• 
• 

Three Part Question

In [patients with acute severe non variceal upper GI bleed] is [the use of vasoconstrictor therapy] indicated [to control bleeding and prevent re-bleeding].

Clinical Scenario

A 65 year old man presents to the ED with a large, fresh upper GI bleed. He has a history of non-steroidal anti-inflammatory drug (NSAID) use and complains of increasing indigestion over the last few months.
On examination, he has no stigmata of chronic liver disease and is unwell with a BP of 80 systolic and tachycardia of 140.
In view of his history and lack of positive examination findings you feel that the most likely diagnosis is a bleeding peptic ulcer.
You wonder if there is any evidence to support the use of vasoconstrictor therapy in non variceal upper GI bleeds.

Search Strategy

MEDLINE (OVID INTERFACE) SEARCHED
1966-2006
{upper gi bleed.mp. OR exp Gastrointestinal Hemorrhage/ or exp Hematemesis/ OR haematemesis.mp. OR hematemesis.mp. OR gastrointestinal adj5 haemorrhage.af. OR gastrointestinal adj5 hemorrhage.af. OR gi adj5 bleed.af. OR peptic ulcer disease.mp. OR exp peptic ulcer/ OR gastric ulcer.af. OR duodenal ulcer.af.} AND {terlipressin.mp. OR vasopressin.mp. OR exp vasopressin/ OR antidiuretic hormone.mp. OR adh.mp. OR exp somatostatin/ OR somatostatin.mp OR somatostatin analogue.mp. OR octreotide.mp. OR exp octreotide/ OR glypressin.mp} Limit to human and English language

Search Outcome

1123 papers found
16 RCTs, 1 lab study, 1 review article and 1 meta-analysis found that were relevant to the three part question

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Avgerinos et al, 2005, Greece	43 patients with malaena/haematemesis with endoscopic signs of stages 2c and 3 Forrest classification. Randomised to SST (15), Pantoprazole (14) or placebo (14) and then had gastric pH monitoring for 24 hrs	Prospective RCT (placebo controlled) -double blind	Mean gastric pH -SST	1.94+/-0.18 to 6.13+/-0.37 p<0.0001	pH study only with no correlation to actual risk of rebleed or need for surgical intervention according to power calculations, study sample not large enough - however study stopped as SST appeared superior to placebo High no of patients excluded from trial (143)
			PAN	1.93+/-0.16 to 5.65+/-0.37 p<0.0001
			Placebo	1.86+/-0.12 to 2.10+/-0.15 p=0.0917
Saruc et al, 2003, Turkey	21 patients with bleeding peptic ulcer - endoscoped within 6 hours. Given SST 250micrograms/hr for 72 hrs after bolus 250.Each patient had superior mesenteric artery velocity (SMA-V), SMA pulsatility index (SMA-PI), portal venous volume flow (PV-F) and renal artery resistance index (RA-RI) measured by doppler on day 1 of infusion infusion and 6 hours post infusion	Observational lab trial -not blind	PV-F - during infusion	33.7+/-12.7 cm3/s	Laboratory trial. Not blind. Small numbers. High no of exclusions including NSAID use ?Correlation between large vessel flow and clinical picture ?Influence of mucosal bleeding
			PV-F - post infusion	56.3+/-16.0 p=0.001
			SMA-V during infusion	39.7+/-13.1
			SMA-V post infusion	64.4+/-15.1p=0.01
			SMA-PI during infusion	2.0+/-0.8
			SMA-PI post infusion	2.8+/-0.8 p=0.02
			Correlation between PV-F and risk of rebleed	r=0.55, p=0.03 (r=0.2 is correlation)
			Correlationbetween SMA-V and risk of rebleed	r=0.22, p-0.22
			No change in RA-RI
Archimandritis et al, 2000, Greece	84 patients over a twelve month period, scoped within the first 24 hours. Randomised to ranitidine and octrotide (50mg iv tds and 100micrograms s/c tds)(40) or ranitidine (50mg ivtds) alone (44).	Prospective RCT- not blind	No of units tx - ranit	1.07+/- 0.24	Not blind Small numbers s/c octreotide and ?wrong dose
			Ranit/octreo	1.7+/-0.37 p= 0.16 (NS)
			Days in hosp - ranit	8.39+/-0.47
			Ranit/octreo	9.20+/-0.53 p=0.25 (NS)
			txd patients -ranit	23/44
			Ranit/octreo	21/40 p=1.0 (NS)
			Emergency op - ranit	3
			Ranit/octreo	3 p=1.0 (NS)
Imperiale et al, 1997, USA	1829 patients from 14 randomised clinical trials comparing SST or octreotide with H2 blockers or placebo in patients with clinical or endoscopic evidence of acute nonvariceal upper GI haemorrhage. 7 trials placebo controlled, 7 used cimetidine, 5 used ranidine. 8 trials blinded	Meta-analysis of 14 trials - database search of English language articles between 1966 and 1996 and the bibliographies of all related articles and textbook chapters- MEDLINE Jan 66 -Oct 96, EMBASE 1980-1996	Continued bleed or rebleed(all trials)	RR 0.53 (95% CI 0.43-0.63)NNT5	Very heterogeneous studies re protocols and doses of drugs,outcome measures,interventions and study groups (some high risk only, some excluded high risk. proportion of patients with active bleeding ranged from 13-100%) ?Publication bias -though calculations show would need 18 trials showing no difference for p>0.05 Are h2 blockers equivalent to placebo? ?What is the ideal treatment duration
			Surgery (all trials)	RR 0.71 (CI 0.61-0.81)NNT8
			Continued bleed or rebleed (SST only)	RR 0.50 (CI 0.39-0.60) NNT5
			Continued bleed or rebleed (octreotide only)	RR 0.72 (CI 0.35-1.08) NS
			Continued bleeding (SST only)	RR 0.41 (CI 0.29-0.53) NNT4
			Surgery (SST only)	RR 0.70 (CI 0.58-0.82) NNT7
			Investigator blinded trials:-
			Ccontinued bleeding or rebleed	RR 0.73 (CI 0.64-0.81) NNT11
			Surgery	RR 0.94 (CI 0.87-1.001)
			For PU bleeding alone, continued/rebleed	RR 0.48 (CI 0.39-0.59) NNT 4
			Non peptic ulcer bleeding, continued/rebleed	RR 0.62 (CI 0.39-1.002)

Comment(s)

In approximately 80% of non-variceal upper GI bleeds, bleeding stops spontaneously. However, the remaining 20% will require treatment for either continued bleeding or rebleed.(the majority will survive the primary bleed). This high risk group mainly comprise those patients with continued oozing at endoscopy or non-bleeding visible vessel. Despite the use of diagnostic and therapeutic endoscopy and improved medical treatment over the last 40 years, the mortality for patients with non variceal UGI bleeds remains at 6-7%. Hence, the search for other effective medical interventions, such as the use of vasoconstrictors. SST has a number of effects on the GI tract including inhibition of gastric acid secretion, pancreatic secretion and biliary secretion. It also reduces gatsric mucosal blood flow, gastric perfusion and stimulates mucus production. Octreotide has a similar secretory effect but it is unknown whether it elicits the same effect on the mucosa or blood flow. The meta-analysis covering 14 RCTs show good evidence for the use of SST in acutely bleeding peptic ulcer to reduce the risk of continued bleeding, and a trend for reducing the need for surgery. The pH study and doppler studies theoretically support its use, as SST is shown to increase gastric pH (and thus allow optimum platelet function and decrease fibrinolysis) and decrease arterial blood flow, though with no date re clinical correlation. The two octreotide studies included in the meta-analysis have completely different conclusions - the blinded trial showing no effect on outcome, the non-blinded trial concluded that octreotide stopped PU bleed, decreased Tx requirements and need for surgery. The further non blinded RCT by Archimandritis concluded that octreotide is not superior to ranitidine. There were no published studies on the use of terlipressin. In conclusion, there is an evidence base for the use of SST in severe, acute, non variceal peptic ulcer bleeding, but not for other vasoconstricors at present. What is needed is a large study looking at the efficacy of SST in specific patient groups, defined by source of, and severity of bleeding (ie active/non bleeding visible vessel).

Clinical Bottom Line

SST should be considered in unwell patients who are likely to be bleeding secondary to PUD until definitive endoscopy, or in situations when endoscopy is contraindicated or unavailable. There is no definitive evidence for the length of time treatment should continue.

References

1. Avgerinos et al. Somatostatin inhibits gastric acid secretion more effectively than pantoprazole in patients with peptic ulcer bleeding:A prospective randomised, placebo controlled trial. Scandinavian Journal of Gastroenterology 2005;40:515-522.
2. Saruc et al. Somatostatin infusion and hemodynamic changes in patients with non-variceal upper gastrointestinal bleeding: a pilot study. Med Sci Monit 2003; 9(7):P184-87.
3. Archimandritis at al. Ranitidine versus ranitidine plus octreotide in the treatment of acute non-variceal upper gastrointestinal bleeding:A prospective randomised study. Current Medical and Research Opinion 2000 - vol 16, No3, 178-183.
4. Imperiale et al. Somatostatin or Octreotide compared with H2 antagonists and placebo in the management of acute non variceal upper GI haemorrhage. Annals of internal medicine 1997: 127(12) p1062-71.