Fig. 1 Radiograph of a loose cementless tibial component with subsidence and associated osteolytic lesion. Reproduced from Hockman DE, Engh GA. Large Structural Allografts, in Whiteside LA (ed). Revision Total Knee Arthroplasty, Rosemont, IL, American Academy of Orthopaedic Surgeons, 2003, p. 18.


Published 6/1/2014
Eric C. Makhni, MD, MBA; Christopher J. Dy, MD, MPH; the AAOS Washington Health Policy Fellows

Why TKAs Fail

Clinical studies, administrative claims, and registry data can help identify the reasons

Clinicians rely on various data sources to synthesize information for evidence-based medicine. Adhering to evidence-based medicine enables us, as orthopaedic surgeons, to help decrease costs and reduce unnecessary testing and diagnostics. Continual review of data may provide important information about how to improve the systems within which we deliver care.

This article examines the benefit of scientific review of three sources in identifying common etiologies for failure of total knee arthroplasty (TKA). By studying these trends, we can more effectively focus on clinical and research efforts, as well as on resource allocation, to optimize patient outcomes.

Data sources
Clinical studies, claims-based data, and national and international registry data are the three main data sources for information about why TKAs fail. Traditional retrospective or prospective clinical studies measure outcomes from either a single surgeon or a group of surgeons. These studies are very powerful in that the patients’ charts are often directly reviewed for data collection and analysis; however, they are limited in generalizability (eg, volume, geography).

Claims-based data come from administrative agencies such as the Centers for Medicare & Medicaid Services (CMS), along with statewide reporting systems, and rely on claims submissions by hospitals across the country.

A variety of databases draw from these claims, including the Healthcare Cost and Utilization Project (HCUP). HCUP is a collection of databases sponsored by the Agency for Healthcare Research and Quality and includes data from federal, state, and industry sources. For example, the Nationwide Inpatient Sample (NIS) represents data from 8 million hospital stays per year, derived from a weighted sample of all hospital discharges (40 million) to provide national estimates. The NIS database has played an integral role in recent years in outcomes reporting in total joint arthroplasty (TJA).

One disadvantage of these samples, however, is that they rely on administrative codes intended primarily for billing purposes; functional outcomes and other measures are not captured. Another disadvantage of administrative data is the absence of detailed clinical characteristics to study preoperative risk factors. However, as coding definitions for diagnoses and procedures are refined, the utility of these samples will similarly be enhanced.

“Despite its limitations, administrative claims data can provide useful insights regarding large populations of patients from diverse practice settings around the country,” said Kevin J. Bozic, MD, MBA, who chairs both the AAOS Council on Quality and Research and the California Joint Replacement Registry.

Registries are another valuable data source. Registries have traditionally served a monitoring role in TJA, helping clinicians identify which implants require revision more often than others. This has been especially helpful in expediting the removal of problematic implants. European registries, which have been active for several years, have data from millions of patients. Registries typically produce detailed reports on an annual basis, providing powerful information on TJA outcomes, such as revision rates, causes for revision, and outcomes following revision.

Although they are able to provide broad, population-based statistics, many registries lack the patient-reported outcome measures (PROMs) that are necessary when considering TJA outcomes. Implementing these PROMs can be extremely challenging due to constraints stemming from costs, logistics, and overall sustainability. However, successful incorporation of PROMs into traditional registries could provide invaluable data to clinicians, researchers, and policy makers.

“Patient-reported outcome measures are invaluable tools to help us understand the impact our interventions have on our patients’ quality of life,” said Dr. Bozic.

What the data sources tell us
Understanding causes for TKA failures may ultimately help focus clinical and economic resources to optimize outcomes. Additionally, examining these trends may provide insight into efficacy of focus, such as success of eradicating infection with enhanced aseptic measures.

Clinical trials
Several clinical studies documenting rates and etiology of TKA failure have been published. A 2002 study by Sharkey et al reported on a series of more than 200 TKA revisions, collected over a 3-year period (2007–2010) from a single institution. The authors implicated polyethylene wear (25 percent) and aseptic loosening (24 percent) as the two most common reasons for revision (Fig. 1), followed by instability (21 percent) and infection (18 percent). Additionally, they found that more than half of the revisions were performed within 2 years of the index operation.

However, a more recent study reporting on 844 revisions from 2010–2011 revealed a shift in causes for failure. It found that the main causes for failure were aseptic loosening (31 percent), instability (19 percent), infection (16 percent), and polyethylene wear (10 percent). Additionally, the authors reported that only 35 percent of these revisions occurred within the first 2 years after the initial arthroplasty.

This shift in failure etiology suggests that the mechanical properties of implants (ie, polyethylene durability) have improved. However, infection rates remained similar, indicating that progress still needs to be made in this area.

Administrative claims
In recent years, administrative claims data—especially the NIS database—has provided valuable insight into outcomes following TKA. A study from 2010 based on information from more than 60,000 revision TKAs in the NIS from 2005–2006 reported that more than 25 percent of revisions were due to infection and that 16 percent were due to implant loosening. Given the nature of the claims data, the report went on to detail total costs ($49,360) and length of stay (5.1 days) from these episodes.

Claims-based data also allow for correlation of outcomes to known patient factors, such as comorbidities. In a follow-up study, also using the 5 percent Medicare sample, Bozic et al identified congestive heart failure, chronic pulmonary disease, preoperative anemia, and diabetes as independent risk factors for periprosthetic joint infection. Other studies from claims data have provided insight into reasons for readmission following TKA and forecasted future demands for TJA, along with risk factors for early revision and increased complication rates when patients change hospitals for revision surgery.

Registry data
Many total joint replacement registries are in Europe and have been made possible due to the single-payer health system in these countries. These large registries are able to provide population-based data concerning TKA outcomes.

Long-standing registries have proven insightful in tracking outcomes of patients who have undergone TKA following prior procedures such as tibial osteotomy. Registries have also provided information on failure rates following TKA according to specific health system characteristics. For example, a recent study of the Norwegian Arthroplasty Register showed decreased revision rates in patients who received surgery at high-volume centers. Such a finding would have been difficult to produce based solely on claims data or individual clinical studies.

Synthesizing information from the “triple threat” of data sources—clinical studies, administrative claims, and registries—may help clinicians, researchers, and policy makers better direct their efforts toward achieving optimal TKA outcomes. Trends in the data may help assess the overall success of various interventions on outcomes. Each source of data contributes in different ways toward this synthesis, and the limitations of these sources are important to keep in mind when assessing their findings.

Eric C. Makhni, MD, MBA, and Christopher J. Dy, MD, MPH are Washington Health Policy Fellows.


  1. Badawy M, Espehaug B, Indrekvam K, Engesaeter LB, Havelin LI, Furnes O: Influence of hospital volume on revision rate after total knee arthroplasty with cement. J Bone Joint Surg Am 2013;95(18):e131.
  2. Bozic KJ, Kurtz SM, Lau E, et al: The epidemiology of revision total knee arthroplasty in the United States. Clin Orthop Relat Res 2010;468(1):45-51.
  3. Bozic KJ, Lau E, Kurtz S, Ong K, Berry DJ: Patient-related risk factors for postoperative mortality and periprosthetic joint infection in medicare patients undergoing TKA. Clin Orthop Relat Res 2012;470(1):130-137.
  4. Dy CJ, Bozic KJ, Padgett DE, Pan TJ, Marx RG, Lyman S: Is changing hospitals for revision total joint arthroplasty associated with more complications? Clin Orthop Relat Res 2014 Mar 11. [Epub ahead of print].
  5. Dy CJ, Marx RG, Bozic KJ, Pan TJ, Padgett DE, Lyman S: Risk factors for revision within 10 years of total knee arthroplasty. Clin Orthop Relat Res 2014;472(4):1198-1207.
  6. Franklin PD, Harrold L, Ayers DC: Incorporating patient-reported outcomes in total joint arthroplasty registries: Challenges and opportunities. Clin Orthop Relat Res 2013;471(11):3482-3488.
  7. Kurtz SM, Lau E, Ong K, Zhao K, Kelly M, Bozic KJ: Future young patient demand for primary and revision joint replacement: National projections from 2010 to 2030. Clin Orthop Relat Res 2009;467(10):2606-2612.
  8. Niinimäki T, Eskelinen A, Ohtonen P, Puhto AP, Mann BS, Leppilahti J: Total knee arthroplasty after high tibial osteotomy: A registry-based case-control study of 1,036 knees. Arch Orthop Trauma Surg 2014;134(1):73-77.
  9. Schairer WW, Vail TP, Bozic KJ. What are the rates and causes of hospital readmission after total knee arthroplasty? Clin Orthop Relat Res 2014;472(1):181-187.
  10. Schroer WC, Berend KR, Lombardi AV, et al: Why are total knees failing today? Etiology of total knee revision in 2010 and 2011. J Arthroplasty. 2013;28(8 Suppl):116-119.
  11. Sharkey PF, Hozack WJ, Rothman RH, Shastri S, Jacoby SM: Insall Award paper: Why are total knee arthroplasties failing today? Clin Orthop Relat Res 2002(404):7-13.