• 
• 

Three Part Question

[In adult patient presenting with acute GI bleed], does [time to endoscopy] affect [outcomes such as mortality/re-bleeding/length of stay/need for blood transfusion?]

Clinical Scenario

A 56-year-old man who presented to the ED with acute upper GI bleed was enlisted for urgent endoscopy after assessment by a senior registrar. You knew that patient would have better outcome if they receive early endoscopy but wonder if there was any difference in outcome between 12 hours and 24 hours.

Search Strategy

Medline via Ovid Interface 1946 to July Week 1 2016

Embase via Ovid Interface 1974 to 2016 Week 28
{exp gastrointestinal hemorrhage OR upper gastrointestinal haemorrhage.mp. OR upper gastrointestinal bleed$.mp. OR upper gastrointestinal hemorrhage.mp. OR gastrointestinal hemorrhage.mp. OR exp hematemesis OR hematemesis.mp. OR haematemesis.mp. OR exp melena OR melena.mp. OR malaena.mp.} AND {exp endoscopy OR exp endoscopy, digestive system OR exp endoscopy, gastrointestinal OR endoscopy.mp.} AND {time.mp. OR tim$.mp.} limit to humans, English language, last 10 years

{exp upper gastrointestinal bleeding OR upper gastrointestinal haemorrhage.mp. OR upper gastrointestinal hemorrhage.mp. OR exp gastrointestinal hemorrhage OR gastrointestinal haemorrhage OR upper gastrointestinal bleed$.mp. OR exp melena OR malaena.mp. OR exp hematemesis OR haematemesis.mp.} AND {exp endoscopy OR exp gastrointestinal endoscopy OR exp digestive tract endoscopy OR endoscopy.mp.} AND {exp time OR tim$.mp. OR time.mp.} limit to human, English language and last 10 years

Search Outcome

3307 papers were identified from the search result. Only 6 papers which consisted of 1 review paper, 1 RCT, 2 retrospective analyses, 1 cohort study and 1 comparative study were used to answer the 3-part question

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Kumar N.L. et al 2016 USA	Review paper of 9 selected papers which included 3 RCTs, 1 prospective cohort trial and 5 retrospective cohort trials	Review article	Mortality	Only 1 out of the 9 studies showed statistically significant difference between early (< 13h) and standard (<24h) for high risk NVUGIB	Not a systematic review. Studies included differed significantly in their design Different definition of “early”
			Rebleeding	Of the 5 studies, none reported any statistically significant difference
			Surgery	Only 1 out of the 8 studies reported statistically significant difference
			Length of stay	4 out of 8 studies reported statistically significant difference, predominantly in those where early is defined as within 24 hours
Chen et al 2012 Taiwan	101 patients with active oesophageal variceal bleeding were divided into haematemesis vs non-haematemesis (73 vs 28). For patients with haematemesis, 37 received early endoscopy and 36 received delayed endoscopy, whereas for non-haematemesis, 12 received early endoscopy and 11 received delayed endoscopy	Cohort study	6-week mortality	(In haematemesis group, early vs delayed) 27% vs 52.8%, P=0.031. Statistically significant (In non-haematemesis group, early vs delayed) 11.8% vs 9.1%, P=0.861. No statistical significance	Small sample size.
			6-week rebleeding	(In haematemesis group, early vs delayed) 18.9% vs 38.9%, P=0.028. Statistically significant (in non-haematemesis group, early vs delayed) 17.6% vs 18.2%, P=0.994. No statistical significance. No significant difference in haematemesis group who received early endoscopy and non-hametemsis who received either early or delayed endoscopy
Jairath V et al 2012 UK	4478 patients admitted with acute NVUGIB. 834 received endoscopy in <12 hours, 1190 receive endoscopy between 12-24 hours and 2158 received endoscopy >24 hours	Prospective comparative study	Mortality	Adjusted for confounding factors (<24 hours) OR 0.99, 95% CI 0.97-1.02) (>24 hours) OR 0.98, 95% CI 0.88-1.09 P=0.7, not statistically significant	296 patients with missing “time to endoscopy” which was imputed by sensitive analysis
			Re-bleeding	Adjusted for confounding factors (<24 hours) OR 0.99, 95% CI 0.96-1.01; (>24 hours) OR 0.95, 95% CI 0.87-1.04 P=0.27, not statistically significant
			Need for surgical or radiological intervention	Adjusted for confounding factors (<24 hours) OR 0.99, 95% CI 0.94-1.03; (>24 hours) OR 0.95, 95% CI 0.80-1.11 P=0.5, not statistically significant
			Red cell transfusion	Adjusted for confounding factors (<24 hours) OR 0.81, 95% CI 0.74-0.89; (>24 hours) OR 0.66, 95% CI 0.55-0.80 P<0.001, statistically significant
			Length of hospital stay (difference in mean days)	Adjusted for confounding factors (<24 hours) OR 0.44, 95% CI 0.36-0.52; (>24 hours) OR 1.69, 95% CI 1.39-1.99 P<0.001, statistically significant
Liu et al 2012 China	875 patients with peptic ulcer bleeding were analysed after randomization to receive either standard regime (n=456) or intensive regime (n=419) of PPI. They were further subdivided based on early (n=365) or late (n=510) endoscopy Standard, early vs late (182 vs 274) Intensive, early vs late (183 vs 236)	RCT	Re-bleeding	Early vs late, 7.1% vs 10.0%, P=0.14 Not statistically significant even when compared between subgroups receiving either standard or intensive PPI regime	PPI formulation used varied throughout duration of study Not all patients underwent H.pylori testing No standardization for methods of endoscopy haemostasis Only tested on a very specific group of patients as relatively high risk patients were excluded)
			No. of units of blood transfused	Early vs late, 0.90 ± 1.76 vs 2.26 ± 2.48, P<0.001 Statistically significant even when compared between subgroups receiving either standard or intensive PPI regime
			Hospital stay (days)	Early vs late, 6.8 ± 4.6 vs 8.8 ± 5.8, P<0.001 Statistically significant even when compared between subgroups receiving either standard or intensive PPI regime
Sarin et al 2009 Canada	502 patients with suspected UGIB receiving endoscopy within 6h, 6-24h and >24h. 375 patients identified as acute nonvariceal UGIB which was studied for primary outcome	Retrospective analysis	Mortality/ Need for surgery	6h vs >24h (OR 3.6, 95% CI 1.4-9.4, P=0.008) 6-24 h vs >24h (OR 2.8, 95% CI 1.3-6.2, P=0.01)	Unclear management of patient prior to endoscopy i.e how haemodynamically unstable patients were managed as it may affect patient outcome. Outcome result from multivariate analysis only reported patients with acute nonvariceal bleed
			Need for transfusion	Time to endoscopy not significantly associated with need for transfusion
Cheung et al 2009 Canada	210 patients with stable acute variceal bleed (AVB)	Retrospective analysis	Mortality	OR, 1.0; 95% CI, 0.92–1.08; P=0.91 No statistically significant difference across different time to endoscopy (<4h, <8h and <12h).
			Haemostasis	<4h vs <8h vs <12 h (P=0.93 vs P=0.75 vs P=0.67) No statistically significant difference across different time
			Blood transfusion (units)	<4h vs <8h vs <12 h (P=0.08 vs P=0.14 vs P=0.93) No statistically significant difference across different time
			Rebleeding	<4h vs <8h vs <12 h (P=0.46 vs P=0.58 vs P=0.37) No statistically significant difference across different time
			Day 3 creatinine	<4h vs <8h vs <12 h (P=0.69 vs P=0.22 vs P=0.12) No statistically significant difference across different time
			Length of hospitalization (days)	<4h vs <8h vs <12 h (P=0.87 vs P=0.99 vs P=0.41) No statistically significant difference across different time
			TIPS use	<4h vs <8h vs <12 h (P=0.27 vs P=0.11 vs P=0.86) No statistically significant difference across different time
			Balloon tamponade use	<4h vs <8h vs <12 h (P=0.29 vs P=0.29 vs P=0.13) No statistically significant difference across different time

Comment(s)

Multiple Level 3 studies had been done over the years to study the relationship between time to endoscopy and outcomes such as mortality, risk of re-bleeding, need for blood transfusion, length of stay and etc. Majority of the studies reported no statistical significance in mortality and re-bleeding with early endoscopy when compared to the standard recommended time (24 hours for acute nonvariceal UGIB and 12 hours for acute variceal bleed). However, a few studies had depicted slightly different outcomes which could be associated with patients who were more critically unwell on presentation. There may be other risk factors that had more effect on mortality such as age, infection during admission (for varices), presentation with shock, co-morbidities and etc.

Clinical Bottom Line

Early endoscopy compared to standard (24 hours for acute nonvariceal UGIB and 12 hours for acute variceal bleed) does not reduce mortality or risk of re-bleeding but is associated with longer length of stay and increase need for blood transfusion. It is important to provide sufficient pre-endoscopic supportive treatment to ensure good outcome.

References

1. Kumar N, Travis A, Saltzman J Initial management and timing of endoscopy in nonvariceal upper GI bleeding Gastrointestinal Endoscopy 2016;84(1):10-17
2. . Chen P, Chen W, Hou M, Liu T, Chang C, Liao W et al Delayed endoscopy increases re-bleeding and mortality in patients with hematemesis and active esophageal variceal bleeding: A cohort study Journal of Hepatology 2012;57(6):1207-1213
3. Jairath V, Kahan B, Logan R, Hearnshaw S, Doré C, Travis S et al Outcomes following acute nonvariceal upper gastrointestinal bleeding in relation to time to endoscopy: results from a nationwide study Endoscopy. 2012;44(08):723-730.
4. Liu N, Liu L, Zhang H, Gyawali P, Zhang D, Yao L et al Effect of intravenous proton pump inhibitor regimens and timing of endoscopy on clinical outcomes of peptic ulcer bleeding Journal of Gastroenterology and Hepatology . 2012;27(9):1473-1479
5. Sarin N, Monga N, Adams P Time to Endoscopy and Outcomes in Upper Gastrointestinal Bleeding Canadian Journal of Gastroenterology 2009;23(7):489-493.
6. Cheung J, Soo I, Bastiampillai R, Zhu Q, Ma M Urgent vs. Non-Urgent Endoscopy in Stable Acute Variceal Bleeding Am J Gastroenterol 2009;104(5):1125-1129.