Oral opioids, including the less potent tramadol and the more potent hydromorphone and oxycodone, do not provide better pain relief than oral nonsteroidal anti-inflammatory drugs (NSAIDs) for knee osteoarthritis (OA), according to a systematic review published February in Osteoarthritis and Cartilage.
“Patients who failed NSAIDs and are considering opioids should be aware that the pain relief from opioids is likely to be no greater than the pain relief they had with NSAIDs, but may provide ‘an extra chance’ before considering total knee replacement,” senior author Elena Losina, PhD, professor of orthopedic surgery at Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts, told Medscape Medical News.
Nearly all patients with knee OA require long-term pharmaceutical pain control, beginning most commonly with NSAIDs or with the oral low-potency opioid tramadol. However, few head-to-head comparisons of NSAIDs and opioids have been performed in this setting, which inspired Savannah R. Smith, BA, from the Orthopaedic and Arthritis Center for Outcomes Research, Department of Orthopaedic Surgery, Brigham and Women’s Hospital, and colleagues to conduct a systematic review of randomized controlled trials that tested the different therapies.
The researchers identified 17 studies that met inclusion criteria, which included randomized controlled trials published in English between 1982 and 2015 that evaluated oral NSAIDs or opioids for knee OA; were of at least 8 weeks’ duration; were conducted in Western Europe, the Americas, New Zealand, or Australia; and used the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) Pain subscale to assess baseline and follow-up pain. These included 11 studies of NSAIDs only, three of tramadol or tramadol/acetaminophen, two of hydromorphone or oxycodone, and one of celecoxib and tramadol. Together, there were 27 active treatment groups (nine celecoxib, one diclofenac, two naproxen, one piroxicam, 10 tramadol, one tramadol/acetaminophen, two hydromorphone, and one oxycodone). Median trial duration was 13 weeks for NSAIDs and 12 weeks for opioids (P < .01).
Baseline demographics and efficacy withdrawal rates were similar for the NSAID and opioid studies. Baseline pain and toxicity withdrawal rates were slightly lower in NSAID studies.
Corrected for efficacy-related withdrawals, all drug classes were associated with pain reductions of 18 to 19 points on the WOMAC Pain scale. Regression analysis revealed no difference in effectiveness by drug class. Specifically, there were no significant differences in knee OA pain reduction for NSAIDs vs less potent opioids, for NSAIDs vs potent opioids, or for less potent opioids vs more potent opioids.
“Our hypothesis was that opioids would provide a greater amount of analgesic benefits, and we were surprised that we did not see such a difference,” Dr Losina said.
The researchers were also surprised to find that the placebo effect might be influenced by the type of comparator drug being tested, such that the placebo effect appeared to be greater in trials testing opioids than in those testing NSAIDs. “The differential placebo effect of opioids was…a very novel finding that should be confirmed by other studies,” Dr Losina said.
“This is a very interesting observation that complements and expands on growing evidence that ‘placebo’ or ‘intention’ effects of different treatments differ by the intensity of the treatment delivery,” Dr Losina explained. “For example, placebo effect of surgery is greater than placebo effect of injection, which in turn is greater than placebo effect of a pill. What we saw is some evidence that even within the same ‘delivery’ class — pills — the placebo or intention effect depends on anticipated potency of the analgesic regimen.”
She added, “If confirmed, [this] has large implications for understanding of the mechanisms of how analgesics work.”
David S. Jevsevar, MD, MBA, chair, Department of Orthopaedics, The Geisel School of Medicine at Dartmouth, and regional vice-president of orthopaedics, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire, told Medscape Medical News that the systematic review and network meta-analysis performed by Smith and colleagues was a challenging approach and that the authors had not overlooked anything.
“In today’s world, it is impossible to perform head-to-head [randomized controlled trials] on all possible combinations. The [network meta-analysis] approach is the next best thing. The results do support equivalency,” he said.
“In general, the use of opioid analgesics for the treatment of OA should be discouraged. The national health implications of addiction secondary to this trend are significant, especially in states like mine,” he added.
“Unfortunately, NSAIDs have significant side effects, so the perfect answer doesn’t exist.”