Editorial: Association of Antibiotic Exposure With Microscopic Colitis

Abstract

Soon after microscopic colitis (MC) was first described, reports emerged of cases that appeared to be due to the use of certain medications. As case reports accumulated, the topic gained further attention. At one point, a sophisticated imputation score was created in an attempt to assess the likelihood of causality of a particular drug or drug class as a cause of MC with the conclusion that eight drugs or drug classes had a high likelihood of inducing MC and seven had an intermediate risk [1]. Combinations of these drugs, such as non-steroidal anti-inflammatory drugs (NSAIDs) together with proton pump inhibitors (PPIs), may exacerbate the risk [2-4]. In addition to drug-induced MC scoring systems, various other clinical predictive models have been proposed to identify patients unlikely to have MC and in whom colon biopsies can potentially be avoided. More recently, immune checkpoint inhibitors used in oncology have also been associated with the development of MC [5].

A meta-analysis of 12 case–control studies demonstrated modest odds ratios suggesting that exposure to NSAIDs, PPIs, SSRIs and aspirin was associated with the incidence of MC [6]. However, several studies evaluating this association demonstrated that the risk was significantly mitigated when controls with diarrhoea—rather than healthy subjects—were used for comparison [7, 8] indicating that the association with some of these medications was not causative but, rather, due to the medications exacerbating diarrhoea and thereby bringing milder cases to medical attention.

Szilcz and colleagues have reported a modest association with antibiotic use and the development of MC in a large national case–control study of older adults in Sweden [9]. To control for detection bias, they also performed an analysis of the association between antibiotic use and normal colon biopsies. In this control group, the association with antibiotic exposure was even stronger indicating that the observed association with MC was probably due to detection bias. Similar to the studies discussed above, these results reinforce the need to consider the control group when studying associations between drug exposure and the development of MC.

Therefore, accumulating evidence linking medication use to the development of MC indicates that the association for many drugs may be more confounded than causal [10]. This observation has important implications for understanding the underlying pathophysiological mechanisms underpinning the development of MC. However, the implications for clinicians caring for these patients are less clear. Specifically, if a patient is being evaluated for diarrhoea, whether they have biopsy-proven MC or not, a careful review of their medication list, including over-the-counter medications, is essential. If any drug might be temporally associated with the onset of diarrhoea, consideration should be given to stopping it and, if necessary, switching to another medication with similar effects but ideally from a different drug family. In many patients, diarrhoea will resolve or at least improve to be easier to manage with anti-diarrhoeal therapy rather than corticosteroids or immunomodulatory medications.