Introduction

Stress ulcers are gastroduodenal erosions that occur commonly in critically ill patients. Although once thought to be due to excess acid production, they are now thought to result from gastric mucosal ischaemia and the value of using pH altering drugs to prevent them has been questioned.1 Cook et al showed that only critically ill patients who have a coagulopathy or who are ventilated for more than 48 hours are at increased risk of developing serious bleeding due to stress ulceration.2 Prophylaxis is usually with either an H₂ receptor antagonist or sucralfate, and these agents appear equally efficacious in terms of reducing bleeding complications.3 A recent meta-analysis has, however, shown that sucralfate is associated with a lower incidence of pneumonia and mortality than H₂ receptor antagonists.4 In spite of this, patients referred to Guy's from other intensive care units have often received H₂ receptor antagonists for prophylaxis, and this survey was undertaken to quantify the extent of this practice.

Methods and results

In March 1996 we performed a structured telephone survey of 67 intensive care units in the Thames regions. If respondents were too busy to provide accurate information a repeat call was made. Four questions were put to the senior nurse or doctor in charge: (1) Does your intensive care unit have a protocol for stress ulcer prophylaxis? (2) Which agent is used most commonly for prophylaxis in your unit? (3) How many patients are there in your unit today? (4) How many of these patients are receiving sucralfate, an H₂ receptor antagonist, omeprazole, no prophylaxis?

On the day of the audit 312 patients were in the 67 units: 118 (38%) patients were receiving an H₂ receptor antagonist, 82 (26%) sucralfate, 3 (1%) omeprazole, and 110 (35%) no prophylaxis (one patient was receiving both sucralfate and an H₂ receptor antagonist.) The 28 units that had a protocol for stress ulceration prophylaxis were more likely to prefer sucralfate to H₂ receptor antagonists than those without a protocol, as were the 13 teaching hospitals (table 1).

Comment

Serious bleeding due to stress ulceration is defined as a drop in haemoglobin of greater than 20 g/1 accompanied by either haemodynamic instability or the need for blood transfusion.2 Although the incidence of bleeding is declining, it is associated with a mortality approaching 50%. This reduction may be due to pharmacological prophylaxis, but other factors such as improved resuscitation and early enteral nutrition are likely to be equally important.

Several adverse effects have been linked to the use of H₂ receptor antagonists in intensive care units, particularly an increased incidence of nosocomial pneumonia, which is associated with a mortality of 40-70% in critically ill patients.5 By raising intragastric pH, H₂ receptor antagonists promote bacterial colonisation and overgrowth within the stomach, and hence provide a potential reservoir of infection for the respiratory tract. Sucralfate does not raise intragastric pH but enhances mucosal blood flow, stimulates bicarbonate and mucous secretion, and may be bactericidal.

Our survey shows that many intensive care units, particularly those in district hospitals and those without a protocol for stress ulcer prophylaxis, continue to use H₂ receptor antagonists in preference to sucralfate. Given the known detrimental effects of H₂ receptor antagonists, it is disappointing that their use is so prevalent. If the principles of evidence based medicine were followed this would no longer be so, and the morbidity and mortality of critically ill patients might be reduced.