Endoscopic Findings in Renal Patients: Diagnostic Yield By Indication for Gastroscopy and Colonoscopy

Abstract

Introduction: Endoscopy in patients with renal disease has associated morbidity and can heavily add to the burden on endoscopy services. Aims: To determine the diagnostic yield by indication for gastroscopy and colonoscopy in a cohort of patients with end stage renal disease receiving dialysis (ESRD) and renal transplant recipients (RT). Method: A retrospective analysis of patients under the care of the largest renal unit in Europe, who underwent gastroscopy and/or colonoscopy during a 36 month period. Results: 237 patients underwent 263 gastroscopies (median age 59.9 years, M: F ratio 1.73: 1, ESRD: RT ratio 2.2: 1). Commonest indications were anaemia (38.4%), dyspepsia (18.2%), melaena (14.0%) and haematemesis (8.0%). Abnormal findings were reported in 72.2% (190/263) procedures and clinically significant findings (defined as findings resulting in a change in clinical management) in 20.9% of cases (severe oesophagitis 13/263, duodenal ulcer 12/263, gastric ulcer 11/263, oesophageal ulcer 5/263, upper GI malignancy 4/263, varices 4/263, GAVE 3/263, angiodysplasia 3/263). Only haematemesis yielded a significant diagnostic yield (11/21, 52.3%, OR 4.6, 95% CI 1.83-11.51) versus other indications. 106 patients underwent 110 colonoscopies (median age: 62 years, M: F ratio 2.2: 1, ESRD: RT ratio 1.89: 1). Commonest indications were anaemia (43.6%), diarrhoea (27.2%), and pr bleeding (23.6%). Overall caecal intubation rate was 90.0%. Abnormal findings were reported in 62.7% (69/110) procedures and clinically significant findings in 32.7% of cases (Active colitis 17/110, Polyps >10mm 10/110, angiodysplasia 5/110, lower GI malignancy 3/110, external compression 1/110). PR bleeding was associated with a higher yield of significant lesions (13/26, 50%, OR 2.56, 95% CI 1.04-6.27) than anaemia (17/48, 35.4%, OR 1.14, 95% CI 0.53-2.45) or diarrhoea (6/30, 20.0%, OR 0.39, 95% CI 0.14-1.06). Significant right sided colonic pathology was less prevalent in diarrhoea (3.3%) than anaemia (16.6%) or pr bleeding (15.3%). Colonoscopy had a significantly higher yield for significant pathology (33.3%) than gastroscopy (13.8%) in anaemia (p=0.007, matched for age and Hb). No 30 day mortality was reported following colonoscopy and 0.7% (2/263) following gastroscopy. Conclusions: Haematemesis and pr bleeding are significant indications with high diagnostic yields at endoscopy in this patient group. Flexible sigmoidoscopy rather than colonoscopy may be a more appropriate investigation for diarrhoea in these patients. The findings allow for rationalisation of endoscopy resources in this select group of patients.