Prevention of Variceal Bleeding With Band Ligation questionable. A comparison between banding and beta-block-
To the Editor: ers would have been mandatory. The authors do state in the We read with interest the recent paper by Lay et al. 1 The discussion that, although they had considered using betaauthors should be congratulated for having been able to select blockers in the control patients, ''they did not have enough a group of patients with an extremely high risk of first bleeding patients to proceed with such a study''. This is hardly a justifiby means of Beppu's 2 score. This, however, contrasts with the cation. The authors do not indicate how they calculated the prospective validation of Beppu's score that we made in the sample size of the study, and, therefore, we do not know the North-Italian Endoscopic Club study, 3 because in that study 'a priori' hypothesis they made about the rate of bleeding in patients with a Beppu's score of 00.38 or lower had a 2-year the control group and the expected therapeutic gain in the incidence of bleeding of 41%, which is much lower than the banding group. Even if they had admitted that such calculation figure of 60% reported by Lay et al. Moreover, in our series, had been made and if they had shown that a number of papatients with such values of Beppu's score represented only tients larger than that available to the authors was needed to 15% of the whole series, as compared with 30% in the series appropriately compare banding with beta-blockers, this would presented by Lay et al.
not constitute an ethical justification for the authors to per-As it has been shown that the higher the bleeding rate in form a study comparing banding to nothing. Instead, a multithe control group, the higher will be the efficacy of prophycenter design should have been considered to collect enough lactic treatment, 4 it is not surprising that variceal band ligapatients to perform the appropriate study, i.e., a comparison tion was also very effective in preventing first bleeding in between banding and beta-blockers. the study by Lay et al. In such conditions, endoscopic sclerotherapy was also very effective. 5 However, the possibility ex-ROBERTO DE FRANCHIS, M.D. ists for banding, as it does for sclerotherapy, that further MASSIMO PRIMIGNANI, M.D. studies might be less fortunate in selecting very-high-risk
Gastroenterology and Gastrointestinal Endoscopy Service patients. Under such circumstances, the value of band liga-Institute of Internal Medicine tion remains to be established.
University of Milan Another important point concerns the handling of patients
Milan, Italy in the control group. In the study by Lay et al., such patients