A Therapeutic Index That Predicts the Individual Effects of Prednisone in Patients With Cirrhosis

CSL, CSL, CSL, CSL, CSL, CSL, CSL

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

28 Citationer (Scopus)

Abstract

Our aim was to construct an index that accurately predicts the degree of benefit or harm that prednisone therapy holds for patients with liver cirrhosis. The admission and survival data of 488 patients with cirrhosis who participated in a controlled clinical trial of prednisone in a dosage of 10–15 mg daily (251 patients) versus placebo (237 patients) and who were observed for up to 12 yr were analyzed using Cox's multiple regression model. Four variables each provided significant therapeutic information; antinuclear factor (p = 0.02) and large piecemeal necroses (p = 0.02) were associated with a beneficial effect, whereas ascites (p = 0.0004) and large regenerative nodules (p = 0.0007) were associated with a harmful effect of prednisone. From these four variables a therapeutic index was constructed. For a given patient the therapeutic index is a measure of how big the effect will be i f prednisone is given. The gain in survival time obtained by administering prednisone according to the therapeutic index was estimated to be 349 yr, mainly confined to 217 patients with a significant positive (121) or negative (96) therapeutic index. The therapeutic index may prove useful for the optimal administration of prednisone treatment in new patients with cirrhosis.

OriginalsprogEngelsk
TidsskriftGastroenterology
Vol/bind88
Udgave nummer1
Sider (fra-til)156-165
Antal sider10
ISSN0016-5085
DOI
StatusUdgivet - 1985

Bibliografisk note

Funding Information:
Received June 2, 1983. Accepted August 1, 1984. Address requests for reprints to: Erik Christensen, M.D" Department of Hepatology 233, University of Copenhagen, Hvidovre Hospital, 30 Kettegard aile, DK-2650 Hvidovre, Denmark. This study was supported by grants 12-0344 and 12-1678 from the Danish Medical Research Council. The authors thank Niels Keiding for many constructive comments, Richard Kay and J. D. Kalbfleisch for making available the computer programs for analyzing the regression model, and Frank Cederbye and Peter Rytgaard for making most of the calculations. The authors thank as well the Department of Data Processing at Rigshospitalet for generous computer access during the early phase of the study. © 1985 by the American Gastroenterological Association

Citationsformater