Objective  

To conduct a systematic review and meta-analyses of published studies examining the association between myeloid leukemias (ML) and occupational pesticide exposure.

Methods  

Studies were identified from a MEDLINE search through 31 May 2006 and from the reference lists of identified publications. Studies were summarized and evaluated for publication bias. Relative risk estimates for ML were extracted from 17 cohort and 16 case–control studies published between 1979 and 2005. Fixed- or random-effect meta-analysis models were used depending on the presence of heterogeneity between studies. Separate analyses were conducted after stratification for study design, occupational group, ML subtype or gender.

Results  

The overall meta-rate ratio estimate (meta-RR) for the cohort studies was 1.21 (95% confidence interval [CI] 0.99–1.48). Substantial heterogeneity existed among cohort studies (p = 1.064 × 10⁻⁵), mainly reflecting the varying occupational categories examined. The meta-RR was 6.32 (95% CI: 1.90–21.01) for manufacturing workers and 2.14 (95% CI: 1.39–3.31) for pesticide applicators. After stratification of cohort studies by specific ML subtype, an increased risk of acute myeloid leukemia (AML) was found (meta-RR: 1.55; 95% CI: 1.02–2.34). No significant heterogeneity was detected among case–control studies and an increased risk of chronic myeloid leukemia (CML) was found among men (meta-RR: 1.39; 95% CI: 1.03–1.88) and farmers or agricultural workers (meta-RR: 1.38; 95% CI: 1.06–1.79).

Conclusion  

The strongest evidence of an increased risk of ML comes from manufacturing workers and pesticide applicators. Further studies will be needed to correlate reliable exposure data within these specific occupational groups with well-defined subtypes of leukemia to refine this assessment.