Correspondence (letter to the editor): Inpatient Overtreatment

During the discussion of infectious complications of long-term treatment with PPI (1), an increased incidence of C. difficile infection is mentioned. The cited meta-analysis (Kwok CS et al.) did indeed identify PPI as a risk factor for recurrent C. difficile infection but showed that the infection risk under PPI treatment is comparable to that under antibiotic therapy. Moreover, the number needed to harm (NNH) among hospitalized patients is 67, i.e., for every 67 patients started on PPI treatment, one will suffer from recurrent C. difficile infection. A risk score has been developed for upper gastrointestinal bleeding among hospitalized patients outside the intensive care unit; only when this score reaches =10 points does the number needed to treat (NNT) drop to 95 (2). If a score of 10 points had been taken as a threshold for ordering a PPI – e.g. corresponding to a man aged 60 years or older on a general medical ward with acute renal failure and thrombocytopenia <50 000/µL – then PPI could have been saved in 82% of the total patient cohort (n=75 723). This overtreatment in the hospital setting persists when the drugs listed in the discharge summary are indiscriminately continued. A study in the German federal state of Mecklenburg-West Pomerania revealed that 263 (58%) of 506 patients discharged from 35 different hospitals had no clear indication for PPI, yet most of them were still inappropriately taking PPI after discharge (3). Finally, the recently published meta-analysis concerning acid suppression for stress ulcer prophylaxis in critically ill patients did not reveal an unequivocal superiority of PPI with respect to either overall survival or gastrointestinal hemorrhage (4). Although PPI have been in clinical use for more than 25 years, there is still a lack of evidence from clinical trials to justify their use in intensive care units around the world.