This review concluded that C-reactive protein measurement four days postoperatively had a high negative predictive value for infectious complications and allowed safe and early discharge of selected patients after colorectal surgery. Methodological limitations of the included studies mean that test accuracy was likely to be overestimated and the strong conclusions and implications for practice should be treated with caution.

To assess the predictive value of C-reactive protein level for postoperative infectious complications after colorectal surgery.

PubMed was searched without language restrictions to March 2011; the search strategy was reported. Reference lists of retrieved studies were searched.

Studies that assessed the diagnostic accuracy of C-reactive protein for identifying infectious complications at specified timepoints after colorectal surgery were eligible for inclusion. Studies that pooled C-reactive protein data from several postoperative days were excluded. Across the included studies, the prevalence of postoperative infections ranged from 11.2% to 39.1%, laparoscopy was performed in up to 31.3% and where reported 28% to 39% had advanced cancer.

Two reviewers independently selected studies for the review.

Two independent reviewers assessed study quality using an adapted version of the QUADAS tool. Any discrepancies were resolved by discussion or by a third reviewer; the requirement of an adequate description of the index tests and reference standard were omitted from the assessment.

Two independent reviewers extracted data to construct 2x2 tables of test performance; any discrepancies were resolved by discussion. Sensitivity, specificity and diagnostic odds ratios were calculated with 95% confidence intervals (CI). Authors were contacted for missing data.

Analyses were conducted for each postoperative day. Summary estimates of sensitivity, specificity, diagnostic odds ratio and prevalence were calculated with 95% CI using random-effects models. Heterogeneity was investigated using Cochran's Q. Pooled estimates of the positive and negative predictive values were calculated using the summary estimates of sensitivity, specificity and prevalence; 95% CI were calculated using the Wald method. Summary receiver operating characteristic (SROC) curves were produced using the Moses-Littenberg model and the area under the curve was calculated. The bivariate random-effects model was used to calculate summary estimates of sensitivity and specificity as a sensitivity analysis.

Six studies met the inclusion criteria (1,832 participants, range 35 to 1,187). There were two prospective and four retrospective studies. All studies included a representative patient spectrum, used an acceptable reference standard, avoided partial verification bias, provided relevant clinical information and reported uninterpretable results. Incorporation bias was avoided in two studies, differential verification bias was avoided in two studies, selection bias was avoided in four studies and five studies explained withdrawals. None of the studies blinded interpreters of the tests.

The overall rate of postoperative infectious complications was 23.7% (95% CI 16.2 to 32.1; six studies). The highest area under the curve (0.81), specificity (83%, 95% CI 77 to 90 with corresponding sensitivity 68%, 95% CI 57 to 79) and diagnostic odds ratio (11.7 95% CI 6.1 to 22.3) were achieved on day four (five studies). On day four the positive predictive value was 55.7% (95% CI 50.3 to 61.1) and negative predictive value was 89.3% (95% CI 87.1 to 91.5). The mean cutoff value of C-reactive protein on day four was 135mg/L. The highest sensitivity was achieved on day two (75%, 95% CI 58 to 91 with corresponding specificity 58%, CI 46 to 69; four studies). Considerable heterogeneity was observed for all analyses.

Summary estimates of sensitivity and specificity from the bivariate random effects model were reported as being similar.

C-reactive protein on postoperative day four had a high negative predictive value for infectious complications of 89% and allowed safe and early discharge of selected patients after colorectal surgery.

The review addressed a clear research question with reproducible inclusion criteria. The search was limited and there was no attempt to avoid publication bias. Prognostic and diagnostic filters were used in the search so studies may have been missed. Each stage of the review was conducted in duplicate, which reduced risks of error and bias. Study quality was assessed using appropriate criteria but the results were reported only in summary.

A robust SROC model was used to produce summary estimates of sensitivity and specificity but this was conducted as a sensitivity analysis and the results were not fully reported. Prevalence was appropriately considered during production of the pooled predictive values. The individual day analyses were based on four or five of the six included studies and sample sizes were relatively small. Most of the studies were retrospective, none of them reported blinding of interpreters of the tests and most did not avoid differential verification or incorporation bias; all of these factors were likely to lead to an overestimation of test accuracy.

The strong conclusions and implications for practice should be treated with caution.

Practice: The authors stated that as increased C-reactive protein preceded infection by approximately three days and most infectious complications occurred between days five and eight, a test on day four was the ideal time point. They suggested that patients with a C-reactive protein level of below 135 on day four were very unlikely to have postoperative infectious complications but patients above this cut-off should not be discharged and should be followed with clinical examination, laboratory testing and imaging where necessary to actively confirm or exclude imminent postoperative infectious complications.

Research: The authors did not state any implications for research.

None reported.