METAANALYSES OF BRAIN CANCER AND FARMING

Background A series of meta-analyses of peer-reviewed studies of brain cancer and farming were performed, using 33 studies published between 1981 and 1996. Methods Before the meta-analyses, all studies were rev iewed and evaluated for heterogeneity and publication bins. A random-e ffect model was used to estimate the combined relative risk. Results A meta-analysis including all the studies yielded an estimator of relat ive risk equal to 1.30, with a 95% confidence interval (95%CI) of 1.09 , 1.56. The estimator of relative risk obtained from a meta-analysis r estricted to female farmers was 1.04 (95%CI = 0.84, 1.29). A third met a-analysis restricted to studies of farmers residing in the central Un ited States resulted in an estimator of relative risk equal 1.25 (95%C I = 1.09, 1.44). These findings were not influenced by either a public ation bias or a specific study design. Conclusion The consistent signi ficant positive findings suggests that there is a weak association bet ween brain cancer and farming. Exposures commonly experienced by farme rs including infectious microorganisms and pesticides may contribute T o the increased risk of brain cancer: Am. J. Ind. Med. 34.252-260, 199 8. (C) 1998 Wiley-Liss, Inc.