Inclusion criteria for the review were broadly defined. One relevant database was searched. Publication bias was assessed and was not detected. There was potential for language bias as only English-language articles were included. Two authors appeared to be involved in selecting studies and performing statistical analysis, which minimised the risk of error and bias in the analysis. Study quality assessment was not reported, which made determining study quality difficult; however, all studies were retrospective cohorts, which can be prone to high levels of bias. Studies were combined using meta-analysis and heterogeneity was assessed, which was appropriate.
There were some methodological problems with the review, but the authors’ conclusions were based on the evidence and appear appropriate. Given the unknown quality of the included studies, the pooled standardised incidence rates should be interpreted as estimates, rather than absolutes.