Document LYgVXBrQpge5X8ggp5Q78a5z

0 8 9 U * S,,U BIO-MEDICAL RESEARCH DOCUMENT DESCRIPTION FORM 63 68 69 76 j OOOtfrei Duplicate in all cards: --> year as-1961- File number [Right justify [Numeric only] Author(s), as Last Name FS (No Punctuation) and coden for journal as JAMA preceeded by one blank space Ia/a^ui// /? .'1 " 1 ~ , /Pax 77 78 J7d^ Sub-Index Code 60 61 62 n 12 13 Title of Report; end with space-hyphen-hyphen-space. Follow with Index Terms, separated from each other with comma-space. Avoid other punctuation; do not abbreviate* 12 1......... ............ 61 62 21 22 23 y/tr 24 Source (Journal, Vol., Number, Pages, Date ) 61 62 Brief Summary 12 10 SUMMARY: 61 62 61 62 63 64 -S'.' i, -,v: * 1- R. N. WHEELER, JR. PROJECT MANAGER ENVIRONMENTAL PROTECTION ft OCCUPATIONAL HEALTH UNION CARBIDE CORPORATION 437 MAC CORKLE AVE.. S.W. SOUTH CHARLESTON. WV 25303 TELEPHONE: (304) 747-2164 .v ; ' f-*' <i >- / o ' If 7 1. '7' / AN EPIDEMIOLOGIC INVESTIGATION OF AN EXCESS LUNG CANCER RISK IN A SYNTHETIC CHEMICALS PUNT 0000331 by V89U-V- s * y Allen H. Smith National Institute for Occupational Safety and Health Cincinnati* Ohio United States ** Occupational Health Studies Group University of North Carolina Chapel Hill, North Carolina + Cancer and Birth Defects Branch Center for Disease Control Atlanta, Georgia ah EPiEE-nccarrc zwariGKriflN cf ah dxess 112c cjcnt xasfi 3H A SYI.TiimC CHDZCT JS KANT ', ** Wi-TwitS j. vaawailpr, Allen II. Snith, H.A. TVrplcr, Itertrv Folk National Institute for Ctoupaticnal Safety and Health, Cincinnati, Chi.o, United States wrrxr Vm) trmecT risk (39 observed and 27 espeeted) was W observed among a cohort of 49S4 males -employed at a synthetic chemical plant since it opened in 1942. Upon review of the pathologic material-, -the was found to he limited to adenaearcincca and largo cell undifferentiated lung cancer. Many of the .workers ware exposed to vinyl chloride, but other exposures including chlorinated solvents, polyvinylchloride (PVT) dust, acrylates, and acrylonitrile existed. Sous, detailed work histories of each cohort istrhffT along with exposure ratings of each job for each caleider year since 1942 for of nineteen different charlcals (or ehsaical groups) were cfstained. A serially additive ejected dose model was constructed uh-ir-h cccpored the drwg of the chemicals observed for 'the lung cancer faeas to the expected on subcohsrts individually matched to the cases. ^FVC dust appeared to be the most Jikely etiologic * agent. Observed and expected doses were then analyzed by year before death to uncover the relevant latent period. This technique may prove very * useful in other pgq^aHmqi studies where multiple potential carcinogens exist simultaneously in a chemical plant. introduction . 1 Recent epidemiologic investigations of cancer in persons occupation- ally exposed to vinyl chloride monomer (VCM) gas showed excess deaths due to liver, brain, and lung cancer. Animal bioassays also found a relation- . ship between VCM exposure and one or more of these site specific malignan cies. However, the excess lung tumors found In these experiments had unclear dose response gradients. Furthermore, only three of the six retrospective cohort studies f( 1976; Tabershaw, 1974; Waxweiler, 1976) found an excess of lung cancers, and the relative risks were much lower than- for angiosarcoma of the liver (e.g., respiratory system cancer standardized mortality ratios were 168, 138, and 156). Additionally, the synthetic plastics plants-Studied used numerous chemicals other than VCM yet none of these other exposures were assessed in spite of the fact that some of these chemicals have demonstrated or are being tested for carcino genicity and mutagenicity. Thus while the weight of the evidence fell heavily on VCM as being the lung carcinogen at these synthetic plastics plants, the need to study the relationship of other chemicals to the lung cancer risk was significant. One plant in Louisville. Kentucky, made synthetic plastics and rubbers consisting mainly of combinations of styrene, butadiene, and acrylonitrile; and additionally polymerized VCM gas into polyvinyl chloride (PVC) plastic. The employees who were exposed to VCM in this plant were included in a pre viously published mortality study (Waxweiler, 1976). After learning of the association between VCM and angiosarcoma of the liver in 1974, the plant personnel rated each job in each year since 1942 on a zero to five scale of exposure for each of nineteen different chemicals or groups. R&s 111684 The work histories, vital statistics information, death certifi cates, and, in the case of deaths due to cancer, pathology reports and tissue specimens had all previously been collected by the National Institute for Occupational Safety and Health on every one of the approximately 5100 persons who had ever worked at the plant since it * opened in 1942. These latter data, not including the 19 chemicals exposure information, were used as part of the aforementioned retro spective cohort analysis of the mortality experience of VCM exposed workers. The lung cancer excess found among VCM exposed workers in this plant was intriguing for two reasons. First of all, the histologic distribution based on a preliminary review indicated that adenocarcinomas and large cell undifferentiated cancers accounted for a much greater proportion of the lung cancers than usual. Secondly, a number of cases of these histologic types occurred in men who had worked in areas of the plant associated with much lower exposures of VCM than the PVC polymerization area. Hence, there was a Strang incentive to search for occupational lung carcinogens other than VCM. Thus the following research had three objectives: 1) to determine if the excess lung cancer risk among employees exposed to vinyl chloride monomer at a synthetic plastics and rubber plant also existed for the total plant population, including those persons not exposed to VCM, 2) to determine whether an excess lung cancer risk of a particular histologic type was. in force at the plant, and 3) to test whether one or more particular chemicals used at the plant were responsible for either the excess risk of all, or of a specific histologic type (adenocarcinoma and large cell undiffer entiated) of lung cancer. Non-occupational etiologies were also considered and investigated to the extent possible, but the emphasis lay on occupational risk factors. /II. METHODS " I . . r-- Efc*i 3 A. Retrospective Cohort Study The first objective, that of determining whether an overall excess risk of lung cancer existed at the plant, could only be accomplished by the use of an external referrent population. Thus a retrospective cohort study design and standardized mortality ratios (SMR's), using the United States death rates as a referrent, were chosen to answer the question. The population at risk.(PAR), 5,093 employees, consisted of all persons ever employed at this plant since it opened in 1942 until May 1974. Of these, the 4,806 males, 98% of whom were white according to R&S 111685 the company, who began work at the plant by December 31, 1973, were selected for study (Figure 1). Date of birth and detailed work history at this plant in the form of, "department number, job number, date in, date out," for all jobs held by each individual were collected. Follow-up of all study members was undertaken from the first date hired at the plant through December 31, 1973, the termination date of the ensuing mortality analysis. Death certificates were obtained for those individuals known to have died. The underlying cause of death was coded according to the "Seventh Revision of the International Lists of Diseases and Causes of Death" by a qualified nosologist. A life table analysis based on the Cutler-Ederer (1958) technique was used to obtain person- months at risk of dying by five year age, calendar, and time since work onset (latency) groups. So as not to overestimate the risk of mortality in the plant cohort, persons lost to follow-up were considered alive throughout the analyses. United States white male death rates specific for five year age and calendar intervals were used to calculate the expected deaths. Worker Case - Community Compared Case Study Because no histologic specific lung cancer incidence or mortality rates exist, and because histologic classification is somewhat variable between pathologists and over time, a synthetic-plant worker lung cancer case versus conmunity comparand case design was chosen to accomplish the second objective. Medical records and pathology reports were obtained on all deceased members of the cohort whose death certificates mentioned cancer or respiratory disease. If a primary or unspecified lung cancer was reported on any of these three records, then histologic material was obtained, if available, and reread by a panel of three pathologists unaware of the employment histories of the cases. Forty-five cases were so identified. A few cases found to have died after December 31, 1973, the cohort ending date, were included in the wofcker case-community comparand case study. As a comparison group, the lung cancer cases most closely preceding and succeeding the chemical plant worker case in the chronologically ordered hospital pathology logs were selected that matched on age diagnosis (+ or -2 years) sex, race, and county of residence. Although only 673* of the controls met the strict criteria of 2 years, 945 of the controls were within +6 years of age. For four cases, only one matched control could be found. The histologic type distributions were then compared between the worker cases and the conmunity comparand cases after review of the microscopic material by the pathology panel using the Veterans Administration classification scheme. R&S 111686 5 C. Serially Additive Expected Dose Model The third objective required testing hypotheses on a number of chemicals with regard to their possible role in causing lung cancer. Thus, it was decided to analyze the previously mentioned cohort ** data, in a Serially Additive Expected Dose (SAED) Model, obtaining observed and expected doses of each chemical conditional on certain characteristics. . The basis of the serially additive expected dose (SAED) method in descriptive terms is to compare the observed exposure of each case in a study with the exposures of fellow workers close to the case in year of birth, and in age at conmencement of work for the company in the study. If the total work force in the plant being studied is referred to as the cohort, then each case can be thought of as belonging to a sub-cohort of workers with approximately the same year of birth and age at consnencement of work in the plant. In each year that a case worked at the plant, his exposure can be compared with the other members of his sub-cohort who were working in that year. . If the exposure being studied is causally . related to the disease of interest, then the cases should, on average, be exposed for longer periods and/or to higher levels of the causative agent than other workers in their sub-cohorts who did not get the disease, and who. did not get seme other disease attributable to the same exposure.- Z If the exposure is not causally related to the disease then we would .* expect case exposures to average out to be the same as for their sub cohorts. Thus for. each case, we want to use members of his sub-cohort to calculate an expected exposure for him, and compare this with his * observed exposure, taking into account both exposure level and exposure * duration. R&S 111687 R&S 111688 6 The exposure data required for the analysis consist of some measures of exposure to a given chemical for each job in a plant, for each calendar year Involved in the study. The exposure measures should Ideally be Interval -- measures, but can also consist of ranked exposure data. Ranked data were compiled by company personnel with a 0 to 5 scale of exposure for each of 12 chemicals examined (Table 1). Each job was assigned an exposure rank for each calendar year of the study (Table 2). These exposure.data were then linked with work histories identifying the jobs each worker had in the plant and the calendar time involved. The analytical method used is based on estimating exposure / dose by multiplying the exposure level and the duration worked at that * level. These "doses" are accumulated over each calendar year to yield the observed dose for each.year that a case works. If a worker is exposed at level 1 for all of a calendar year then he has accumulated 3'65 dose units. If he spends 200 days at level 2, and the remainder of ]>. the year at level 1 then he has accumulated 565 dose units, - The calculation of the expected dose for each year a case works is ** # * obtained from the cohort dose matrix (Figure 1). This is a,three dimensional -matrix defined by age of first employment, year of first . employment and calendar year. During employment in the Industry, each non-case contributes to the relevant cells of this matrix the number of . deys spent in the given calendar year at each of the possible exposure levels. For example, an employee aged 22.when first employed in 1944 who spent 200 days of .1944 at level 2 and 165 days at level 1 would contribute to the corresponding cell shown in Figure '1, the respective number of days at the two levels. Subsequent employees of the same age when first employed in 1944 would add to the values in this cell in a * similar manner. For computational simplicity this matrix is defined by age at first employment rather than year of birth which was suggested R&S 111689 , < " above to define the sub-cohorts*.^`fEtfch*+,cell is uniquel* y defined whichever approach is used. i Once data from each non-case have been entered into the cohort dose . matrix, the case exposure dose can be compared with the matrix values for each year the case works. However this comparison is somewhat restrictive t as only non-cases with* exactly the same year of birth and age at first i employment are involved. It was indicated earlier that comparisons might be made with sub-cohorts with approximately the same year of birth and age at first employment. This is put into effect in the SAED model by using a moving sum accumulation of data in the cohort dose matrix. In the application to be presented we define a sub-cohort for comparison with a case as consisting of all non-cases whose year of first employ- mentand age at first employment are within two years of those of the case. However in the first year of employment of a case there are ob viously no non-cases working who were first employed in the two subsequent ' years. For this reason the expected values for the first year of employ- * ment are* calculated only from non-cases who started work in the same year as the case. In the second year of work the non-cases first employed in the year before and the year after the year of hire of the * case are * included. In sub* sequent years the estimation is from the whole sub-cohort within two years of first employment and within two years of birth of the case. When, the cohort being studied is small in number the Jk. moving sum accumulation may need to span more years than this* * The cohort dose matrix is then used to calculate the expected dos for each year that a case works. Table 3 presents the first two years of Work for a case who was aged 22 when first employed, and who was first employed in 1944. In his first year he worked 115 days, 50 of them at level 0, followed by 65 at level 5. He thus`accumulates a dose of 50x0 + 65x5 * 325 dose units. The non-case members of his sub-cohort b accumulated 600 days at level 0, 1350 days at level 1, etc. The total number of sub-cohort days of work in 1944 was 3165 days, and the total sub-cohort dose units was 4830. The dose units per work day for the sub-cohort equal 4830/3165 1.53. Now the case worked for 115 days in that year, so his expected dose units (ie. the expectation for his dose units under the null hypothesis that the exposure does not cause the disease) is 115x1.53 * 176 dose units. The actual exposure was 325 dose units so it can be seen that this particular case was exposed to roughly twice the dose units for the chemical as would be expected for that year If he were an average member of his sub-cohort. The calculations so far relate to one year of work history at the plant. As indicated previously, an individual's doses can be summed over his entire work history to give his observed total dose. His expected dose for each year of work can also be summed to give his expected total dose. In addition to giving this test of the hypothesis that cases have more than expected total doses of an exposure, the SAED analysis facil itates an examination of the relationship between dates of exposure and the dates of death of the cases. For each case in the study, one can calculate for each year prior to death his observed dose and his expected dose based on his subcohort. (If the case does not work in a given year then both observed and expected are set to zero.) 069 VW 9 There were three definitions of cases for this SAED study. The first definition consisted of the 45 deceased lung cancer cases identified In the cohort study as detennined by best diagnosis after review.of all medical data and of available microscopic material as stated in the worker casecommunity comparand case study. Because previous investigation just of , the vinyl chloride monomer exposed sub-cohort (Ignoring other exposures) Indicated an excess of adenocarcinomas (Type 3) and more so of large cell undifferentiated lung cancers (Type 4), a second definition of cases was further restricted to just Types 3 and 4 and a third definition was limited to only Type 4. In the SAED model, each of a number of chemicals can be tested separately. However, since It was not known how well the model might discriminate between chemicals, the ability of the SAED Model to identify a carcinogenic agent was tested by first using the liver angiosarcoma cases at this plant and the Cohort Dose Matrix for a number of chemicals including vinyl chloride monomer. The results indicated the model could identify YCM as the'liver carcinogen, but it could not distinguish between highly, correlated chemical exposures. r &S 111691 R&S 111692 10 RESULTS A. Retrospective Cohort Study Ninety-eight percent of the 4,806 male cohort were believed to be white according to conrunication with the plant personnel office. The distribution of year of birth Indicated that by the t end of study date, December 31, 1973, over 801 of the cohort, if alive, would have been over 34 years of age, 635 over 44 years of age, and only 305 over 54 years of age. Most of the employees that have ever worked at the plant were hired before 1954 (Table 4); in fact 635 of the cohort had the opportunity to achieve at least nineteen years latency and 805 had the opportunity to achieve nine years latency. The ages of first employment at this plant are probably fairly typical of a large manufacturing company. Over two thirds of the employees began work before age 30. The other third is more heavily weighted with persons who began work at the plant during the 1940's, ie. during and just after World War II. This trend is probably due to bringing in skilled people and older persons not eligible for the armed forces during the first years the plant was in operation. A number of workers moved through various job titles at the plant but two thirds of the workers had only one or two job titles. This .inplant job stability facilitates the investigation of specific chemical etiologies by limiting the number of exposures a typical worker will experience. By follow-up of the 4806 man cohort uh' tiT i'Hthi- e end of 1973, It was determined that 4174 members of the cohort were alive, 559 deceased. and 73 lost-to-follow-up. The sixteen, deceased persons whose death certificates could not be found were treated as deceased, cause of death unknown. Two separate analyses of the cohort were made. Initially all members were considered at* risk from their first date of employment at the plant. This analysis yielded 556+ observed and 550^expected deaths (Table 5). Death due to malignant neoplasms of the central nervous system **^7* for this rare (27 angiosarcomas are expected to occur in the entire United States in one year} malignancy. A second "over ten year latency" analysis was carried out by beginning person years at risk only after an individual had achieved his tenth anni versary since his first date of employment at the plant, regardless of still working there. The ten year period was chosen to coincide with the original published analysis of just the VCM exposed workers at this plant (Waxweiler, 1976). This analysis demonstrated very similar results to those in the first cohort analysis but with slightly higher SMR`s. Respiratory system cancer displayed an SMR of 156 based on 39 observed cases. Because the analysis of just VCM exposed workers at the plant, described in the intro duction, also found an SMR of 156 after 10 years latency, it was felt that these plant-wide results Implied the existence of an excess lung cancer risk not solely As shown in Figure 1, three of the 559 deaths were rejected from analysis since they had no work histories. O) due to YCM exposure. It oust be reiterated here that the earlier VCM exposed cohort was defined on jnuch cruder exposure criteria than were used in the SAED model. In the earlier reported VCM exposed cohort, exoosure at this plant was determined on a departmental basis following a walk-through survey by National Institute for Occupational Safety and Health industrial hygienists and review of company process, engineering control, and air sampling data. Those specific departments classified as having VCM exposure included VC Monomer production, VCM polymerization PVC compounding, and maintenance. While that cohort study used a dichot omous definition of exposed departments, the SAED model used a 0-5 ranking developed by company personnel. B. Worker Case - Comnunity Comparand Study As previously stated in the methods section, death certificates, medical and pathology records, and re-examinations of histologic specimens were used to make a best diagnosis that resulted in 45 lung cancer deaths. The twenty-seven of these 45 cases for whom histologic specimens were available made up the worker case group. As could be expected, cases with histologic specimens available for reconfirmation occurred slightly more frequently in later years and among younger persons than did cases with no specimens available. There were no differences between the two groups with respect to the first year employed (both medians 1947) or duration of employment (both have 40 percent greater than five years). The results of the majority opinion of the pathology panel are seen in Table 6. The panel found a substantially higher percentage of large cell undifferentiated (Type 4) cancers in the worker cases than in the community comparands (30S vs 105) which was statistically significant (p<.05). R&S 111694 g69WV 13 Starting with the SMR for respiratory system cancer of 149 found in the retrospective cohort study, and the observed and expected histologic distri bution of lung cancer deaths based on the worker-case conrnunity-comparand study. Table 7 calculates histologic specific lung cancer SMR's (column C). It appears the lung cancer excess is limited to types 3 and 4, adenocarcinoma and large, cell undifferentiated, with the greatest rate ratio lying in the large cell category. Furthermore, if one assumes that the 27 cases for whom histologic specimens were available are representative of all 45 lung cancer cases, then the 14.8 excess cases that would exist (given an SMR of 149)', can be broken down by histologic type (Column E). This theoretical breakdown shows that approximately 13.5 of the 14.8 excess lung cancers among the plant workers would be due to adeno and especially large cell undifferentiated carcinoma. Note the total excess cases of each cell type do not add up to 14.8 because of the 22 of the expected distribution, column B, that is due to other types of lung cancer. Thus it appears that there is indeed a histologic specific excess lung cancer risk among workers at this plant and that this differential distri bution of cell types is not an artifact of the geographic region nor the pathologists' techniques. The fact that this risk occurs for adenocarcinomas and large cell undifferentiated lung cancers makes it very unlikely that it is due to cigarette smoking and is further encouragement to seek an occupa tional etiologic agent. C. Serially Additive Expected Dose Model The 398 department number - job number combinations that were originally identified from the plant personnel records were transformed into 72 job categories the company identified as having some exposure at some time in the past to at least one of the nineteen chemicals. Table 8 gives a general idea of how frequently high exposure job categories occurred among the 72 possible ones. This does not . necessarily Imply frequency of person years exposed because some job categories contain many more person years than*others, but one can infer from Table 8 that no single exposure dominates the others. This is especially evident from the raw data which are too voluminous to reproduce here.' Seven of the chemicals were .felt to have so few persons exposed to levels 3, 4 and 5 that they were excluded from consideration in the analysis. These were acrylic acid, acrylamides, blsphenol A, chloro- ethyl vinyl ether, diethyl maleate, phenol, and toluene. * R&s 111696 The distribution of first year employed by age first employed is shown for all lung cancers in Table 9, and for Type 3 .and 4 lung cancers in Table 10. The distributions in these tables are very similar, the only noticeable differences being that the Type 3 and 4 lung cancers started work at earlier ages. The distribution of the eight Type 4 lung cancer cases resembled that of the 15 cases of Type 3 and 4 combined. None of the six lung cancer deaths occurring before I960 were adeno- carcinoma or large cell undifferentiated. This is partially due to the fact that only one of these six cases had available specimens for histologic re-examination. Types 3 and 4 lung cancer deaths were more heavily represented (392 vs 272) than all lung cancers at ages of death below fifty. The duration of employment distribution of the lung cancer cases (Table 4) is delimited by the corresponding distribution for all employees. Thus, it is not unexpected that .282 - 382 of the cases worked at least ten years because only 212 of the cohort worked that long. The SAED Model Analysis resulted In the observed;minus expected cumulative .r * - dose differences per case found in Table 12. Surprisingly, only two chemicals had a positive dose difference for all 27 histologically reviewed cases and no chemical had a positive dose difference for the 18 cases without histo logic reconfirmation. The differences for PVC Dust doses stand out strikingly compared with the other exposures. For the Type 3 and 4 cancers and Type 4 * alone, the PVC Dust differences are three to four times as large as the next most evident chemical, vinylidene chloride. The significance levels of the larger of these differences are found in Table 13 . They are listed for the total cumulative doses shown in Table 12 and also for cumulative doses up until ten years before death. The p-values of these t-tests signal a sig nificant excess dose only for PVC dust. P-values below .05 exist only for all pathologically reviewed cases and for adenocarcinomas and large cell undifferentiated cancers combined. The p-value for large cell undifferenti ated by itself Is just above .05. Regardless of the definition of cases, the p-values become larger when the exposures occurring only 10 or more years before death are considered. The reason for this is illustrated in the latency analysis in Figure 2 which shows a peak positive PVC-Dust dose difference in the five to sixteen years before death range that would contri bute more to the cumulative dose difference over all years than to just the ten years before death sum. As the diagnosis of the disease of interest becomes more narrowly defined'the dose difference per case increases, yet the relevant latent period, stays constant. The p-values associated with the paired t-tests are shown for call types 3 and 4 (adenocarcinoma and large cell undifferentiated) in Figure 3. No more than seven of the 15 Type 3 and 4 lung cancer cases worked in any single year before death. Despite this consistency, a sharply defined cluster of p-values below .01 occurs in the interval seven to thirteen years before death. pp /.69 VW V. ' DISCUSSION The first objective of this study was to determine if the excess lung cancer risk among employees exposed to'vinyl chloride monomer i (VCM) also existed for the total plant population Including those persons not exposed to VCM. The retrospective cohort mortality study showed that respiratory system malignant neoplasms occurred approxi*, mately SOS more frequently among the employees of the entire plant than would have been expected based on age, sex-, race, and calendar year. United States death rates. A similar excess had been previously shown to exist among the subcohort of employees exposed to VCM (Waxweiler, 1976). Thus the excess lung cancer risk at the plant ' appeared to be independent of VCM exposure unless one hypothesized that it was due to extremely low levels of VCM that could have permeated * almost all buildings. R&S 111698 Potentially confounding variables in the study of lung cancer are age, sex, calendar year, race-ethnicity, migration, urban residence, cigarette smoking and socioeconomic status. In the retrospective cohort study ace and calendar year are standardized. Furthermore, all persons in-the study are males and 983 _sre_ white, thus white male specific rates were used for comparison and these variables are con trolled by subject category restriction. Ethnicity and miaratinn (hnt.h inter and intra country),effects are considered minimal for all three study designs because of the clustering of birthplaces of the cases in and around the state of Kentucky, ______ JP'- 17 ,,vc 'r While an urban effect could have accounted forTsome of the 50* excess In 6 6 9 VVV S *** the cohort It probably played a very Insignificant role because Haenszel (1962) found it to exist mainly for epidermoid and small cell lung cancers and the worker case comnunity comparand study showed these cell types not responsible for the excess-rate at the plant. Socioeconomic status could '# potentially effect lung cancer relative risks of the following magnitudes: 1) 1.2 if everyone at the plant had received no more than eight years of education (Kitagawa, 1973) or 2) 1.3 if everyone was classified as a laborer (Guralnick, 1963). Recent information Indicates these gradients may be partially due to smoking patterns (Sterling, 1976; Winklestein, 1975)-. Since many of the employees in most of the departments would be classified at a higher socioeconomic status, there is probably insignificant confounding at the plant considered as a whole. However, it is possible that some exposure areas consisted of extremely homogeneous SES groups that were \ different from the plant as a whole. Even if this were true, SES could only have accounted for a 20-305 excess in lung cancer, much less than was found for cell types 3 and 4 in the hlstopathology study. Unfortunately education data on current employees were not available. Cigarette smoking with its tenfold relative risk Is a more threatening confounder when first considered. * Its relationship can best be elucidated In the worker case community comparand study discussion. The second objective was to determine whether an excess lung cancer risk of a particular histologic type was in force at the plant. The worker case versus comnunity comparand study showed that regardless of whose patholo gic diagnosis one accepts, there was clearly an excess R&S 111700 18 representation of Type 3 and 4 lung cancers within the cases occurring among plant employees compared to other comnunity lung cancer cases. By having the same pathologists review both sets of slides without knowing which were of plant employees, one can feel fairly certain that no diagnostic biases occurred. One could argue that the twenty-seven cases for whom histologic specimens were available were not representative of the entire 45 lung cancer cases at the plant. The major differences between those cases with and without specimens were date and age of death. Histologic specimens were plainly more difficult to find the earlier the date of death and the older the age of death. However, community comparand cases were matched on date of diagnosis and had to have available specimens themselves. x One could also argue that less common types of lung cancer, such as large X* cell undifferentiated, would more often tend to have specimens taken or saved for research of diagnostic interests. For such bias to have an effect on the results, one would have to conceive of a selection that caused different types of lung cancers to be biopsied or autopsied among comnunity members than among plant employees. Such a bias is difficult to imagine, particularly after the age matching procedure. The choice of community comparand cases makes it unlikely that a community wide pollutant was responsible for the excess risk at the plant specifically for Types 3 and 4 lung cancer. Under the assumption that the 27 lung cancer cases with histologic R&S 111701 I 19 specimens available were representative of all 45 lung cancer cases among plant employees. It can then be Inferred from Table 7 that the excess lung cancer risk In the cohort was limited to adenocarcinoma and large cell undifferentiated cancer. Adenocarcinomas, accounting for a minor proportlon.of this excess risk,-have been shown at most to be only weakly related* to cigarette smoking (Doll, 1964; Weiss, 1972) and some studies show no association at all (Yesner, 1973; Auerbach, 1973). Large cell undifferentiated carcinoma of the lung, accounting for the vast majority of the excess lung cancer risk, is the only major histologic type that has always appeared In epidemiologic studies not to be related to cigarette smoking. Thus it Is felt that cigarette smoking was not a major confounding variable in any of the three study designs In this dissertation. However, the role of smoking as a \ promoter or co-carcinogen cannot be ruled out. Such an effect appeared among uranium miners and asbestos workers. Interestingly enough such an effect did not appear to alter the characteristic histologic specificity of the lung cancer risk for oat cell carcinomas among the uranitan miners (Archer, 1974). Furthermore in this study of chemical plant workers 71JTof the histologically reviewed cases and 752 of the large cell undifferentiated cases were smokers, while all the rest had no smoking history data available. If fewer cases than comparands had smoked it would reflect an Independent effect due to occupation, other wise an Interactive effect could be Inferred. With such a large propor tion of smoking histories missing for the lung cancer cases it is diffi cult to discern what role, if any, smoking played in the excess lung cancer risk at the plant. Nevertheless, the conclusion of this phase of the Investigation was that an excess risk occurred among plant 20 employees for Types 3 and 4 lung cancer* especially Type 4. The third objective was to test whether one or more particular chemicals used at the plant were responsible for either the excess risk of all lung cancer or of the Types 3 and 4 or just Type 4 lung.cancer. The SAED model was specifically developed for this Objective. It proved to be very useful in confirming the relationship between angiosarcoma of the liver and VCM and was subsequently used in the lung cancer investi gation. The major hypothesis was tested for each chemical in the SAED model by using the one sided t-test of the observed minus expected cumulative doses over 1) all years before death and 2) just ten or more years before death (Table 13)'. PVC-dust was the only chemical that was or even came near being statistically significant. It was not near statistical significance for all 45 lung cancer cases but attained p-values of .026 and .047 for the 27 cases with histologic specimens available. More importantly for Types 3 and 4 cases in the p-values were .037 and .061 and for Type 4 were .068 and .177. In epidemiology one must look beyond straightforward statistical significance at times to other criteria such as strength of the association, consistency of results, and biological plausibility. The strength of the association was four-fold for the large cell undifferentiated SMR for the entire plant (Table 7), and a 1.7 dose ratio existed for PVC-dust. Con sistency of results was evident in that for all sets of lung cancer cases PVC-dust, out of all nineteen chemicals, showed; 1) the largest positive cumulative dose difference and the largest positive dose ratio (1.2-1.7), and 2) a generally unimodal R&S 111702 yearly dose difference distribution over a conceivably biologically relevant latent period. These results lead me to believe that this research suggests that an excess of Type 3 and 4 lung cancer cases exists at this plant and Is related to PVC-dust exposure. More extensive exposure and smoking data are currently being collected by researchers at the University of Louisville and will allow them to retest this hypothesis. These results differ from the earlier reported study at this plant (Waxweiler, 1976) which implicated VCM as the likely lung carcinogen because other chemicals and workers exposed to those chemicals were not considered at that time. The more detailed exposure data developed by the company also contributed to the differing results. Biologically the suggestion of PVC dust being a lung carcinogen is very plausible. PVC-dust particles are often in the respirable range. Almost all PVC particles produced by the emulsion system, one of the systems at this plant, are in the respirable range (Jones, 1978). These particles could easily settle in the lungs and conceivably by themselves cause lung cancer. In fact one case of supposedly PVC-dust induced pneumoconiosis has been found in a worker (Szende, 1970), and pulmonary granulomas (Frongia, 1974; Agarwal, 1978) have been induced'in animals exposed to PVC-dust. ' Additionally, a large proportional mortality study of 4341 deaths that occurred among PVC fabricators, persons expected to be exposed to VCM and PVC dust was reported (Chiazze, 1977) to find no angiosarcomas of the liver but a slight (PMR=117) excess lung cancer risk. R&S 111703 R&S 111704 * 22 Yinyl chloride monomer gas is easily inhaled and possibly only a short duration of tissue contact time would occur before the gas is either exhaled or absorbed into the bloodstream. However, it Is known that YCM becomes entrapped in the PVC-dust and can be released slowly over time.. Thus there exists the possibility that PVC-dust particles in the lung slowly-release VCM to small adjacent areas of the tissue, prolonging the contact time of that chemical to tissue. If this latter hypothesis is true then it begs the question of why almost all the angiosarcoma cases occurred among polymer reactor cleaners who received extremely high YCM doses, but the lung cancer cases occurred quite often among the less heavily exposed workers. In fact VCM showed no relationship with lung cancer in the analysis. One would expect the lung to be the major route of entry for VCM regardless of the cancer site. However, the etiologic mechanism Is a problem to be answered in laboratories and is only speculated upon here in the context of biologic relevancy and as a lead to further Investigation. Finally, it Is hoped that this SAED Model could be refined to sort out confounding or interactive chemical carcinogens perhaps in a multiple regression adaptation and used in both an industrial health surveillance system to test its utility in a prospective basis and also in a case-control research situation. REFERENCES Agarwal, D.K., Raw. J.L. and Srivastava, S.P., et al. Some biochemical and histopathological changes Induced by poly vinyl chloride dust in rat lungs. Ind. Tox. Res. Center Report, Lucknow, India, 1978. Archer, V.E., Saccomanno, G. and Jones, J.J. Frequency of different histologic types of bronchogenic carcinoma as related to radiation exposure. Cancer 34: 2056-2060, 1974. Auerbach, 0., Garfinkel, L. and Parks, V. Histologic type of lung cancer in relation to smoking habits, year of diagnosis and sites of metastasis. Chest 67: 382-387, 1975. Byren, 0., Engholm, G. and Englund, A., et al. .Mortality and 9 cancer morbidity in a group of Swedish VCM and PVC production workers. Environ. Health Persp. 17: 167-170, 1976. Chiazze, Jr., J., Nichols, W.E. and Wong, 0. Mortality among employees of PVC fabricators. J. Occ. Med. 19: 623-628, 1977. Cutler, S.J. and Ederer, F. Maximum utilization of the life table method in analyzing survival. J. Chron. Dis. 8: 699-709, 1958. Doll, R. and. Hill, B. Mortality in relation to smoking: ten years' observations of British doctors. Br. Med. J. 5395: 1399-1410; 5396: 1460-1467, 1964. Frongia, N., Spinazzola, A., and Bucarelli, A. Experimental lung damage from prolonged inhalation of airborne PVC dust. Med. Lavoro 65: 321-342, 1974. gOiLl-V Guralnick, L. Mortality in 1950 by occupation and industry. Vital StatisticsrSpecial Reports 53, 1-5, U.S. Dept. HEW, 1961-1963. Haenszel, W., Loveland, D.B.. and Sirken, M.G. Lung cancer mortality as related to residence and smoking histories. I. White males. t J. Natl. Cancer Inst. 28: 947-1001, 1962. Jones, O.H. Worker exposure to vinyl chloride in vinyl chloride and polyvinyl chloride production and fabrication. NIOSH Technical Report Draft, January, 1978. Kitagawa, E.M. and Hauser, P.M. Differential Mortality in United States: A Study in Socioeconomic Epidemiology. Harvard University Press, Cambridge, Mass., 1973. Sterling, T.D., and Weinkam, J.J. Smoking characteristics by type of employment. J. Occ. Med. 18: 743-754, 1976. Szende, B., Lapis, K., and Nemes, A., et al. Pneumoconiosis caused by the inhalation of polyvinyl chloride dust. Med. Lav. 61: 433-436, 1970. * Tabershaw, I.R., and Gaffey, W.R. Mortality study of workers in the . manufacture of vinyl chloride and its polymers. J. Occup. Med. 16: 509-518, 1974. Waxweiler, R.J., Stringer, W., Wagoner, U.K., et al. Neoplastic risk among workers exposed to vinyl chloride. Ann. N.Y. Acad. Sci 271: 40-48, 1976. R&S 111706 Weiss, W., Boucot, K.R., and Seidman, H., et al. Risk of lung cancer according to histologic type and cigarette dosage. J.A.M.A. 222: 7*9-801, 1972. Winkelstein, Jr. W. Contemporary perspectives on prevention, p. 27-28. Vol. SI, No. 1, Bull. N.Y. Acad. Med., 1975. Yesner, R., Gelfman, N.A., and Feinstein, A.R. A reappraisal of histopathology in lung cancer and correlation of cell types with antecedent cigarette smoking. Am. Rev. Resp. 01s. 107: 790-797, 1973A. TABLE 1 Exposure ratings Ho exposure Minimal exposure to low levels (Chemical in building--not handled, low vapor pressure and dust level, probably works on-different floor) Moderate exposure (Works around the chemical, but `exposure is' minimal) Works in area subject to occasional hi oh excursions (Normally exposure is minrimal but occasional spills, leaks, or dust exposure may occur) Works in areas where level is hi oh (Exposure levels in the area are frequently high. Might consider that some risk is involved if-chemical Is very tox* ic) * Intimate contact -- skin or high inhalation (Such as poly cleaners in the old days -- handling slurry) " : '4 :s*4. X TABLE 2 Chemical Exposure Ratings Specific For Job Identification Number And Calendar Year For A Given Chemical CHEMICAL #1 Job Identification Number 1942 1943 Calendar Year 1944 1945 1946 1 00 0 2 2* 2 55 5 55 3 22 1 10 84 44 4 1973 2 5 0 1 t R&S 111709 I II I. V. r i ji#r- ju ^ m i l l SSU TABLE 3 Example of Data Used to Calculate the Observed and Expected Dose Units for a Case in One Calendar Year Year 1944 TOTAL Exposures level 0 1 2 3 4 5 \ Case days 50 0 0 0 0 65 115 Case dose units 0 0 0 0 0 325 325 Sub-cohort days 600 1350 540 400 175 100 3165 Sub-cohort dose units 0 1350 1080 1200 700 500 4830 TABLE 4 First Year Employed by Age First Employed Distribution of Male Cohort of Workers Ever Employed Before 1974 First Year Employed 1941-1949 1950-1959 1960-1973 total _______________________Age First Employed 16-29 1103 30-39 528 40-49 192 50-66 73 1120 325 97 10 1049 230 53 7 3272 (685) 1083 (225) 342 (75) 90 (25) Total 1896 (405) 1552 (335) 1339 (285) 4787* (1005) R&S 111712 * Of the 4806 males, 19 were missing either date of birth or detailed work hx, leaving 4787 In the above table. it. TABLETS , LOUISVILLE PLANT COHORT MORTALITY Cause (ICO Code) All Observed Expected SMR Over Ten Years Latency Observed Expected SMR All Causes 556* 550.2 101 450 All Malignant Neo plasms (140-205) 1) Central Nervous System (193) 109 9 92.5 118 4.3 209+ 94 j 8 11) Respiratory* System (160-164) 42 28.2 149++ 39 111) Digestive System 24 25.6 94 21 (150-159 excl. 1565} 1v) Lymphatic & Hemat. (200-205) . 9 11,4 79 7 v) Other Cancers 25 23.0 109 19 Includes 3 non lung cancer deaths with ICO 161. t one tailed Poisson p value <.05 ++ one tailed Poisson p value <.01 421.7 107 76.4 123 3.1 256+ 24.9 156++ 21.0 100 8.5 82 18.8 101 R&S 111713 TABLE 6 Worker Case and Matched Community Comparand Histologic Distributions -Histologic - Type Veterans Administration Classification Code' . Frequency Worker Consnunity Cases Comparand Cases Epidermoid 1 6(225) 15 (305) Small Cell Undifferentiated Adenocarcinoma 2 3 6(225) 7(265) IS (305) 14 (28%) Large Cell Undifferentiated 4 8(305) 5 (105) Other Total - 0 1 (2S) . 27(1005) 50 (100S) R&S 111714 *. '. * TABU 7 a . ' : : >, ` ` Histologic Specific Lung Cancer Risk Among Plant Personnel Histologic Distribution Histologic Observed Expected Specific SMR f * (A) (B) (C)*149xA B u CO cn Total Epidermoid Small Cell Adenocarcinoma Large Cell 10W 22.25 22.25 25.95 29.65 305 305 2S5 105 149 no no 138 ' 441 . Excess Cases Total Among "hose Excess Pathologically Cases Reviewed (D)+ (E)++ 8.9 14.8 .6 .9 .6 .9 1.9 3.2 6.2 10.3 .1 *r jj r fio ; w: tD * (A x 27) - (B x 27 x 100) * 149 *j*p _ /a h\ TTL a \*f\ X HD) .4 A "7" 4 TABLE 10 First Year Employed by Age First Employed Distribution of Adenocarcinoma and Large Cell Undifferentiated Lung Cancers (Type 3 and 4) First Year Employed 1941-1949 1950-1959 1960-1964 Total 20-29 Age First Employed 30-39 40-49* 251 221 1 00 5 (332) 7 (475) 2 (145) ItVno Total 1 9 (60S) 0 5 (33S) 0 1 (7%) 1 (75) 15 (100S) R&S 111718 4* J. <1 . TABLE T1 V ' Duration of Employment Distribution of Lung Cancer Deaths Years 0 - .5 1-4.9 5 - 9.9 >10 TOTAL A11 Lung Cancer Deaths 21(465) 7(165) 4(K) 13(285) 45(1005) Adenocarcinoma and Cell Cell Undifferentiated (Type 3 and Type 4) 6(405) 2(135) 2(135) 5(335) 15(1005) Large Cell Undifferentiated (Type 4) 3(385) 0(05) 2(255) 3(385) 8(1005) e u vu s*a TABLE,. 12 Jl. Observed - Expected Cumulative Dose DifferencesPer Lung Cancer Case Chemical All Lung Cancers Acrylonitrile -652 Acetylene -353 Acrylates -322 Butadiene -330 Caprylyl Chloride -547 Clorinated Solvents -868 Mercuric Chloride -507 Methanol -430 Ylnyl Chloride Monomer -1428 Ylnylidene Chloride -341 Vinyl Acetate -707 PYC-Dust 763 Pathologically Reviewed Cases -402. -289 -251 -207 -258 -672 -425 -618 -795 210 I48O 2448 Adenocarcinoma Large & Laroe Cell Cell -446 -128 291 -167 -83 -339 -406 -622 -270 -1139 -150 749 -211 -62 -456 -1078 26 907 804 1525 -217 328 3225 4526 R&S 111720 TABLE -13hJ4; Probability Value of Paired T-Tests:of Cumulative Differences Between Lung Cancer Cases' Observed and Expected Doses All Cell Types PVC-Dust All Pathologically Reviewed PVC-Dust Adenocarcinoma & Large Cell (Types 3 and 4) Vinylidene Chloride PVC-Dust Large Cell (Type 4) Chlorinated Solvents Vinylidene Chloride PVC-Dust Totalt (All Years) .185 .* .026* .267 . .037* .360 .201 .068 10 Years Before Death .253 * .047* .333 .061 .435 .322 .177 * p<.05 + Twelve chemicals were tested for each lung cancer group thus under an ` assumption of independence of tests one would expect one test to be significant at the .08 level. R&S 111721 V* 1 f 3D (/) -to4 to -I*-*" 1' FIGURE 2 Observed Minus Expected Dose Differences Per Case Among Lung Cancer Cases for PVC Dust figure z Zl p-Values of Paired T-Tests of Observed Versus Expected Annual Doses of PVC Dust Among Type 3 and 4 Lung Cancer Cases (Number of Cases Working Each Year is Noted- at Each Point) R&S 111723 * Central Toxicology Laboratory Alderley Park Nr Macclesfield Cheshire SK10 4TJ * Teleohone Alderley Edge (0625) 582711 Imperial Chemical Industries Limited Dr R J Waxwiler 'National Institute for Occupational Safety and Health Cinciimati Ohi SA J . vert \JC n - R&S IIV 4 Your ref f^IUr Our ref ifhpiSMS Tel ext 116 Date 9 Oct 78 Dear Dr Waxweiler .^ /,.' Thank you for letting me have a copy of your paper, "An epidemiologic investigation of an excess lung cancer risk in a synthetic chemicals plant". Unfortunately, but not unexpectedly(I), I was unable to find you again ax the Congress and so I thought it best tc write to you to give you some of my preliminary impressions of your paper. Firstly, may I say that I found it an extremely thought-provoking paper and you seem to have developed a powerful new tool in the SAED model. I will be most interested to see it used in other situations. As you may know, I am a biologist not a statistician, and any comments I make are from that point of view. The statistics you use are far too complex for me to comment on! There are two mair. comments which stem from a biological appreciation of the chemical induction of cancer which I believe have a direct-bearing on' the conclusions of your paper. The first depends on the relationship between dose and time on the one hand, and incidence of cancer on the other. Druckery has studied the induction cf cancer in rats with a whole scries of nitrosamines and has produced quantitative data on the dose/ time/cancer incidence relationship. The best publication of this data is: H Druckery. "Quantitative aspects in chemical carcinogenesis" p.60, in "Potential carcinogenic hazards frem drugs" R Truhaut ed. UICC Monograph Series, vol 7. Basically he finds that incidence is proportional to daily dose x time*1-. In most cases the value of n is about 3. Coroboration of this .relationship comes from a study by Jones and Grenaon (Food Cosmet Toxicol 3^ 251-268 1975) using partly Druckery's data but also data from other studies, and again the relationship between time and incidence is a 3rd-power relationship. There are other examples in the literature which support this observation. In your study the time of exposure is treated as if it had "a 'constant influence on cancer incidence over the whole life of the plant. This is contrary to the way in which the relationship = known to operate, where earlier exposure should have a weighting which acknowledges its much greater contribution to cancer incidence. In your study the only way that this problem -is tackled is via latsr.rv 3- = -* r o' From: 'I F H Purchase To: Dr R J Waxweiler Page 2 9 Oct 78 R&S 111725 all the exposure data and also on the exposure data minus the last ten years. It seems to me that this overlooks one of the basic and important characteristics of the biology of cancer. The second problem I have relates to the estimation of dose in the SAED technique. I understand that the only technique available is to grade it from 0-5 using the opinion of the plant personnel, and clearly this is a good technique with which I do not have any quarrel. The problem is in regard to the biological significance of the doses and the best way I can explain it is by an hypothetical example. Let us assume that there are two chemicals, exposure to which is estimated by the SAED technique. The chemicals have the following properties:- A Cl) relatively non-toxic in the acute sense; (2) exposure undetected by human senses, ie odourless, non-irritant, allowing a relatively high TLV (say 1,000 ppm); (3) carcinogenic to man at a high dose.' B Cl) same physical properties as A; C2) acutely toxic; C3) irritant and with pungent odour, requiring a very low TLV Csay 1 ppm); C4) carcinogenic to man at the same dose as A. In this example, the estimation of exposure of A and B, in the absence of knowledge of carcinogenicity, could well be very different. Thus, figures might be:- Category of exposure Exposure In ppm Chemical A Chemical B 0 00 1 77 2 77 3 77 4 1,000 ppm 1 ppm 5 >>>>1,000 ppm >1 ppm not acceptable because of smell etc Remembering that our assumption is that the chemicals are of equal carcinogenicity, but that the TLV is not based on knowledge of carcinogenicity (as would be the case with PVC dust if your deductions are right), you can see that a grade 4 exposure might be allocated by the company personnel to doses (in ppm) of chemicals which are 1,000-fold different, and this would also be reflected in terms of carcinogenicity. The reason for this is that the company personnel are grading exposure on criteria which are definitely different from those assumed in the further statistical treatment of the grades. From: I F H*?urchase Tt Dr R J Waxweiler Page 3 9 Oct 78 Expressing the concept outlined above has proved to be very difficult, and I hope that it is well enough laid out for you to understand. There are other ways of expressing it and possibly you might like to compare, in ppm terms, the grading which would be given to ethylene and vinyl chloride or ethylene and sulphur dioxide. All are gases, vinyl chloride and ethylene being controlled in the old days because of flammability hazards and S02 being very irritant. A grade 5 exposure for ethylene and SO2 might easily be 1,000-fold different with vinyl chloride more like ethylene, but in addition being a carcinogen. How does the SAED technique take account of these different biological properties? In conclusion there are possible flaws in both of the parameters used in computing the serially additive dose which, I believe, will not be corrected by the sophisticated statistics which are later applied to the data. I am also forced to conclude that the arguments which connect tumour incidence specifically to exposure to any of the chemicals on this plant are not particularly persuasive, but I would be pleased to hear from you that I have overlooked something. These issues are so complex, that writing a letter is a poor substitute for personal discussion - something I hope we will be able to arrange in the not too distant future. Yours sincerely R&S 111726 I F H Purchase Assistant Director r`*v ./ ,r Imperial Chemical Industries Limited Central Medical Group , teUtf&ge. PO Box 3 Fulshaw Hall Wilmslow Cheshire SK9 1QB JJ/ Telephone Alderley Edge (0625) 582828 Telex 66841 1 Your ref Our ref JS/MJE/DS0-107 GMP/CMC Tel ext 5354 Copies to Dr K S Williamson Dr D M Ferguson Dr W G F Adams Dr D F Duffield Sir C Lawrence*Jones Dr F W Best Mr G J Sleddon Dr G Pigott Date 11 October 1978 Deer John FVC DUST Thank you for sending me Waxweiler's thesis; it is a fascinating piece of work. As an item of research 1 think it is excellent. Indeed, had I sufficient staff available, it is precisely the sort of approach I would like to have investigated! `myself. It also manages to throw in several reminders to budding epidemiologists which are well worth recording for future reference. For instance, the misguided linking of vinyl chloride with lung cancer without checking foralternative causes, and the sensible division of lung cancers into those which show an exces and those which do not, expose flaws in my own simplistic approach to problems. But you ask for an expert view on the hypothesis that FVC dust causes lung cancers. You will have noticed that the calculations are rather complicated and that they have to be performed by a computer program. A sound assessment f the validity of the calculations requires some intermediate data which are not provided. Lacking these data, one is forced to say that there are flaws in the technique, and that it is impossible to assess the importance of these flaws. The significance level at which the hypothesis 'FVC dust exposure is unrelated to lung cancer' is rejected is certainly far from the 1% or 5% figures which are to be found in the report. It is possible that it approaches the 50%, or toss of a coin# level. Some of the flaws are subtle, some less so, while others are improvements on greater flaws in other techniques. It is all rather complicated. As this paper is likely to be widely quoted, and as this technique is now available as a computer program and is a novel way of creating hypotheses, I feel an attempt should be made to put the technique on a firmer footing. Is there any satisfactory way of going about such a task? Yours sincerely R &s 111727 6 M paddle (Dr) i