J. Gallaher1, G. Mulima2, J. Qureshi1, C. Shores1, B. Cairns1, A. Charles1,2 1University Of North Carolina At Chapel Hill,Surgery,Chapel Hill, NC, USA 2Kamuzu Central Hospital,Surgery,Lilongwe, , Malawi
Introduction: Upper gastrointestinal (UGI) bleed is a significant public health problem in sub-Saharan Africa especially in resource-poor environments, where there is a lack of diagnostic adjuncts such as endoscopy. This study sought to characterize UGI bleeding at a tertiary care hospital in sub-Saharan Africa and the role of endoscopy in management.
Methods: A prospective analysis of adult patients (age ≥ 18 years) presenting with a clinical diagnosis of UGI bleed to Kamuzu Central Hospital (KCH) in Lilongwe, Malawi over two years was performed (October 2011 – September 2013). Patient characteristics and both short and long-term outcomes were recorded. Long-term outcomes were recorded in an outpatient clinic or via communication with family. Bivariate and logistic regression analyses were used to compare endoscopy and non-endoscopy patient cohorts.
Results: 293 patients were included in the study. Mean age was 41.8 years (SD ± 15.8) with an overall male preponderance (62.9%). 38.9% (n=114) received endoscopy. There were no differences between the endoscopy and non-endoscopy cohorts in clinical history, physical exam, vital signs, laboratory studies, or imaging findings. However, patients who received endoscopy received more blood transfusions (mean 1.9 vs. 1.5 units, p=0.0108) and were more often medicated with beta-blockers (71.1 vs. 55.3%, p=0.007). In the endoscopy cohort, 64.9% (n=74) had findings of esophageal or gastric varices and 43.0% (n=49) of these had endoscopic banding. Length of stay was longer for patients who received endoscopy (14.9 vs. 8.7 days, p<0.001) but mortality was substantially lower in the endoscopy cohort (4.4 vs. 12.9%, p=0.016). The adjusted odds ratio for mortality for patients not receiving endoscopy was 3.53 (CI 1.25-9.99, p=0.017). Outpatient follow-up rates were similar between the two cohorts (31.3 vs. 29.0%, p=0.671). At follow-up, there were similar rates of repeat upper gastrointestinal bleed (5.6 vs. 6.1%, p=0.843) and post-hospitalization mortality (5.0 vs. 6.1%, p=0.683) between the endoscopy and the non-endoscopy cohorts.
Conclusion: Diagnostic endoscopy with or without therapeutic intervention had a significant in-hospital mortality benefit for patients presenting with upper gastrointestinal bleed even with a relatively low utilization rate. Varices were the most common cause and these patients responded well to banding. Prioritizing the improvement of endoscopy capacity in resource-poor environments would likely have a significant impact on mortality.