Patient-Reported Outcomes and Endoscopic Appearance of Ulcerative Colitis: A Systematic Review and Meta-analysis

Abstract

Background & Aims: We aimed to evaluate the association of the patient-reported outcomes for rectal bleeding and stool frequency among patients with ulcerative colitis (UC) in endoscopic remission. Methods: We performed a systematic review of studies reporting the association of patient-reported outcomes (the patient-reported components of the Mayo score: rectal bleeding and stool frequency) and endoscopic remission (Mayo endoscopic subscore of 0 or 1). We performed a meta-analysis of diagnostic accuracy using the hierarchical bivariate method. Results: Our meta-analysis consisted of 5 studies comprising 2132 participants. A rectal bleeding subscore of 0 identified patients in endoscopic remission with a pooled sensitivity value of 81% (95% CI, 73%–86%), a specificity value of 68% (95% CI, 61%–75%), a positive likelihood ratio (LR) of 2.5 (95% CI, 2.2–3.0), and a negative LR of 0.28 (95% CI, 0.22–0.37). A stool frequency subscore of 0 identified patients in endoscopic remission with a pooled sensitivity value of 40% (95% CI, 25%–58%), a specificity value of 93% (95% CI, 86%–97%), a positive LR of 6.0 (95% CI, 3.7–9.7), and negative LR of 0.64 (95% CI, 0.50–0.82). A combined rectal bleeding and stool frequency subscore of 0 identified patients in endoscopic remission with a pooled sensitivity value of 36% (95% CI, 22%–54%), a specificity value of 96% (95% CI, 91%–98%), a positive LR or 8.4 (95% CI, 5.5–12.8), and a negative LR or 0.66 (95% CI, 0.53–0.84). Conclusions: In a meta-analysis of 5 studies, we found that most patients with UC and normal rectal bleeding and stool frequency subscores have attained endoscopic remission. Many patients in endoscopic remission from UC have no rectal bleeding. Normal stool frequency associates with endoscopic remission, but many patients have abnormal stool frequencies despite endoscopic remission. These results should be carefully considered given the high heterogeneity and wide confidence intervals for some outcomes.