Benjamin Lebwohl1, Peter H R Green2, Louise Emilsson3, Karl Mårild4, Jonas Söderling5, Bjorn Roelstraete6, Jonas F Ludvigsson7. 1. Celiac Disease Center, Department of Medicine, Columbia University Medical Center, New York, New York; Department of Epidemiology, Mailman School of Public Health, Columbia University, New York, New York. 2. Celiac Disease Center, Department of Medicine, Columbia University Medical Center, New York, New York. 3. School of Medical Science, University of Örebro, Örebro, Sweden; Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden; Värmlands Nysäter Health Care Center and Centre for Clinical Research, County Council of Värmland, Sweden; Department of General Practice, Institute of Health and Society, University of Oslo, Oslo, Norway. 4. Department of Pediatrics, Institute of Clinical Sciences, Sahlgrenska Academy, Gothenburg, Sweden; Department of Pediatric Gastroenterology, Queen Silvia Children's Hospital, Gothenburg, Sweden. 5. Clinical Epidemiology Division, Department of Medicine Solna, Karolinska Institutet, Stockholm, Sweden. 6. Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden. 7. Celiac Disease Center, Department of Medicine, Columbia University Medical Center, New York, New York; Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden; Department of Pediatrics, Örebro University Hospital, Örebro University, Örebro, Sweden. Electronic address: jonasludvigsson@yahoo.com.
Abstract
BACKGROUND & AIMS: Celiac disease (CD) is associated with increased mortality, in part due to cancer. Most studies investigating this cancer risk involved patients diagnosed before widespread increases in CD diagnosis rates and access to gluten-free food. We performed a population-based study of the risk of cancer in CD. METHODS: We identified all patients in Sweden with CD as defined as duodenal/jejunal villus atrophy, using the Epidemiology Strengthened by histoPathology Reports in Sweden cohort. Each patient was matched to ≤5 controls by age, sex, and county. We used stratified Cox proportional hazards model, following patients from diagnosis until first cancer, or by December 31, 2016. RESULTS: Among 47,241 patients with CD, 30,080 (64%) were diagnosed since 2000. After a median follow-up of 11.5 years, the incidence of cancer was 6.5 and 5.7 per 1000 person-years in CD patients and controls, respectively. The overall risk of cancer was increased (hazard ratio [HR], 1.11; 95% confidence interval [CI], 1.07-1.15), but it was only significantly elevated in the first year after CD diagnosis (HR, 2.47; 95% CI, 2.22-2.74) and not subsequently (HR, 1.01; 95% CI, 0.97-1.05), although the risks of hematologic, lymphoproliferative, hepatobiliary, and pancreatic cancers persisted. The overall risk was highest in those diagnosed with CD after age 60 years (HR, 1.22; 95% CI, 1.16-1.29) and was not increased in those diagnosed before age 40. The cancer risk was similar among those diagnosed with CD before or after the year 2000. CONCLUSIONS: There is an increased risk of cancer in CD even in recent years, but this risk increase is confined to those diagnosed with CD after age 40 and is primarily present within the first year of diagnosis.
BACKGROUND & AIMS: Celiac disease (CD) is associated with increased mortality, in part due to cancer. Most studies investigating this cancer risk involved patients diagnosed before widespread increases in CD diagnosis rates and access to gluten-free food. We performed a population-based study of the risk of cancer in CD. METHODS: We identified all patients in Sweden with CD as defined as duodenal/jejunal villus atrophy, using the Epidemiology Strengthened by histoPathology Reports in Sweden cohort. Each patient was matched to ≤5 controls by age, sex, and county. We used stratified Cox proportional hazards model, following patients from diagnosis until first cancer, or by December 31, 2016. RESULTS: Among 47,241 patients with CD, 30,080 (64%) were diagnosed since 2000. After a median follow-up of 11.5 years, the incidence of cancer was 6.5 and 5.7 per 1000 person-years in CD patients and controls, respectively. The overall risk of cancer was increased (hazard ratio [HR], 1.11; 95% confidence interval [CI], 1.07-1.15), but it was only significantly elevated in the first year after CD diagnosis (HR, 2.47; 95% CI, 2.22-2.74) and not subsequently (HR, 1.01; 95% CI, 0.97-1.05), although the risks of hematologic, lymphoproliferative, hepatobiliary, and pancreatic cancers persisted. The overall risk was highest in those diagnosed with CD after age 60 years (HR, 1.22; 95% CI, 1.16-1.29) and was not increased in those diagnosed before age 40. The cancer risk was similar among those diagnosed with CD before or after the year 2000. CONCLUSIONS: There is an increased risk of cancer in CD even in recent years, but this risk increase is confined to those diagnosed with CD after age 40 and is primarily present within the first year of diagnosis.