Search in the Abstract Database

Search Abstracts 2015

* = Presenting author

P366 Withdrawing or continuing maintenance treatment with thiopurines in patients with Crohn's Disease in sustained clinical remission: a lifetime risk-benefit analysis

J. Kirchgesner*1, 2, F. Carrat2, 3, J. Cosnes1, M. Schwarzinger4, L. Beaugerie1, For the BERENICE Study Group1

1Hopital Saint Antoine, Department of Gastroenterology, Paris, France, 2INSERM, University Paris 06 , UMR S 1136 , Paris, France, 3Hopital Saint Antoine, Department of Public Health, Paris, France, 4THEN, Translational Health Economics Network, Paris, France


Long-term treatment of Crohn's disease (CD) with thiopurines modifies the risk of various cancers, depending on gender and age for lymphomas, and presence of extensive colitis (EC) for colorectal cancer. We evaluated risks and benefits of withdrawing or continuing thiopurines in patients with CD in sustained clinical remission.

35 year-old woman without extensive colitisW65 year-old woman without extensive colitisW35 year-old man with extensive colitisW65 year-old man with extensive colitisW
Life expectancy, age at death (years)79.8679.8482.9883.0373.9673.8779.3879.28
Events for 1,000 patients-years
Severe relapse0.1690.3490.1430.2700.2050.4280.1460.258
Colorectal cancer1.021.030.7630.7681.571.752.463.71
Relative risk compared to general population
Colorectal cancer0.991.000.990.991.


We developed a Markov model assessing risks and benefits of withdrawing or continuing thiopurines in a lifetime horizon. We simulated a cohort of CD patients in stable clinical remission for 5 years under thiopurines. The model was stratified by age, gender and presence of EC. Parameter estimates were taken from French hospital Diagnosis Related Groups (DRGs) databases, the French cancer and death national registries (FRANCIM), the CESAME cohort, the Saint-Antoine hospital database (MICISTA)), and from the literature. The primary outcome was life expectancy. Secondary outcomes included the probability of CD relapse, cancers, and causes-of-deaths. We conducted threshold analyses on age to assess potential age-related changes in best life expectancy according to the two strategies.


The continuing strategy was globally associated with decreased lifetime risk of flares and increased risk of lymphoma. However, life-expectancy and risks associated with the two strategies differed markedly according to age at decision, gender and presence or absence of EC, as illustrated in Table 1. In terms of life expectancy, in patients without EC, withdrawal strategy was associated with increased life expextancy after the age of 41.7 years for men, and 47.0 years for women. In patients with EC, continuing thiopurines was associated with increased life expectancy, regardless of age.


Factors determining life expectancy associated with withdrawal or continuation of thiopurines in patients with CD and 5-year sustained clinical remission vary substantially according to gender, age and presence of EC. Individual decisions to continue or withdraw thiopurines in patients with CD in sustained remission should take into account age, gender, and presence of EC.