PEPTIC Study

Effect of Stress Ulcer Prophylaxis With Proton Pump Inhibitors vs Histamine-2 Receptor Blockers on In-Hospital Mortality Among ICU Patients Receiving Invasive Mechanical Ventilation

Design

This was a multi-centre cluster randomised trial involving 50 ICUs in Australia, New Zealand, Canada, Ireland, and the UK.  The units were randomised (as opposed to individual patient randomization) to one type of stress ulcer prophylaxis for 6 months and then switched over to the other medication.

Study Population

Included: The study population included adult patients 18 years and older who underwent invasive mechanical ventilation in the first 24 hours of ICU admission.

Not included: Patients who were admitted to the ICU with upper gastrointestinal bleeding

The PEPTIC trial was an open-label, unblinded trial. A proton pump inhibitor or a histamine-2 receptor blocker was administered according to the ICU randomization status among those who were deemed to require stress ulcer prophylaxis according to the treating clinician,. Regardless of assigned therapy, either treatment could be administered based on clinician judgement. The assigned treatment was continued until death, ICU discharge, presence of clinically important upper gastrointestinal bleed, or until stress ulcer prophylaxis was considered to be not indicated anymore. In case upper gastrointestinal bleed occurred, a proton pump inhibitor was administered at clinician discretion.

The investigators assumed an absolute reduction in hospital mortality (the primary outcome) of 2.4% from a baseline mortality of 15%; corresponding to a relative risk reduction 16%. This required a mean cluster period size of 310 patients. Sensitivity analyses was carried out for missing data in the following manner. In the worst-best scenario for a binary outcome like mortality, a worst outcome event (e.g., in-hospital death) was assigned to all patients with missing data for the outcome in one treatment group, and a best outcome event (in-hospital survival) was assigned to all patients with missing data for the outcome in the other treatment group. The best-worst scenario was carried out in the reverse manner.

26 828 patients enrolled à13 436 assigned to proton pump inhibitors à 13 392 to histamine-2 receptor blockers. After exclusion of patients who presented with upper GI bleed, those who opted out, and those with missing mortality data,

• 13, 415 Patients in PPI group were included in primary outcome analysis
• 13, 356 Patients in H2RB group were included in primary outcome analysis

Patients were matched at baseline.

Results

Primary outcome

In-hospital, all-cause mortality during the index hospitalization on follow-up until 90 days from the date of the initial ICU admission. PPIs vs. H2-receptor antagonists:

2459/13 415 (18.3) vs. 2333/13 356 (17.5); RR, 1.05 (1.00 to 1.10); p = 0.054

Secondary outcomes

Clinically important upper GI bleed PPIs vs. H2-receptor antagonists:

1.3 vs. 1.8%; RR, 0.73 (0.57 to 0.92); p = 0.009

Clostridium difficile infection PPIs vs. H2-receptor antagonists: 0.3 vs. 0.43%; RR, 0.74 (0.51 to 1.09); p = 0.13

ICU and hospital days among survivors: Not different

Tertiary outcomes

Hours on mechanical ventilation

Ventilator-associated conditions

Subgroup analysis

Post-operative cardiac surgical patients: PPI vs histamine-2 receptor blocker group, the RR for the primary outcome: 1.27 (95% CI, 1.04-1.57)

Based on severity of illness according to the APACHE II score:

APACHE II score up to 17, no difference

APACHE II score range: 18-23: 17.5 vs.14.7%; RR: 1.15 (1.05-1.25)

APACHE II score range: 24-61: 45.3 vs. 43.5%; RR: 1.05 (1.00-1.11)

Limitations

• Cluster randomization though more convenient to carry out, may not be as powerful as an individual patient randomization.
• Open-label, clinicians knew what they were giving.
• Higher risk patients (who some evidence of bleed) were excluded.
• A significant number of patients got the drug other than assigned to them, or both drugs. A large number of patients appear to have not adhered to protocol.
• The mortality advantage of H2 receptor antagonists was not statistically significant.
• Clinically important upper GI bleeding was significantly lower with PPIs risk ratio: 0.73 (95% CI 0.57 to 0.92). This begs the question, should we really choose mortality as the endpoint in a study that addresses the efficacy of prevention of upper GI bleed. This is particularly so considering the lack of a putative mechanism that may cause death with PPI use. Do remember, there was no increase in the rates of C. difficile infection or pneumonia with PPI use, which are purported mechanisms of possible harm from PPI use.  
• No difference in the ICU LOS and the duration of mechanical ventilation.
• The data does not reveal a consistent relationship between the adverse effects of PPIs and illness severity (the sickest patients do not seem to demonstrate a difference in mortality).

Summary –

-Thought-provoking and the landmark trial

-Unable to change practices due to unavailability of suitable H2 receptor antagonists.

-Need large RCT comparing acid suppression therapy vs standard practices in ICU.