Colonoscopy Timing Dilemma in Lower Gastrointestinal Bleeding: A Nationwide Inpatient Sample Study

Introduction: Lower gastrointestinal bleeding (LGIB) remains a common reason for hospitalization in the United States (∼25/100,000 in 2021). Although bleeding stops spontaneously in most cases, patients presenting with severe hematochezia and acute hemoglobin drop require colonoscopy to identify the source of bleeding and achieve hemostasis. We aim to observe if the timing of colonoscopy affects the outcomes of LGIB. Methods: Utilizing the Nationwide Inpatient Sample database 2016-2018, we included patients with a primary diagnosis of LGIB who underwent at least one colonoscopy. Propensity scores were calculated using patient demographic factors, Elixhauser comorbidity index, bleeding severity indicators, and hospital characteristics. The Average Treatment Effect on the Treated weight (ATT) was calculated based on the propensity scores and used to adjust for confounders. Results: Out of 329,310 patients admitted for LGIB, 129,895 underwent early colonoscopy within 48 hours and 199,415 after 48 hours. The baseline characteristics and bleeding severity indicators of both groups are listed in Table 1. The adjusted mean difference of length of stay in early colonoscopy decreased by 1.66 days (95% CI 1.61-1.72, P < 0.001), and total charges by 10,099.53$ (95% CI 9,198.12-11,000.94, P < 0.001).[Figure1 A, B] The early colonoscopy group had a lower odds ratio of having more than one colonoscopy 0.68 (95% CI 0.64-0.71, P < 0.001), getting discharged to rehab 0.75 (95% CI 0.71-0.79, P < 0.001), and discharged a visiting nurse 07.0 (95% CI 0.67-0.73, P < 0.001). [Figure 1C, E, F] The adjusted odds ratio of death was higher in the early group by 1.67 (95% CI 1.33-2.11, P < 0.001), and the crude mortality was 755 (0.6%) compared to the late group 1015 (0.5%). [Figure 1D]. Conclusion: Patients who underwent early colonoscopy < 48 were associated with lower healthcare resources, including their hospital stay, charges, number of procedures, and disposition. Interestingly, the early colonoscopy group was linked to higher mortality; however, the unadjusted overall death between the two groups was unremarkable, and the overall mortality was low.Figure 1.: Length of stay, procedure, and discharge. Table 1. - Baseline Characteristics and Bleeding Severity Characteristics Early Colonoscopy Late Colonoscopy P-value Total number (Weighted) 129895 199415 Age (mean (SD)) 96.7 (14.95) 70.79 (14.37) < 0.001 Female (%) 61,165 (47.1) 102,075 (51.2) < 0.001 Race (%) < 0.001 1. White 83510 (66.1) 83510 (66.1) 2. Black 21415 (16.9) 38650 (19.8) 3. Hispanic 12840 (10.2) 19180 (9.8) 4. Others 8620 (6.8) 9795 (5.0) Expected primary payer (%) < 0.001 1. Medicare 88070 (67.9) 143440 (72.0) 2. Medicaid 9890 (7.6) 16050 (8.1) 3. Private insurance 25420 (19.6) 30400 (15.3) 4. Self-pay/ Others/No charge 6375 (4.9) 9320 (4.7) Weighted Elixhauser’s score (mean (SD)) 6.21 (9.88) 8.62 (10.92) < 0.001 Anticoagulation use 20965 (16.1) 42415.0 (21.3) < 0.001 Aspirin use 28810 (22.2) 41835 (21.0) 0.001 NSAID use 9425 (7.3) 15070 (7.6) 0.158 Weekend admissions (%) 25395 (19.6) 53340 (26.7) < 0.001 Location/teaching status of hospital < 0.001 1. Rural 10895 (8.4) 15475 (7.8) 2. Urban nonteaching 36845 (28.4) 53830 (27.0) 3. Urban teaching 82155 (63.2) 130110 (65.2) Bleeding Severity Indicators Shock 4370 (3.4) 5430 (2.7) < 0.001 Acute kidney injury 16095 (12.4) 37335 (18.7) < 0.001 Acute respiratory failure 2575 (2.0) 6250 (3.1) < 0.001 Blood transfusion prior to colonoscopy 32360 (24.9) 69745 (35.0) < 0.001 FFP and coagulation factors transfusion prior to colonoscopy 4280 (3.3) 8010 (4.0) < 0.001