Clinical Question

Are antidepressants safe and effective for the treatment of hip and knee osteoarthritis (OA)?

Evidence-Based Answer

Antidepressants provide a 50% or greater pain reduction in patients with hip and knee OA compared with placebo (number needed to treat [NNT] = 6; 95% CI, 4 to 11). (Strength of Recommendation [SOR]: A, consistent, good-quality patient-oriented evidence.) Antidepressants cause more adverse effects (number needed to harm [NNH] = 7; 95% CI, 5 to 11) compared with placebo.¹ (SOR: A, consistent, good-quality patient-oriented evidence.)

Practice Pointers

The estimated worldwide prevalence of OA of the hips and knees is 0.85% and 3.8%, respectively.² OA can lead to pain that affects physical and social function and mood, and it can be a financial burden. Although nonsteroidal anti-inflammatory drugs and intra-articular injection combined with nonpharmacologic interventions (e.g., structured exercise, weight loss) are commonly used to relieve pain, most patients are not satisfied with their pain control.^3,4 Antidepressants may inhibit sensory transmission in nociceptive fibers of the spinal cord through serotonin and norepinephrine reuptake inhibition.⁵ The authors of the review sought to determine if antidepressants improve outcomes in people with hip or knee OA compared with placebo.

The Cochrane review included nine double-blind randomized controlled trials with a total of 2,122 adult patients.¹ Participants 40 years or older who had knee and hip OA were recruited in the United States, Canada, Europe, Asia, and New Zealand. The review included only placebo-controlled trials with or without nonsteroidal anti-inflammatory drugs. Antidepressants included serotonin-norepinephrine reuptake inhibitors (duloxetine [Cymbalta], 60 to 120 mg; milnacipran [Savella], 200 mg), selective serotonin reuptake inhibitors (fluvoxamine, 150 mg), and tricyclic antidepressants (nortriptyline, 25 to 100 mg). Duration of the trials ranged from 10 to 14 weeks. Individuals with prior joint prostheses or concurrent psychiatric disorders (e.g., depression, anxiety, mania) were not included. Primary outcomes were pain at eight to 16 weeks (on a scale of 0 to 10 points, with 0 = no pain), responder rate (defined as a 50% or greater reduction in 24-hour mean pain level), physical function (total score on the Western Ontario and McMaster Universities Arthritis Index on a scale of 0 to 100, with 0 = best function), quality of life (score on the EuroQol 5-Dimension scale of −0.11 to 1, with 1 = perfect health), adverse effects, and study withdrawals due to adverse effects.

Compared with placebo, antidepressants provided slight and clinically unimportant pain relief at eight to 16 weeks (mean difference [MD] = −0.59; 95% CI, −0.88 to −0.31; nine trials; n = 2,038; high-certainty evidence). Antidepressants yielded an increased responder rate (NNT = 6; 95% CI, 4 to 11; six trials; n = 1,904; high-certainty evidence) compared with placebo. Another analysis revealed that antidepressants slightly improved physical function (MD = −5.65; 95% CI, −7.08 to −4.23; six trials; n = 1,909; high-certainty evidence) and quality of life (MD = 0.04; 95% CI, 0.01 to 0.07; three trials; n = 815; moderate-certainty evidence) compared with placebo, but this effect has questionable clinical meaning.

Participants in the antidepressant group withdrew from a study due to adverse effects (NNH = 17; 95% CI, 10 to 35; six trials; n = 1,977; moderate-certainty evidence) more often than those taking placebo. Antidepressants also caused more adverse effects (NNH = 7; 95% CI, 5 to 11; nine trials; n = 2,101; high-certainty evidence) but did not increase serious adverse effects compared with placebo (low-certainty evidence).

Limitations of the review included extensive exclusion criteria regarding concurrent psychiatric conditions, underrepresentation of hip OA (only two out of nine trials included hip OA), and the use of antidepressants other than duloxetine (only one trial each used milnacipran, fluvoxamine, and nortriptyline). Six of the nine trials were funded by a pharmaceutical company (all six funded trials were for duloxetine).

Among the antidepressants included in the studies, duloxetine is the only medication approved by the U.S. Food and Drug Administration for the treatment of chronic musculoskeletal pain.⁶ The recommended starting dosage of duloxetine is 30 mg daily; the dosage can be titrated up to 60 mg daily.⁶ Common adverse effects of duloxetine include nausea, constipation, xerostomia, dizziness, and fatigue. Although anti-depressants may provide modest pain relief, their use is associated with more adverse effects compared with placebo; therefore, family physicians should carefully consider antidepressants as a treatment option for OA with usual care options. Further research is needed to better understand the clinically important effectiveness and safety of antidepressants in patients with OA and concurrent psychiatric disorders.

The practice recommendations in this activity are available at https://www.cochrane.org/CD012157.