Current issues of ACP Journal Club are published in Annals of Internal Medicine


Review: Sclerotherapy to prevent first variceal hemorrhage reduces bleeding and mortality, but complications are high

ACP J Club. 1992 Nov-Dec;117:65. doi:10.7326/ACPJC-1992-117-3-065

Source Citation

Van Ruiswyk J, Byrd JC. Efficacy of prophylactic sclerotherapy for prevention of a first variceal hemorrhage. Gastroenterology. 1992 Feb;102:587-97.



To analyze the complications, bleeding rates, and mortality for prophylactic sclerotherapy in adults with esophageal varices.

Data sources

Index Medicus (1975 to 1989), bibliographies of relevant papers, and searches of the MEDLINE (1966 to 1989) and BIOSIS Previews (Biological Abstracts, 1980 to 1989) data bases, using the terms esophageal varices, sclerotherapy, and prophylaxis.

Study selection

All English-language, randomized controlled trials of adults receiving prophylactic sclerotherapy if they had no previous bleeding. The most recent and complete data were used.

Data extraction

8 trials with 742 patients were included. All trials reported ≥ 13 months of follow-up. Mortality and first bleeding rates were calculated along with complication rates and mortality related to the procedure. Major complications included variceal bleeding during or immediately after sclerotherapy, bleeding from sclerotherapy-induced esophageal ulcers requiring transfusion, esophageal stenoses requiring dilation, and sclerotherapy-related mediastinitis or pleural effusion.


The mean age range of study patients was 42 to 62 years. 0% to 90% of patients had alcoholic cirrhosis. All studies used a similar sclerotherapy schedule. In 7 of the 8 studies, patients received sclerotherapy until varices were obliterated. Treatment patients had follow-up endoscopies at 2- to 6-month intervals, and in 6 studies the control group follow-up was less frequent. Mortality was lower in the treatment compared with the control group (17% vs 27%). This absolute risk reduction (ARR) of 10% means that {10 patients would need to be treated (NNT) with prophylactic sclerotherapy (rather than control treatment) to prevent 1 additional death, 95% CI 6 to 25}*; the relative risk reduction (RRR) was 41%. This reduction in mortality remained significant when sensitivity analyses included the interim results from abstracts and foreign-language articles. Death from bleeding was lower in the treatment group (ARR 15%, {NNT 7, CI 4 to 34}*). 1% to 20% had major complications, and the pooled complication rate was 9% (CI 6% to 12%). The mortality rate was positively correlated with first variceal hemorrhages and deaths and negatively correlated with complication rates.


Although complication rates are high in adults who have sclerotherapy to prevent first variceal hemorrhage, both bleeding and mortality are reduced.

Source of funding: Not stated.

For article reprint: Dr. J. Van Ruiswyk, Clement J. Zablocki Veterans Affairs Medical Center, 5000 West National Avenue, Milwaukee, WI 53295, USA.

*Numbers calculated from data in article.


Some of the limitations of meta-analysis are evident in the study by Van Ruiswyk and Byrd. First, although the pooled study results showed homogeneity, the test for heterogeneity has low sensitivity. Second, many parameters such as referral, enrollment, exclusion criteria, and sclerotherapy schedules could not be pooled, and agreement among endoscopists was not evaluated. Also, severity assessment of liver disease was not uniform, and follow-up and alternative therapy for bleeding varied, confounding assessment of survival. Third, although overall results reached statistical significance, 95% confidence intervals were large, and therefore the clinical significance of prophylactic sclerotherapy is still uncertain (1).

The authors deserve to be commended for undertaking a difficult study. Their analysis and conclusions appear premature, however, because 6 other studies have subsequently been published. Sclerotherapy was marginally beneficial in 1 study; control and sclerotherapy did not differ in 3; and propranolol was superior to sclerotherapy in 2. 1 study in which control group patients had sham endoscopies (1) showed the harmful effects of sclerotherapy. This result raises the issue of whether efficacy estimates of sclerotherapy in other studies were inflated by less intense follow-up in the control groups. The results of the meta-analysis do not clarify our uncertainty about prophylactic sclerotherapy. This can be overcome only when we are able to accurately assess and include, in controlled studies, patients with esophagogastric varices who are at imminent risk for bleeding and in whom sclerotherapy can be effectively achieved with little morbidity.

Jacob Korula, MD
Rancho Los Amigos Medical CenterDowney, California, USA

Jacob Korula, MD
Rancho Los Amigos Medical Center
Downey, California, USA


1. Bulpitt J. Lancet. 1987;1:494-7.

2. The Veteran Affairs Cooperative Variceal Sclerotherapy Group. N Engl J Med. 1991;324:1779-84.