Reply

We appreciate the interest of Nahon and colleagues in our analysis of the “weekend effect” for upper gastrointestinal bleeding (UGIB) from the Nationwide Inpatient Sample (NIS).1 In their multi-center cohort from 53 general non-university hospitals in France,2 they found similar mortality for weekend (3.8%) and weekday (6.3%, P = .17) admissions for UGIB-related peptic ulcers. Interestingly, they found a shorter mean time to endoscopy for their weekend compared to weekday admissions. There are several possible reasons for the differences in results between the studies. First, it could represent patient tendencies for earlier or delayed presentation. Second, it could be due to institutional, or even national, differences in management algorithms and supportive care for UGIB; it would be interesting to know if the 53 hospitals had similar or standard UGIB management algorithms. Third, it could be due to the shorter endoscopy time for weekend admissions and the greater rate of endoscopic intervention noted in their study; in our analysis, we found that the excess mortality was mostly in the group that did not undergo endoscopic intervention. Interestingly, the overall mortality, need for surgery, and length of stay was much higher in their study, suggesting a more severely ill cohort. Several additional subgroup analyses would be of interest from the study by Nahon et al;2 specifically if there was a “weekend effect” in mortality in any subgroup, such as those with active bleeding, who received endoscopic intervention, or after stratification by the Rockall score. In another study from our institution, we identified that hospitals with a high volume of non-variceal UGIB had lower mortality, suggesting institution-specific differences in outcomes.1 The “weekend effect,” while significant, still offers only a modest contribution to mortality compared to other patient factors. Nevertheless, it represents a modifiable risk factor for adverse outcomes in UGIB. Future studies from different settings with more detailed information on patient-level, provider-level, and systems-level variables are essential to identify actionable factors in order to improve patient outcomes.