DENVER – Using antidepressants to treat osteoarthritis pain can benefit some individuals but appears to have a clinically unimportant reduction in pain when looking at all patients who have tried them, according to a study presented at the OARSI 2023 World Congress. The review was also published in the Cochrane Database of Systematic Reviews in October 2022.
In terms of implications for clinical practice, the findings “seem to suggest there is a subgroup that is more likely to respond to antidepressants,” Anita Wluka, PhD, MBBS, a professor in the School of Public Health and Preventive Medicine at Monash University in Melbourne, told attendees. The findings also raise an important research question: “How can we identify the patient phenotype likely to benefit so we can [minimize the] risk of those adverse events and effects?”
Osteoarthritis pain is heterogeneous, and an estimated 30% of the pain is neuropathic-like, likely including central and peripheral sensitization, Dr. Wluka said. Given that antidepressants affect multiple sites along these pathways, multiple organizations have issued a conditional recommendation for duloxetine in their osteoarthritis guidelines, including OARSI, the European Alliance of Associations for Rheumatology, and the American College of Rheumatology.
The Cochrane Collaboration therefore conducted a systematic review and meta-analysis of research on the benefits and harms of using antidepressants to treat symptomatic knee and hip osteoarthritis. The review included studies through January 2021 whose participants had knee and/or hip osteoarthritis and which compared antidepressant therapy with placebo or another intervention for at least 6 weeks. The authors looked at seven outcomes: overall pain on a 0-10 scale, clinical response (at least a 50% reduction in 24‐hour mean pain), physical function using the Western Ontario and McMaster Universities Arthritis Index (WOMAC), quality of life using the EQ-5D, the proportion of participants withdrawing because of adverse events, the proportion who experienced any adverse events, and the proportion who experienced serious adverse events.
The researchers considered a change on the pain scale of 0.5-1 points to be “slight to small,” a difference above 1 up to 2 to be “moderate,” and a difference greater than 2 points to be “large.” In assessing quality of life function on a scale of 0-100, a slight to small difference was 5-10, a moderate difference was 11-20, and a large difference was above 20.
Of the 18 articles the researchers identified for qualitative synthesis, 9 met the criteria for qualitative synthesis in the meta-analysis, including 7 studies only on the knee and 2 that included the knees and hips. All nine studies compared antidepressants with placebo, with or without NSAIDs. Most focused on serotonin and norepinephrine reuptake inhibitors (SNRIs) – six studies on duloxetine and one on milnacipran – while one included fluvoxamine and one included nortriptyline.
The trials included a combined 2,122 participants who were predominantly female with an average age range of 54-66. Trials ranged from 8 to 16 weeks. Five of the trials carried risk of attrition and reporting bias, and only one trial had low risk of bias across all domains.
In five trials with SNRIs and one trial with tricyclics (nortriptyline) totaling 1,904 participants, 45% of those receiving antidepressants had a clinical response, compared with 29% of patients who received placebo (risk ratio, 1.55; 95% CI, 1.31-1.92). This absolute improvement in pain occurred in 16% more participants taking antidepressants, giving a number needed to treat (NNT) of 6. Average improvement in WOMAC physical function was 10.5 points with placebo and 16.2 points with antidepressants, indicating a “small, clinically unimportant response,” the researchers concluded.
Withdrawals because of adverse events included 11% of the antidepressant group and 5% of the placebo group (RR, 2.15; 95% CI, 1.56-2.87), putting the NNT for a harmful outcome at 17.
For all nine trials together, however, the mean reduction in pain from antidepressants was 2.3 points, compared with 1.7 points with placebo, a statistically significant but ”clinically unimportant improvement,” the researchers concluded. Adverse events occurred in 64% of the antidepressant group, compared with 49% of the placebo group (RR, 1.27; 95% CI, 1.15-1.41), which put the NNT for a harmful outcome at 7. No significant difference in serious adverse events occurred between the groups.
The analysis was limited by the low number of trials, most of which were sponsored by industry and most of which used duloxetine. Further, few of the studies enrolled patients with osteoarthritis of the hip, none assessed medium- or long-term effects, and none stratified the participants for different types of pain (neuropathic-like or central or peripheral pain sensitization).
“My general impression is that there was a statistically significant difference found in favor of duloxetine and the antidepressants,” David J. Hunter, MBBS, PhD, MSc, of the University of Sydney, said after the presentation. “There is a real risk of harm, which I think is important to take into consideration, but at least for me as a clinician and in advising other clinicians, it’s one tool in our armamentarium. I think it’s really important to allow patients to make an informed decision about the potential benefit, the real risk of harm, and the fact that it is quite useful in some patients, and I use it in my clinical practice.”
Jeffrey N. Katz, MD, MS, of Brigham and Women’s Hospital in Boston, said he uses antidepressants in the same way in his practice and that other types of medications, such as TNF inhibitors, also carry risk of harm that may exceed that of antidepressants.
“I’ve had lots of people start duloxetine, and if they stop it, it’s usually because they just don’t tolerate it very well,” Dr. Katz said.
“We don’t want to throw too many things away,” Dr. Hunter added. “Our patients don’t necessarily have a lot of choices here from a pharmacologic perspective, so I think it’s one of those options that I want to keep in my tool kit, and that’s not necessarily going to change.”
The research did not involve outside funding, and Dr. Wluka reported having no industry disclosures. Disclosure information was unavailable for Dr. Katz and Dr. Hunter. The Congress was sponsored by the Osteoarthritis Research Society International.