Letter: peptic ulcer bleeding in patients with cirrhosis

Abstract

SIRS, The study by Rudler et al. on the aetiology and outcomes of peptic ulcer bleeding in cirrhotic and noncirrhotic patients has several limitations. First, the study recruited only 203 patients, the majority (85%) of whom were noncirrhotic. These were compared with only 29 (14%) cirrhotic patients, who were loosely defined using clinical, biochemical and radiological parameters. Based on the very small numbers in the cirrhosis group, and as the patients in the two groups were not matched, the study was not sufficiently powered to address the questions they set out to answer. The gold standard test for diagnosing cirrhosis is a liver biopsy, although hepatic elastography is an acceptable alternative. Reliance on tests such as radiology, blood parameters and endoscopic findings can be misleading and lack specificity. For instance, noncirrhotic idiopathic portal hypertension can mimic cirrhosis both clinically and radiologically. Hence, it is conceivable that some of the patients they considered cirrhotic were not. Second, the authors throughout the manuscript refer to ‘idiopathic’ peptic ulcer disease (PUD), meaning those patients who had no evidence of Helicobacter pylori infection or nonsteroidal anti-inflammatory drug (NSAID)/aspirin use. This gives the erroneous impression that these two factors are the only causes of PUD. Before an ulcer can be labelled as idiopathic, one has to exclude other known but rare causes of peptic ulcers (such as other ulcerogenic drugs and diseases including Crohn’s and Zollinger–Ellison syndrome). The authors made no such attempts in their study; hence, it is impossible to know whether other aetiologies were involved. Moreover, recent alcohol intake in cirrhotic patients has been implicated in the pathogenesis of gastroduodenal ulcers; hence, it is misleading to disregard alcohol as a causative factor for ulcers in these patients. Third, the use of H. pylori serology to diagnose H. pylori infection in this study is questionable. It is well known that serological tests for H. pylori are unreliable in patients with cirrhosis and portal hypertension, with high sensitivity but low specificity (83.6 and 55.9, respectively). 5 Gastric biopsies for H. pylori histology and/or rapid urease test would have been preferable due to their higher accuracy in this setting. Thus, the figures they quote for H. pylori infection rate in the study are likely to be inaccurate. Moreover, the authors’ make no mention of whether the patients had received any prior H. pylori eradication therapy, which makes it impossible to determine whether the patients who were positive had current or prior infection, as H. pylori antibodies can take several months to years to disappear following eradication. In addition, noncirrhotic patients were much older than cirrhotic patients, and the differences seen in H. pylori infection rate between the two groups could have been influenced by the age difference. It is well known that the prevalence of H. pylori infection increases with age. Peptic ulcer disease, particularly gastric ulcer, is found more commonly in patients with portal hypertension. Although the exact pathogenetic mechanisms are unknown, it is generally believed that portal hypertension leads to impairment of gastric mucosal defences, hence rendering the gastric lining more susceptible to injury by aggressors such as NSAIDs alcohol and bile salts, culminating in peptic ulcers. Although a high alcohol intake alone has not been unequivocally identified as a risk factor for PUD in noncirrhotic patients, Aurox et al. found that in cirrhotic patients, recent alcohol intake prior to endoscopy was independently associated with gastroduodenal erosions. Almost all cirrhotic patients in this study had a history of excessive alcohol intake. Thus, the only