There have been concerns raised about the potential adverse effects of proton pump inhibitors, especially with long-term use. In particular, their potent action can suppress the features and delay the diagnosis of gastric cancer, while prolonged exposure may hasten the development of gastric carcinoids. To examine the use of proton pump inhibitors in patients at the major teaching hospital in Tasmania, Australia, principally to determine the appropriateness of the therapy according to published guidelines. A retrospective review of the medical records of all patients prescribed any of the proton pump inhibitors at the hospital over a 7-month period, was performed. An extensive range of demographic and clinical variables was recorded for each patient. The patients were also asked a series of questions during their hospitalization to extract some of the relevant information - in particular, if and when they had undergone endoscopy. The 200 patients (52% males) had a mean age of 69 +/- 16.4 years. The most common indications for using proton pump inhibitors were acute gastrointestinal bleeding (20.9%), severe refractory ulcerating oesophagitis (17.3%), mild/moderate oesophageal reflux (17.3%) and refractory peptic ulcer (11.7%). A large number of patients were using a proton pump inhibitor for 'other' indications (39.6%). The prescribing of proton pump inhibitors satisfied the approved indications, as outlined in the Australian Schedule of Pharmaceutical Benefits, in only 37.1% of cases. Endoscopy had been performed in 54.1% of patients prior to commencing therapy with a proton pump inhibitor and within the next 7 days in another 12.8% of patients. Only 59% of patients had previously been treated with an H2-receptor antagonist before commencing therapy with a proton pump inhibitor. Even worse, only 58.5% of patients had used an H2-receptor antagonist before a proton pump inhibitor for mild/moderate oesophagitis. The median duration of proton pump inhibitor therapy for patients admitted to the hospital and already receiving one of the drugs was 450 days. Over half of the patients were being concurrently treated with other drugs which are known to cause or exacerbate gastro-oesophageal disease, and 18% were smokers. Whereas the proton pump inhibitors are undoubtedly effective agents, studies of their prescribing in practice consistently suggest over-use prior to endoscopy, use in patients who do not fit the approved criteria, and prescribing for indications in which 'less powerful' agents should have been sufficiently effective for the patient's symptoms. This poses economic and safety concerns, particularly in light of the suggestion that these drugs could delay the diagnosis of gastric cancer.