P663 The effect of vitamin D deficiency, and its correction, on healthcare utilisation among patients with inflammatory bowel disease (IBD)
N. Chou1, J. Gubatan2, O. H. Nielsen3, A. Moss*1
1Beth Israel Deaconess Medical Center, Gastroenterology, Boston, USA, 2Stanford University Medical Center, Gastroenterology, Stanford, USA, 3University of Copenhagen, Clinical Medicine, Copenhagen, Denmark
Vitamin D (25(OH)D) deficiency is frequent in patients with inflammatory bowel disease (IBD). Observational studies have associated this with an increased risk of flaring disease, although under-sized supplementation studies have not shown improved clinical outcomes. Here, we undertake a case–control study to compare healthcare utilisation in patients with sustained vitamin D deficiency to those whose 25(OH)D levels have been corrected.
An electronic medical record (EMR) search query was used to identify all patients at a single medical centre with a diagnosis code for IBD and two recorded 25(OH)D levels (at T0 and T1) within an 18-month time window. The EMR of this cohort was searched for all clinic, emergency room (ER), and hospital visits in the 24 months after T0. Patients were grouped by 25(OH)D levels to three groups; ‘low’ ( < 20 ng/ml), ‘high’ ( >30 ng/ml), and ‘restored’ (T0 < 20 ng/ml, T1 >30 ng/ml). A point was assigned for each IBD-related healthcare visit to determine healthcare utilisation. Logistic regression modelling was used to find propensity scores between ‘low’ and ‘restored’ groups and to estimate a treatment effect of vitamin D correction on healthcare utilisation, when controlled for confounding variables.
Data on 332 patients were analysed; mean age 58 years, 65% female, and 83% white. Ulcerative colitis (UC) and Crohn’s disease (CD) were distributed 44% and 56%. The ‘low’ cohort consisted of 31 patients (M = 14 ng/ml), the ‘high’ cohort had 105 patients (M = 48 ng/ml) and the ‘restored’ included 51 patients (M = 41 ng/ml after correction). Outcomes from the rest of the cohort (those who had insufficient change in 25(OH)D or had decreased levels) were not analysed for the purpose of this study. Mean number of IBD-related clinic visits was 3.33, and 15% had more than one IBD-related ED visit or hospitalisations over 24 months. ‘Low’ vitamin D status was not associated with an increased risk of IBD-related clinic, ER, or hospital visits, when compared with ‘high’ vitamin D status patients. Patients who converted from ‘low’ to ‘high’ vitamin D status (‘restored’) increased their mean 25(OH)D from 15 ng/ml to 41 ng/ml. There was no statistical difference in any outcome between matched controls (‘low’) and cases (‘restored’) over 24 months follow-up. Power for these analyses was confirmed to be >0.8 for dichotomous outcomes based on this sample size and proportions.
Persistently ‘high’ (>30 ng/ml) vitamin D levels, or restoration of levels from < 20 ng/ml to >30 ng/ml, did not reduce IBD-related healthcare utilisation over 24 months in this observational study of more than 300 patients. A higher goal for levels may be necessary to determine whether there is an effect of vitamin D replacement on IBD outcomes.