Are NSAIDs Genuinely Too Dangerous in Inflammatory Bowel Disease?

At least some patients with inflammatory bowel disease (IBD) can safely use common drugs for musculoskeletal aches and pains, a large study of insurance claims data suggested, contradicting a widespread concern that these products can trigger IBD flares.

With records of nearly 350,000 IBD patients available, risk of developing a disease flare requiring hospitalization was just slightly increased among those filling prescriptions for nonsteroidal anti-inflammatory drugs (NSAIDs) compared with matched nonusers (HR 1.07, 95% CI 1.04-1.10), according to Adam Mayer, MD, MSCE, of Hackensack Meridian Health in Nutley, New Jersey, and colleagues.

But the risk wasn’t the same for the two major IBD subtypes, the group reported in Arthritis Care & Research. Those with ulcerative colitis showed no increase at all in flare risk with NSAID use (HR 0.97, 95% CI 0.93-1.02) whereas Crohn’s disease patients had a correspondingly greater likelihood of IBD-related hospitalization (HR 1.16, 95% CI 1.12-1.21).

In comparing NSAID use with nonuse, the study protocol had defined noninferiority as a 95% confidence interval not exceeding 1.20 at the upper bound. Thus, it could not be said that NSAID use in Crohn’s disease patients was noninferior, although it was for ulcerative colitis and the IBD cohort as a whole.

“While further prospective study of NSAID safety in IBD is needed, the findings in this study suggest that the risk of NSAID exposure may be acceptable in select cases with strong indications for their use, such as those with significant musculoskeletal disease,” the group concluded.

Mayer and colleagues opened their report by noting that joint pain and inflammation are common in IBD patients. NSAIDs are typically the first-line treatment for most people with mild to moderate symptoms. But small studies conducted long ago found increased rates of disease flares in IBD patients using NSAIDs as opposed to acetaminophen, leading to a widespread belief that NSAIDs should be avoided in this population.

But while some of those studies were prospective, they weren’t randomized or placebo-controlled, and many were quite small, Mayer’s group indicated. Meanwhile, some later studies suggested that flare risk following NSAID use may be greater in Crohn’s disease versus ulcerative colitis. And most of the prior research didn’t distinguish among different categories of NSAIDs such as those selective for cyclooxygenase-2 (COX-2).

Therefore, Mayer and colleagues turned to a large insurance claims database maintained by Optum to get a sense of how prescription NSAID use may affect flare risk both for the different IBD subtypes, different NSAID classes, and other subgroups not previously examined with sufficient power.

Data for IBD patients spanning 2000-2022 were pulled, covering drug codes, prescription fills, emergency department visits, hospitalizations, and gastrointestinal surgeries. NSAID exposure was defined as a first filled prescription. Users were matched by age and year of diagnosis with two or three nonusers, who were followed via the records from the same date as the user’s initial prescription fill (i.e., the index date). All patients were followed until they experienced an IBD-related hospitalization lasting more than 1 day, left the Optum system, or stopped “continuous” NSAID use (defined as refilling prescriptions every 6 months or less). Exposure was considered to last until 6 months after the last filled prescription.

Mean patient age was 52 at the index date and just under 60% were women. Median follow-up was 0.6 years for users and 1.2 years for nonusers. About 44% had ulcerative colitis, 42% Crohn’s disease, and the rest an indeterminate form of colitis.

Half of NSAID users had diagnostic codes for joint pain and 56% had codes for chronic pain; corresponding figures for nonusers were 35% and 37%.

Among NSAID users, 87% took nonselective products such as ibuprofen; the remainder were given COX-2 selective agents. About one-third used opioids, and a similar number were taking corticosteroids. Of course, many were also on IBD-specific drugs such as sulfasalazine or biologic agents.

After adjusting for factors such as baseline clinical and demographic characteristics, Mayer and colleagues found, as noted above, that IBD-hospitalizations were slightly more common with NSAID use across all IBD patients taken together, and somewhat more so in the Crohn’s disease subgroup, whereas no risk increase was seen in ulcerative colitis. The researchers calculated that treating 71 Crohn’s patients with NSAIDs would lead to one IBD-related hospitalization (95% CI 56-100). No “number needed to harm” could be estimated for the ulcerative colitis group, although a lower bound to the 95% confidence interval of 333 was found.

No significant increases in gastrointestinal surgeries were seen with NSAIDs for Crohn’s disease or ulcerative colitis patients. On the other hand, risk of all-cause hospitalizations was definitely higher, by 29% in both disease subtypes. Mayer and colleagues said they weren’t surprised by the latter finding, insofar as NSAIDs are well-known to raise risk of problems such as acute kidney injury and cardiovascular events unrelated to IBD. They also observed that COX-2 selective products had the same risk profile for IBD-related hospitalizations as nonselective agents.

Probably the most important limitation to the study was that it couldn’t account for use of NSAIDs bought over-the-counter (OTC), although the authors said they “anticipate OTC use to be low given national society guidance against NSAID use in this population and that IBD patients are generally counseled by their gastroenterologists as such.” Other limitations included the possibility that users may have been (correctly) deemed lower risk by their clinicians, and that insurance claims data don’t capture all relevant factors and may contain errors.