Weight loss is associated with reduced risk of knee and hip replacement: a survival analysis using Osteoarthritis Initiative data

Key points

1. 01Every 1% weight loss linked to 2% reduced knee replacement risk (adjusted HR 1.02 per 1% weight gain; 95% CI 1.00-1.04)
2. 02Every 1% weight loss linked to 3% reduced hip replacement risk, but only in hips that were persistently painful at baseline (adjusted HR 1.03; 95% CI 1.01-1.05)
3. 03A 10% weight loss was associated with roughly 20% reduced knee replacement risk and 30% reduced hip replacement risk (painful hips)
4. 04The benefit was seen across all BMI categories, not just in overweight or obese individuals
5. 05Population-level modelling suggested 10% weight loss among people with overweight/obesity could reduce knee replacements by 17% and hip replacements by 28%

How it was conducted

Design

Time-to-event survival analysis (Cox proportional hazard models) using data from the Osteoarthritis Initiative (OAI), a prospective population-based cohort

Participants

4081 participants (8069 knees) for knee analysis; 4064 participants (8076 hips) for hip analysis; 59.3% female; mean age 60.8 years; mean BMI 28.7 kg/m2

Setting

Four clinical research centres in the United States (OAI data)

Exposure

Percent body weight change from baseline, measured at repeated follow-up visits over 8 years (baseline, 1, 2, 3, 4, 6, and 8 years)

Primary outcome

Incidence of primary knee or hip replacement during 8-year follow-up

Follow-up

Median 7.9 years (IQR 7.8-8.0)

What they found

• Adjusted HR for knee replacement per 1% body weight increase from baseline: 1.02 (95% CI 1.00-1.04), p=0.03; each 1% weight loss associated with approximately 2% reduced risk of knee replacement
• Adjusted HR for hip replacement per 1% body weight increase, in hips persistently painful at baseline: 1.03 (95% CI 1.01-1.05), p<0.01; each 1% weight loss associated with approximately 3% reduced risk of hip replacement
• No significant association between body weight change and hip replacement in hips that were not persistently painful at baseline (adjusted HR 1.00, 95% CI 0.97-1.03, p=0.86)
• No significant interactions between body weight change and baseline BMI category for either knee or hip replacement
• There were 423 knee replacements in 8069 knees (5.2%) and 181 hip replacements in 8076 hips (2.2%) during follow-up
• Population attributable fraction: 5% weight loss (BMI >=25 kg/m2) could reduce knee replacement risk by 9% (95% CI -17%) and hip replacement risk in painful hips by 16% (95% CI -27%)
• Population attributable fraction: 10% weight loss (BMI >=25 kg/m2) could reduce knee replacement risk by 17% (95% CI -30%) and hip replacement risk in painful hips by 28% (95% CI 8%-44%)
• Sensitivity analyses (multiple imputation, complete case, competing risk for death, exclusion of underweight BMI) yielded consistent results

Limitations

• Observational design prevents causal inference; residual confounding cannot be excluded
• Weight change was recorded only at scheduled OAI visits, so gradual weight fluctuations between visits may not be fully captured
• The hip benefit was only observed in those with persistent hip pain at baseline, limiting generalisability to people without hip pain
• Participants were recruited from clinical research centres in the United States, which may limit generalisability to other populations or healthcare settings

Why it matters

For patients

People with knee osteoarthritis or at risk of it may substantially lower their chance of needing a knee or hip replacement by losing even modest amounts of weight, and this benefit appears to apply regardless of starting BMI.

For clinicians

Weight loss should be recommended as a joint-preservation strategy not just for obese patients but across BMI categories; for hip osteoarthritis, the replacement-risk benefit of weight loss appears confined to patients who already have persistent hip pain.

For readers

This large, well-adjusted survival analysis strengthens the evidence base for weight management in osteoarthritis guidelines and provides quantified population-level projections that could inform health-economic decisions about weight-loss interventions.