To the EditorWe would like to raise some concerns about the methods used to 'adjust' for the effect of subfertility in the metaanalysis of ART and major birth defects in the recent paper by Rimm et al. [1] Having also published a systematic review and meta-analysis of ART and birth defects (BDs) [2], we are concerned both by the methods used and some of the statements made in the Rimm et al. paper.Assumptions underlying the adjustment method:In their recent meta-analysis, Rimm et al.[1] attempt to 'adjust' the study-specific estimates of BD risk for couples conceiving by ART for the proportion they estimate to be due to underlying subfertility, by subtracting the risk for couples who conceived naturally after more than 12 months (based on Zhu et al. [3]). In doing so, they have assumed that the underlying causes of infertility in couples who managed to conceive spontaneously after 12+ months of trying were the same (and of equal severity) as those in couples who required ART to conceive. We think this is unlikely since many couples who proceed to ART treatment, such as those where the female partner has blocked or absent fallopian tubes or is unable to ovulate, or where the male partner has very low or no sperm in his ejaculate, would have little or no chance of conceiving spontaneously.Meta-analysis is a statistical technique used to summarise the results from many different studies identified through systematic review on the premise that we should look at the body of evidence rather than any study in isolation. Rimm et al. have adjusted the results of the many ART studies they identified (with OR estimates ranging from 0.54 to 15.4) based on an estimate of subfertility derived from a single study, ignoring any of the uncertainty around the single estimate they use. The fact that there are limited data on subfertility and BD risk does not make this valid or appropriate.Inconsistencies in data presented: There appear to be errors in some of the adjustments made to the ORs in this meta-analysis.Although the OR estimates for individual studies are not listed in the original , the ORs presented in the forest plot increase in size from left to right. When these ORs are compared with those in the more recent paper, it is evident that the OR for Place et al. [5] is lower than the one included in the first meta-analysis. Using information in the Place et al. paper, we estimate the OR for major BDs among IVF and ICSI children combined to be 1.36, compared with the OR of 1.24 used in the second Rimm meta-analysis. It is not clear how this was calculated. Moreover, many of the adjusted ORs in Table 1 of the second meta-analysis appear to have been incorrectly calculated (eg MRC, Isaksson, Place, Sutcliffe 1995) so that the hierarchical position of the studies on the second forest plot has changed even further from the original. Some studies also appear to have been incorrectly labelled on the forest plot (eg the OR for Ludwig appears to have been labelled Zadori and vice-versa).Capsule Clinicians who counsel prospective...