Endoscopy 2011; 43(10): 925
DOI: 10.1055/s-0030-1256571
Letters to the editor

© Georg Thieme Verlag KG Stuttgart · New York

Timing of endoscopy in high-risk patients with nonvariceal upper gastrointestinal bleeding

G.  D.  Corbett, E.  A.  B.  Cameron
Further Information

Publication History

Publication Date:
07 October 2011 (online)

We read with interest the study by Lim et al. [1], which suggests that earlier endoscopy is associated with better outcomes for high-risk patients with nonvariceal upper gastrointestinal bleeding (NVUGIB). NVUGIB has a significant mortality. Whilst endoscopic therapy has been shown to significantly improve this, there is also good evidence that most of the mortality in patients presenting with NVUGIB is not directly related to bleeding [2] and little evidence to suggest that very early endoscopy affects outcomes although consensus suggests that it should be performed within 24 hours [3].

This study concludes that performing endoscopy within 13 hours for patients with a high pre-endoscopy Glasgow–Blatchford score is associated with reduced mortality. It is important to interpret these results in the context of the weaknesses of the study, in particular that it is a retrospective observational study with a very small sample size of high-risk patients. The high-risk group had only 97 patients and 11 of these did not receive endoscopy within 24 hours (for reasons including severe co-morbidities requiring optimization and delay in next-of-kin consent). Of the 86 remaining, 47 underwent endoscopy within 13 hours and 39 between 13 and 24 hours. The multivariate analysis did show a statistically significant effect of presentation-to-endoscopy time on mortality (P = 0.012) but the size of this effect is very small (odds ratio 1.092). On the other hand, one or more significant co-morbidities was not statistically significant (P = 0.12) but the odds ratio of 10.087 is considerably more sizeable and the confidence intervals are exceptionally wide (0.547 – 185.988), suggesting a type I error as a result of sample size. It is most likely that the confounder of co-morbidities leading to delay in endoscopy in the high-risk group, rather than early endoscopy itself, has produced the findings in this study.

We do agree that designing an appropriate randomized controlled study to determine the optimal timing of endoscopy is a major challenge. However, it is important that the results of this study do not encourage a practice of very early endoscopy at the expense of adequate resuscitation and optimization of co-morbidity.


G. D. Corbett, MD 

Addenbrooke's Hospital – Gastroenterology
Box 293

Hills Road
Cambridge CB2 0QQ
United Kingdom

Fax: +44-1223-348729

Email: [email protected]