Diabetes mellitus and colorectal carcinoma outcomes: a meta-analysis.

Abstract

The impact of diabetes mellitus (DM) on colorectal cancer (CRC) outcomes remains unknown. We studied this by conducting a meta-analysis to evaluate (1) CRC outcomes with and without DM and (2) treatment patterns. We searched PubMed, EMBASE, Google Scholar, and CINAHL for full-text English studies from 1970 to 12/31/2017. We searched keywords, subject headings, and MESH terms to locate studies of CRC outcomes/treatment and DM. Studies were evaluated by two oncologists. Of 14,332, 48 met inclusion criteria. In accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses method, we extracted study location, design, DM definition, covariates, comparison groups, outcomes, and relative risks and/or hazard ratios. We utilized a random-effects model to pool adjusted risk estimates. Primary outcomes were all-cause mortality (ACM), disease-free survival (DFS), relapse-free survival (RFS), and cancer-specific survival (CSS). The secondary outcome was treatment patterns. Forty-eight studies were included, 42 in the meta-analysis, and 6 in the descriptive analysis, totaling > 240,000 patients. ACM was 21% worse (OR 1.21, 95% CI 1.15-1.28) and DFS was 75% worse (OR 1.75, 95% CI: 1.33-2.31) in patients with DM. No differences were detected in CSS (OR 1.10, 95% CI 0.98-1.23) or RFS (OR 1.12, 95% CI 0.91-1.38). Descriptive analysis of treatment patterns in CRC and DM suggested potentially less adjuvant therapy use in cases with DM and CRC. Our meta-analysis suggests that patients with CRC and DM have worse ACM and DFS than patients without DM, suggesting that non-cancer causes of death in may account for worse outcomes.