Do polychlorinated biphenyls cause cancer? A systematic review and meta-analysis of epidemiological studies on risk of cutaneous melanoma and non-Hodgkin lymphoma.

In 2015 a IARC Working Group upgraded the classification of PCBs to Group 1 "Carcinogenic to humans", also on the basis of evidence from epidemiological studies showing an excess risk for melanoma. Increased risks for non-Hodgkin lymphoma (NHL) and breast cancer were also reported though the evidence was limited. However, some recent reviews of studies on PCB exposure and risk of cancer provided discrepant findings. Therefore, we re-evaluated the association between exposure to PCBs and risk of melanoma and NHL by a systematic review and meta-analysis. We retrieved 11 independent cohort studies on occupationally exposed workers. About half of them showed increased standardized mortality or incidence ratios (SMRs or SIRs) for melanoma and none for NHL. The pooled SMRs were 1.32 (95% CI: 1.05-1.64) for melanoma and 0.94 (0.73-1.23) for NHL. Among population-based cohort and case-control studies with individual measures of PCB exposure, one only study was carried out on PCB exposure and melanoma, showing an odds ratio (OR) of 6.0 (2.0-18.2) for the highest compared to lowest quartile of PCB distribution. 13 cohort and case-control studies evaluated the association between NHL and PCB concentration in blood or subcutaneous fat, with summary OR = 1.5 (1.1-1.7) for the highest vs lowest quantile of PCB distribution. However, two cohort studies on people intoxicated by rice oil containing PCBs found no excess of deaths for skin cancer and inconsistent results for NHL. In conclusion, these findings do not provide a strong evidence that PCB exposure can increase the risk of melanoma and NHL in humans.