Long-Term Mortality in People With Celiac Disease Diagnosed in Childhood Compared With Adulthood: A Population-Based Cohort Study

Masoud Solaymani-Dodaran, M.D., Ph.D., M.P.H.; Joe West, M.B., Ph.D., M.R.C.P.; Richard F.A. Logan, M.B., M.Sc., F.R.C.P.


Am J Gastroenterol. 2007;102(4):864-870. 

In This Article


General Characteristics of the Cohort

Our study cohort consisted of 625 patients with celiac disease that included 252 (40%) males diagnosed between 1943 and 1983. The diagnosis had been made in childhood (before the age of 15) in 285 (46.6%) of the participants, and the remaining 340 (54.4%) patients were diagnosed with celiac disease as adults ( Table 1 ). The median age at diagnosis of the children was 1.5 yr compared with 46.3 for the adults, and there were more women diagnosed as adults (65%) compared with diagnosed as children (55%) (χ2 = 5.32, P = 0.02). Of the 218 deaths that occurred up until the end of our study, we included the 195 that occurred during the 14,926 person-years of follow-up between 1970 and 2004 in our analysis. While there was little difference in the loss to follow-up between children and adults, the children did contribute more person-years to our analyses. In children there was a median follow-up of 34 yr and in adults 23 yr.

Cause-specific Mortality

When all follow-up was included a childhood diagnosis of celiac disease was associated with a higher overall mortality risk (SMR 2.62, 95% CI 1.62-4.00) than adulthood diagnosis (SMR 1.55, 95% CI 1.33-1.80). The excess mortality in childhood-diagnosed celiac disease was mainly accounted for by deaths from accidents, suicide, and violence, malignancies, including two lymphomas, and cerebrovascular diseases resulting in two deaths, both of which were recorded as subarachnoid hemorrhage ( Table 2 ). One of the lymphomas was a T-cell lymphoma of the small bowel while the other was a cerebral lymphoma.

In adults the excess mortality was primarily because of malignant neoplasms. The excess deaths were mostly either gastrointestinal cancers or lymphomas. Overall, of the 14 lymphomas, seven were described as either small bowel lymphoma or as histiocytic lymphoma of the bowel, two non-Hodgkin's lymphoma, one Hodgkin's disease, two disseminated lymphoma, one ulcerative lymphoma of stomach, and one cerebral lymphoma. There were no deaths resulting from breast cancer in women when at least three would have been expected, but this deficit was not significant at the 5% level. In the seven deaths in childhood-diagnosed celiac disease that resulted from external causes, three were from drug overdoses.

Overall Mortality by Period After Diagnosis

All-cause mortality more than 5 yr after diagnosis was increased threefold in children (SMR 3.32, 95% CI 2.05-5.07) in contrast to only a 38% increase in adults (SMR 1.38, 95% CI 1.16-1.63). For adulthood diagnosis the overall risk decreased gradually over time with an SMR in the first year following diagnosis of 4.43 (95% CI 2.13-8.16). After 15 yr or longer of follow-up the point estimates showed only approximately a 30% increase of mortality but the 95% CIs all crossed unity ( Table 3 ). In childhood-diagnosed disease there were no deaths within the first 5 yr of our study but overall mortality risk was increased thereafter and remained over threefold 25 yr or more after diagnosis.

Cause-specific Mortality by Period After Diagnosis

Mortality from various causes by period diagnosis is shown in Table 4 and Figure 1. In children the main increased risks were for accidents, suicides, and violence, malignancies, and cerebrovascular disease, whereas in adult-diagnosed disease the small excess that we observed was mainly because of malignancies. After 25 yr or more of follow-up the relative increase of mortality from lymphoma was particularly raised in children (SMR 34.32, 95% CI 4.16-120.00) because of two deaths during this period, but also in adults, although not so markedly (SMR 4.64, 95% CI 0.12-26.00). The increase of deaths in children as a result of accidents, suicides, and violence was in the period 5-14 yr after diagnosis (SMR 7.49, 95% CI 1.54-22.00) and 25 yr or more (SMR 3.16, 95% CI 0.65-9.25).

Figure 1.

Standardized mortality ratios for all-cause and cause-specific mortality more than 5 yr after diagnosis in people diagnosed with celiac disease in childhood (A) and adulthood (B).

By Decade of Diagnosis

When analyzed by period of diagnosis the overall SMRs varied very little (overall SMR diagnosed before 1970 1.61, 95% CI 1.30-1.98, SMR diagnosed in 1970-1980 1.60, 95% CI 1.30-1.94). There was also only minor variation when further stratified by sex, after 5 yr of follow-up and childhood or adulthood diagnosis.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: