GLP-1 Drugs Cut Need for Knee Replacement

People with knee osteoarthritis (OA) who used GLP-1 receptor agonists were less likely to undergo knee replacement surgery, an analysis of a large records database indicated.

Compared with otherwise similar OA patients not taking GLP-1 drugs, the 8-year probability of undergoing knee arthroplasty was 2.8 percentage points lower among those who did use them for 1 year, according to Jay Karri, MD, MPH, of the University of Maryland School of Medicine in Baltimore, and colleagues.

People taking newer GLP-1 drugs such as tirzepatide (Mounjaro, Zepbound) or semaglutide (Ozempic, Wegovy, Rybelsus) for 3 years had an even greater reduction in arthroplasty risk: 4.7 percentage points by year 8, the group reported in Regional Anesthesia and Pain Medicine.

“These sustained and duration dependent associations, supported by preclinical evidence of joint tissue modulation and clinical evidence of analgesic benefits, are consistent with the possibility of effects beyond symptomatic relief or weight loss alone,” Karri and colleagues wrote.

But this possibility remains just that. The data, from the TriNetX storehouse of millions of U.S. electronic health records, did not include changes in individual patients’ weight over time nor could it be determined why the drugs were started in the first place. It’s possible that users were more interested in healthy behaviors generally and found other ways to reduce their need for arthroplasty besides GLP-1 drugs.

Indeed, Karri’s group acknowledged that “more robust prospective studies incorporating defined patient phenotypes, GLP-1 [agonist] dosing and duration, and objective measures of knee OA progression are required to confirm these associations and guide clinical care.”

Their study relied on data from some 4.1 million people whose records included diagnostic codes for knee OA; of these, 42,062 also had codes indicating that they were prescribed GLP-1 drugs for at least 1 year, starting within 3 years of a first OA diagnosis. For the control group, the researchers picked out 42,062 unexposed people who best matched the GLP-1 users through age, sex, race, body mass index (BMI) category (<20, 20-29.9, 30-39.9, ≥40), and broad categories of comorbidities and social determinants of health (ICD-10 codes Z00-Z99, e.g., healthcare encounters, "social environment"-related problems).

Mean age in both groups was 60, two-thirds were women, and 70% were white. About 60% had codes for obesity-related disorders, which are separate from those for BMI categories.

Among GLP-1 users, 28,599 were prescribed either tirzepatide or semaglutide. They did not differ substantially from the overall group of 42,062 by age, sex, or other factors. This was also the case for the subset of users whose prescriptions extended to 3 years or beyond (30,981), although fewer than half of these had been prescribed the newer GLP-1 agents.

Karri and colleagues didn’t report absolute rates of subsequent knee replacement surgeries, only differences between users and non-users as expressed by hazard ratios and percentage points. So, for use of any GLP-1 drug for at least 1 year, the hazard ratio for 8-year risk of arthroplasty was 0.90 (95% CI 0.83-0.98) for users versus non-users and the difference between rates was 2.80 percentage points. Other comparisons were as follows:

• Any GLP-1 drug, 1-year use, 3-year risk: HR 0.81 (95% CI 0.74-0.88), difference 1.39 points
• Any GLP-1 drug, 3-year use, 3-year risk: HR 0.85 (95% CI 0.80-0.92), difference 2.33 points
• Any GLP-1 drug, 3-year use, 8-year risk: HR 0.85 (95% CI 0.80-0.90), difference 3.25 points
• Newer GLP-1 drugs, 1-year use, 3-year risk: HR 0.72 (95% CI 0.62-0.83), difference 2.31 points
• Newer GLP-1 drugs, 1-year use, 8-year risk: HR 0.74 (95% CI 0.64-0.85), difference 3.44 points
• Newer GLP-1 drugs, 3-year use, 3-year risk: HR 0.65 (95% CI 0.59-0.70), difference 2.99 points
• Newer GLP-1 drugs, 3-year use, 8-year risk: HR 0.72 (95% CI 0.67-0.78), difference 4.71 points

These analyses did not include longitudinal tracking of factors that could influence decisions to undergo arthroplasty, particularly body weight. Nevertheless, the researchers concluded that GLP-1 drugs, and especially semaglutide and tirzepatide, may have exerted “sustained metabolic and anti-inflammatory” effects and thus have “potential disease-modifying effects in knee OA,” even though this was not explicitly demonstrated.

But if their speculations are correct, use of GLP-1 drugs for patients with knee OA would have a massive impact on U.S. health economics. Karri and colleagues noted that the 1-year risk reduction of 1.44 percentage points seen with newer GLP-1 drugs used for 3 years “translates to approximately 14,400 fewer [arthroplasties] annually in the U.S., with corresponding reductions in healthcare expenditures and surgical morbidity.”

Besides the limitations noted above, the study results could have been affected by unmeasured factors such as OA severity as well as the potential for patient data recording errors and omissions. Changes over time in OA management and GLP-1 drug prescribing practices could also have influenced the results.