Appropriateness of the indication for colonoscopy: systematic review and meta-analysis

Application of appropriate indications for colonoscopy (OC) should conserve limited endoscopic resources. To perform a systematic review and meta-analysis to assess the accuracy of ASGE and EPAGE guidelines in selecting patients referred for OC, relative to the detection of neoplastic and non-neoplastic relevant endoscopic findings.

Studies comparing the appropriateness of OC indication according to ASGE or EPAGE guidelines and the detection of cancer, adenomas, and benign relevant endoscopic findings were identified by searching MEDLINE (1982 - June 2009). Predefined outputs of the meta-analysis were sensitivity, specificity, positive and negative likelihood ratios (LR+, LR-), and the diagnostic odds ratio (DOR).

We included twelve cohort studies comprising 14,160 patients; 10,056 OC indications were categorized as appropriate, and 3,522 (26%) as inappropriate. For cancer detection, the weighted sensitivity, specificity, LR+, LR- and DOR were 89% (95% CI, 82-93%), 26% (95% CI, 21-31%), 1.2 (95% CI, 1.1-1.3), 0.45 (95% CI, 03-0.7), and 3 (95% CI, 1-5), respectively. For adenomas, the adjusted sensitivity, specificity, LR+, LR- and DOR were 85% (95% CI, 77-91%), 27% (95% CI, 22-32%), 1.14 (95% CI, 1-1.2), 0.6 (95% CI, 0.4-0.9), and 1.9 (95% CI, 1.2, 2.9), being for relevant findings equal to 89% (95% CI, 82-93%), 26% (95% CI, 21-31%), 1.16 (95% CI, 1-1.3), 0.44 (95% CI, 0.25-0.8), and 2.6 (95% CI, 1.2-5.6).

Appropriateness guidelines appeared to have a suboptimal sensitivity and a poor specificity for colorectal cancer, being also characterized by a similar accuracy for the diagnosis of benign relevant endoscopic findings. Better strategies are required to select patients with significant pathology for OC.