Article Text

Download PDFPDF

Mortality with oesophageal varices: different things to different people
  1. K S Henley
  1. Department of Internal Medicine, Division of Gastroenterology, University of Michigan Health System, Ann Arbor, Michigan 48109, USA; Khenley{at}

Statistics from

LeBrec (

) has summarised reports showing that mortality from bleeding varices has significantly decreased in the past few decades. Because of the limited follow up in some of these publications, we have examined the published mortality from bleeding varices where follow up data were available for at least one year (group BV). For comparison, similar data were collected from studies in patients with known varices who had not bled (group NBV).

For inclusion, patients had to be reported in a randomised controlled trial, published in full between 1984 and 2001. Management under the protocols had to be initiated promptly and survival data had to be recorded for at least one year. The few patients reported as transplanted were counted as survivors. In most instances there was no statistical difference in survival between treatment arms and the data were combined. Where this was not the case (three in the BV group and four in the NBV group) only data from the treatment arm showing better survival were included in the analysis.

Thirteen publications in the BV group representing 1321 patients of mean age 53.4 (SEM 1.13) years met the inclusion criteria. The NBV group comprised 2472 patients of mean age 54.9 (0.90) years. Mortalities are shown in table 1.

There was no statistically significant difference in mortality between the BV and NBV groups for the two years for which comparisons were possible. There was no correlation between mortality and the number of patients included in each report or the date of publication.

These data show that patients with oesophageal varices, although in their mid-fifties, have a life expectancy of normal individuals in their mid-eighties. The presence of varices, whether they have bled or not, is an ominous prognostic sign regardless of how they are treated.

There are limitations to this analysis. The prevalence of varices is not known. The fate of patients who die before reaching a treating facility or where the outcome is unreported is also unknown. Yet, as pointed out by Graham and Smith,1,2 the timing of randomisation for any treatment programme has a major impact on outcome. Again, the relatively large number of patients included in this analysis and the wide range of mortalities reported may obscure real differences in as yet unidentified subgroups. None the less these data are sobering. As pointed out by Smith and Graham more than 20 years ago, “….we have learnt much and accomplished little”.1

Table 1

Mortality (%) in patients with bleeding varices (group BV) compared with those in patients with known varices who did not bleed (group NBV)


Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.