Maternal smoking during pregnancy and the risk of childhood brain tumors: a meta-analysis of 6566 subjects from twelve epidemiological studies.

Objective: Prior epidemiological studies suggest a possible association between maternal smoking during pregnancy and risk of childhood brain tumors. A meta-analysis was performed statistically pooling all available observational studies on this topic in order to evaluate this suspected association.

Methods: Using previously described methods, a protocol was developed for a meta-analysis examining the association between maternal smoking during pregnancy and subsequent development of primary brain tumors in their offspring. Literature search techniques, study inclusion criteria and statistical procedures were prospectively defined. Data from epidemiological studies were pooled using a general variance-based meta-analytic method employing confidence intervals previously described by Greenland. The outcome of interest was a summary relative risk (RRs) reflecting the risk of childhood brain tumor development associated with mother's smoking during the index pregnancy. Sensitivity analyses were performed when necessary to explain any observed statistical heterogeneity and/or to evaluate the impact of demographic or study characteristics on the summary estimate of effect.

Results: Twelve observational studies meeting protocol specified inclusion criteria were obtained via a comprehensive literature search. These studies enrolled a total of 6566 patients. Analysis for homogeneity demonstrated that the data were homogeneous (P > 0.50) and could be statistically combined. Pooling all twelve reports yielded an RRs of 1.05 (0.90-1.21), a non-statistically significant result suggesting no clear association between maternal smoking during pregnancy and risk of childhood brain tumor development. Numerous sensitivity analyses examining the possible effect of study design and various patient characteristics failed to show any influence on the RRs further supporting the observed lack of association.

Conclusion: The available epidemiological data do not support a clear association between maternal smoking during pregnancy and pediatric brain tumor development. Although it appears likely that no association exists, limitations in study designs limit definitive conclusions based on available data.