ICJR ROUND TABLE: Should a Study on the Effectiveness of TKA Change Practice?

    We asked 3 surgeons to weigh in on a recently published paper whose authors concluded that total knee arthroplasty has little impact on quality of life and, therefore, is not a cost-effective option for treating all patients with knee osteoarthritis.

    Orthopaedic surgeons generally consider total knee arthroplasty (TKA) to be an effective procedure that relieves the pain of osteoarthritis and restores physical functioning.

    A study published earlier this year by the BMJ came to a different conclusion: TKA provides only minimal improvement in quality of life and is economically unattractive because the procedure is performed on too many patients. [1]

    However, If TKA were limited to patients with more severe symptoms, the study authors said, its effectiveness – and economic attractiveness – would rise. Click here to read a summary of the study’s findings and here to read the full article from the BMJ.

    The study authors, from the Icahn School of Medicine at Mount Sinai and Erasmus University Medical Center in Rotterdam, The Netherlands, included data from 2 US cohort studies in a comprehensive marginal structural modeling and cost-effectiveness analysis:

    • 4,498 participants from the Osteoarthritis Initiative (OAI) who were ages 45 to 79 years, had knee OA or were at high risk of developing it, and who were followed for up to 9 years; this was the initial analysis
    • 2,907 patients from the Multicenter Osteoarthritis Study (MOST) who were followed for up for 2 years; this was the secondary analysis done to validate the effect estimates of the initial findings

    Quality of life was measured using a recognized score of physical and mental function – the SF-12 – and osteoarthritis-specific quality of life scores.

    Are the conclusions made by the study authors valid? If so, should surgeons who perform TKA change their practice?

    To find out, we asked 3 surgeons – Denis Nam, MD; Antonia F. Chen, MD, MBA; and James A. Browne, MD – to read the paper and share their thoughts on its significance to the orthopaedic surgery community. Here’s what they had to say.

    What are the take-home messages of this study?

    Denis Nam, MD, Rush University Medical Center, Chicago, Illinois: The authors found that TKA resulted in an improvement in quality of life for patients with knee osteoarthritis, with greater improvements seen in patients with more severe preoperative symptoms.

    Furthermore, they indicated that TKA would be optimally cost-effective if limited to patients with a SF-12 PCS score less than 35, thus implying a guideline for appropriateness of performing a TKA.

    However, despite these seemingly positive findings, the authors emphasized that the current practice of TKA as performed in patients with knee osteoarthritis has minimal effects on quality adjusted life years at the group level.

    Thus, they recommend that cost-effectiveness of TKA would be improved if it was restricted to more severely affected patients.

    Antonia F. Chen, MD, MBA, The Rothman Institute, Philadelphia, Pennsylvania: As Dr. Nam indicates, the authors found patients who have lower baseline functional status (SF-12 PCS scores less than 35) can have greater improvements in quality of life metrics after undergoing TKA compared with patients with higher baseline functional status.

    In addition, hospitals should strive to bring costs under $14,000 per admission for primary TKA to maximize cost-effectiveness of the procedure.

    James A. Browne, MD, University of Virginia School of Medicine, Charlottesville, Virginia: This study attempts to identify those patients for whom TKA is a cost-effective option based on their preoperative disease severity. As Dr. Nam states, the authors found – not surprisingly – that TKA was most cost-effective in those patients with lower functional status at baseline.

    The controversy comes with the authors’ conclusion that current practice of TKA in the US had minimal effects on quality of life at the group level and did not appear to be economically justifiable.

    What are the strengths of this study?

    Dr. Browne: Cost-effectiveness analysis is a form of economic analysis that compares the relative costs and outcomes of different courses of action. Over the past few decades, these analyses have been increasingly used in medicine as a tool to assist in allocating limited resources. These analyses can help physicians understand the overall impact of a procedure on patient care at the societal level and can guide policy making decisions.

    Although TKA is known to be one of the most successful and cost-effective of all surgical interventions, there are no doubt some patients for whom TKA is not cost-effective, particularly those with minimal symptom burden and limited radiographic arthritis. Non-operative management is recognized as the appropriate treatment for patients with early knee osteoarthritis and minimal symptoms.

    Attempting to identify those patients who benefit the most from an intervention such as TKA is a worthy exercise.

    Dr. Nam: The primary strengths of the study are the number of patients included and the comprehensive manner in which the authors built their model. They used numerous statistical methods to both build the model and generate their conclusions. The use of marginal structural models is known to estimate causal treatment effects. In addition, the authors account for a number of variables related to the presence of knee osteoarthritis, such as the use of pain medications or non-pharmacologic treatments.

    Dr. Chen: In addition to the above comments, the authors utilized 2 robust datasets (OAI and MOST) to perform their analysis, using very comprehensive economic modeling.

    What are its weaknesses?

    Dr. Chen: Although the authors reported marginal improvements in outcomes, patients exhibited minimal clinically important difference (MCID) in outcomes for WOMAC and KOOS quality of life scores, regardless of functional status. The authors did not highlight this in the paper.

    They also assumed that the disproportionate increase in TKA was due to the “expansion of eligibility to people with less severe symptoms.” However, they did not consider that the increase in TKA utilization may parallel the increase in the number of individuals entering a certain age group (for example, baby boomers) who more likely to need TKA due to cartilage degeneration seen with age.

    In addition, workforce issues should have been considered. The authors only looked at quality of life and did not consider potential wages that could be accrued if TKA were performed in younger individuals with debilitating knee osteoarthritis, thus allowing them to reenter the workforce.

    Although this study is interesting, I believe the findings should be taken with a grain of salt because they are based on modeling. Individual patients may have different scenarios (for example, OARSI grade 3 radiograph, but high SF-12 PCS score) that may lead to improved outcomes with TKA.

    Dr. Nam: The weaknesses of this study can be attributed to the inherent flaws in using statistical modeling to extrapolate the potential impact of an intervention on clinical outcomes.

    As with any model, it is important to consider the inputs that were used (ie, costs of interventions and patient population samples) and assumptions made to generate the model. First, regarding the patient population, the authors recognize that patients from the OAI and MOST investigations may not be generalizable to patients in the US population.

    Second, they also recognize that they may have underestimated the true severity of patients’ preoperative symptoms because they were unable to obtain outcome measures on these patients immediately before undergoing surgery.

    The authors attempted to account for the cost of pharmacologic and non-pharmacologic interventions (such as oral anti-inflammatory drugs, massage, and chiropractor care) in their model, but they failed to account for costs associated with intra-articular injections and the use of opiates. These costs likely were not available in the data sets they used.

    In addition, the authors assumed that the use of medication to treat osteoarthritis would remain stable in this cohort of patients over 8 years. In most clinical scenarios, however, patients tend to increase their medication use as their symptoms worsen. Thus, the true costs of these interventions may not be accounted for.

    Furthermore, the authors’ model assumes the lifetime risk of undergoing TKA to be 39.9% (95% uncertainty level 34.5 to 45.3). This number is much higher than in the orthopaedic literature. Weinstein et al, [2] for example, estimated the lifetime risk of TKA to be 9.0% for females and 7.0% for males. The discrepancy between these numbers and its impact on the model must be considered.

    I agree with Dr. Chen; the authors did not consider the potential impact of untreated knee osteoarthritis on younger, working patients who cannot continue in their occupation due to the pain and mobility issues caused by osteoarthritis.

    Dr. Browne: As my colleagues have noted, the authors of this study constructed a complex, seemingly sophisticated model that relies on many assumptions and expert opinion. Most clinicians, such as myself, will have a hard time evaluating the accuracy and validity of the statistical methods.

    Some of the methods used by the authors, however, such as measuring pain and function a year or more prior to TKA and assuming they remained constant leading up to the surgery, seem flawed. And although the sensitivity analysis provides a method for evaluating some of these assumptions, the level of precision of the authors’ conclusions may be unwarranted and misleading.

    One major concern specific to this study is that a global physical function measure (SF-12 PCS score) is used to assess the effect of TKA. This score is not specific to the knee that has undergone TKA and is influenced by co-morbid conditions as well as arthritis elsewhere in the body. A knee-specific measure (such as the KOOS score) would have been a much better way to assess the pain relief and functional gain due to TKA.

    Will this study change your current clinical practice?  If so, how…and if not, why not?

    Dr. Nam: This study essentially concludes that TKA remains a cost-effective intervention when it is performed in patients with severe preoperative symptoms. Of course, if the actual cost of the intervention were to decrease – or the costs used in the model were decreased – then the potential cost-effectiveness of the procedure could be more generalizable.

    This study will not change my current practice as I agree with their primary finding that TKA should be reserved for those patients with the most severe symptoms. I agree with their overall assessment that the appropriateness of TKA in patients with less-severe preoperative symptoms should be questioned.

    However, the overall conclusion that current practice of TKA in the US may not be cost-effective should not be misconstrued to mean that TKA is an ineffective procedure. TKA remains an effective procedure for the management of knee osteoarthritis, and the aforementioned limitations must be considered when interpreting the results of this statistical model.

    Dr. Chen: This study will help with counseling my patients, although it will not change my clinical practice.

    For patients with a high SF-12 physical component scores (more than 35 to 40), I will use this study to warn them that their outcomes may not be as favorable as other patients with worse scores. However, they should still experience overall functional outcome improvements, and undergoing TKA can still improve their quality of life.

    Dr. Browne: I agree with my colleagues; this study will not change my current clinical practice. I have concerns about the validity of the methodology and would not take their conclusions to the bank just yet.

    Furthermore, at the level of the individual practitioner, the situation is much more complex. A cost-effectiveness analysis compares the relative costs and outcomes of different courses of action at the societal level, which can be helpful for public policy decisions. At the clinical level, however, providing the best possible care for an individual patient with knee osteoarthritis comes down to much more than an SF-12 PCS score and needs to be individualized to the patient.


    1. Ferket BS, Feldman Z, Zhou J, Oei EH, Bierma-Zeinstra SM, Mazurdam M. Impact of total knee replacement practice: cost effectiveness analysis of data from the Osteoarthritis Initiative. BMJ 2017 Mar 28;356:j1131. doi: 10.1136/bmj.j1131.
    2. Weinstein AM, Rome BN, Reichmann WM. Estimating the burden of total knee replacement in the United States, J Bone Joint Surg Am, 2013 Mar 6; 95(5) 385-92. doi: 10.2106/JBJS.L.00206.