Reply to Corbett & Cameron

We would like to thank Corbett & Cameron for their letter. Our study [ 1 ] was an observational study with prospectively collected data. It showed that urgent endoscopy was associated with lower mortality in high-risk but not low-risk nonvariceal upper gastrointestinal bleeding. Our findings are consistent with the recommendations of an international consensus [ 2 ] which states that there is no additional benefit conferred by urgent endoscopy (< 12 hours) compared with early endoscopy (> 12 hours) in unselected nonvariceal upper gastrointestinal bleeding. In the majority of our cohort, urgent endoscopy was not associated with decreased mortality. Endoscopy within 13 hours was associated with lower mortality only in a very small subset of patients with severe nonvariceal upper gastrointestinal bleeding, as triaged using the Glasgow-Blatchford score. In our study, multivariate analysis showed a statistically significant effect of presentation-to-endoscopy time on mortality ( P = 0.012) in high-risk patients. In this subgroup, the odds ratio of 1.092 for the presentation to endoscopy time was a per-unit change. This meant that for each hour of delay, the risk of mortality increased by 9.2 %. For example, if the delay was 10 hours, then the corresponding odds ratio would be 10.92. Multivariate analysis had corrected for the possible confounding caused by co-morbidities leading to delay in endoscopy in the high-risk group. Corbett & Cameron rightly suggest that the results of this study should not encourage a practice of very early endoscopy at the expense of adequate resuscitation and optimization of co-morbidity. In fact, we emphasized this in the discussion section of our paper, and we are glad that Corbett & Cameron reinforce our point that proper resuscitation is the mainstay in the management of gastrointestinal bleeding. Finally, risk stratification using the Glasgow-Blatchford score to evaluate patients presenting with nonvariceal upper gastrointestinal bleeding can potentially help ensure appropriate service delivery without unduly straining limited resources, as the majority will not require urgent endoscopy.