The yield of colonoscopy in patients attending a tertiary hospital in Cape Town, South Africa

Abstract

Background: Colonoscopy is widely used for diagnosis, therapeutic intervention of conditions of the lower gastrointestinal tract and colorectal cancer screening. Given its widespread use and complications, colonoscopy should be performed for appropriate indications. The aim of this study was to determine the yield and appropriateness of colonoscopy at a tertiary hospital. Methods: Demographic, clinical and endoscopic data was retrieved from a prospective endoscopy database from 01 Jan 2014 to 31 December 2019. The variables collected were patient details (age, sex), clinical indication, symptoms, urgency of colonoscopy, sedation used, quality of bowel preparation, and findings. Results: The total number of patients was 4033 with median age of 56 (IQR 18-85), of whom 59.6% were female. Colonoscopy was performed for appropriate indications in 98% of cases (according to EPAGE II criteria). Older patients were more likely to be investigated for anaemia (OR=0.55, CI 0.43-0.70, p=0.000), and surveillance postpolypectomy (OR=0.57, CI 0.37-0.85, p=0.005), while patients under 50 had a colonoscopy for an IBD flare (OR=1.98, CI 1.49-2.6, p=0.000), or surveillance for HNPCC (OR=3.0, CI 2.1-4.3, p=0.000). Rectal bleeding was the commonest symptom (16.5%). Patients younger than 50 were more likely to present with abdominal pain (OR=1.3, CI 1.07-1.6, p=0.006), and diarrhoea (OR=1.8, CI 1.5-2.3, p=0.000), whilst those older than 50 were more likely to present with alteration of bowel habit (OR=0.63, CI 0.43-0.9, p=0.012), and loss of weight (OR=0.45, CI 0.33-0.60, p=0.000). Polyps were the commonest finding (15.4%); the prevalence of colorectal cancer and diverticular disease was 3.6% and 6.5% respectively. Patients younger than 50 were more likely to have inflammation (OR=2.6, CI 2.18-3.15, p=0.000), whilst those older than 50 were more likely to have diverticulae (OR=0.08, CI 0.05-0.14, p=0.000), polyps (OR=0.43, CI 0.35- 0.53, p=0.000) or a tumour (OR=0.47, CI 0.29-0.72, p=0.000). Although 69% of the bowel preparation was reported as good or adequate, there was no difference in the detection of lesions when compared to poor bowel preparation. The adjusted caecal intubation rate was 95.1%. Conclusion: Colonoscopy was performed in a relatively young cohort, predominantly for symptoms not screening. The study provides some epidemiologic data on common lower gastrointestinal conditions in a referral center, but also highlights some inefficiencies in the system. Furthermore, this study serves as an audit on the clinical service, a baseline from which improvements can be made.