Search in the Abstract Database

Abstracts Search 2014

* = Presenting author

P598. Proton pump inhibitor use and risk of microscopic colitis - A nationwide Danish case–control study with 5751 cases


The etiology of microscopic colitis (MC) is unknown but it has been reported that drug consumption may increase the risk of MC. Proton pump inhibitor (PPI) may trigger or cause the disease. This presumptive association is based on small case–


control studies. Based on Danish national registries we wanted to explore the potential association between exposure to PPI and the risk of MC.


All patients with a recorded diagnosis of either collagenous colitis (CC) or lymphocytic colitis (LC) in the Danish Pathology Register during the period January 2005 to December 2011 were identified. As controls 100 persons per case, matched on sex and age were randomly selected. None of the controls could have microscopic colitis. In the control group a subgroup of persons with a normal colonic biopsy were registered. The total sale of PPI in the study period expressed as defined daily dose (DDD)/1000 individuals/day was registered. For both cases and controls we also registered the individual PPI exposure from the Danish Prescription Registry in a one-year period preceding the case index date. Odds ratios of CC and LC according to prescription of PPI were calculated.

To assess the association between drug exposure and the risk of having an endoscopic examination with normal colonic biopsies, we compared PPI use within the control group with or without a colonic biopsy recorded.


The cases consisted of 5751 patients, hereof 3474 (60%) with CC and 2277 (40%) with LC. In the seven-year period the annual incidence of diagnosed cases of CC increased from 7.2/100.000 to 14.9/100.000 and of LC from 4.1/100.000 to 9.8/100.000. The sale of PPI increased from 24.8 to 53.3 DDD/1000inhabitants/day.

Compared to controls the observed OR between PPI and both CC (OR 8.15; 95% CI: 7.60–8.73) and LC (OR 4.11; 95% CI: 3.77–4.49) were highly significant.

In our control group a total of 2202 persons - 1332 were selected as controls for CC and 820 for LC - were registered with a normal colonic biopsy in the Pathology Register.. The association between PPI exposure and having a normal biopsy was significant for both groups (OR 3.64; 95% CI: 3.25–4.08 and OR 3.77; 95% CI: 3.27–4.35, respectively). Incorporating the results from this subgroup in our final model attenuated the association between PPI exposure and risk of MC. It remained significant in case of PPI and CC (adjusted OR 2.03; 95% CI: 1.77–2.33) but disappeared for LC (adjusted OR 0.95; 95% CI: 0.80–1.12).

  • Written by:

    O. Bonderup1, T. Wigh2, M. Fenger-Grøn3, G. Lauge Nielsen4, 1Diagnostic Center, Regional Hospital Silkeborg, Denmark, Section of Gastroenterology, Silkeborg, Denmark, 2Regional Hospital Randers, Medical Department, Randers, Denmark, 3Aarhus University, Research Unit for General Practice and Department of Clinical Epidemiology, Aarhus, Denmark, 4Hospital Himmerland, Farsø, Medical Department, Farsø, Denmark