You are here
Some patients typically contraindicated for prescription nonsteroidal anti-inflammatory drug (NSAID) use may be able to take them with no increased risk of harm, according to a large observational study.
After a primary care visit for a musculoskeletal disorder, 9.3% of high-risk patients -- those with hypertension, heart failure, or chronic kidney disease -- age 65 and older in Ontario started taking prescribed NSAIDs (mostly traditional rather than selective ones).
NSAID use in this group was not associated with any impact on clinical outcomes out to 37 days following the outpatient visit, reported R. Sacha Bhatia, MD, MBA, of Toronto's Women's College Hospital, Ontario, and colleagues online in JAMA Internal Medicine.
According to an analysis of 35,552 propensity score-matched pairs within the overall cohort of 814,049, rates were:
- Cardiac complications: 0.8% for both NSAID users and non-users
- Renal complications: 0.1% for both
- Death: 0.1% for both
- Cardiovascular or renal-safety related outcomes: 0.9% versus 0.8% (P=0.74)
"The similarity in risk between users and non-users, each group primarily consisting of patients with hypertension, suggests that the short-term association of NSAIDs in high-risk patients with musculoskeletal pain may not be as dangerous as initially thought," Bhatia's group concluded. "Considering present concerns regarding opioid use for noncancer pain, the ability of physicians to prescribe NSAIDs to manage musculoskeletal pain in the short term could be an important clinical option in this patient population."
International nephrology societies currently recommend against NSAID use in patients with hypertension, heart failure, or chronic kidney disease. The American Society of Nephrology has a Choosing Wisely campaign against the practice, which Bhatia and colleagues suggested could be revisited (particularly when it comes to short-term NSAID use).
"Clinicians already recognize that many patients with heart failure, kidney disease, or hypertension tolerate a short course of NSAIDs without obvious harm," commented Jonathan Zipursky, MD, and David Juurlink, MD, PhD, both of Sunnybrook Health Sciences Centre in Toronto.
In an accompanying editorial, they said that Bhatia's study helps "soften our stance" on the use of NSAIDs in patients with some contraindications.
The retrospective cohort study relied on linked databases providing information on patients who visited primary care in 2012-2016 in Ontario. This group had a mean age of 75.3 years, and 61.1% were women.
NSAID use was gleaned from prescription claims data from the Ontario Drug Benefit program.
Physicians had widely varying prescribing patterns for NSAIDs in high-risk patients, ranging from 0.9% to 69.2% (median 11.0%).
Overall, there was a decline in prescription NSAID use over time province-wide (2.0% decrease per quarter, RR 0.98, 95% CI 0.98-0.99), according to Bhatia's group.
"A major strength of this study is that it provides a real-world estimate of the extent to which NSAIDs are prescribed to patients notionally at increased risk of harm. This long-standing perception originates mainly from case reports, observational studies, and older trial data of patients taking long-term NSAIDs," Zipursky and Juurlink said.
The study was not without limitations, however, as "readers are left questioning whether patients who were not prescribed NSAIDs were simply taking over-the-counter medications (including NSAIDs), or were in fact prescribed other analgesics such as opioids," the editorialists cautioned.
Moreover, the databases used lacked critical information such as presence of symptoms, Bhatia and colleagues noted, adding that propensity matching may not have fully account for unmeasured differences between NSAID users and non-users.
"Done well, observational studies can provide valuable information about the real-world safety of drugs," Zipursky and Juurlink maintained, noting that national initiatives such as the 21st Century Cures Act are spurring more observational studies that may challenge long-standing assumptions about treatments or help researchers discover new uses for old therapies.
"Yes, the findings will need to be contextualized and viewed with more skepticism than randomized controlled trials, but in some instances, they can be thoughtfully integrated into our treatment decisions," they emphasized.
Bhatia, Zipursky, and Juurlink disclosed no conflicts of interest.