Abstract

2,4-dichlorophenoxyacetic acid (2,4-D), an herbicide used widely to control weeds, was recently classified as “possibly carcinogenic to humans” (Group 2B) by IARC in June 2015. The evaluation was based on limited evidence for carcinogenicity in experimental animals, strong mechanistic evidence that 2,4-D induces oxidative stress in humans as well as moderate evidence of immunosuppression, and inadequate evidence for carcinogenicity in humans. The IARC Working Group conducted a meta-analysis to assess the association between exposure to 2,4-D and risk of non-Hodgkin lymphoid neoplasms (including non-Hodgkin lymphoma (NHL), chronic lymphocytic leukaemia, hairy cell leukaemia, and multiple myeloma) to inform its evaluation of the epidemiological evidence. In a primary analysis that prioritised inclusion of estimates adjusted for other pesticides, when available, the meta-relative risk (meta-RR) was 1.04 (95% CI: 0.88–1.22; I2 = 6%) whereas the meta-RR was 1.31 (95% CI: 1.10–1.56; I2 = 37.6%) when using estimates unadjusted for concomitant pesticide use from the same studies. However, the risk estimates were elevated in persons with higher exposure to 2,4-D in the 8 studies reporting on some metric of exposure-response (frequency, intensity or duration of use or lack of protective equipment). Various sensitivity analyses were conducted to explore the robustness of the data. The meta-RR was 0.97 (95% CI: 0.82–1.15; I2 = 0%) for the 10 retrospective case-control studies, 1.11 (95% CI: 0.84–1.47; I2 = 41.1%) for the 6 studies conducted in the United States, 0.99 (95% CI: 0.71–1.39; I2 = 0%) for the 3 studies reporting results for multiple myeloma and 1.06 (95% CI: 0.80–1.40; I2 = 36.3%) for the 9 studies reporting on NHL. Adjustment for confounding modified the meta-RRs, indicating the importance of conducting sensitivity analyses.