Methotrexate Fails in Knee Osteoarthritis Save

We'll be blunt: methotrexate flopped in a randomized placebo-controlled trial among patients with osteoarthritis (OA) of the knee.

Contrary to results from a previous trial in hand OA, the old-line disease-modifying anti-rheumatic drug (DMARD) failed to bring any improvement relative to placebo in either pain or an objective measure of joint inflammation after a year of treatment, according to Changhai Ding, PhD, MD, of Southern Medical University in Guangzhou, China, and colleagues.

Mean change from baseline in patient-reported pain scores differed by just 0.3 points on a 100-point scale (95% CI -6.7 to 7.3), the researchers reported in JAMA Internal Medicine. And improvement in maximal effusion-synovitis area was nearly identical as well, at -0.2 cm² in the methotrexate group versus -0.3 cm² with placebo (difference 0.1, 95% CI -0.8 to 1.0).

Ding and colleagues noted that among the few patients with severe knee pain at baseline, with initial scores of 80 out of 100 or higher, methotrexate did seem to provide more relief than placebo, and thus they stopped short of closing the door on the drug as an OA therapy.

But in an accompanying commentary, Nancy Lane, MD, of UC Davis Health in Hillsborough, California, did slam it shut. She pointed to several other trials of methotrexate in both hand and knee OA that showed little to no benefit, and that in those that did suggest efficacy, the doses were high enough to prompt safety worries. 

"Given the lack of efficacy of MTX [methotrexate] across knee OA studies and the known potential adverse events, it is not recommended for the treatment of painful, inflammatory knee OA," Lane wrote. "We now need to focus our attention on treatments that can both inhibit joint inflammation and stimulate chondrocytes within the cartilage to synthesize replacement matrix."

She added, "The future of pharmaceuticals for knee OA needs to move past MTX."

Many in the field will be disappointed, as finding a truly effective DMARD for osteoarthritis has long been a goal. Current treatments are only palliative; the only real remedies are painful surgeries that clinicians try to delay as long as possible, especially in younger people.

Methotrexate's prospects got a boost 2 years ago when researchers reported that the drug produced an average 15.2-point reduction in pain scores after 6 months in hand OA patients, compared with 7.7 points in the study's placebo group. Other subjective measures also showed a benefit with methotrexate.

Knee OA, however, is arguably a more clinically significant problem (in that it impairs patients' mobility and physical activity) and the literature on methotrexate's potential was more mixed. Thus, Ding and colleagues organized a randomized placebo-controlled trial with 215 knee OA patients from several regions in China.

To be eligible, patients needed to show clear signs of inflammation in the affected joint and report pain scores of at least 40 out of 100. Also, MRI scans had to show effusion-synovitis of at least grade 2.

Mean age for enrolled patients was about 61, and some 90% were women. More than 80% had Kellgren-Lawrence scores of 3 or 4; pain scores averaged 60. Mean maximal effusion-synovitis area at baseline stood at 3.3 cm² in the methotrexate group and 3.6 cm² among those assigned to placebo.

Besides pain and effusion-synovitis area assessments, outcome measures also included Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) scores, both total and for the three subdomains of pain, stiffness, and function, as well as evaluations of general health status. Methotrexate dosing started at 5 mg/week orally and was raised in steps over 4 weeks to a maximum of 15 mg/week.

In short, none of these measures showed even a hint of benefit for methotrexate over placebo. The protocol did include a prespecified analysis of patients with baseline pain scored at ≥80; among the nine members of this group on methotrexate, mean pain scores declined by 47 points after 1 year, compared with a mere 6-point average decline in the seven members on placebo. Similar differences were seen for other outcomes, too.

Ding and colleagues acknowledged, however, that with only 16 such patients, chance could account for these results. "Further studies are required to confirm" whether this was a real effect, they wrote.

Other limitations to the study included its conduct in China, where many patients were also using traditional herbal therapies, and these and other nonprescription products were not disallowed. As well, there were too few men in the study to allow conclusions about their outcomes. It remains possible, too, that methotrexate could still be effective in non-knee forms of OA.