Background and aims <p>Small bowel bleeding poses a significant diagnostic challenge for clinicians, and capsule endoscopy is currently considered the investigation of choice for evaluating suspected cases. However, evidence regarding the role of second-look endoscopy before capsule endoscopy is limited. This study aimed to evaluate capsule endoscopy findings after a negative second-look endoscopy in patients with suspected small bowel bleeding at a tertiary care center in Sri Lanka.</p> Method <p>This retrospective cross-sectional study analyzed data from patients who underwent capsule endoscopy for suspected small bowel bleeding at the National Hospital of Sri Lanka from June 1, 2018, to March 31, 2023. All patients had negative results on second-look endoscopies and abdominal imaging (CT, MRI, Enterography). The principal investigator reassessed all the capsule recordings at 3–6 frames per second in single-image viewing mode. The study primarily evaluated demographic data, capsule endoscopy findings, and mucosal visualization. Continuous variables were presented as mean ± standard deviation or median. Comparisons for continuous variables were performed using independent-sample t-test and categorical variables were compared using the chi-square test. A two-sided p-value &lt; 0.05 was considered statistically significant.</p> Results <p>The sample size was 71, with a mean age of 54.41 ± 18.44 years. 60.6% of patients were males. Overt bleeding occurred in 52.1% of patients, while occult bleeding occurred in 47.9%. The mean small intestine transit time was 5:16:30 ± 1:48:01&#xa0;h. Hemoglobin was significantly lower (<i>p</i> &lt; 0.001) in patients with overt bleeding, with a mean level of 9.27 ± 1.27&#xa0;g/dL. Capsule endoscopy identified 30 patients (42.3%) with a definite cause for small bowel bleeding and 22 patients (31.0%) with a probable cause, resulting in an overall diagnostic yield of 73.2%. The proximal small intestine was the most commonly affected segment of the small bowel in cases of small bowel bleeding, accounting for 33 patients (46.5%). None of the patients had lesions in the endoscopically accessible sites of the upper or lower gastrointestinal tract. We noted vascular findings in 18 patients (25.4%), with inflammatory findings accounting for the majority (<i>n</i> = 30/42.3%). The predominant treatment option was medical management (71.8%), followed by endoscopic interventions (26.8%) and interventional radiological interventions (2.8%). The number of polyethylene glycol sachets used for bowel preparation did not affect the diagnostic yield or mucosal visualization quality.</p> Conclusion <p>Capsule endoscopy demonstrated a high diagnostic yield in Sri Lankan patients with suspected small bowel bleeding who had normal initial and second-look endoscopies. The proximal small intestine was the commonest involved segment, and, contrary to previous studies, inflammatory lesions were the predominant finding. We recommend low-volume bowel preparation as the number of polyethylene glycol sachets did not affect the diagnostic yield or quality of mucosal visualization. This study supports integrating capsule endoscopy into routine clinical practice for evaluating gastrointestinal bleeding and recommends performing a second-look endoscopy before proceeding to capsule endoscopy.</p>