James Nott CT1 ACCS

 Epidemiology and Definition  Pathophysiology of Stress ulcers  Clinical presentation  Risk factors  Prophylaxis agents available – mechanisms  Meta-Analysis: PPIs vs H 2 RA for stress ulcer prophylaxis in critically ill patients.

 Common  ~ 1.5 – 8.5% GI bleeding in all ICU patients  15% - 25% of ICU patients not on prophylaxis  75% mucosal abnormalities of ICU patients <72hrs (major burn/cranial trauma)  Ulcer: ‘Lesion of mucosal membrane accompanied by oedema and necrosis of surrounding tissue’  Loci: Stress ulcers - Fundus (Proximal) Peptic ulcers - antrum (Distal) / proximal duodenum Presentation range: Asymptomatic – Acute haemorrhage depending on depth of ulcer

 Imbalance: mucosal protection vs gastric pH  Multi-factorial:

 Range depends on depth of ulcer Superficial: Asymptomatic Deep: Haemorrhage (Haematemesis /Melena)

 Intubated >48hrs OR 15.6 (3) Cook. DJ et al ’94  Coagulopathy OR 4.3  Additional risks identified: SHOCK Everyone! Sepsis Hepatic and Renal failure Multiple trauma Burns of >35% BSA Glucocorticoid therapy

Pharmacological mechanisms: 1) Block acid secretion  Competitive H 2 antagonists (Ranitidine)  Proton pump inhibitors (Omeprazole) 2) Neutralise stomach acid contents Antacids (Gaviscon) – Bicarbonate neutralises pH

3) Protecting stomach mucosa – nil buffering Sucraflate - polysaccharide + Aluminium hydroxide 4) Prostaglandin analogues Misoprostol – inhibit parietal cells to generate cAMP, thus reduce stomach acid secretion

 Nosocomial pneumonia (HAP/VAP) ↑ pH – Less hostile environment for bacteria Not so with Sucralfate (not affects pH)  Ways to reduce risk: (Preventing aspiration) - Patient position – semirecumbant not supine - Mouth care – Chlorhexidine mouthwash/gel - Subglotic drainage - ? Residual gastric monitoring

 Aim: ‘Determine efficacy and safety of proton pump inhibitors verses H 2 receptor antagonists for the prevention of upper GI bleeding in ICU’  Methodology: Search strategy – MEDLINE (1948-March 2012) EMBASE (1980-March 2012) Two researchers independently ACPJC (1991-March 2012) extracted data Cochrane (central) database CINHAL.  Any disagreements resolved by discussion or consensus

 Types of study: - Randomised Control Trials (RCTs)  Population: - ICU Adults (Medical and Surgical included)  Intervention: - Control= H 2 RA Intervention=PPIs - para-enteral/enteral - regardless of dose, frequency and duration

 Primary outcomes : - clinically important GI bleed (?Meaning Hb drop/instability) - overt upper GI bleeding (coffee ground emesis, melena, fresh blood PR from UGI )  Secondary outcomes : - Nosocomial pneumonia (VAP/HAP) - All-cause mortality - ICU length of stay - C. Diff infection

 All trials assessed for risk, depending on domains: - Low = risk low in ALL domains - Unclear =risk is unclear in >1 domain - High = risk is high in >1 domain  Domains: Sequence Generation Allocation concealment Blinding Incomplete outcome data Selective reporting bias Free of other bias Plus – Overall risk of bias

 Data analysed using RevMan 5.1 model.  Pooled: Dichotomous outcomes - RR 95% CI Continuous outcomes – Mean  Methods of assessing heterogeneity (? Bias): Subgroup analysis – Size, Egger’s test /funnel plot to assess publication bias

Primary objectives: 1) Clinically important bleeding (12 Trials n=1614) Significantly lower RR with PPIs vs H 2 RA: (RR % CI p=0.002)

2) Overt Bleeding ( 14 Trials n= 1720) Significantly lower RR with PPIs vs H 2 RA: (RR 0.35; 95%CI p<0.0001)

 1) Nosocomial Pneumonia ( 8 Trials, n= 1100) No significant difference: RR % CI ( ) p=0.76

2) Mortality ( 8 Trials n= 1196) No significant difference: RR % CI ( ) p=0.91

3) ICU Length of stay ( 5 Trials n=555) No significant difference :CI ( ) p=53 4) Clostridium difficile infection No trials reported on C. Difficile infection

‘ Significantly ↓ risk of both 1 0 outcomes with PPIs - Clinically important GI bleeding – RR 0.36 ( ) - Overt UGI bleeding – RR 0.35 ( ) ‘No significant ↓ risk of 2 0 outcomes with PPIs vs H 2 RA’ Nosocomial pneumonia – RR 1.06 ( ) ICU mortality – RR 1.01 ( ) ICU length of stay – RR 0.54 ( )

 Primary outcomes (14 Trials): Low risk of bias – 3 Inflates benefit of PPIs Unclear – 5 High risk of bias – 6 (lack of blinding)  Publication bias (Funnel plot): Larger trials (y axis) are more reliable – closer to mean RR (x axis) Inverted funnel shape Clinically important bleeding Overt bleeding outcome

 Definitions varied (Pneumonia)  No randomisation of nutritional strategies  Age of studies (1948) – ? relevance to today’s  Accepts - ‘PPIs may lower risk of GI bleeding’  Welcomes further research: - Effect of early nutrition - Cost-benefit analysis - C. Diff infection

 NO!  Meta-analysis does not show significant evidence that PPIs are better than H 2 RAs.  Rantidine is cheap and as effective as PPIs for ulcer prophylaxis, so why change?  Establish enteral nutrition important – no statistical evidence.

Thank you, any questions!?!

1) Cook.DJ, Stress ulcer prophylaxis. Best evidence synthesis Scand J Gastroenterol Supple 1995; 210:48 2) Stollman. N, Pathophysiology and prophylaxis of stress ulceration in ICU, Journal of critical care, vol 20, March ) Cook. DJ et al, risk factors for gastrointestinal bleeding in critical ill patients, Canadian Critical Care Trials Group. N Engl J Med 1994; 330:377 5) Marik P.E et al, Stress ulceration prophylaxis in the new millennium, meta-analysis. Critical Care Med 2010; 38:2222