Medications and Risk for Microscopic Colitis: A Nationwide Study of Older Adults in Sweden

Several medications have been identified as potential risk factors for microscopic colitis (MC), but evidence so far is hampered by methodological limitations.

To examine the potential causal effects of previously implicated medications on risk for MC.

Emulation of 6 target trials.

Sweden.

All residents in Sweden aged 65 years or older between 2006 and 2017 who met eligibility criteria (n = 191 482 to 2 634 777).

The primary outcome was biopsy-verified MC. The date of diagnosis was obtained from the nationwide histopathology cohort ESPRESSO (Epidemiology Strengthened by histoPathology Reports in Sweden). Twelve- and 24-month cumulative incidences and absolute risk differences for MC were estimated via inverse probability weighing.

The 12- and 24-month cumulative incidences of MC were less than 0.5% under all treatment strategies. Estimated 12-month risk differences were close to null under angiotensin-converting enzyme versus calcium-channel blocker (CCB) initiation, angiotensin-receptor blocker versus CCB initiation, nonsteroidal anti-inflammatory drug initiation versus noninitiation, proton-pump inhibitor initiation versus noninitiation, and statin initiation versus noninitiation. The estimated 12-month risk difference was 0.04% (95% CI, 0.03% to 0.05%) for selective serotonin receptor inhibitors (SSRIs) versus mirtazapine. Results were similar for 24-month risk differences. Several medications were also associated with increased risk for receiving a colonoscopy with a normal colorectal mucosa biopsy result.

There is possible residual bias due to differential health care utilization or surveillance. Lack of primary care data limited measurement of, and adjustment for, symptoms and medical diagnoses that increase risk for receiving a colonoscopy.

No evidence of a causal relationship between most previously suspected pharmacologic triggers and risk for MC was found. Previously reported associations and persistent associations with SSRI initiation may be due to surveillance bias.

The National Institutes of Health and the Swedish Research Council.