[Eradication of Helicobacter pylori infection in Hungary (1993-2002): a meta-analysis]

Meta-analysis is a useful tool in obtaining sound data in evidence-based medicine. The guidelines of eradication have been proposed in the I. Maastricht consensus (1996), which was modified in 2000.

The scope of present paper is the meta-analysis of eradication studies performed in Hungary between 1993-2002.

Articles/abstracts dealing with eradication, published in peer-reviewed Hungarian and international journals, meetings of the Hungarian Society of Gastroenterology and its Endoscopic Section and international congresses were identified from major databases and by manual search. The data were classified into groups based on eradication schedules, type of antibiotic used, and provenience of the study. The pooled eradication rates were calculated and compared with the international data.

The pooled eradication rate of proton pump inhibitor-based triple therapies was 82.9% (confidence interval: 72.1-93.7%); omeprazole, pantoprazole and lansoprazole-based therapies given b.i.d. were equally efficient. H2-blockers based triple combinations lasting 10-14 days obtained a pooled eradication rate of 86.3% (confidence interval: 72.2-99.8%), insignificantly higher (p = 0.07) than the rate obtained with proton pump inhibitors. Classical triple therapy's (bismuth + 2 antimicrobials) pooled eradication rate is significantly lower as that of proton pump inhibitor (p = 0.01) or H2 blocker-based combinations (p = 0.008). Clarithromycin, nitroimidazoles, amoxicillin and doxycyclin-based therapies obtained pooled eradication rates of 79.4%, 74.7%, 79.3% and 73.0%, respectively. The results obtained in randomized controlled trials and university centers were better than those achieved in county and urban hospitals.

The results are in part contrasting with the recommendations of European and national consensus. Proton pump inhibitor-based triple regimens are proposed as first-line treatment, which is in agreement with our results, but H2-blockers based regimens seem to be of same efficacy in open studies, although these represent a lower level of evidence. Classical triple therapies, also recommended, did not reach acceptable rates of eradication. Antibiotic-based analysis did not reflect the local resistance profile. A multicenter, randomized Hungarian study is warranted to assess the real efficacy of the eradication schedules.