Registration Dossier

Toxicological information

Epidemiological data

Currently viewing:

Administrative data

epidemiological data
Type of information:
migrated information: read-across based on grouping of substances (category approach)
Adequacy of study:
key study
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
other: Human population study, published in peer-reviewed literature, minor restrictions in design but adequate for assessment
Reason / purpose:
reference to other study

Data source

Referenceopen allclose all

Reference Type:
Risk of benzene-induced leukemia: a sensitivity analysis of the Pliofilm cohort with additional follow-up and new exposure estimates.
Crump KS
Bibliographic source:
J Toxicol Environ Health 42, 219-242
Reference Type:
Benzene exposure and haematopoietic mortality: a long-term epidemiologic risk assessment
Rinsky RA, Hornung RW, Silver SR and Tseng CY
Bibliographic source:
Am J Ind Med, Vol 42, pp 474-480

Materials and methods

Study type:
cohort study (retrospective)
Endpoint addressed:
Test guideline
no guideline available
Principles of method if other than guideline:
Previous studies of a cohort of ~1200 rubber hydrochloride workers (the Pliofilm cohort) exposed to benzene have indicated a significant association between benzene exposure and mortality from leukaemia (Rinsky et al, 1981, 1987; Paxton et al, 1994a). To determine whether risks remain elevated with increasing time since plant shutdown, this study extended follow-up from 1981 to 1996, although Paxton et al (1994a) has reported on the mortality of essentially the same cohort through 1987. Risk was evaluated using standardized mortality ratios (SMR) and generalised Cox proportional hazards regression models.
GLP compliance:
not specified

Test material

Details on test material:
- Name of test material (as cited in study report): Benzene
No other details reported


Type of population:
Ethical approval:
not specified
Details on study design:
HYPOTHESIS TESTED: Does the leukaemia mortality risk from benzene exposure remain elevated with increasing time since exposure.

METHOD OF DATA COLLECTION: Record review / Work history
- Details: To update vital status through to 1996, records were used from the Social Security Administration and the National Death Index (which provided coded causes of death). Job-exposure matrices were created based on available air sampling data. Cumulative exposures (ppm-years) for each worker were derived from the matrices by summing daily exposure values using detailed work histories, as described in the previous publication (Rinsky et al, 1987) and dividing the result by 365.25.

STUDY PERIOD: Follow-up from January 1, 1950 through to December 31, 1996.

SETTING: Rubber hydrochloride production workers employed in two plants in Ohio, USA for at least 1 day between January 1, 1940 and December 31, 1976.

- Total population (Total no. of persons in cohort from which the subjects were drawn): 1845
- Selection criteria: At least 1 ppm-day of exposure between January 1, 1940 and December 31, 1976 who were alive as of January 1, 1950.
- Total number of subjects participating in study: 1845 (exposed and unexposed non-salaried workers)
- Sex/age/race: Race was only known definitively for deceased employees but since 99% of these were white race was not modelled. The effect of gender was examined. As of 31 December 1996, about 94% of the living members of the cohort were at least 50 years old and 53% were age 70 or over.
- Smoker/non-smoker: Not reported.
- Total number of subjects at end of study: 869 workers (exposed and unexposed) alive at end of study.

- Type: Control or reference group
- Details: 554 unexposed workers who were alive as of January 1, 1950

- Disease(s): Underlying cause of death was used for all analyses.
- Analysis restricted to deaths classified under the seventh or later revisions of the ICD.

Person-years were stratified by increasing levels of cumulative exposure, each strata representing a 40 year working-lifetime exposure to less than 1, 1-4.99, 5-9.99 and 10 or more ppm benzene.
Information about each case of leukaemia and multiple myeloma, including age at death, year of death, latency, diagnosis, cause of death, cumulative exposure and location and duration of employment was reported.
Exposure assessment:
Details on exposure:
Exposed workers included all those (1,291) with at least 1 ppm-day of exposure between January 1, 1940 and December 31, 1976, employed at one of three rubber hydrochloride plants in two Ohio locations. Operations began at Location 1 in 1939 and ceased in April 1976. At location 2, plant 1 operated between 1936 or 1937 until approximately 1949 and plant 2 operated between mid 1949 until 1965. A more detailed description of the rubber hydrochloride operations is presented by Rinsky et al. (1981, 1987). Job exposure matrices were created based on available air sampling data and cumulative exposures (ppm-years) were derived from the matrices by summing daily exposure values using detailed work histories (Rinsky et al., 1987).
Statistical methods:
Standardised mortality ratios (SMRs) with 95% confidence intervals for all causes of death were calculated using NIOSH's PC-Life Table Analysis System (PC-LTAS), Version 1.0c. US population death rates were used for comparison with death rates in the cohort.
To evaluate the effects of benzene on risk of leukaemia and multiple myeloma, as well as potential confounders and effect modifiers, a generalised form of Cox proportional hazards regression model was used.

Results and discussion

For all exposed cohort members, overall mortality during the years 1950-1996 was similar to that of the US population for that time period (Table 1) with an SMR of 0.98 (95% CI= 0.90-1.05). Mortality from all malignant neoplasms was also as expected but deaths from lymphatic and haematopoietic neoplasms were elevated, with a combined SMR of 1.64 (95% CI = 1.06-2.42, 25 deaths) and the SMR for leukaemia was 2.47 (95% CI = 1.38-4.07, 15 deaths). The leukaemia cases had widely varying times to death since first exposure (latency) with 6 occurring at least 30 years after the workers first exposure. Only one new death from leukaemia had occurred since the update through 1987 by Paxton et al. (1994a). There were no deaths from lymphatic and haematopoietic neoplasms among exposed female subjects.
Among exposed white males, SMRs for leukaemia were elevated for all exposure categories and increased with increasing exposure (Table 2) but only reached statistical significance for the highest exposure category (> 400 ppm years).
Beginning observation in 1960, and using 1960-1999 rates for multiple myeloma yielded an SMR of 2.04 (95% CI 0.66-4.76) for exposed subjects. One case of multiple myeloma had occurred in exposed subjects since 1981 bringing the total to 5 cases. Three cases had occurred in the smaller unexposed persons cohort but an SMR was not reported. Of the 5 cases of multiple myeloma occurring in exposed workers in this cohort, 4 were in the lowest exposure category (1 ppm-day to 30.9 ppm years of exposure) with the 5th case being in the highest exposure category (> 400 ppm-years).
Cox proportional hazards modelling showed significant associations between leukaemia and cumulative benzene exposure for models with no lag, a 10 year lag and a 20 year lag. There was no corresponding association between multiple myeloma and cumulative benzene and the coefficient in the twenty year lag model was negative
Confounding factors:
The impact of gender, exposure rate (cumulative exposure divided by duration of exposure) and peak exposure (highest exposure level experienced by each worker for any length of time) were also examined.
Strengths and weaknesses:
Strengths: The Pliofilm cohort has the big advantage that exposure to other chemicals (and the resulting potential confounding effects) can be neglected (EU RAR, 2008).
Weaknesses: The study is small (~1200 subjects) and no information on potential confounders such as smoking or social class was available. There is debate about the exposures of workers in the cohort and two other sets of estimates were used by Crump (1994) to model the relation between leukaemia and cumulative benzene exposure.

Any other information on results incl. tables

TABLE 1. Observed Deaths* and SMRs 1950-1996 in Rubber Hydrochloride Workers Exposed to at Least 1 ppm-day of Benzene (Table based on Rinsky et al, 2002; Table III).


All subjects (n=1291)

White males (n=1165)



SMR (95% CI)


SMR (95%CI)

All causes





All malignant neoplasms


1.03 (0.88-1.20)



Lymphatic and haematopoietic cancers combined:


1.64 (1.06-2.42)


1.70 (1.10-2.51)






Multiple myeloma*


2.04 (0.66-4.76)



Non-Hodgkin lymphoma*


0.96 (0.31-2.25)



*Risk begins in 1950 and expected deaths are based on U.S. population rates 1950-1996 for all outcomes except multiple myeloma and non-Hodgkin Lymphoma (NHL).

For multiple myeloma and NHL, risk begins in 1960, the first year separate rates are available. Expected deaths for these outcomes are based on 1960-1996 U.S. rates.


TABLE 2. SMRs for Leukaemia in White Males Exposed to Benzene, by Exposure Category(Table based on Rinsky et al, 2002; Table IV).


Exposure category

1 ppm day-30.99 ppm-years

40 - 199.99 ppm-years

200 - 399.99 ppm-years

> 400 ppm-years











SMR (95%Cl)



5.55 (0.62-24.08)



Applicant's summary and conclusion

Benzene exposure is associated with an increased risk of mortality from leukaemia. The Pliofilm cohort provides the best data set to perform a quantitative exposure analysis and analyses performed by Crump (1994) and Paxton et al. (1994b) remain relevant. Overall, occupational risk estimates based on these data are in the range 0.02 to 5.1 additional leukaemia deaths per 1000 individuals with 45 ppm-years of cumulative benzene exposure. Risk estimates for AMML derived by Crump (1994) are in a range of 0.02 to 4.3 additional AMML deaths per 1000 individuals. With regard to general population exposure, the models derived by Crump (1994) and Paxton et al. (1994b) remain relevant, but lifetime risk estimates calculated by Crump (1994) appear conservative when compared to estimates calculated by TCEQ (2007) using the same models derived by Crump (1994), but assuming 70 years exposure and calculated using much more recent US mortality and survival rates
Executive summary:

A cohort of rubber hydrochloride workers has been investigated in an update of a retrospective cohort mortality study.

The latest findings reaffirm the leukaemogenic effects of benzene exposure. Only 1 new leukaemia death had occurred in the additional 9 years of follow up and the study suggests that excess risk diminishes with time. Risk estimates such as Crump (1994) based on the previous update would be unlikely to change much if they incorporated additional follow up information from the latest update of the cohort mortality study.