Question special
Residents and Fellow

As part of the first secondary outcome, overt GI bleeding was defined as hematemesis, coffee ground emesis, melena, hematochezia or bloody nasogastric aspirate, based on page 7 of the supplementary appendix. Based on Table S8, 88 patients in the pantoprazole group, and 148 patients in the placebo group had overt GI bleeding. Of those 41 patients in the pantoprazole group, and 69 in the placebo group fulfilled further predefined criteria to be considered clinically important gastrointestinal bleeding.

· Was there any effort attempted to further distinguish the cause of hematochezia, from being upper vs lower bleed such as ischemic colitis in the setting of vasopressor use?

· Presumably, such patients with clinically important gastrointestinal bleeding, were part of the patients who had indications for open-label acid suppression. How were these pts managed medically (dose escalation vs increase in the frequency of PPIs) and how did that affect the data collection and results? Were these patients excluded from further data collection?

· Even though the Relative risk for GI bleeding in the pantoprazole vs the placebo group was 0.58 and the 95% confidence interval did not cross 1, the authors noted that there are limited inferences that could be drawn from these finding given the absence of adjustment for multiple comparisons. Could you please elaborate further?