The authors addressed a clear research question, with broad but clear inclusion criteria. Several sources were searched; language restrictions were applied and there was no specific search for unpublished studies so bias could not be ruled out. There was no reporting on duplication in the review process, so potential for error and bias could not be assessed. It was stated that study quality was assessed, but it seemed that a level of evidence was assigned, and only an overall assessment of the evidence available given; there was no detailed assessment of individual studies.
Though the surgical techniques and consequently the risk of infection were considered homogeneous, the types of surgery spanned orthopaedic, gastrointestinal and cardiac procedure (no further detail was provided). The method for producing the summary estimates of sensitivity and specificity were not reported, and may have been simple means. The SROC model used appeared to be the Moses-Littenberg model; more robust models were available that maintained the within-study relationship between sensitivity and specificity. There was no investigation into either clinical or statistical heterogeneity across the studies. Although the authors considered this to have been "an extremely homogeneous review", there appeared to be considerable clinical heterogeneity across the studies, and the pooled results contained studies that conducted CRP test preoperatively and postoperatively. There was also a contradiction between the authors' overall conclusion and recommendations for further research.
Overall, this review has several limitations and it is unlikely that the summary estimates, or the overly strong conclusion, can be deemed reliable.