Registration Dossier

Toxicological information

Epidemiological data

Currently viewing:

Administrative data

Endpoint:
epidemiological data
Type of information:
other: Case Control Study
Adequacy of study:
key study
Reliability:
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
Justification for type of information:
A discussion and report on the read across strategy is given as an attachment in Section 13 of the dossier.
Cross-reference
Reason / purpose:
read-across: supporting information

Data source

Referenceopen allclose all

Reference Type:
publication
Title:
The Health Watch case-control study of leukaemia and benzene
Author:
Glass DC, Gray CN, Jolley DJ, Gibbons C and Sim MR
Year:
2006
Bibliographic source:
Ann NY Acad. Sci.1076, 80-89
Reference Type:
publication
Title:
Leukemia risk associated with low-level benzene exposure
Author:
Glass DC, Gray CN, Jolley DJ, Gibbons C, Sim MR, Fritschi L, Adams GG, Bisby JA, and Manuell R
Year:
2003
Bibliographic source:
Epidemiology 14(5):569–577.
Reference Type:
publication
Title:
Benzene exposure and leukemia - Letter to the editor
Author:
Schnatter AR
Year:
2004
Bibliographic source:
Epidemiology15, 509
Reference Type:
publication
Title:
Benzene exposure and leukemia - authors reply
Author:
Glass DC, Gray CN, Jolley DJ, Sim MR and Fritschi L
Year:
2004
Bibliographic source:
Epidemiology 15, 510-1.
Reference Type:
publication
Title:
Update of mortality and cancer incidence in the Australian petroleum industry cohort.
Author:
Gun RT, Pratt N, Ryan P, Roder D
Year:
2006
Bibliographic source:
Occup Environ Med 63, 476-81.
Reference Type:
publication
Title:
Health Watch exposure estimates: do they underestimate benzene exposure?
Author:
Glass DC, Gray CN, Jolley DJ, Gibbons C, Sim MR.
Year:
2005
Bibliographic source:
Chem Biol Interact 153-154, 23-32.

Materials and methods

Study type:
case control study (retrospective)
Endpoint addressed:
carcinogenicity
Test guideline
Qualifier:
no guideline available
Principles of method if other than guideline:
This study was a case-control study with 79 cases (33 leukaemias, 31 non-Hodgkin lymphoma and 15 multiple myeloma) each matched to 5 controls, nested in the Health Watch cohort, to investigate whether the observed excess of lymphohaematopoietic cancers was related to occupational benzene exposure. Health Watch is a prospective cohort study of employees in the Australian petroleum industry who have worked for more than 5 years (16,547 males and 1,356 females). It compares the mortality and cancer incidence of the cohort with that of the Australian population. The cohort includes employees from offices, upstream extraction and processing sites, refineries, terminals and airports all over Australia.
GLP compliance:
no

Test material

Reference
Name:
Unnamed
Type:
Constituent
Details on test material:
- Name of test material (as cited in study report): Benzene
- Physical state: vapour
- No further details reported.

Method

Type of population:
occupational
Ethical approval:
not specified
Details on study design:
HYPOTHESIS TESTED: Whether the excess of lymphohaematopoietic cancers, identified among male members of the Health Watch cohort was associated with benzene exposure (the most recent update of the Health Watch study reports virtually no excess of leukaemias).

METHOD OF DATA COLLECTION: Type: Interview, Questionnaire, Record review , Work history:
- Details: A job history was prepared for each subject based on data collected from 4 cohort interviews between 1980 and 1999. This was then checked by the employing company against their records. Using appropriate job-specific questionnaires, contemporary colleagues were interviewed identifying the time spent on each task, the technology used and the products handled (benzene, diesel, petrol). The interviews were carried out by one of 2 experienced occupational hygienists, blinded as to case status.

STUDY PERIOD: 1981-1999.

SETTING: Occupational exposure in the Australian petroleum industry, including employees from offices, upstream extraction and processing sites, refineries, terminals and airports, all over Australia.

STUDY POPULATION
- Total population (Total no. of persons in cohort from which the subjects were drawn): 18,000 employees (1300 of which are women) in the Health Watch cohort.
- Total number of subjects participating in nested case-control study: 494 of which 79 were cases (33 leukaemias including 11 acute non lymphocytic leukaemia (ANLL), 31 non-Hodgkin lymphoma and 15 multiple myeloma).
- Sex/age/race: Males. Age and race not reported.
- Smoker/non-smoker: not reported
- Total number of subjects at end of study: not reported
- Matching criteria: matched on age and sex to 5 controls from the cohort.
- Other: subjects worked in the industry from an average of 20 years (range 4-42 years), the majority started employment after 1965.

COMPARISON POPULATION
- Type: Control or reference group:
- Details: Controls were drawn from the cohort, with replacement, and matched to cases on age and sex.

HEALTH EFFECTS STUDIED
- Disease(s): lymphohaematopoietic (LH) cancer: non-Hodgkin lymphoma, multiple myeloma or leukaemia
Exposure assessment:
estimated
Details on exposure:
TYPE OF EXPOSURE: Occupational exposure to benzene in the petroleum industry in Australia, in refineries and terminals.

TYPE OF EXPOSURE MEASUREMENT: other: Exposure was estimated retrospectively for each occupational history using an algorithm in a relational data base. Benzene exposure measurements, supplied by Australian petroleum companies, were used to estimate exposure for specific tasks. The tasks carried out within the job, the products handled and the technology used, were identified from interviews with contemporary colleagues.
Jobs were also identified where there might have been infrequent high exposure events as a result of spillages or work practices no longer carried out and so not represented by the base estimates. The exposure that might have resulted from such occurrences was added to the mean exposure for that subject in some analyses.

EXPOSURE LEVELS: Exposure was low, the estimated lifetime cumulative benzene exposures ranged from 0.005 to 57.3 ppm years with a mean of 4.9 ppm years. 85% of the cumulative exposure estimates were <10 ppm years.
For cases the mean exposure was 7.27 ppm years (10.63 ppm years, for leukaemia cases) for controls it was 3.86 ppm years. These figures do not include exposure from high exposure events and no quantitative information on mean cumulative exposures after the inclusion of high exposure events is given, but it is clear from Figure 1 that it may have doubled the estimates of cumulative exposure of many workers in the highest exposure group (>16 ppm years). Overall, the addition of occasional high exposures increased exposure for 25% of subjects but for most, the increase was less than 5% of the total exposure.

EXPOSURE PERIOD: Average of 20 years (range 4-42 years).

POSTEXPOSURE PERIOD: Health Watch commenced in 1981 and recruitment closed in 1999. The cancer incidence and mortality of this cohort has been compared with that of the Australian population up to 2005.

DESCRIPTION / DELINEATION OF EXPOSURE GROUPS / CATEGORIES: Cases were defined as male members of the Health Watch cohort who had: 1st diagnosis of lymphohaematopoietic (LH) cancer after entering the Health Watch cohort; and diagnosis confirmed by pathology report, cancer registration, letter from medical practitioner, or death certificate; and had reported LH cancer to Health Watch either by self or by family, unless they were lost to contact by Health Watch, or were deceased.
Statistical methods:
Odds ratios for leukaemia, non-Hodgkin lymphoma and multiple myeloma were calculated from conditional logistic regression using STATA. In categorical analyses for leukaemia, the lowest exposure group (described as < 1 ppm year) was used as a reference group by Glass et al, 2003 but Glass et al. (2005,2006) give results for a new reference group which combines the two lowest exposure groups used by Glass et al, 2003. This was done to give a more stable analysis as the lowest exposure group only contained 3 leukaemia cases. Cumulative exposure in the new reference group is stated to be ≤1 ppm year by Glass et al (2006), although it was described as ≤ 2 ppm year by Glass et al 2003, 2005.

Results and discussion

Results:
Matched analyses showed that non-Hodgkin lymphoma and multiple myeloma were not associated with benzene exposure.
Leukaemia risk was significantly increased for the subjects with greater than 16 ppm years cumulative exposure, odds ratio (OR) 51.9 (95% CI 5.6-477) but reduced to an OR of 7.79 (95% CI 2.34-25.89) when high exposure events were included (see Table). When cumulative exposure was treated as a continuous variable in a matched analysis, the OR = 1.10 (95% CI 1.04-1.16) per ppm year, but reduced to OR = 1.03 (95% CI 1.01-1.05) per ppm year when high exposure events were included. Glass et al. (2003) used a different reference group and did not include high exposure events, and reported an OR of 98.2 (95% CI 8.8-1090) for subjects with greater than 16 ppm years cumulative exposure and OR = 1.65 (95% CI 1.25-2.17) per ppm year when cumulative exposure was treated as a continuous variable. For ANLL, Glass et al (2003) reported an OR of 7.17 (95% CI 1.27-40.4) for subjects with greater than 8 ppm years cumulative exposure compared to those with ≤ 4 ppm years. Increased risk of leukaemia was reported for subjects whose highest exposed job had greater than 0.8 ppm intensity, but it is not stated whether this analysis included high exposure events. Excluding high exposure events, the estimated lifetime cumulative benzene exposures ranged from 0.005 to 57.3 ppm years, with a mean of 4.9 ppm years. A comparison of the distribution of lifetime cumulative exposure (excluding high exposure events) with those of subjects in the two comparable petroleum industry case–control studies from the UK and Canada, showed exposures were similar in the three studies.
Confounding factors:
Tobacco and alcohol consumption were considered as confounders. No other confounding exposures were considered likely.
Strengths and weaknesses:
Strengths: Exposure estimates were detailed, based on job histories collected prospectively and verified by colleagues and on exposure data collected by the companies' occupational hygienists. Job histories were checked against company records and interviews were case blind and structured.
Weaknesses: Potential for recall bias in exposure assessments.
- Small number of cases.
- The range of cumulative exposures of cases and controls is very small given the likely precision of exposure estimates.
- Study results were very sensitive to choice of reference group and choice of cut off points for categorical groupings (see Schnatter et al., 2004).
– It isn’t clear what level of benzene the reference group was exposed to.
- Study results were also very sensitive to assumptions made about high exposure events.
- Insufficient information is supplied about the distribution of cumulative exposures to benzene when high exposure events were incorporated in the exposure assessment.

Any other information on results incl. tables

The most recent analyses incorporating exposure from high exposure events and utilising a different reference group indicate a much weaker association between leukaemia and benzene exposure than originally reported by Glass et al. (2003). However, they still predict a large excess of leukaemia in the Health Watch cohort as just over half of control subjects had cumulative exposures greater the cut off point for the reference group. There is no corresponding increase in leukaemia incidence in the Health Watch cohort. The latest update by Gun et al. (2006) reported an SIR of 1.07 (95% CI 0.74-1.49, 34 observed) for leukaemia and an SIR of 0.97 (95% CI 0.42-1.91, 8 observed) for acute myeloid leukaemia.

Conditional (fixed-effects) logistic regression, leukaemia by cumulative exposure to benzene (ppm-years) (Table based on Glass et al, 2005; Table 2).

Benzene exposure
(
ppm-years)

Number of
controls

Number of
cases

Cumulative exposure
OR (95% CI)

Cumulative exposure including

high exposure events OR (95% CI)

≤2

84

9

1.00

1.00

>2–4

30

8

2.89 (0.97–8.52)

3.07 (1.02–9.28)

>4–8

27

3

1.17 (0.27–4.98)

1.22 (0.28–5.22)

>8–16

21

6

3.11 (0.91–10.56)

2.68 (0.71–10.08)

>16

3

7

51.88 (5.64–477)

7.79 (2.34–25.89)

 

Applicant's summary and conclusion

Conclusions:
The case-control study indicates that occupational exposure to benzene is associated with increased incidence of leukaemia but not non-Hodgkin lymphoma or multiple myeloma. However, the dose response curve is inconsistent and very sensitive to the choice of reference group, cut off points for categorical groupings and assumptions made about high exposure events.
Executive summary:

A nested case-control study investigated a possible association between benzene exposure and lymphohaematopoietic cancers noted in male members of the Health Watch cohort. Exposure was retrospectively estimated for each occupational history and shown to be low - 85% of cumulative exposure estimates were < 10 ppm years.

A strong association between leukaemia and modest benzene exposure was demonstrated, but non-Hodgkin lymphoma and multiple myeloma were not associated with benzene exposure. Leukaemia risk was increased for subjects with greater than 16 ppm years cumulative exposure when high exposure events were incorporated (OR 7.79) or with greater than 0.8 ppm intensity of highest exposure. However, the dose response curve is inconsistent and very sensitive to the choice of reference group, cut off points for categorical groupings and assumptions made about high exposure events. Only ANLL was significantly associated with benzene exposure, but there was limited information available to assess associations for leukaemia subtypes and no analyses incorporating high exposure events.