The accompanying article by Bor-Seng-Shu et al.1 attempts to address a rather straightforward question: Does decompressive craniectomy reduce intracranial pressure (ICP) and increase cerebral perfusion pressure (CPP) in the setting of traumatic brain injury? While conventional wisdom would hold this as a foregone conclusion, studies that have included this information have not been entirely conclusive. Moreover, most studies on the topic have been rather small and have not had sufficient power to arrive at a conclusion. The authors of this pooled analysis attempt to address the issue of statistical power by pooling available studies. The outcomes examined by Bor-Seng-Shu are rather straightforward, but the method does illustrate some of the limitations of doing a meta-analysis of widely heterogeneous studies. As the authors themselves indicate, there is wide variability in almost every aspect of the studies utilized in the meta-analysis—from patient inclusion criteria, to ICP measurement techniques, to medical treatment intensity, to surgical timing, to surgical technique. In fact, in a true meta-analysis, such important sources of variability are typically incorporated in the inclusion/ exclusion criteria. While the authors of this meta-analysis excluded some analyses (such as case reports, reviews, nonhuman studies, non-English studies, and so forth), very important sources of variability remain in the included studies. We should always exercise caution when interpreting the results of this type of pooled analysis. The assumption that pooling patient data from multiple studies will result in a valid pooled comparison is not always a reliable one. Disparity in patient numbers between studies with confounding variables will not infrequently lead to a contradictory conclusion (Yule-Simpson effect). Given that there is wide variation in patient accrual in the present pooled analysis (4–100 patients in the included studies) as well as a number of likely confounders, the possibility of a Simpson paradox should not be discounted. Because of the study heterogeneity in this pooled analysis, it is difficult to have much confidence in the degree of ICP reduction or CPP elevation afforded by decompressive craniectomy. However, the conclusion of the analysis—that cranial decompression reduces ICP and increases CPP—is believable. It remains to be seen whether the degree of ICP reduction and its durability hold up as more studies on the topic are published. (http://thejns.org/doi/abs/10.3171/2012.1.JNS1283)