DOP019. The relevance of population based IBD biobanks: a meta-analysis and introduction of the IBD-SL biobank cohort
T. van den Heuvel1,2, D. Jonkers1,2, S. Jeuring1,2, W. Hameeteman1, L. Oostenbrug3, M. Romberg-Camps4, A. Masclee1,2, M. Zeegers5,6, M. Pierik1,2, 1Maastricht University Medical Center+, Internal Medicine - Division Gastroenterology-Hepatology, Maastricht, Netherlands, 2Maastricht University Medical Center+, NUTRIM, School for Nutrition, Toxicology and Metabolism, Maastricht, Netherlands, 3Atrium Medical Center, Internal Medicine and Gastroenterology, Heerlen, Netherlands, 4Orbis Medical Centre, Gastroenterology and Hepatology, Sittard-Geleen, Netherlands, 5Maastricht University Medical Center+, Complex Genetics, Cluster of Genetics and Cell Biology, Maastricht, Netherlands, 6University of Birmingham, Public Health, Epidemiology and Biostatistics, Birmingham, United Kingdom
To investigate IBD pathophysiology and markers for predicting disease course, biobanks with extensively phenotyped data of unselected IBD populations are warranted. Since no such biobanks are available, we started a prospective population based (PB) biobank cohort in South Limburg (IBD-SL), The Netherlands. Here we report the design of the IBD-SL biobank cohort, and the characteristics of included ulcerative colitis (UC) patients. Additionally, we aimed to test the widely-used assumption that hospital based (HB) settings comprise a different subset of IBD phenotypes than PB settings, by using the UC characteristics of IBD-SL in a meta-regression analysis.
Since 1991, incident IBD patients attending one of 3 regional hospitals (1 academic, 2 general hospitals) were enrolled in IBD-SL. Only adult IBD patients residing in South Limburg were eligible. Missed patients were captured by scrutinizing the National Pathology Database and cross-checks with GPs, resulting in a completeness of 93%. Data on IBD phenotype, medication, surgery, extra intestinal manifestations, and demographics are collected at time of diagnosis and during follow-up. Since 2009, plasma, serum, faeces, and biopsies are being biobanked. Additionally, we performed a meta-regression analysis, including IBD-SL data, to calculate the pooled proportion of proctitis (E1), left-sided colitis (E2) and extensive colitis (E3) in PB settings (N = 10) versus HB settings (N = 13; DL random effect, Ln-transformation).
A total of 1748 UC patients (53% male) patients were included in IBD-SL. Mean age at diagnosis and mean disease duration were 44.2 (SD 16.8) and 9.6 years (SD 6.0), respectively. Proportions of E1, E2 and E3 at diagnosis were 35%, 47%, and 18%, respectively. Ninety-eight percent ever used mesalamine, 32% immunomodulators and 10% anti-TNF agents. Eight percent had undergone (partial) colectomy. Ten percent ever had an extra intestinal manifestation (skin, eyes, joints and/or liver) during disease course, of which artropathy (3%) was most frequent. Additionally, pooled E1 proportion was higher for PB (39%, 95% CI 33–45) than for HB (21%, 95% CI 17–27, p < 0.001). Pooled E3 proportion was lower for PB (24%, 95% CI 21–28) than for HB (35%, 95% CI 30–42, p < 0.001).
Biobanks should be representative for the entire IBD population when studying pathophysiology and markers for predicting disease course. Observed phenotypic differences between population based and hospital based settings support the relevance of population based biobanks, such as IBD-SL.
- Posted in: DOP Session 3 - Epidemiology in practice