Author's reply.

We thank Zullo et al for submitting their criticism on our study published in Annals of Gastroenterology [1]. A point-by-point response to their comments can be found below.

Firstly, the overall performance of commercially available serology kits that measure IgG antibodies for the diagnosis of Helicobacter pylori (H. pylori) infection showed that serology has accuracy >90% in diagnosing H. pylori infection. These are widely available and cheap, and particularly helpful in cirrhotic patients taking antibiotic and/or antisecretory (proton pump inhibitor, PPI) treatments which may affect culture, histology, rapid urease test, urea breath test or stool test [2]. This is relevant especially in our cirrhotic patients as most of them were on PPI treatment. Our study shows that patients with liver cirrhosis have a seroprevalence of H. pylori 35.7%, comparable to the Indian data reported by Batmanabane et al [3] in which the prevalence of H. pylori was 43%. The decreasing seroprevalence may be due to improved sanitation and hygiene as a result of a rapidly growing economy.

Secondly, this study demonstrates not only a significant association of H. pylori with portal hypertensive gastropathy (PHG) in cirrhosis but also with the severity of PHG. Being a retrospective study selection bias cannot be completely ruled out.

Thirdly, one of the major limitations of the study is the absence of histological data for diagnosis of PHG, which was diagnosed on the basis of upper gastrointestinal endoscopic appearance only, an accepted method for the diagnosis of PHG in previous studies as well [3,4]. However this might have been attributed to inclusion of 30 CHILD C cirrhotic patients as controls.