ai 4 a5 14 22 a 5t----13 1- SMR' 72 82 a6 134 30 45 a3 69 1a3 71 37 47 61 3a C.I. of SMRt 59 - a6 54 120 ~ 3a - 170 73 - 228 2 . 151 14 - 109 64 - 106 49. 94 110 - 287 35 - 131 9 101 ~ 15 - 114 35 101 ~ 14 - a5 'SMR = Observed/Expected x 100 (computed from values expressed to two decimal places) 795% exact confidence intervals computed by iteration based on mid-p values' $Including two with missing death certificates -1 Table 3. Observed and Expected Deaths from Selected Causes According to Plant. I Cause of Oeath All causes Malignant neoplasms Cancer of digestive organs and peritoneum Cancer of lung Ischemic heart disease Cerebrovascular disease External causes 746 Marshall Memphis Observed ExDectsd SMR Dbserved Expected 76 107 1 71 37 5 0 1 13 20 1 65 11 9 1 5 5 9 a5 2 2 3 74 121 88 7 6 1 115 5 2 9 174 26 3a I 68 11 15 6 70 11 7 0 158 6 2 3 260 - 9 12 3 73 5 10 4 4a Mortality of Workers Manufacturing Chlordane and Heptachlor/Wang and MacMahon 1 I 1 I I i According to Age at Observation and Age at -Table 6. Observed/Expected Deaths from Cerebrovascular Disease According to Age at Observation and Age at Entry into Occupation. Age at Observation <50 50-64 65 + Total Age at Entry into Occupation <35 35-49 50+ Total 2/1 1 0/07 - 211 8 110 4 3/20 3/1 8 7/4 3 2/05 6/27 8/32 3/1 5 5/33 914 5 17/9 3 and significantly so, since i t s confidence interval does not in the subgroup with intermediate durations. It is n o t e ! include 100 However, the value (72) is typical of that observed in an employed population This deficit of worthy that in the person-years with 20 or more years of latency there is a statistically significant deficiency of observed deaths relative to expected seen in working lung cancer deaths. The lack of any clear pattern in this populations is generally attributed to the selective factors table casts doubt on the meaningfulness of the relation- associated with employment ship seen in Table 4. A - Observed deaths from cancer are also fewer than expected, but not significantly so Only one type of cancer - cancer of the lung - shows more observed deaths ---fnmtaranalyses -of- deaths from cerebrovascular disease are given in Tables 6 and 7. Within the variability inherent in the small numbers in these tables the excess than expected, but again the difference is not statistically appears to be distributed more or less evenly throughout significant There was one death attributed to carcinoma and there is no clear relationship with any of the four vari- of the liver This occurred in a man who worked for the ables. Although 3.2 deaths from cerebrovascular disease company for five years, beginning in 1944 when he was 68 would have been expected among currently employed years old, and who died in 1958, at the age of 81, of "car- workers, there were none. Six deaths from this cause oc- cinoma of the liver " The carcinoma was not specified as curred within five years after termination; four occurred primary but was so coded by us The expected number of 5-9 years after termination; and seven, 10 or more years deaths from cancer of the liver and biliary tract (ICD8 after termination. The expected values for the latter three #l55, 156) was 059 There was one death each from periods were 1.7, 1.5 and 2.9. cancer of the bladder, the prostate and the central ner- The effects of worker selection are explored for deaths y w s system, and one with primary site unspecified from two principal causes and from all causes in Table 8. While diseases of the circulatory system as a whole ac- All three categories show substantial deficits of observed count for fewer deaths than expected, there is a substan- deaths in the first 10 years of the follow-up period. For tial and statistically significant excess of deaths due to deaths from malignant disease, the effects of the selec- cerebrovascular disease No other causes of death are in tive factors which are presumed to be responsible for excess. these deficits seem to have disappeared by 10 years after Table 3 gives similar data separately for the two plants the beginning of the follow-up For deaths from ischemic The features outlined above are present in both sets of heart disease and for all deaths, the deficits persist data throughout all three categories of duration of follow-up, Table 4 shows observed and expected lung cancer though not so markedly as in the first ten years. deaths according t o age at entry into occupation and age . at observation Almost the entire lung cancer excess oc- Discussion curs among workers who were aged less than 35 at entry To the authors' knowledge this is the only published into occupation and less than 50 at observation In this study of causespecific mortality among workers occupa- particular category the difference between observed and <expected is statistically significant (p 001) Because of tionally exposed to chlordane or heptachlor and there are, therefore, no other data with which these can be small numbers in the individual cells of this table it is not compared directly possible to tell whether the excess is a function of young The impetus to the study was a search for evidence of age at entry, young age at observation, or both Relation- carcinogenicity of these compounds to man None has ships to duration of follow-up ('llatency") and duration of been found, although the study population is too small employment are explored in Table 5 Again numbers are and the period of followup too short to translate this into small but there is no pattern suggesting a real relationship a statement that there is no excess risk of cancer with either variable For both variables the excess appears associated with exposure in man There was no excess of -T8bk 5. ObservedfExpected Deaths from Lung C a m r According to Duration of Employmenl and Duntian d F~OW-UD. 31-1.5 311.5 211.3 61-2 1 013.4 011 4 010.8 6f2. 111 0 1/18 114.0 -12f90 -Tabb 7. Obsmed/Expected Deaths from Cerebrmrcutar Disease According to Duration of EntpbVmnI an8 Duration of Follow-up. Duration d EmP(OYmeA1 <lo 2100-+19 Total <10 -3/24 - 3t2.0 Duratbn d Follow-up 10-19 20+ 311 4 21-2.1 513 5 311 3 611.3 011.3 913 9 Total 914.7 813.4 W1.3 17193 1 ccrvud OISm All causa I Malignant neoplasms Ischemic bearl disease a 21l44.4 31 6.6 4f10.8 ..- a+ -Dvntkr d lbbmr) . I-` 1QlS 45f58.4 14/ 11.0 16/ 21.3 47f51.7 ?/11.6 1?/21.8 18bl 113f157.5 24/ 29.2 37/ 53.8 deaths from cancer compared to the number expected, even among persons followed 20 and more years after first employment. There was only one death from liw -cancer the tumor associated with chlordane exposure in mice.' The only tumor in excess in these data is cancer of the lung. For this tumor the excess is not statistically significant and does not follow any pattern of association with duration of exposure or latency that suggests an etiologic relationship between occupational exposure and lung cancer. The fact that there are no data on cigarette smoking in this study group is further reason for caution in interpreting the observations on lung cancer. A surprising observation is the statistically significant excess of deaths from cerebrovascular disease. This a p pears not to have been observed or suspected previously and should be confirmed in other data or further followup of these workers before it is accepted as evidence of a consequence of exposure to chlordane or heptachlor. There is, however, some equivocal evidence of increased blood pressure among pesticide workers. Thus, no tn- crease in blood pressure was reported in three surveys of 'workers specifically exposed to chlordane.' In one of these surveys two cases of hypertension were found among 24 workers but there was no control group and the finding was considered incidental.' Increased systolic and diastolic pressures have been reported among workers with mixed pesticide exposures9 lo but lack of increase has also been reported." Hypertension was reported in two studies of workers exposed to DDT, another chlorinated hydrocarbon, but in the absence of a comparison group in either study the findings are difficult to interpret.'* lJObviously, research in this area is in an un- satisfactory state. The relationship of pesticide exposure to blood pressure and to cerebrovascular disease serves more investigation. The authors are indebted to Sdney Shindell, M.D.of the MedKal Wwomrn for supplementing the informationon deaths in the rtudv Socul kcunty Admmistratmn carrled out the matching of m d w d cohat with Social kcurity records References 1 U S Department of Health,Educaton and Welfare Eighth R s a n International Classification of Diseases Adapted for Use in United States Public Health Service Pub1 No 1693 Superintendent of Documents, U 5 Government Printing Office. Washington. D. 1%7 2 Monson RR Analysis of relative survlval and proportional tality Comput Bromed Res 7 325-332, 1974 3 Wang HH and Miettinen 0 Evaluation of three asymptotic methods of interval-estimation for the odds ratio In preparation. 4 Data provided by Veisicol Chemical Corporation 5 Epstein SS Carcinogenicity of heptachlor and chlordane Sd Total nvrron 6103-154, 1976 6 Princi F and Spurbeck C H A study of workers exposed to the in- secticldes chlordane, aldrin. dieldrin Arch Ind Hyg Occup M e d 3 6972. 1951 7 Alvarez WC and Hyman S Absence of toxic manifestations in workers exposed to chlordane Arch Ind Hyg Occup Med 8 480.483. 1953 8 Fishbein WI. White JV.and lsaacs H I Survey of workers exposed to chlordane lnd Med Surg 33 726727, 1964 9 Sandifer SH, Keil IE. Finklea IF. and Cadsden RH Pesticide ef- fects on occupationally exposed workers A summary of four yean observation of industry and farm volunteers in South Carolina /d Med 41 9-12, 1972 10 Warnick SL and Carter I E Some findings in a study of worken occupatlonally exposed to pesticides Arch Lnvrron Health 25 265-270, 1972 11 Morton WE, Crawford ED, Maricle RA. et al Hypertension in Oregon pesticideformulating workers 1 Occup M e d 17 182-185.1975. 12 Ortelee MF Study of men with prolonged intensive occupational exposure to DDT Arch Ind Health 18433-440. 1958 13 Laws ER. Curley A, and Biros FJ Men with intensive occupational exposure to DDT Arch Envrron Health 15 766775. 1967 * f I t I i I 1 1 1 1 i I 1 Warning on Intubation Passing a nasogastric tube is such a common everyday task in medicine that physicians rarely s t o p to t t n ~ u e t r m i u o u s - l o o k i n cgatheter is a potentially lethal instrument. But indeed it is. There have been at least five reports in the last few years of catheters that went directly into the brain rather than into the stomach. What's more, the tubes went in so easily the clinicians passing them thought they were going right where they should. , . . And all the reports in which the patients died - there were four deaths - had several things in common: massive frontal-impact injuries with nasal bleeding; blind insertion of a nasogastric tube through the nose; returns resembling gastric contents when the position of the tube was checked by aspiration; deterioration of neurologic status after intubation; discovery of the tube position on skull x-ray; and cribriform plate fractures noted at autopsy. - From "When N C lubes Go T o The Head." in m e r g e m y Medicme. July 15, 1979 748 Mortality of Workers Manufacturing Chlordane and Heptachlor/Wang and MacMahon