Ask the Expert: Do NSAIDs Cause More Deaths Than Opioids?
Question: I have recently read a lot on Twitter that NSAIDs cause more deaths than opioids. Can you clarify if that is true?
Answer: Balancing the benefits of well-controlled pain with the risks of analgesic use is complicated and often challenging for health care professionals and patients alike. Although there are a number of therapeutic options for the treatment of pain, none are without risk. Concerns of hepatotoxicity with the use of acetaminophen, overdose and dependence with opioids, and gastrointestinal (GI)-related events with nonsteroidal anti-inflammatory drugs (NSAIDs) force prescribers to critically evaluate the risks and benefits for every patient. The Food and Drug Administration (FDA) has developed initiatives to improve the safe use of each of these analgesic types; however, the comparative safety of these agents is difficult to determine.1,2 Recently, debate has ensued over whether drug-induced mortality is higher with NSAIDs or opioids.
In the US, overdose mortality data are collected in the National Vital Statistics System.3,4 According to an analysis of these data by the Centers for Disease Control and Prevention (CDC), 36,450 deaths were caused by drug overdoses in 2008.3 Nearly three-quarters of the 20,044 prescription drug-related deaths (14,800) were linked to opioid pain relievers. This equates to 4.8 deaths due to opioid overdose per 100,000 people in the US. By 2010, an estimated 16,651 overdose deaths were linked to opioid pain relievers; 6,631 in women and 10,020 in men.4 This represents an alarming increase from 1999—415% for women and 265% for men.
Mortality reported with NSAID use is generally linked to NSAID-associated GI bleeding, and the reported incidence is quite variable. Much of the literature reports 16,500 deaths annually as a result of NSAID-induced GI bleeding. However, it is important to put this number into perspective. This estimate was published in a 1999 observational study evaluating data from the Arthritis, Rheumatism, and Aging Medical Information System (ARAMIS).5 The database captured 19 NSAID-related GI bleeding deaths; the authors then extrapolated this number to the overall US population with arthritis. The 16,500 deaths has been considered overestimated or inaccurate because it was based on a small number of actual deaths and extrapolated inappropriately.6,7 A much smaller number was reported by Tarone and colleagues.7 They reported 3,200 deaths annually as a result of NSAID-induced GI bleeding based on US mortality data reported in the 1990s.
Some authors have inappropriately used any combination of the above numbers to directly compare the safety of NSAIDs and opioids. The above estimates were made using different patient populations at different periods of time and should not be directly compared. A more accurate comparison comes from an observational study by Solomon and colleagues.8 In this study, the authors sought to compare the relative safety of analgesic medications in older patients with arthritis. The authors identified all-cause mortality for Medicare beneficiaries with a diagnosis of rheumatoid arthritis or osteoarthritis who were receiving a non-selective NSAID, a selective cyclooxygenase-2 (COX-2) inhibitor, or an opioid.
All-cause mortality was higher in patients receiving opioids than other analgesics. An overall mortality incidence rate of 48/1,000 person-years was reported for patients taking non-selective NSAIDs compared with 75/1,000 person-years with opioids. Coxibs were associated with an elevated risk of cardiovascular disease, but a reduction in GI bleeding compared to non-selective NSAIDs. Coxibs did not raise the risk of all-cause mortality.8
Although this is a more accurate comparison of mortality between opioids and NSAIDs, it is not without limitations. First, despite propensity score matching for a variety of disease states, it is possible that patients receiving opioids were in poorer health than those receiving NSAIDs. Second, there are some unknown factors including the use of nonprescription pain relievers, the type of NSAID or opioid used, as well as the dose or duration of the analgesic. The authors provided a list of medications used in supplemental materials, which included 10 different NSAIDs, 2 selective COX-2 inhibitors (coxibs), and 8 opioids. No further analysis was performed based on individual agents.
Finally, although this study is a more direct comparison, it was conducted in a very specific patient population (elderly patients with arthritis), and the mortality rates are not applicable to the general population with pain.
In summary, it does not appear that NSAIDs cause more deaths than opioids; however, the evidence is poor. The comparative safety of NSAIDs and opioids is difficult to assess without a well-designed clinical trial. Since no trial of this scope is likely to be conducted, it is imperative that prescribers consider the patient-specific risks and benefits of each class of medication prior to prescribing. The dramatic increase in opioid-related overdose deaths is worthy of consideration, but should not preclude appropriate prescribing of these agents for patients who require this level of analgesia.
Clinical Assistant Professor
Department of Pharmacy Practice
University of Illinois at Chicago