Reply to Dr. Krag et al.

We thank Dr. Krag et al. for their interest in our meta-analysis of the endoscopic therapy of acute variceal bleeding. We were very careful in what evidence we considered to be supported by the pooled data given by meta-analysis interpreted in a reasonable manner. Thus our summary in the abstract states: ”The conclusive evidence for substituting banding ligation or the combination of vasoconstrictors with sclerotherapy as better therapeutic approaches is not present in randomized trials. Sclerotherapy can remain a gold standard in variceal bleeding but there is scope for further studies of ligation and vasoactive drugs.“

The reasons for this claim can be summarized as follows. Firstly, in the randomized trials of vasoactive drug combined with sclerotherapy, the median efficacy of sclerotherapy alone was only 69 %, which is not explained by differences in patient type, or duration or dose of therapy in this group of studies, compared with sclerotherapy versus drugs alone, in which the median efficacy of sclerotherapy was 83 %. Secondly, turning to the particular issue which troubles Dr. Krag and colleagues regarding the interpretation of our meta-analysis of emergency sclerotherapy vs. ligation, the issues are the following (as detailed in our paper): first, there is no difference in mortality between the two techniques (which they acknowledge); second, the average risk difference in control of bleeding is only 2.6 %; third, the interpretation of the slight superiority (meta-analytically) of ligation must be colored by (a) the use of the overtube to perform ligation in 8/12 studies (this is not currently usual practice), (b) the fact that most data were taken from long-term studies of ligation in which the ”acute“ data are not given in great detail, and (c) the heterogeneity in the reported efficacy of ligation, for example from the same geographical area (e. g. 97 % at 72 hours [1] and 62 % at 48 hours [2]); fourth, a comparison of complications cannot be made as one cannot separate complications associated with acute therapy from those associated with long-term therapy.

The issue of double intubation remains, unless the authors suggest that standard practice should be to perform diagnostic endoscopy with the ligating apparatus already attached. Although we agree there are no data that suggest that complications are increased as a result of double intubation, it is plausible (contrary to the opinion of Krag et al.) that they could be. This makes the ligation less applicable, given the small gain in efficacy, even disregarding our caveats. We raised the issue of double intubation as it has not been formally evaluated in the setting of acute variceal bleeding. It should be, as there is a potential increased risk of aspiration pneumonia, unless tracheal intubation under anesthesia is used universally.

We did assess complications (as stated in the Methods section) but, as we pointed out, there was no consistent reporting of these (they were variously reported per patient, or by absolute numbers, or as major vs. minor complications), so that only a descriptive analysis could be made (see Tables e2, e4, and e6 in our article).

While we agree with their comment that repeated ligation (for prevention of rebleeding) results in fewer complications than long-term sclerotherapy, the issue here is the complications of emergency endoscopic sclerotherapy or ligation and not of long-term therapy. There are no good data available for complications related to emergency therapy, so, for example, the stricture rates mentioned by them are not applicable to emergency sclerotherapy.

Considering the above, therefore, we concluded that there was scope for further studies of ligation and vasoactive drugs in acute variceal bleeding and that at the initial diagnostic endoscopy, emergency sclerotherapy can remain a gold standard therapy. We disagree with how the Baveno IV recommendation is formulated - that ligation is the ”default“ best endoscopic emergency therapy for acute bleeding varices. We believe that sclerotherapy retains its role and our meta-analysis provides evidence for this. However, we accept that because ligation is used for long-term therapy, then experience with injection therapy will be limited and so an operator will always choose the procedure that he or she is most adept at. However, this is a different justification for the use of emergency ligation, which while acceptable, cannot be used to ”penalize“ emergency sclerotherapy. This is particularly so because the average difference between ligation and sclerotherapy in terms of efficacy is only 2.6 %, with no difference in mortality between the two techniques.

The study by Villanueva et al. [3] evaluates the addition of both endoscopic techniques to somatostatin started before endoscopy. Accepting that somatostatin is acting in the same way in both groups, then the addition of this trial to our meta-analysis results in a 3.3 % risk difference for efficacy and no difference in mortality. Notably, and in contrast to the interpretation of the superiority of ligation by Krag et al., the two techniques had similar failure rates in those with active bleeding and in patients with Child Pugh C disease, the groups in whom better efficacy is needed. In our article we cited a paper by Masci et al. (Gastrointest Endosc 1997; 45 : 847A) which compared banding ligation with sclerotherapy of esophageal varices, but excluded it from the meta-analysis because it contained no data on acute bleeding.

We therefore stand by our interpretation of the results of our meta-analysis of ligation vs. sclerotherapy based on the current published data. Clinicians will not be at fault if they are competent sclerotherapists, able to inject varices at diagnostic endoscopy, thus avoiding re-intubation.

Competing interests: None

References

• 1 Lo G H, Lai K H, Cheng J S. et al . Emergency banding ligation versus sclerotherapy for the control of active bleeding from esophageal varices.  Hepatology. 1997;  25 1101-1104
• 2 Sung J J, Chung S C, Yung M Y. et al . Prospective randomised study of effect of octreotide on rebleeding from oesophageal varices after endoscopic ligation.  Lancet. 1995;  346 1666-1669
• 3 Villanueva C, Piqueras M, Aracil C. et al . A randomized controlled trial comparing ligation and sclerotherapy as emergency endoscopic treatment added to somatostatin in acute variceal bleeding.  J Hepatol. 2006;  45 560-567