Registration Dossier

Toxicological information

Epidemiological data

Currently viewing:

Administrative data

Endpoint:
epidemiological data
Type of information:
experimental study
Adequacy of study:
key study
Study period:
1948-1983
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
other: The study was not conducted according to guideline/s and GLP but the report contains sufficient data for interpretation of study results

Data source

Referenceopen allclose all

Reference Type:
publication
Title:
Unnamed
Year:
1982
Reference Type:
publication
Title:
Unnamed
Year:
1990

Materials and methods

Study type:
cohort study (retrospective)
Principles of method if other than guideline:
Retrospective cohort epidemiologic study of workers at Shell plants in Deer Park, Texas and Norco, Louisiana.

Test material

Reference
Name:
Unnamed
Type:
Constituent
Details on test material:
Production grade material.

Method

Details on study design:
The mortality experience was observed through 1983 among men believed to have been exposed to ECH for more than one quarter during the years 1948-65 at Shell's Deer Park, Texas plant and more than 6 months during the years 1955-1965 at Shell's Norco, Louisiana plant.
Exposure assessment:
estimated
Details on exposure:
The mortality experience was observed through 1983 among men believed to have been exposed to ECH for more than one quarter during the years 1948-65 at Shell's Deer Park, Texas plant and more than 6 months during the years 1955-1965 at Shell's Norco, Louisiana plant.

See below
Statistical methods:
See below.

Results and discussion

Results:
Table I shows observed deaths and SMRs for the Deer Park and Norco cohorts combined for four follow up periods. These represent three earlier follow up reports to the Shell Oil Company and the follow up reported here. The cause of death of primary interest was cancer of the respiratorv system.
Table I shows that the SMR for this condition was raised in the report for the period up to 1977 but that the SMR has since declined; in the report respiratory cancer was a cause for concern. Since the overall death rate at that time was fairly low respiratory cancer contributed heavily to total deaths.

The all cause SMR rose from 44.9 up to 1975 to 67.3 up to 1983. This type of increase appears to be due to an initial selection against unhealthy workers with the effect of this on SMRs diminishing as time progresses." SMRs for the leukaemias have been consistently raised, although numbers of deaths are small. There is a sharp increase in the "all other cancer" category in the current follow up period and in deaths from stroke. Of the five additional all other cancer deaths in the current follow up one was a brain cancer, one a melanoma, one prostatic, and two were of unspecified site.

Of the 93 deaths in table 1, 90 occurred during 1960-83. To use the MPDS fully to calculate expected deaths in the material that follows, the follow up period will be limited to 1960-83. Use of the MPDS permits local area mortality to be used to calculate expected deaths. This is thought to be desirable since it tends to correct the SMRs for geographic variations in cultural, social, and economic factors associated with mortality. Thus the mortality
experience of the workers is being compared with the mortality experience of a local population from which the workers were drawn. Expected deaths are based on white male death rates since the study population consisted only of men, almost all of whom were white.

Table 2 shows observed deaths and SMRs for a detailed cause of death Iisting by time since first exposure to ECH. The total number of workers in table 2 is 859 as compared with 863 in table 1, reflecting the three deaths that occurred before 1960 plus one worker lost to follow up before 1960. For the total cohort, the death rate for all causes of death is statistically significantly low and for no cause of
death is there a statistically significant excess. There are no deaths from non-malignant disease or cirrhosis of the liver and a low death rate for external causes and for heart disease. The deficits for non-malignant respiratory disease, cirrhosis of the liver, heart disease, and external causes are statistically significant. Taken together the deficits in these four causes of death suggest a group of workers with a fairly healthy lifestyle.

There is an increase in SMRs by time since first exposure for all causes of death and for all malignant neoplasms and a statistically significant excess in leukaemia 20 years or more since first exposure. For respiratory cancer, however, the SMR declined with time since first exposure and this is true when workers from Deer Park and Norco are examined separately (not shown). The SMR also declined with time since first exposure for cerebrovascular disease. The fact that the SMR for respiratory cancer was highest less than 20 years since first exposure argues against a causal relation for ECH. For Deer Park the possibility that exposures before exposure to ECH are involved in an excess has been explored elsewhere. Curiously, all deaths from cancer that occurred less than 20 years from first exposure were deaths from lung cancer.

The pattern of mortality in the two plants was similar, although numbers for most of the causes shown are small. SMRs for all causes of death are statistically significantly low and almost identical; those for prostatic cancer and leukaemia are in excess in both plants but these are based on small numbers of deaths.

For workers with 20 years or more since first exposure the pattern of SMRs is similar in Deer Park and Norco. The category all malignant neoplasms is slightly in excess in both plants and the excess is in prostatic cancer and leukaemia. Death rates for all causes, heart disease, and external causes are low in both plants

Of 444 workers from Deer Park observed 20 years or more since first exposure, 376 (85%) could be classified by level of exposure, whereas of 370 workers from Norco observed 20 years or more since first exposure, 342 (92%,) could be so classified. At Deer Park, of 376 workers classified as to exposure level, 62% were classified as moderate to heavy whereas for Norco, of 342 workers that could be classified, only 33%, were classified as moderate to heavy. Since there was an attempt to make each exposure classification comparable this may reflect lower exposures at Norco.

Table 7 shows the relation between probable exposure to epichlorohydrin and selected causes of death when the two study plants are combined. There is a direct relation between level of exposure and most major causes of death except external causes. The differences for all causes and for heart disease approach statistical significance (p < 0.07). The respiratory cancer relation with exposure is due almost entirely to the relation at Deer Park. The respiratory cancer SMRs for the higher exposure groups do not differ much from a SMR of 92.1 for respiratory cancer for white men 20 years or more from onset of employment for the entire Deer Park chemical complex (19 deaths) (G Marsh et al, unpublished report). Thus for Deer Park, at least, it does not appear that respiratory cancer is unusually high in the higher exposure group. While in the lower exposure group a SMR of 54.8 seems low this is based on only two deaths and may have little meaning. There is a weak relation between exposure to ECH and SMRs for leukaemia. As shown in table 2, the overall excess of leukaemia was statistically significant.
Confounding factors:
Workers who had probable exposure to ECH at two facilities operated by the Shell Oil Company had a favourable overall mortality experience, although with the passage of time their death rate is rising and approaching that expected in a general population. There is some evidence that these workers had a healthy lifestyle. The fact that there was a statistically significant deficit (no deaths) in non-malignant respiratory disease suggests that these workers may have smoked less than the general population. There was a statistically significant deficit (no deaths) in cirrhosis of the liver, which suggests that these workers may have drunk less alcohol than the general population. Death rates from heart disease accidents and other violence were low, suggesting other behavloural factors associated with healthy living.

Part of the cohort worked in the glycerine plant were potential exposure to allyl chloride occurred. Workers with heart disease were also exposed to allyl chloride. Thus the authors suggest allyl chloride exposure may explain the observed heart disease dose response relationship.
Strengths and weaknesses:
Not specified in report.

Any other information on results incl. tables

Table 1 Comparison of observed deaths and SMRs in four reports on workers with probable exposure to epichlorohydrin

   1948  -75  1948  -77  1948  -79

 1948

 -83
 Cause of death  Obs  SMR  Obs  SMR  Obs  SMR  Obs  SMR
 All causes  37  44.9**  52  53.4**  65  56.2**  93  67.3**
 All cancers  7  50.8  14  80.0  16  73.6  28  91.9
  Buccal cavity & pharynx  0  -  0  -  0  -  0  -
  Digestive system  0  -  2  52.5  2  42.3  3  46.6
  Respiratory system  5  97.8  9  131.6  10  114.4  14  110.7
  Urinary system  0  -  0  -  0  -  1  80.0
  Leukemias  1  125.0  2  224.7  2  194.2  3  243.9
  Lymphomas  0  -  0  -  0  -  0  -
  All other cancer  1  35.6  1  29.8  2  50.4  7  116.9
 Stroke  2  63.7  2  50.5  3  60.6  7  130.4
 All heart disease  16  58.0*  23  69.2  28  68.7*  37  72.2*
  Coronary heart disease  16  70.0  22  83.9  27  84.3  33  75.7
  Other heart disease  0  -  1  14.3*  1  11.4**  4  52.4
 Accidents  5  31.5**  6  35.3**  9  49.5*  10  65.9
 Suicide  2  57.0  2  47.4  2  41.9  2  38.1
 All other causes  5  34.6**  5  29.6**  7  34.6**  9  34.1**

*Significant at 5%, level; ** significant at 1% level.

Table 2 Observed deaths and SMRst by time since first exposure for workers wit)^ probable exposure to epichlorohydrin, Deer Park and Norco, 1960-83

   Total    <20    > 20  
 Cause of death (ICD 8th rev codes)  Obs  SMR  Obs  SMR  Obs  SMR
 All causes of death  90  64.7*  31  51.3**  58  75.6*
 All malignant neoplasms (140 -209)  28  90.8  6  53.4  22  112.2
  Prostate (185)  2  215.0  0  -  2  270.3
  Kidney (189.0 -189.2)  1  135.1  0  -  1  217.4
  Melanoma of skin (172.0 -172.4, 172.6 -172.9  1  138.9  0  -  1  256.4
  Central nervous system (191, 192)  1  85.5  0  -  1  161.3
  All lymphatic & hematopoietic tissue (200 -209)  3  100.0  0  -  3  181.8
  Leukemia & aleukemia (204 -207)  3  254.2  0  -  3  500.0*
  All other cancer  3  101.7  0  - 3  174.4
  Cerebrovascular disease (430 -438)  7  128.9  4  168.1  3  98.4
 Ulcer of stomach and duodenum (531 -533)  1  172.4  0  -  1  476.2
             
 Motor vehicle accidents (810 -823)  6  86.3  3  61.7  3  142.8
 Total No. of workers  859    859    814  
 Person-years  19681.7    13144.6    6537.1  

*p < 0.05; **p < 0-01.

Only data with SMRs greater than 150 are presented.

Table 7 Observed and expected deaths and SMRs of 20 years or more since first exposure for workers with probable exposure to

epichlorohydrin by estimated level of exposure, Deer Park and Norco

Level of exposure

     nil to light      moderate to heavy  
 Cause of death (ICD 8th rev codes)  Obs  Exp  SMR  Obs  Exp  SMR
 All causes of death  17  31.25  54.4**  38  39.83  95.4
 All malignant neoplasms (140 -209)  5  8.06  62.0  15  10.17  147.5
  Digestive system & peritoneum (150 -159)  0  1.66  -  3  2.12  141.5
  Respiratory system (160 -163)  2  3.65  54.8  5  4.70  106.4
  Prostate  1  0.30  333.3  1  0.41  243.9
  Leukemia & aleukemia (204 -207)  1  0.24  416.7  2  0.32  625.0
  All other cancer  1  2.21  45.2  4  2.62  152.7
Cerebrovascular disease (430 -438)  1  1.22  82.0  2  1.64  121.9
 Heart disease (380 -398, 400.1, 402, 404, 410 -414, 420 -429)  5  12.75  39.2*  17  16.13  105.4
 All external causes (800 -999)  2  3.35  59.7  2  4.13  48.4
 All other causes  3  5.88  51.0  2  7.76  25.8
 Unknown causes  1      0    
 Total No. of workers  373      345    
 Person-years  2734.1      3148.4    

*p < 0.05; **p < 0 .01.

Applicant's summary and conclusion

Conclusions:
There appears to be a slight increase in leukemia in workers after 20 years of exposure as well as a slight increase in cardiovascular disease. Overall, the heart disease SMR 20 years or more after first exposure was 39.2 (five deaths) for low exposure and 105-4 (17 deaths) for high exposure
Executive summary:

An epidemiological study was undertaken to determine whether the animal carcinogen, epichlorohydrin (ECH), produces cancer in man. A total of 863 workers with probable exposure to ECH at two chemical plants during 1948-65 were followed up for deaths up to 1983. Twenty years or more after first exposure the all cancer SMR was 112.2 (22 deaths) and the SMR for leukaemia was 500.0 (three deaths), which is statistically significant. All cancer, leukaemia, and most other causes of death were related to estimated levels of exposure to ECH, except violence. The most consistent (both plants) relation was between exposure level and heart disease. Overall, the heart disease SMR 20 years or more after first exposure was 39.2 (five deaths) for low exposure and 105-4 (17 deaths) for high exposure. Limited evidence of a cardiovascular disease relation to ECH production in one other epidemiological study is supported by this study. Ally1 chloride used in the production of ECH may play a part. The relation of heart disease and exposure does not appear to be an artifact, although the fact that many other causes of death were also related to exposure argues against a causal relation.