Environment and Breast Cancer: Science Review

Evidence From Humans
Print this page
Exposure to benzene and risk of breast cancer among shoe factory workers in Italy
Costantini, A. S., Gorini, G., Consonni, D., Miligi, L., Giovannetti, L., Quinn, M. Tumori. 2009. 95:1, 8-12.
Topic area
Environmental pollutant - Occupation Benzene
Study design
Retrospective cohort
Funding agency
Not reported
Study Participants
Menopausal Status
The menopausal status of women included in this study is listed here.
No analyses by menopausal status
Number in Cohort
Cohort: 679
Participant selection: Inclusion and exclusion criteria
Criteria used to select participants in the study.
The cohort included all women at work on January 1st, 1950 and women hired through December 31st, 1963 at a shoe-factory in Florence, Italy, who had specific job information. Cases were women with incident breast cancer who lived in the Province of Florence, identified from the regional cancer registry from 1985-2000. Women who moved out of Florence before 1985 were excluded. Person-years spent outside of Florence were also not counted in the exposure metric.
Comment about participation selection
The selection of this occupational cohort of women exposed to benzene may be useful for understanding the relationship of benzene and breast cancer. The authors report that 80% of the cohort has information on work history. It is unclear how women missing work history may have been different from others in the cohort.
Exposure Investigated
Exposures investigated
Cumulative occupational exposure (duration of work x intensity from job exposure matrix) to benzene estimated from work history at a shoe factory where benzene-based glues were used until 1964.
How exposure was measured
Job exposure matrix
Exposure assessment comment
The exposure assessment was not able to include environmental samples of benzene, and instead used an industry-based job exposure matrix to quantify each participants' estimated intensity of exposure. This exposure measure may or may not correctly rank the exposures relevant to disease. The range of the exposure was between 0 and 488 ppm-years, while the median was 4.4 ppm-years suggesting that the majority of cohort members experienced relatively low exposure levels.
Breast cancer outcome investigated
Primary incident breast cancer
Confounders considered
Other breast cancer risk factors, such as family history, age at first birth, and hormone replacement therapy use, that were taken into account in the study.
Latency period was considered. No information on individual reproductive factors or other possible risk factors were obtained.
Genetic characterization included
If the study analyzed relationships between environmental factors and inherited genetic variations, this field will be marked Yes. No, if not.
Strength of associations reported
Breast cancer mortality compared to Tuscany population:
>40 ppm-yrs cumulative exposure: SMR 131.0 (95% CI 54.5-314.8) (5 cases)
>40 ppm-yrs cumulative exposure, <30 yr latency: SMR 71.1 (95% CI 10.0-504.9) (1 case)
>40 ppm-yrs cumulative exposure, ≥30 yr latency: SMR 166.0 (95% CI 62.3-442.2) (4 cases)

Breast cancer incidence compared to Florence population:
>40 ppm-yrs cumulative exposure: SIR 130.1 (95% CI 62.0-273.0) (7 cases)
>40 ppm-yrs cumulative exposure, <30 yr latency: SIR 211.9 (95% CI 29.9-1504) (1 case)
>40 ppm-yrs cumulative exposure, ≥30 yr latency: SIR 122.3 (95% CI 42.4-245.0) (6 cases)
≤40 ppm-years cumulative exposure: SIR 116.0 (95% CI 72.1-186.6)
≤40 ppm-years cumulative exposure, <30 yr latency: SIR 135.8 (95% CI 10.7-261.1) (9 cases)
≤40 ppm-years cumulative exposure, ≥30 yr latency: SIR 99.7 (95% CI 49.8-199.3) (8 cases)
Results Comments
When lower cumulative exposures were considered, nonsignificantly elevated SMRs and SIRs were observed. The authors were unable to adjust for individual reproductive factors and risk factors as potential confounders. Two models with different latency periods (> or <30 years latency) were explored, but the category of <30 years latency was limited by the cancer registry's inception in 1985.
Author address
Occupational and Environmental Epidemiology Unit, ISPO Cancer Prevention and Research Institute, Florence, Italy. a.seniori@ispo.toscana.it
Reviewers Comments
Because the general population was chosen as the comparison population, these results may be underestimated due to the healthy worker effect.