Abstract / Summary
Type 2 diabetes mellitus (T2DM) is an independent risk factor for inadequate bowel preparation. Sodium–glucose cotransporter-2 (SGLT-2) inhibitors promote osmotic diuresis and natriuresis and might therefore impair preparation quality during the fluid-restricted, fasting state of bowel cleansing; conversely, their prokinetic potential could leave cleansing unaffected. We evaluated whether SGLT-2 inhibitor use is associated with bowel preparation quality, assessed by the Boston Bowel Preparation Scale (BBPS), in patients with T2DM undergoing colonoscopy. In this single-center, prospective comparative observational study, 150 adults with T2DM undergoing elective colonoscopy were enrolled and divided into two equal groups according to SGLT-2 inhibitor use (users, n = 75; non-users, n = 75). All patients followed an identical standardized preparation comprising a low-residue then clear-liquid diet, a split-dose sennoside A + B regimen, and daily sodium-phosphate enemas. SGLT-2 inhibitors were continued without interruption throughout the preparation period. Bowel preparation was scored with the BBPS. Peri-procedural serum electrolytes, glucose, urea, creatinine, and calculated osmolarity were compared before and after the procedure. Secondary analyses examined diabetes duration, concomitant motility-affecting drugs, and the specific SGLT-2 inhibitor agent. The cohort was 56.7% male with a mean age of 63.17 ± 9.04 years; groups were comparable in age, weight, BMI, and diabetes duration. Mean total BBPS did not differ significantly between SGLT-2 inhibitor users and non-users (4.88 ± 2.58 vs. 5.39 ± 2.38; mean difference − 0.51, 95% CI − 1.31 to 0.29; p = 0.153), a result unchanged after multivariable adjustment (adjusted difference − 0.40, 95% CI − 1.23 to 0.43), and the categorical distribution of preparation quality was also similar (p = 0.219). Diabetes-duration subgroups, concomitant beta-blocker, calcium-channel-blocker and antidepressant use, and the specific agent (empagliflozin vs. dapagliflozin) showed no association with BBPS (all p > 0.05). Peri-procedural sodium, potassium, urea, and osmolarity were comparable between groups (all p > 0.05); glucose was higher in users at both timepoints, without a significant within-group temporal change. The pre-to-post rise in creatinine was smaller in SGLT-2 inhibitor users (group×time interaction p = 0.047). No detectable association was found between SGLT-2 inhibitor therapy and bowel preparation quality, and the confidence intervals exclude a large detrimental effect, although smaller differences cannot be ruled out; peri-procedural electrolyte and osmolar changes were likewise comparable between groups. With a well-applied standardized protocol, adequate cleansing is achievable irrespective of SGLT-2 inhibitor use, supporting an individualized, risk-stratified rather than routine-discontinuation approach.