Open Access Research article

Comparing the functional impact of knee replacements in two cohorts

Jingbo Niu1, Michael Nevitt2, Charles McCulloch2, James Torner3, C Elizabeth Lewis4, Jeffrey N Katz5, David T Felson1* and the Multicenter Osteoarthritis Study Group

Author Affiliations

1 From the Clinical Epidemiology Unit, Boston University School of Medicine, 650 Albany Street, Suite X200, Boston, MA 02118, USA

2 Department of Epidemiology and Biostatistics, University of California, San Francisco, USA

3 Department of Epidemiology, University of Iowa, Iowa City, Iowa, USA

4 Department of Preventive Medicine, University of Alabama, Birmingham, England

5 Department of Orthopedic Surgery and the Division of Rheumatology, Immunology and Allergy at the Brigham and Women’s Hospital, Boston, USA

For all author emails, please log on.

BMC Musculoskeletal Disorders 2014, 15:145  doi:10.1186/1471-2474-15-145

Published: 5 May 2014



To examine if different rates of total knee replacement (TKR) in two similar cohorts with symptomatic knee osteoarthritis (OA) were associated with different functional impact of disease.


Subjects from the Multicenter Osteoarthritis Study (MOST) and the Osteoarthritis Initiative (OAI), persons with or at high risk of OA, had knee radiographs, completed Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) surveys and had TKRs confirmed at each visit. At each visit, subjects were defined as having symptomatic OA (SxOA) if ≥ one knee had pain and radiographic OA or if they had a TKR. WOMAC function scores at each visit were compared by analysis of covariance adjusting for age, sex, body mass index, race, site, depression, comorbidity, painful leg joints and knees affected. Post-TKR function scores were imputed to estimate scores that would have been present without TKR.


Subjects with SxOA (n > 750 in MOST and in OAI) had a mean age 66 to 67 years; most were women and were White. Subjects were followed 4–5 years. Among those with SxOA, more TKRs were done in MOST (35%) than OAI (19%). Adjusted mean WOMAC function (0–68, 68 = worst) improved from 26.9 to 21.9 in MOST and from 24.5 to 22.0 in OAI (difference between MOST and OAI in change in WOMAC function, p = .01). Estimates of function without TKRs showed function would not have changed in MOST (23.2 at baseline to 22.4).


Functional status of subjects with knee OA in MOST improved more than in OAI, probably because of higher rates of TKRs. The decline suggests that TKR diminishes the functional impact of OA in the community.

Total knee replacement; Function; Epidemiology