Methotrexate Around Joint Replacement: Should We Really Stop It? New 10-Year Data Says Probably Not.

If you’ve spent any time in pre-op clinics for joint replacement, you’ve seen the hesitation. The orthopedic surgeon glances at the medication list, sees methotrexate, and the question immediately lands on the rheumatology consult: Do we hold this? For decades, the perioperative management of methotrexate has been more art than science—driven by a mix of historical caution, single-center experiences, and a nagging fear that immunosuppression plus a prosthetic joint is a recipe for disaster.

A new study published in Scientific Reports this year offers some of the most reassuring long-term data we’ve had on this front, and it’s worth a close look.

Ursini and colleagues leveraged the Emilia-Romagna Orthopedic Arthroplasty Implants Register (RIPO)—a large, well-established Italian registry with a 95% capture rate—to track patients with inflammatory arthritis (rheumatoid arthritis, psoriatic arthritis, and ankylosing spondylitis) who underwent primary total hip or knee arthroplasty while receiving perioperative methotrexate monotherapy. The definition of “perioperative” was pragmatic: at least one MTX prescription within 90 days before or after surgery. They then propensity-score matched these patients to OA controls undergoing the same procedures.

Importantly, they excluded anyone on biologics, JAK inhibitors, or other conventional DMARDs. This was intentionally a clean look at MTX alone.

The cohorts were substantial: 225 THA and 215 TKA patients in the MTX-treated IA group, matched 1:2 to OA controls. And here’s where it gets interesting.

At 10 years, implant survival was essentially identical.

For THA, survival was 93.9% in the MTX IA group versus 96.1% in OA—no significant difference (p = 0.302). For TKA, it was 93.7% versus 93.9% (p = 0.827). Even after accounting for death as a competing risk using Gray’s test, the curves stayed tightly overlapped.

But perhaps more striking was the steroid story. Over 40% of IA patients had received glucocorticoids in the 90 days pre-op, compared to under 7% of OA patients. We know steroids are a real risk factor for periprosthetic joint infection and revision. Yet despite this heavier burden of risk in the IA cohort, their long-term outcomes didn’t diverge. If anything, that’s an argument that MTX itself isn’t the bogeyman we sometimes treat it as.

The authors also ran an exploratory Cox model stratifying by when MTX was prescribed—only before surgery, only after, or both. No significant differences emerged. Granted, this is a secondary analysis and the study wasn’t powered specifically for this question, but it adds to the sense that obsessing over the exact stop-and-restart window may not be clinically meaningful for long-term implant survival.

These findings align neatly with the 2022 ACR/AHKS guidelines, which already conditionally recommend continuing MTX through the perioperative period. But guidelines based on limited evidence always feel a bit shaky in practice. What this registry study adds is real-world, decade-long durability data at scale. It’s not a randomized trial, and the authors are appropriately cautious about causal inference—they can’t tell us whether continuing MTX is definitively safe, only that being exposed to it perioperatively doesn’t appear to curse the prosthesis.

There are limitations, of course. The cohort was overwhelmingly RA (about 80%), so generalizing to PsA and AS is tentative. Dosing and exact discontinuation patterns were invisible to the researchers. And because they relied on prescription records, adherence is a black box. Still, in the messy world of perioperative medicine, this is a meaningful signal.

For clinicians still reflexively stopping methotrexate two weeks before surgery “just to be safe,” this study is another nudge to reconsider. The risk of disease flare from holding MTX is real and well-documented. If the trade-off is a slightly higher chance of a rheumatoid flare versus no demonstrable long-term penalty on implant survival, the math starts to favor continuation.

Of course, every patient is different. But the era of automatically scrubbing methotrexate from the surgical calendar may be quietly ending.

Ursini, F., Bordini, B., Ciaffi, J. et al. Ten-year hip and knee arthroplasty implant survival in patients with inflammatory arthritis receiving methotrexate monotherapy compared with osteoarthritis: a registry-based data linkage study. Sci Rep 16, 17481 (2026). https://doi.org/10.1038/s41598-026-48865-w