Background: The annual volume of major cardiovascular and oncologic procedures performed in hospitals and by surgeons has been inversely associated with the rates of perioperative mortality and complications. The relationship between hospital and surgeon volume and perioperative outcomes following total knee replacement has received little study.

Methods: We analyzed claims data for Medicare patients who had elective primary total knee replacement between January 1 and August 31, 2000. Hospital and surgeon volumes were defined as the number of primary and revision total knee replacements performed in the hospital or by the surgeon in Medicare recipients in 2000. We examined the associations between the annual volumes of total knee replacement performed in the hospitals and by the surgeons and the rates of mortality and complications (infection, pulmonary embolus, myocardial infarction, or pneumonia) in the first ninety days postoperatively. The analyses were adjusted for age, gender, comorbid conditions, Medicaid eligibility (a marker of low income), and arthritis diagnosis. Analyses of hospital volume were adjusted for surgeon volume and vice versa.

Results: Twenty-five percent of the primary total knee replacements were done by surgeons who performed twelve of these procedures or fewer in the Medicare population annually, and 11% were done in hospitals with an annual volume of twenty-five of these procedures or fewer. Compared with the patients who had a primary total knee replacement in hospitals with an annual volume of twenty-five procedures or fewer, those managed in hospitals with an annual volume exceeding 200 procedures had a lower risk of pneumonia (odds ratio, 0.65; 99% confidence interval, 0.47 to 0.90) and any of the adverse outcomes examined (death, pneumonia, pulmonary embolus, acute myocardial infarction, or deep infection) (odds ratio, 0.74; 99% confidence interval, 0.60 to 0.90). Similarly, patients who had a primary total knee replacement done by surgeons who performed more than fifty such procedures in Medicare recipients annually had a lower risk of pneumonia (odds ratio, 0.72; 99% confidence interval, 0.54 to 0.95) and any adverse outcome (odds ratio, 0.81; 99% confidence interval, 0.68 to 0.98) compared with patients of surgeons with an annual volume of twelve procedures or fewer.

Conclusions: Patients managed at hospitals and by surgeons with greater volumes of total knee replacement have lower risks of perioperative adverse events following primary total knee replacement. Patients and clinicians should incorporate these findings into discussions about selecting a surgeon and a hospital for total knee replacement. These data should also be integrated into the policy debate about the advantages and drawbacks of regionalizing total joint replacement to high-volume centers.

Level of Evidence: Prognostic study, Level II-1 (retrospective study). See Instructions to Authors for a complete description of levels of evidence.