BOA1
A systematic review and meta-analysis of occupational exposure to benzene: leukemia in relation to benzene exposure
Abdul Khalade1, Maritta Jaakkola1, Eero Pukkala2, Jouni Jaakkola1
1University of Birmingham, Birmingham, UK, 2Finnish Cancer Registry, Institute for Statistical and Epidemiological Cancer Research, Helsinki, Finland
Background
The original description of occupational exposure to benzene and risk of leukemia was introduced be Le Noire in 1897. Occupational exposures have been implicated in the aetiology of Leukemia but the rarity of the disease, combined with the limited size of most occupational cohorts; have made difficult the study of the association between these exposures and leukemia. Previous studies investigating the risk of cancer from exposure to benzene have shown inconclusive findings suggesting further research is required as there is accumulating evidence in the risk of cancer from benzene exposure.
Method
We conducted a through systematic literature review simultaneously using two databases ‘MEDLINE’ and ‘EMBASE’ from 1966 through to April 2008. The search category commands we used were Benzene; occupational exposure and Cancer. The list of potential papers was then screened by initially checking the abstracts. The papers were then analyzed further following a priori inclusion criteria. We selected articles which provided information that can be used to estimate any relation between benzene exposure and cancer risk (effect size) in terms of odds ratio, standardized mortality ratio (SMR), Standardized relative risk (SRR), p value or Cumulative incidence ratio (CIR).
Results
In all 17 studies the effect size was greater than 1, and 16 studies had the lower limit of the 95% CI at greater than 1. The summary OR from the fixed-effects model was 2.552 (95% CI, 2.276-2.794). The random-effects model that allowed for heterogeneity yielded a summary OR of 2.300 (95% CI, 1.753-3.019), suggesting an even greater risk of leukemia in relation to benzene exposure.
Conclusion
Our study provides a benchmark to work on revealing publication bias present in studies conducted earlier than most new studies. Evidently the exposure type and size of study have varied considerably amongst individual studies giving rise to significant heterogeneity.
