A Matter Of Life Or Death: Mortality In Crohn's Disease

Mortality and causes of death in Crohn's disease: follow-up of a population-based cohort in Copenhagen County, Denmark. Jess T, Winther KV, Munkholm P, Langholz E, Binder V. Gastroenterology 2002;122:1808–14. Jess et al. present data on overall survival and cause-specific mortality in an inception cohort of Crohn's disease (CD) patients from Copenhagen County, Denmark. Three hundred seventy-four patients who had been diagnosed with CD between 1962 and 1987 were observed until 1997 for overall survival and causes of death. Observed deaths were compared with expected deaths to calculate standardized mortality ratios (SMRs). The expected number of deaths was derived from age- and gender-specific mortality rates in Denmark. The Kaplan-Meier method was used to calculate actuarial survival, which was compared with that of a matched background population. Follow-up evaluation was available for 98.9% of the cohort. Eighty-four deaths were observed in the cohort compared with 67 expected (SMR, 1.3; 95% CI, 1.01–1.56). The increase in mortality was limited to women (SMR, 1.4; 95% CI, 1.03–1.89). In men, the number of observed deaths was not significantly higher than expected (SMR 1.1; 95% CI, 0.8–1.5). Women diagnosed before the age of 50 years had a mortality rate triple than that expected, whereas women diagnosed after the age of 50 years had a mortality rate similar to the general population. Twenty years after diagnosis, the overall survival was 76% compared with 82% expected. Virtually all the excess mortality could be explained by an increase in CD-related deaths (n = 15) or deaths due to gastrointestinal tract cancer (n = 7). Early studies of the natural history of CD and ulcerative colitis (UC) demonstrated such an increased risk of mortality that many of our patients continue to encounter difficulty in obtaining life insurance at a reasonable price. Most recent population-based studies of mortality in patients with UC, from widely disparate locales such as Denmark, the United Kingdom, Italy, and the United States, have not demonstrated decreased survival in patients with UC relative to the general population (Gastroenterology 1992;103:1444–51, Dig Dis Sci 1993;38:538–41, Eur J Epidemiol 1997;13:189–94, Gut 1998;42:175–9, Gut 2000;46:336–43, Am J Gastroenterol 2001;96:501–7, Digest Liver Dis 2001;33:686–92). Presumably this decrease in mortality is due to improved medical and surgical therapies for UC. On the other hand, recent studies of mortality in CD have yielded conflicting results, with reports from Uppsala (Gastroenterology 1992;103:954–60), Stockholm (Gastroenterology 1996;110:1339–45), Florence (Gut 1998;42:175–9), and Olmsted County (Gastroenterology 1998;114:1161–8) indicating higher mortality, and reports from Italy (Scand J Gastroenterol 1996;31:372–5) and the United Kingdom (Gut 1992;33:1226–8;Am J Gastroenterol 2001;96:501–7) demonstrating no differences in survival compared with the general population. This population-based inception cohort of CD from Copenhagen County, Denmark is perhaps the most extensively studied in the published inflammatory bowel disease (IBD) literature. In an earlier study of the same cohort, Munkholm et al. found no difference in overall survival compared with the general Danish population (Gastroenterology 1993;105:1716–23). In that study, subgroup analysis suggested that CD patients diagnosed in the third decade of life and those with extensive small bowel disease had a higher than expected mortality in the first 5 years after diagnosis. The present study by Jess et al. extends follow-up evaluation of the Copenhagen County cohort by 10 years. Follow-up participation was extraordinarily high (> 98%). In contrast to the previous study, the authors found that CD is associated with a mortality rate approximately 30% higher than that expected in the general population. Interestingly, a significant mortality increase was seen in women but not men. Among 188 women who were diagnosed before the age of 50 years, 25 deaths were observed, while only 7.3 were expected. Among the deaths unrelated to CD, there appeared to be slightly more deaths than expected due to infection or genitourinary disease among women, while there was an excess of non–CD-related gastrointestinal disease deaths among men. The results of the study seem to confirm the general impression that mortality in patients with CD is slightly increased. As nicely illustrated by the difference in the 1993 Copenhagen study and the present one, the differences in results may be more related to length of follow-up period of a cohort rather than true differences in the natural history of disease. Perhaps the overall difference in survival between patients with CD and the general population is small enough to require a lengthy follow-up period to detect it. (A similar phenomenon was seen in Olmsted County, Minnesota, where a series of reports could not demonstrate a decreased survival until the cohort was extended to a 54-year period.) The absolute difference in survival at 20 years, 6%, is similar to the difference noted in Olmsted County (7% at 20 years). Although these differences may seem small at first glance, they are obviously clinically relevant when extrapolated to the entire CD population. Virtually all of the excess mortality in this study could be explained by an increase in CD-related deaths or deaths due to gastrointestinal tract cancer. When these deaths were excluded, survival curves of CD patients and the background population were similar. Approximately one fourth of all deaths were thought to be disease related or due to gastrointestinal cancer, similar to the percentage seen in other cohorts. We must continue to remain vigilant about the development of gastrointestinal tract cancer in our CD patients. The finding of a decreased relative survival in women but not men is a novel one and remains unexplained. Most incidence studies of CD suggest a slight female predominance (50–60%). Case-control studies have suggested a stronger association between cigarette smoking and CD among women than men, and studies of postoperative recurrence of ileal CD have revealed that women who continue to smoke are much more likely to experience recurrence than men who smoke. It would be interesting to know if there were gender-related differences in smoking rates in the Copenhagen County cohort. Although additional subgroup analyses were performed in this study, based on age at diagnosis, duration of IBD prior to surgery, and years of follow-up evaluation, they must be interpreted with caution, due to the relatively small size of the cohort. Moreover, it is not clear that such analyses had been planned a priori. This study continues to add to our understanding of the clinical course and natural history of CD. Now that our therapeutic armamentarium has improved with wider use of immunosuppressive agents, the advent of biologic therapies, and more judicious use of surgery, we must plan to continue such studies to determine if we are truly altering the natural history of this debilitating disease.