Chapter  86:  Total Knee Replacement

Preface

The Agency for Healthcare Research and Quality (AHRQ), through its Evidence-Based Practice Centers (EPCs), sponsors the development of evidence reports and technology assessments to assist public- and private-sector organizations in their efforts to improve the quality of health care in the United States. This report on Total Knee Replacement was requested and funded by the Office of Medical Applications of Research, National Institutes of Health. The reports and assessments provide organizations with comprehensive, science-based information on common, costly medical conditions and new health care technologies. The EPCs systematically review the relevant scientific literature on topics assigned to them by AHRQ and conduct additional analyses when appropriate prior to developing their reports and assessments.

To bring the broadest range of experts into the development of evidence reports and health technology assessments, AHRQ encourages the EPCs to form partnerships and enter into collaborations with other medical and research organizations. The EPCs work with these partner organizations to ensure that the evidence reports and technology assessments they produce will become building blocks for health care quality improvement projects throughout the Nation. The reports undergo peer review prior to their release.

AHRQ expects that the EPC evidence reports and technology assessments will inform individual health plans, providers, and purchasers as well as the health care system as a whole by providing important information to help improve health care quality.

We welcome written comments on this evidence report. They may be sent to: Director, Center for Outcomes and Evidence, Agency for Healthcare Research and Quality, 540 Gaither Road, Rockville, MD 20850.

Carolyn M. Clancy, M.D.

Director

Agency for Healthcare Research and Quality

Barnett S. Kramer, M.D., M.P.H.

Director, Office of Medical Applications of Research

National Institutes of Health

Jean Slutsky, P.A., M.S.P.H

Acting Director, Center for Outcomes and Evidence

Agency for Healthcare Research and Quality

The authors of this report are responsible for its content. Statements in the report should not be construed as endorsement by the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services of a particular drug, device, test, treatment, or other clinical service.

Acknowledgments

We would like to thank the people who worked on the abstractions themselves including: Tyler Bailey; Jesse Botker; Kevin Bozic, MD; Kevin Broder; Gideon Burstein, MD; Jason J. Caron; Stephanie Cintora; Todd Gengerke; Brad Hilger; Richard Iorio, MD; Craig Israelite, MD; William Macaulay, MD; Charles Nelson, MD; and Alexander W. Stricker III. We also wish to thank William M. Woodhouse for his role as data manager and Krystal Wiesenberg for data entry. Marilyn Eells was responsible for editing and formatting the report.

Structured Abstract

Context: The projected growth in the population with arthritis is likely to expand the future demand for elective arthroplasty. At present, there is no strong empirical base for the indicators in current use for what criteria should be used to identify potential candidates for Total Knee Arthroplasty (TKA)^a; nor is there professional consensus around such indications. An NIH consensus conference has been planned to address these questions. This report summarizes the literature as part of the background for that conference.

Objectives: A systematic review of the literature was undertaken to address four questions:

1. What are the current indications for, and outcomes from, primary total knee replacement?

2. How do specific characteristics of the patient, material and design of the prosthesis, and surgical factors, affect the short-term and long-term outcomes of primary total knee replacement?

3. Are there important perioperative interventions that influence outcomes?

4. What are the indications, approaches, and outcomes for revision total knee replacement?

5. What factors explain disparities in the utilization of total knee replacement in different populations?

6. What are the directions for future research?

Data Sources: The primary TKA literature search was performed by the National Library of Medicine, which searched PubMed from 1995 to April 2003. The access search was done using PubMed and covering the period from 1990 through April 2003. The literature search on revisions was done in two stages. A prior Medline search covering the period from 1996 through 2000 was the basis for a meta-analysis. An updated search using PubMed covered 2001 through April 2003.

Study Selection: The nature of this topic required heavy reliance on observational studies. The major criteria for identifying studies for inclusion in the indications for TKA search required that they address primary TKAs, have at least pre and post surgery data using at least one of four standard functional measures (Knee Society [KS] score, Hospital for Special Surgery [HSS] score, WOMAC, or SF-36), have a sample size of at least 100 total knee replacements, be published in English, and utilize tricompartment TKA. Sixty-two studies met the full inclusion criteria. The selection of studies on access required that they examine the relationship of at least gender or race to the performance of primary TKAs. Six articles were included. The same inclusion criteria applied to primary TKAs were applied to the update of the TKA revision study. Fourteen articles met the criteria.

Data Extraction: Data were abstracted by trained abstractors using a standardized abstraction tool that had been pilot tested and reviewed by the Technical Expert Panel. For the indicators search, the original abstractions were reviewed to assure reliability. All articles meeting the inclusion criteria were independently re-reviewed by each of the three principals. Information related to study and patient characteristics, baseline and followup functional status measures, perioperative complications, and revision rates were extracted using a standardized abstraction tool that had been pilot tested. The access data was abstracted by a subset of the original abstractors using another standardized tool. The TKA revision update was abstracted by an abstractor and one principal using a modification of the primary TKA tool.

Data Synthesis: Both TKA and total knee arthroplasty revision (TKAR) are associated with improved function. The strongest evidence exists over a followup period of up to two years, but the studies that extend to five and even ten years of followup show positive results as well. The average age of patients undergoing TKA in these reports was 70 years with few over aged 85. Two-thirds were female, one third were considered obese, and nearly 90% had osteoarthritis. No studies provided data on racial/ethnic status. The mean effect size (expressed as numbers of standard deviations) is considered large in magnitude and varies from 1.6 to 3.9 depending on the functional measure used and the duration of followup. There is no evidence that age, gender, or obesity are strong predictors of functional outcomes. Patients with rheumatoid arthritis show more improvement than those with osteoarthritis, but this may be related to their poorer functional scores at the time of treatment and hence the potential for more improvement. The revision rate through five or more years is 2.0% of knees and 2.1% of patients. Complications as defined by the investigator occurred in 5.4% of patients and 7.6% of knees. Patients with rheumatoid arthritis show more improvement than those with osteoarthritis. With regard to access, nonwhites receive TKAs less often than whites despite higher rates of osteoarthritis. Women receive TKAs more often than men, but the pattern is not as consistent as with race. TKA revisions are associated with consistent improvement in function on an order of magnitude similar to primary TKAs.

Conclusions: In general, the outcomes research on TKAs emphasizes before and after studies that are variations on case series of various techniques and prostheses with little attention to the role of other factors or to attrition. Although demographic and clinical factors are recorded, they are rarely used in the analysis. A consistent body of evidence suggests substantial improvement in function associated with TKA and TKAR. The follow-up periods vary but the mean is greater than five years. More informed decision making about indicators for TKAs will require stronger research designs. These need to be planned as prospective studies with multivariate analysis. Such analyses will require larger samples and more consistent and comprehensive data collection than was found in this review.

Chapter 1. Introduction

Throughout this report the term total knee arthroplasty will be used in lieu of total knee replacement because the abbreviation of the latter term may be confused with total knee revision.

At present, approximately 43 million individuals suffer from arthritis. Because this condition becomes increasingly prevalent with advancing age,^{1, 2} given the population projections, the Centers for Disease Control estimate that by 2030 over 41 million persons aged 65 and older will have arthritis or chronic joint symptoms.³ In particular, arthritis of the knee and accompanying joint symptoms result in considerable morbidity, loss of functional status, independence, and quality of life. The high prevalence of arthritis in the population is reflected in the high cost of treatment, which has been estimated at $95 billion per year.⁴ These figures do not include the additional costs due to lost job productivity. Treatment options are primarily designed to relieve pain and improve functional status.

Standardized instruments have been developed in order to assess the severity of the symptoms and evaluate outcomes related to treatment. For example, Callahan et al., defined a generic global knee score (GKS) as “an instrument that measured patient outcomes in the domains of pain, function, and range of motion and combined these domains in a summary scale.”⁵ Widely used scales include the Hospital for Special Surgery score (HSS),⁶ Knee Society (KS) score,⁷ and Western Ontario and MacMaster University (WOMAC) Osteoarthritis Index.⁸ (Copies of these scales are shown in Appendix A.) These scales typically cover aspects of pain and function (usually emphasizing walking). The HSS and KS are completed by clinicians; the WOMAC and SF-36 are designed to be completed by patients. They are intended to provide a score of 0 to 100, where a higher score implies a better outcome. For at least the HSS and KS scores, less than 60 is considered poor pain and function status; 60–69 represents fair pain and function status; 70–84 is considered good; 85–100 is considered excellent pain and function status.

Treatment options include physical therapy, analgesic and/or anti-inflammatory medications, and surgical therapy. The primary surgical treatment for patients is replacement of the native knee joint with a prosthesis (Total Knee Arthroplasty—TKA). A wide variety of prostheses and surgical techniques have been utilized but all are considered under the category of TKA. Total knee arthroplasty is one of the most common orthopaedic procedures performed. In 2001 171,335 primary knee replacements and 16,895 revisions were performed.⁹ Medicare paid approximately $3.2 billion in 2000 for hip and knee joint replacements. Because these procedures are elective and expensive and because the prevalence of arthritis is expected to grow substantially as the population ages, these procedures are likely to come under increasing scrutiny. By 2030, it is estimated that there will be an 85 percent increase in TKA.¹⁰ With this growth in mind, as well as the uncertainty related to the indications for, and outcomes associated with TKA, the Minnesota EPC was asked to conduct a systematic review of the literature to address four specific questions:

1. What are the current indications for, and outcomes from, primary total knee replacement?

2. How do specific characteristics of the patient, material and design of the prosthesis, and surgical factors, affect the short-term and long-term outcomes of primary total knee replacement?

3. Are there important perioperative interventions that influence outcomes?

4. What are the indications, approaches, and outcomes for revision total knee replacement?

5. What factors explain disparities in the utilization of total knee replacement in different populations?

6. What are the directions for future research?

The Total Knee Replacement evidence report will help inform the deliberations of the Consensus Conference Panel.

Previous reports suggest that TKA improve functional status, relieve pain, and result in relatively low perioperative morbidity. A systematic review and meta-analysis of 130 studies evaluating 154 cohorts published in 1994 by Callahan and colleagues evaluated patient outcomes following tricompartmental total knee replacement. They noted that global rating scale scores improved by 100% for the typical patient and that 89% of patients reported good or excellent outcomes after a mean followup of 4.1 years. The weighted mean complication rate was 18.1% and the mean mortality rate per year of followup was 1.5%. The overall rate of revision during 4.1 years was 3.8%.⁵

Table 1. Studies of beliefs about indications, referrals, (more...)

Table 1. Studies of beliefs about indications, referrals, and thresholds for total knee arthroplasty

Study	Journal	Population focus, N	Objective	Results
Wright et al., 199512	Can Med Assoc J	All orthopaedic surgeons in Ontario, Canada n=325	Determine extent of agreement on indications for TKA and how perceptions differ according to the number of procedures performed	- Clinical agreement (>90%) in 14 of 34 patient characteristics (38%) in determining need for TKA
				- Clinical disagreement (<60%) with 7 of 34 (21%) patient characteristics
				- No agreement in treatment with 3 hypothetical case scenarios with varying degrees of osteoarthritis (n=205) (highest agreement was 86.8%)
				- High volume orthopaedists disagreed with low volume orthopaedists in 7 of 34 patient characteristics as indication for TKA (21%)
				- Speculated causes for disagreement:
				1. may reflect limitation of available knowledge
				2. may reflect controversy within orthopaedic literature
				3. information may not be adequately disseminated to, or adopted by, practicing orthopedists despite the fact that the factor's effect on outcome of TKA has been clearly demonstrated in the medical literature
				4. surgeons may choose to treat patients based on personal experience or training
Coyte et al., 199614	J Rheum	Rheumatologists and family practitioners n=98 Rheumatologists, 250 family practitioners (66 & 99 in final analysis respectively)	Assess agreement for indications for TKA, outcomes of TKA, and non-surgical management of osteoarthritis between family practitioners and rheuma-tologists. These results were to be compared with data on orthopaedists	- Clinical agreement (>90%) for BOTH rheumatologists and family practitioners with 2 of 32 patients factors
				- Rheumatologists clinical agreement (>90%) with 6 of 32 (13%) patients factors
				- Family practitioners Clinical agreement (>90%) with 4 of 32 (19%)
				- Clinical disagreement (<60%) with 10 of 32 factors for family practitioners
				- Clinical disagreement (<60%) with 10 of 32 factors for rheumatologists
				- Disagreement among specialties: Family practitioners > rheumatologists > orthopaedists (family practitioners & orthopaedists P<0.0001, rheumatologists & orthopaedists P<0.04).
Wright et al., 199915	Medical Care	Orthopaedists and primary care physicians n=(Provider data from Wright et al., 1995 in Can Med Assoc J and Coyte et al., 1996 in J Rheum) [See both studies above for provider numbers]	Identify factors that might be amenable to intervention by investigating determinants of regional variation in the use of knee replacement surgery	- Surgeon opinion or “enthusiasm” was “the dominant modifiable determinant of area variation” in the utilization of TKA
				- Surgeons propensity to operate (based on responses to the survey in the article cited above) and opinions on patient outcome were both positively correlated with the total # of procedures performed in the study period (p<0.0001)
Hadorn & Holmes, 1997a, 1997b19,20	BMJ	New health policy description	Describes New Zealand's new priority criteria for major joint replacement (TKA & THA)	- Checklist utilizes 4 major components incorporating both clinical and social factors in determining order for receiving TKA: Pain (40% of scale), Functional Activity (20%), Movement and deformity (20%), Other factors (20%)
				- Checklist created to assess where patients would be placed on list for elective surgeries prior to New Zealand moving away from waiting list format to booking appointments
Mancuso et al., 199613	J Arthroplasty	Orthopaedists n=328 (80 in final analysis)	Survey of all orthopaedists in specific geographic area regarding their indications and modifying factors for primary TKA and THA	- Clinical agreement (>90%) with 6 of 24 (25%) factors related to determining need for TKA
				- Clinical disagreement (<60%) with 3 of 24 (13%) factors related to determining need for TKA
				- They found no correlation with # of years in practice and agreement
Dieppe et al., 199918	Rheumatology	Review article: consensus panel of professionals to examine problems re: use of TKA in management of osteoarthritis	“review literature of effectiveness of TKA for osteoarthritis of the knee, the evidence of practice variation and underutilization, and the publications on possible indications for TKA”	- Primary care MDs likely to lack confidence in the exam of the knee joint leading to delays in diagnosis and inability to assess severity of joint damage due to little exposure in training
				- 4 potential problems:
				1. persistent negative attitudes towards osteoarthritis in general and towards value of TKAR in particular amongst the public and primary care MDs
				2. the lack of simple tools to help assess severity and impact of knee osteoarthritis that can be used in the community
				3. the absence of any clear guidelines or agreed evidence based indications for TKA
				4. the absence of any studies that compare the efficacy of TKA with that of non-surgical intervention strategies
				- 3 useful variables for surgical decision making in TKA:
				1. severity of joint damage (pain at night, severity of pain, function)
				2. other patient related variables (psychosocial, patient motivation)
				3. the environment (socio-economic status - availability of surgeons, economic status of patients)
				Consensus panel conclusions and recommendations:
				1. no clear evidence-based indications for TKA
				2. no comparisons with other forms of treatment
				3. no understanding of which patients are particularly likely to benefit from the procedure
				4. the absence of any studies that compare the efficacy of TKAR with that of non-surgical intervention strategies
Malmlin et al., 199816	Arch Fam Med	Family practitioners and general internists n=300 each (70 and 72 in final analysis)	Description and comparison of the self-reported practice patterns of family practitioners and general internists for the evaluation and management of severe osteoarthritis of the knee, including factors that might influence referral for TKA	- Combining family practitioners and general internists, clinical agreement (>90%) with 6 of 26 patient factors (23%) determining need for TKA
				- Clinical disagreement (<60%) with 5 of 26 patient factors determining need for TKA
Tierney et al., 199411	Clin Ortho	All orthopaedists in Indiana, USA. n=280 (188 in final analysis)	To understand reasons for variation of who gets TKAs using orthopaedists' perspectives of indications and outcomes and comparing them with self-reported annual number of TKAs they performed	- Clinical agreement (>95%) in 7 of 34 patient factors (21%)
				- Agreement (=95% and >60%) with 21 of 33 factors
				- No agreement (<60%) with 5 of 34 (15%) patient factors
				- When correlated with # of TKAs in prior year, significant factors were:
				Patient Characteristics: female gender (r=0.17, p=0.02), non-compliant patient (r=0.20, p=0.008), unstable knee (r=0.20, p=0.008)
				Continuous parameters: old age (r=0.16, p=0.03), varus deformity (r=0.16, p=0.03), valgus deformity (r=0.17, p=0.02) - Independent variables associated with reported # of TKAs in prior year
				Independent Variable	Fraction of Variance explained	P-value
				Female gender	0.06	0.0009
				Unstable knee	0.02	0.488
				Patient can be too old	0.01	0.076
Naylor & Williams, 199617	Quality in Health Care	Consensus Panel (n=11) 4 orthopaedic MDs 2 rheumatoid MDs 2 general practitioners 1 “general physician” 1 epidemiologist 1 physiotherapist	Consensus finding using 120 case scenarios to try and gain agreement on priorities and appropriateness for hip and knee replacement surgery. Consensus findings also with 42 case scenarios for urgency of replacement	- Found that key determinants to prioritize surgery were: pain at rest, severity of functional impairment, problems with care-giving, perceived likely improvement in function
				- Panel agreement statistics:
				• Agreement of =9/11 panelists occurred 61% (73/120) of appropriateness scenarios for referral for TKA (not appropriate, uncertain, appropriate) and in 17% of urgency categories
				• Agreement of =10/11 panelistsoccurred in 92% of appropriateness scenarios and 74% (31/42) of urgency scenarios

Table 2. Summary of studies of clinical agreement about (more...)

Table 2. Summary of studies of clinical agreement about patient factors for either referral or surgery (set at >90% for significant agreement)

Study	Journal / Population	Pro	Neutral	Con	Clinical Factors of Disagreement (<60% agreement)
Wright et al., 199512	Can Med Assoc J	Pain despite meds	Male	Peripheral vascular disease	Patient is >80 years old
	Orthopaedic surgeons		Female	Isolated patellofemoral arthritis	Nursing home resident
			White race	Alcohol/Drug Abuse	Severe hip osteoarthritis
			Non-white race	Local active skin infection	Local psoriasis
				Major psychiatric disorder	Quadriceps lag
				Patient non-compliant	Weak quads
				Age <55 years old	Sensation of instability
				High physical demands at work
				Septic arthritis >1 year ago
Mancuso et al., 199613	J Arthoplasty	Be independent		Poor soft tissue coverage	Age >80 years old
	Orthopaedic surgeons			Dementia	Weight >200 pounds
				Poor patient motivation	Wants psychiatric benefit
				Hostile personality
				Unreal expectations
Malmin et al., 199816	Arch Fam Med	Pain despite meds	Male		Septic knee arthritis > 1 year ago
	Family practitioners and general internists	Persistent weight bearing knee pain	Female		No health insurance
			White race		Isolated patellofemoral arthritis
			Non-white race		Patient demands TKA
					Painful feet
Coyte, Hawker et al., 199614	J Rheum	Pain despite meds - family practitioner/rheumatologist	Male - Family practitioners/ rheumatologists	Peripheral vascular disease (Rheumatologist)	Family practitioner/rheumatologist:
	Rheumatologists and family practitioners	Limited walking <1 block - family practitioner	Female - Family practitioners/ rheumatologists	Isolated patellofemoral arthritis (Rheumatologist)	non-compliant patient
				Local active skin infection (Rheumatologist)	obese patient
					septic knee >1 year ago
					Varus or Valgus deformity
					High physical demands at work
					Family practitioner:
					<55 years old, severe hip osteoarthritis,
					Quadriceps lag, weak quads
					Rheumatologist:
					Nursing home resident
					Patient demands TKA
					Limited active flexion/extension
					Sensation of instability
Tierney et al., 199411	Clin Ortho	Persistent weight bearing pain	Female	Alcohol/drug abuse	Nursing home resident
	Orthopaedic surgeons		Race (white or black)	Major psychiatric disorder	Painful feet
				Local active skin infection	Patient demands TKA
					Unstable knee
					Severe hip osteoarthritis

Chapter 2. Methods and Analytic Framework

This review has three major components, which correspond to the questions posed in the charter. The major effort was directed at examining the indications for (or at least the outcomes of) primary TKA. The second component is a report of a meta-analysis of total knee revisions, which has already been published,²¹ and an update of the literature since that work was completed to be sure no new developments had affected the initial conclusions. The third component was a review of the literature on access to care, especially the effects of gender and age.

The principal analytic framework for the first review (the outcomes of primary TKA) was based on the fundamental principles of outcomes research.²² The underlying model can be briefly expressed as:

Outcomes = f(baseline status, clinical factors, demographic factors, treatment)

In general, the goal of outcomes research is to identify the effect of treatment on outcomes, adjusting for the other factors that might affect outcomes. In this case, however, we use the same model to address the predictive role of various patient characteristics on outcomes when all are treated similarly. Interpreting this relationship is somewhat more complex because factors associated with good outcomes are not necessarily indications for treatment. For example, a person with no problems may have a very good outcome, but one would not want to treat such a patient. The true test of an indication for surgery is a factor that gets worse without treatment and better with it. In effect, one would want to randomly assign patients with the specified condition to receive either TKA or medical management and then compare the clinical course with and without the treatment under study. Those factors that produced the greatest difference associated with treatment would be the strongest indicators for such treatment.

Where randomized clinical trials are available, many of the relevant confounding clinical and demographic factors can be assumed to be randomly distributed, or they may be controlled by elements of the study design that specific inclusion and exclusion criteria, and thus any differences between two groups can likely be attributed to the intervention. However, in the absence of RCTs, as is the case in most of the orthopaedic literature, strong quasi-experimental designs are needed, wherein multivariate analysis is employed to isolate the effect of treatment and address issues related to selection bias. The literature review was thus initially targeted at identifying those studies that had at least the rudiments of such a design. However, given the studies uncovered, we were forced to revise our criteria to assess a broader array of studies that provided at least some baseline and followup information.

Based on consultations with the technical expert panel (members are shown as Appendix B) and discussion with OMAR, AHRQ, and the Chair of the Consensus Panel, we determined that functional measures would be used as the primary outcome measures. We identified several demographic and clinical variables of primary interest: age, gender, baseline status (with regard to pain and function), arthritis type, and body mass index/obesity. The analysis for demographic factor effects, which correspond to the question about access, was conducted separately.

TKA Indicators

The literature search strategy for clinical predictors of TKAs was developed in consultation with the National Library of Medicine, which conducted the search. The literature search was done using a combination of MeSH headings, keywords, and publication types shown in Appendix C.

The search was limited to studies published between 1995 and April 2003. This start date was chosen because a previous review was published in 1994.⁵ Animal studies were excluded, as were non-English language references and references on unicompartmental (unicondylar) knee replacement. Although unicondylar knee replacements (UKR) share many features with total knee replacement (tricompartment), these studies were excluded from our search because UKRs have 1) more specific indication ie, unicompartmental tibio-femoral arthritis with minimal involvement of the patello-femoral and 2) different patient demographics, primarily male population, low activity, minimal deformity, and good range of motion. Additionally, indications for UKRs appear to be in a transition phase. Surgeons have only recently gained experience with this reportedly less invasive procedure. Thus it was felt too early to adequately assess outcomes.

The titles and abstracts of the resulting 3,519 references were then screened, using our inclusion criteria (primary total knee arthroplasty studies; more than 100 knees per study; baseline data and post-op standardized symptom scale outcomes data provided; experimental or quasi-experimental study design).

All articles that appeared to meet the screening criteria were abstracted by trained abstractors. Extracted data included study and patient characteristics, baseline and followup symptom scale scores, revision rates, and perioperative complications as defined by the authors and occurring within six months of surgery. This workforce included medical students, two review staff, an orthopaedics fellow and several volunteer orthopaedic surgeons. A 10 percent subsample of all the abstracts was reviewed independently by a second abstractor to assure consistency. All of the studies that met the minimal criterion of having pre- and post-surgery data were re-reviewed independently by all three of the study principals.

The abstracting form (see Appendix D) included a long list of potential prognostic factors, developed with the assistance of our technical advisory committee. These included co-morbidities, x-ray evidence of joint destruction, bone loss, extensor mechanism integrity, pre-operative range of motion, alignment, tibio-femoral angle, and ligament integrity, as well as the characteristics of the operating surgeon, such as volume and experience.

Figure 1. TKA article inclusion/exclusion flow chart (more...)

   Figure 1. TKA article inclusion/exclusion flow chart

Of the original results, 611 references either met the inclusion criteria or needed further screening of the full article to determine if they met inclusion. The reasons for exclusions are shown in Figure 1, which traces the flow of the articles retained. Of these, 62 studies reported pre- and post-TKA functional data using at least one of the four established measures we relied on (Knee Society score, Hospital for Special Surgery score, WOMAC, or SF-36). All but 15 studies were conducted in the US or Canada.

One of the problems that made summarizing this area difficult was the inconsistent use of patients and knees as the unit of analysis. The reason for this practice is related to the performance of bilateral procedures, either simultaneously or sequentially, but the result is an inconsistent count. Some studies provide both units; some only one. For some types of analysis knees seem like the best measure, but for many (including function and demographics) the data apply more reasonably to patients. Wherever feasible, we present the analysis using both patients and knees.

We conducted a meta-analysis on the functional outcomes data. Meta-analysis methodology assumes that to estimate the combined effect we compute the weighted mean of the results observed in different studies. In the simplest approach weights are based on the sample size but more sophisticated methods account for the precision of the studies and thus adjust for different standard deviations. The effects in this meta-analysis were normalized by dividing to combined standard deviation of two (baseline and followup) measures. Therefore the statistical results of the meta-analysis are expressed in the units of standard deviation and reported as an “effect size.” An effect size greater than 1 SD is considered to be large in magnitude. An additional benefit of this approach is that various effects obtain the same measurement scale and therefore can be compared. In modeling the effects we could use either fixed or random effect models. Because the data at baseline and followup was not consistent, we selected the model with random effects to simplify the interpretation. This model assumes that all studies come from a common population. That is, if the sample size in each study were infinite, then the effect size in all studies would be identical and the standard error of the estimate would approach zero. Because we did not have precise information from all studies, we treated each pre- and post-pair as if they were separate data sets. This is a conservative approach. An analysis using pairs would have produced even more dramatic results. All calculations were implemented using the trial version of the Comprehensive Meta Analysis™ software.²³

Chapter 3. Results

Baseline Characteristics of Patients

The 62 studies that had pre- and post-functional data using one of the four established outcome measures (ie, the Knee Society score, the Hospital for Special Surgery score, the WOMAC, or the SF-36) are summarized in Appendix F, Evidence Table 1. All were simple pre- and post-comparisons. Although various demographic information is provided to describe the sample, that data were rarely and inconsistently used in the reported analyses.

Table 3. Descriptive statistics on 62 studies

Table 3. Descriptive statistics on 62 studies

Patent Characteristics	Average	SD	Number of Studies Reporting
Mean Age (years)
Average	67.5	4.4	57
Weighted by patients	69.1		50
Weighted by knees	69.2		56
Percent Female
Average	65.4	17	47
Weighted by patients	64.6		41
Weighted by knees	64.5		47
Percent Obese (BMI >30)
Average	36.7	3.5	3*
Weighted by patients	37.7		3
Weighted by knees	37.3		3
Percent Osteoarthritis
Average	86.8	13	45
Weighted by patients	86.7		40
Weighted by knees	85.6		43

*

One study of all obese patients not included

Table 3 presents a summary of selected patient and clinical characteristics. We used the full sample from each study whenever possible. Because of the variation in reporting practices here and elsewhere, the mean rates were calculated using means weighted separately on the basis of the numbers of knees and patients in the studies. The data here used weights for numbers of patients and knees, as well as the raw averages. The weightings made little difference. Two studies did not report the numbers of patients. The discrepancy between the numbers of patients and knees reported is an artifact of which studies reported knees.

The average age of patients was approximately 75 years. Very few were over 85; about two-thirds were female; about one-third were considered obese (using a criterion of a Body Mass Index (BMI) of 30 or higher). Nearly 90 percent of patients had osteoarthritis. One-third of subjects underwent bilateral TKA. None of the studies provided information regarding racial/ethnic status. We did not separately address outcomes for patients undergoingbilateral TKAs from those undergoing unilateral procedures. However, we conducted separate analyses by numbers of knees and numbers of patients.

The most commonly used functional measures were the Knee Society score and the Hospital for Special Surgery scale. A major factor in their greater usage is likely the fact that they have existed longer. The WOMAC Arthritis Scale is considered by many in the field to be a psychometrically better measure, but it has only been used since 1991.⁸ The physical function component of the SF-36 is a generic functional outcomes measure, not specific to knees.

Table 4. Mean followup duration according to functional (more...)

Table 4. Mean followup duration according to functional assessment scale (in months)

	Number of Studies	Mean Weighted Followup Time (months)
		Baseline Patients	Followup Patients	Baseline Knees	Followup Knees
KS	46	66.2	79.3	89.8	65.7
HSS	24	66.9	63.0	61.2	61.4
WOMAC	8	44.5	68.1	67.7	72.7
SF-36	9	18.0	23.6	59.2	61.6

Table 4 presents the summary data on the mean duration of study followup periods according to the type of functional outcome assessment scale used. The results are shown using various approaches to weighting the numbers of cases used. They were weighted separately by the numbers of patients and the numbers of knees. Because there is substantial sample loss, we then divided each of these categories to weight by the numbers at baseline and at followup. Several studies used more than one scale. In comparison to the demographic data cited above, there is greater variation when the different weights are applied. When weighting by numbers of patients, the generic measure (the SF-36) was used for shorter followup periods. In general, determining the sample sizes at different points in time was difficult. A substantial number of studies failed to provide adequate data to identify how many patients (or knees) were available at followup.

The longer established measure KS score is associated with longer followup periods, perhaps because it was in use earlier, allowing more time to elapse for such followup. For example, weighting for baseline patients the mean followup for KS and HSS is 66 and 67 months, compared to 45 months for WOMAC. However, weighting for baseline knees, KS has a mean followup of 90 months and WOMAC is 68 months, but HSS is only 61 months. The longest mean followup time was 90 months (KS score weighted for baseline knees), well less than the ten years that has been suggested in order to evaluate long term functional results. Only ten studies had followup time of at least ten years.

Note: Appendixes and evidence tables cited in this report are provided electronically at http://www.ahrq.gov/clinic/epcindex.htm

Some information on attrition rate was reported for 49 studies. Of these the median percentage of subjects lost to followup was 2%, the range was 0–28%. In five studies more than 10% were lost to followup. If death and other exclusions are added to the definition, the range increases to 0–56% with a median of 12%. Five studies had a total loss rate of more than 40%; another five lost 30–40%; and another seven studies lost 20–30%

The issues of outcomes addressed here looked at only the aggregate outcome in the context of having had a TKA. No special efforts were made to distinguish the relative contribution of rehabilitation or type of procedure. Although the latter was the major focus of many studies, few actually compared alternative approaches.

What is the Magnitude of Effect of Primary TKA?

Table 5. Weighted baseline and followup scores for (more...)

Table 5. Weighted baseline and followup scores for TKA outcomes measures

Outcome Measure	Number of Studies Reporting Pre/post Scores Based on Number of Subjects	Baseline Score Based on Number of Subjects	Followup Score Based on Number of Subjects	Number of Studies Reporting Pre/post Scores Based on Number of Knees	Baseline Score Based on Number of Knees	Followup Score Based on Number of Knees
Knee Society (KS)	30 / 7* (n=12,261)	39.8	80.0	27 / 5** (n=15,454)	41.1	82.4
Hospital for Special Surgery (HSS)	17 / 3* (n=2,546)	54.2	89.2	16 / 2** (n=3,333)	52.8	88.7
Western Ontario and McMaster Osteoarthritis Index (WOMAC)	7 (n=2,925)	48.3	76.8	NA †
SF-36 physical function	8 / 7* (n=2,166)	27.6	43.8	2 / 1**	22.4	47.1

*

subjects only

**

knees only

†1 study only

Table 5 summarizes the raw data on change from pre- to post-TKA functional scores (albeit with widely varying followup periods). In each scale the range has been defined as 0–100. In general, a higher score is better, although the WOMAC was standardized such that a lower score is better. In each case there is strong evidence of improved function (and decreased pain). Of the 46 studies using KS scores only 30 provided pre- and post-intervention results according to the number of subjects enrolled (n = 12,261 subjects) (27 provided this information based on number of knees (n = 15,454 knees). There were 17 studies using HSS scores (2,546 patients) Seven studies, representing 2,925 patients reported results with the WOMAC.

Table 6. Functional outcomes by measure and followup (more...)

Table 6. Functional outcomes by measure and followup interval based on number of subjects

Study	Scale	Group	Base Score	n	SD	Followup Score	n	SD	Followup in Years
HSS / patients
0–2 years
Ververli et al., 199552	HSS	Group 1	63.5	42	10.7	84.5	42	12.1	2
		Group 2	65	41	9.5	81.3	41	11.1
Worland et al., 199853	HSS	Continuous passive machine	62.9	37	7	95.3	37	2.8	0.5
	HSS	Physical therapy	61.7	43	10	95.7	43	3
				163			163
				MEAN=63.3			MEAN=89.1
2.1–5 years
Hasegawa et al., 200254	HSS	With heterotopic ossification	48	9	16	91	9	7	3
	HSS	Without heterotopic ossification	48	131	14	93	131	6
Hsu et al., 199855	HSS		64	113	13.2	90	113	10.2	4.8
Larson et al., 200156	HSS		58	82	10.25	89	82	8.75	4
Liu & Chen, 199857	HSS	Group 1	42	64	9.58	84.1	64	4.81	2
		Group 2	47.4	24	11.7	85.3	24	4.51
Moskal & Diduch, 199858	HSS		48	514	15	89	488	22.25	4.3
Pereira et al., 199859	HSS	PCL sacrificing	51.12	67	NR	92.16	67	NR	3
		PCLsparing	56.08	40	NR	90.2	40	NR
Rand & Gustilo, 199660	HSS		59	182	10	88	182	8	2.3
				1226			1200
				MEAN=51.9			MEAN=89.3
>5 years
Diduch et al., 199761	HSS		55	88	11	92	80	6	8
Evanich et al., 199762	HSS		58	251	NR	98	169	NR	7.6
Healy et al., 200263	HSS	1992	57.68	56	11	86.92	56	10.5	5,8 years
	HSS	1995	60.64	103	15	88.06	103	9.25
Ikejiani et al., 200064	HSS	With patellar resurfacing	56	45	13.4	91	45	7.4	6.5
		Without patellar resurfacing	54.8	140	12.7	89.1	140	9.5
Malkani et al., 199565	HSS		55	118	12	81	84	9	10
O'Rourke et al., 200266	HSS	Modular tibial component	59	106	12	87	92	9.5	6.4
		All polyethylene tibial components	71	28	9.25	87	22	5
Regner et al., 199767	HSS		42	120	10	82	103	10	6.8
Schroder et al., 200168	HSS		52	102	12	91	52	8	10
				1157			946
				MEAN=55.4			MEAN=89.1
KS / patients
0–2 years
Bert et al., 200169	KS	KS	41	264	15	85	90	16	1
		Function	45	264	17	71	90	19
Bourne et al., 199570	KS clinical	Patella resurfaced	37	50	15	81	50	15	2
		Patella not resurfaced	41	50	14	87	50	8
	KS function	Patella resurfaced	41	50	13	67	50	26
		Patella not resurfaced	44	50	13	76	50	19
Cohen et al., 199771	KS	Unilateral group	55	100	NR	87	100	NR	0.5
		Bilateral group	53	86	NR	89	86	NR
Deshmukh et al., 200227	KS	KS score	23	177	16	79	130	19	1
		KS function	42	177	17	63	130	24
Heck et al., 199872	KS	KS score	34.7	291	22.2	68.4	268	19.6	2
		KS function	41.2	291	18.8	69	268	26.2
Lin et al., 200273	KS score	Preclinical pathway	43	53	12.54	93.5	36	4.77	2
		Clinical pathway	40.56	69	16.86	93.68	42	2.71
	KS function	Preclinical pathway	34.14	53	22.76	84.72	36	9.47
		Clinical pathway	46.67	69	13.18	84.21	42	10.71
Matsueda & Gustilo, 200074	KS score	Parapatellar	52	143	18.5	90	143	10	0.5
		Subvastus	51	148	15.25	90	148	12.5
	KS/ function	Parapatellar	46	143	20	74	143	17.5
		Subvastus	47	148	17.5	75	148	17.5
				2676			2100
				MEAN=42.3			MEAN=77.6
2.1–5 years
Bullens et al., 200175	KS		32.9	108	16.3	83.5	86	12.9	4.9
Elke et al.,199576	KS score	OA	30	300		87	187
		RA	21	43		77	27
	KS function	OA	50	300		65	187
		RA	40	43		67	27
Jenny & Jenny, 199877	KS score	ACL replacing	50	32	12	89	32	11	2.5
		ACL retaining	41	93	16	90	93	11
	KS function	ACL retaining	41	32	20	80	32	13
		ACL replacing	38	93	23	79	93	21
Konig et al., 199830	KS score		28.7	249	NR	82.3	249	NR	3.3
Meding et al., 200178	KS score		37.3	1888	NR	84	1888	NR	2.5
	KS function		42.9	1888	NR	78	1888	NR
Ranawat et al., 199779	KS score		44	118	15	93	96	10.75	4.9
	KS function		40	118	17.5	78	96	25
Rand & Gustilo, 199660	KS/pain		40	195	15	89	182	11	2.3
	KS function		46	195	17	81	182	20
Rodriguez et al., 199680	KS score		28	99	NR	55	67	NR	4.3
Yang et al., 200181	KS score		37	86	13.5	79	86	11	3
	KS function		44	86	16.25	64	86	14.25
				5966			5584
				MEAN=37.2			MEAN=80.6
>5 years
Brown et al., 200182	KS	Symmetric	51	250	12	90	250	12	6.4
		Asymmetric	54	18	14	91	18	10
Clouter et al., 200183	KS	KS score	33	130	10	90.7	89	8.5	10
		KS function	44	130	16	82	89	21
Duffy et al., 199884	KS score	Cement	32	47	17.9	92.4	47	8.2	10
		Cementless	33	46	18.9	87.8	46	13.8
	KS/ function	Cement	45.4	47	22.4	72.4	47	25.9
		Cementless	52.3	46	20.7	66.3	46	29.1
Ewald et al., 199985	KS score		42	180	NR	82	180	NR	10 to 14
	KS function		37	180	NR	68	180	NR
Gill et al., 199986	KS score		39	223	17	90	223	25	16.8
	KS function		44	223	20	58	223	25
Healy et al., 200263	KS score	1995	51.58	103	23.25	92.11	103	10	5,8
		1992	43.61	56	15.25	90.75	56	13.75
	KS function	1995	49.9	103	25	75.11	103	20
		1992	45.18	56	20	74.69	56	25
Indelli et al., 200287	KS score		41	91	18	94	85	11	7.5
	KS function		48	91	24	79	85	18
Martin et al., 199788	KS score		28	231	19.2	88	231	7.6	6.5
	KS function		49	231	NR	72	231	NR
Miyasaka et al., 199789	KS score		28.1	83	14.7	88.7	46	9.7	14
	KS function		30.2	83	22.2	69.2	46	28.6
Mokris et al., 199790	KS/pain		50	90	16.75	97	90	8.25	4.25
	KS function		41	90	18.75	88	90	15
Mont et al., 199991	KS score		52	101	13	94	101	8.5	5.4
	KS function		42	101	20	70	101	25
O'Rourke et al., 200266	KS score	Modular tibial component	30	106	15	85	92	15.25	6.4
		All polyethylene tibial component	34	28	13	87	22	11.5
	KS function	Modular tibial component	50	106	17.5	79	92	17.5
		All polyethylene tibial component	64	28	11.25	79	22	17.5
Rinta-Kiikka et al., 199692	KS score		48.5	97	NR	76.9	89	NR	5.3
	KS function		42.6	97	NR	64.2	89	NR
Sextro et al., 200193	KS score		32.8	61	16	87.9	50	14.2	15.7
	KS function		48.7	66	16.5	51.3	50	32.9
				3619			3368
				MEAN=42.3			MEAN=80.5
WOMAC / patients
0–2 years
Bachmeier et al., 200194	WOMAC	Physical function	38.3	108		54.8	48		0.8
Fortin et al., 199928	WOMAC function	High function	24.3	59		39.1	59		0.5
		Low function	44.2	47		65.4	47
Beaupre et al., 200195	WOMAC	Slider board pain	46	40	13	85	32	15	0.4
		Continuous passive machine pain	47	38	14	76	34	15
		Control pain	51	39	15	79	34	16
		Slider board stiffness	50	40	22	73	32	19
		Continuous passive machine stiffness	44	38	15	65	34	21
		Control stiffness	49	39	18	69	34	19
		Slider board function	41	40	13	81	32	15
		Continuous passive machine function	51	38	14	74	34	15
		Control function	53	39	15	77	34	18
Jones et al., 200196	WOMAC pain	<80 years	44	221	18	78	221	19	0.5
		≥80 years	41	35	16	73	35	20
	WOMAC function	<80 years	43	221	18	72	221	18
		≥80 years	38	35	12	66	35	17
	WOMAC stiffness	<80 years	39	221	21	64	221	22
		≥80 years	43	35	21	65	35	23
Stickles et al., 200197	WOMAC	Body Mass Index <25	57	146	NR	77.5	146	NR	1
		25–30	53.7	304	NR	77.1	304	NR
		30–35	49.9	271	NR	73	271	NR
		35–40	46.8	149	NR	72.1	149	NR
		>40	46.9	92	NR	73.6	92	NR
				2295			2184
				MEAN=46.2			MEAN=71.9
2.1–5 years
Clark et al., 200198	WOMAC	Posterior stabilized	50.4	76		78	57		3
		Cruciate retaining	47.2	67		75.9	51
				143			108
				MEAN=48.9			MEAN=77.0
>5 years
Hawker et al., 199829			58.2	487		98.4	487
				MEAN=58.2			MEAN=98.4
SF-36 / patients
0–2 years
Bachmeier et al., 200194	SF-36/ physical function		25.2	108	17.2	49.7	45	27	0.8
Beaupre et al., 200195	SF-36/ physical function	Slider board	31	40	19	53	32	24	0.4
		Continuus passive machine	31	39	15	46	36	20
		Control	31	40	22	55	34	27
Bert et al., 200099	SF-36/ physical function		29	254	7	41	158	11	1
Fortin et al., 199928	SF-36/ physical function	High function	37.3	59	22.2	63	59	25	0.5
		Low function	14.9	47	12.2	47	47	26.8
Heck et al., 199872	SF-36/ physical function		24.2	291	17.01	50.9	268	26.2	2
Jones et al., 200196	SF-36/ physical function	<80 age	21	221	18	47	221	25	0.5
		≥80 age	17	35	17	35	35	23
Kiebzak et al., 2002100	SF-36/ physical function		27	70	NR	50	70	NR	2
Stickles et al., 200197	SF-36/ physical function	Body Mss Index <25	32.2	146	NR	40.2	146	NR	1
		25–30	30.7	304	NR	40	304	NR
		30–35	30	271	NR	38.3	271	NR
		35–40	27.8	149	NR	37.3	149	NR
		>40	28.1	92	NR	37.9	92	NR
				2166			1967
				MEAN=27.58			MEAN=43.76

Table 7. Functional outcomes by measure and followup (more...)

Table 7. Functional outcomes by measure and followup interval based on number of knees

	Study	Scale	Group	Base Score	Number of Knees	SD	Followup Score	Number of Knees	SD	Followup Years
HSS / KNEES 0–2 years
	Worland et al., 199853	HSS	Continous passive motion	62.9	49	7	95.3	49	2.8	0.5
			Physical therapy	61.7	54	10	95.7	54	3
				TOTALS	103			103
					MEAN=62.3			MEAN=95.5
2.1–5 years
	Baldwin & Rubinstein, 1996101	HSS	Group A	55	272	NR	90	272	NR	4
			Group B	48	74	NR	87	74	NR
	Liu & Chen, 199857	HSS	Group 1	42	128	9.58	84.1	128	4.81	2.6
			Group 2	47.4	48	11.7	85.3	48	4.51
	Hsu et al., 199855	HSS		64	140	13.2	90	140	10.2	4.8
	Larson et al., 200156	HSS		58	118	10.25	89	118	8.75	4
	Moskal & Diduch, 199858	HSS		48	646	15	89	617	22.25	4.3
	Pereira et al., 199859	HSS	PCL sacrificing	51.12	93	NR	92.16	93	NR	3
			PCL sparing	56.08	50	NR	90.2	50	NR
	Rand & Gustilo, 199660	HSS		59	251	10	88	251	8	2.3
				TOTALS	1820			1791
					MEAN=52.4			MEAN=88.8
> 5 years
	Diduch et al., 199761	HSS		55	114	11	92	103	6	18
	Harwin, 1998102	HSS	OA	49	241	6	93	241	10	5.1
			RA	42	109	8	84	109	10
	Ikejiani et al., 200064	HSS	Patellar resurfacing	56	45	13.4	91	45	7.4	6.5
			without patellar resurfacing	54.8	140	12.7	89.1	140	9.5
	Malkani et al., 199565	HSS		55	168	12	81	119	9	10
	O'Rourke et al., 200266	HSS	Modular tibial component	59	145	12	87	128	9.5	6.4
			All-polyethylene tibial component	71	31	9.25	87	25	5
	Regner et al., 199767	HSS		42	144	10	82	106	10	6.8
	Schroder et al., 200168	HSS		52	114	12	91	58	8	10
	Healy et al., 200263	HSS	1992	57.68	56	11	86.92	56	10.5	5,8 years
		HSS	1995	60.64	103	15	88.06	103	9.25
				TOTALS	1410			1233
					MEAN=52.7			MEAN=88.0
KS 0–2 years
	Cohen et al., 199771	KS	Unilateral group	55	172	NR	87	172	NR	0.5
			Bilateral group	53	100	NR	89	100	NR
	Matsueda & Gustilo, 200074	KS score	Parapatellar	52	169	18.5	90	169	10	0.5
			Subvastus	51	167	15.25	90	167	12.5
		KS/function	Parapatellar	46	169	20	74	169	17.5
			Subvastus	47	167	17.5	75	167	17.5
				TOTALS	944			944
					MEAN=50.5			MEAN=83.8
2.1–5 years
	Bullens et al., 200175	KS		32.9	126	16.3	83.5	100	12.9	4.9
	Gioe & Bowman, 2000103	KS score	All-polyethylene tibial component	38.1	103	15.4	84.3	103	14.2	4.1
			Metal-backed tibial component	35.4	97	16.1	85.4	97	11.8
		KS/function	All-polyethylene tibial component	55.9	103	15.4	74.4	103	19.6
			Metal-backed tibial component	57.2	97	17.2	72.1	97	22.1
	Hube et al., 2002104	KS		52.3	297	NR	90.6	276	6.25	3
	Jenny & Jenny, 199877	KS score	ACL retaining	50	32	12	89	32	11	2.5
			ACL replacing	41	93	16	90	93	11
		KS/function	ACL retaining	41	32	20	80	32	13
			ACL retaining	38	93	23	79	93	21
	Jordan et al., 1997105	KS score		29	472	12.25	93	410	3.25	4.7
		KS/function		34	472	11.25	92	410	22.5
	Konig et al., 1997106	KS score		28.7	276	NR	82.3	276	NR	4.7
	Meding et al., 200178	KS score		37.3	2759	NR	84	2759	NR	2.5
		KS/function		42.9	2759	NR	78	2759	NR
	Mokris et al., 199790	KS/pain		50	105	16.75	97	105	8.25	4.25
		KS/function		41	105	18.75	88	105	15
	Ranawat et al., 199779	KS score		44	150	15	93	125	10.75	4.9
		KS/function		40	150	17.5	78	125	25
	Rand & Gustilo, 199660	KS/pain		40	277	15	89	251	11	2.3
		KS/function		46	277	17	81	251	20
	Rodriguez et al., 199680	KS score		28	145	NR	55	104	NR	4.3
	Title et al., 2001107	KS score	Total condylar prosthesis	43.4	74	NR	95.4	74	3.5	4.3
			Press-fit condylar prosthesis	44	74	NR	96.7	74	3.2
		KS/function	Total condylar prosthesis	31	74	NR	85.5	74	20.8
			Press-fit condylar prosthesis	30.4	74	NR	92.2	74	19.5
	Yang et al., 200181	KS score		37	109	13.5	79	109	11	3
		KS/function		44	109	16.25	64	109	14.25
				TOTALS	9534			9220
					MEAN=39.6			MEAN=82.8
> 5 years
	Brown et al., 200182	KS	Symmetric	51	500	12	90	500	12	6.4
			Asymmetric	54	36	14	91	36	10
	Cloutier et al., 200183	KS	KS score	33	163	10	90.7	163	8.5	10
			KS function	44	107	16	82	107	21
	Duffy et al., 199884	KS score	Cemented	32	51	17.9	92.4	51	8.2	10
			Cementless	33	55	18.9	87.8	55	13.8
		KS/function	Cemented	45.4	51	22.4	72.4	51	25.9
			Cementless	52.3	55	20.7	66.3	55	29.1
	Ewald et al., 199985	KS score		42	306	NR	82	306	NR	>10
		KS/function		37	306	NR	68	306	NR
	Gill et al., 200186	KS score		39	254	17	90	254	25	16.8
		KS/function		44	254	20	58	254	25
	Harwin, 1998102	KS score	OA	42	241	6.75	92	241	4.5	5.1
			RA	32	109	8	86	109	5
		KS/function	OS	52	241	6.5	90	241	6.5
			RA	28	109	11	68	109	7
	Indelli et al., 200287	KS score		41	100	18	94	92	11	7.5
		KS/function		48	100	24	79	92	18
	Martin et al., 199788	KS score		28	306	19.2	88	306	7.6	6.5
		KS/function		49	306	NR	72	306	NR
	Miyasaka et al., 199789	KS score		28.1	108	14.7	88.7	60	9.7	14.1
		KS/function		30.2	108	22.2	69.2	60	28.6
	Mont et al., 199991	KS score		52	118	13	94	118	8.5	5.4
		KS/function		42	118	20	70	118	25
	O'Rourke, et al., 200266	KS score	Modular tibial component	30	145	15	85	128	15.25	6.4
			All-polyethylene tibial component	34	31	13	87	25	11.5
		KS/function	Modular tibial component	50	145	17.5	79	128	17.5
			All-polyethylene tibial component	64	31	11.25	79	25	17.5
	Rinta-Kiikka et al., 199692	KS score		48.5	102	NR	76.9	100	NR	5.3
		KS/function		42.6	102	NR	64.2	100	NR
				TOTALS	4658		TOTALS	4496
					MEAN=42.4			MEAN=81.4
WOMAC No Studies
SF-36 0–2 years
	Jones et al., 200196	SF-36/ physical function	<80 years	21	221	18	47	221	25	0.5
			≥80 years	17	35	17	35	35	23
				TOTALS	256			256
					MEAN=20.5			MEAN=45.4
2.1–5 years
	Giow & Bowman, 2000103	SF-36/ physical function	All-polyethylene tibial component	25	103	36	45	103	47	4.1
			Metal-backed tibial component	25	97	18	54	97	23
				TOTALS	200			200
					MEAN=25			MEAN=49.4
> 5 years No Studies

Table 6 shows the mean scores at baseline and followup for each of the four major scales, organized by length of followup, analyzed in terms of patients; Table 7 shows the same data analyzed by knees. Baseline scores were highest in studies using the HSS and lowest in studies using the KS. This may reflect differences in severity of pain and function among subjects enrolled in these studies. HSS scores improved by about the same order of magnitude for each followup period; baseline scores were in the mid 50s and followup scores were in the high 80s and low 90s. The same general pattern applied to the KS scores but the results were a little less dramatic. The baseline values were in the high 30s and low 40s and the mean followup scores were high 70s and low 80s. The WOMAC scores showed more variation; the studies addressing followup at less than five years showed baseline mean values in the high 40s and followup values in the 70s, but the single study with more than five years of followup showed a mean baseline of 58.2 and a followup mean score of 98.4. The SF-36 mean functional scores increased from the mid-20s to the mid 40s, a level that still shows substantial limitations. Although there is no formal basis for translating the size of a change in the scores, the generally accepted rule of thumb for the KS and HSS scales is that a score of less than 60 is considered poor; 60–69 represents a fair results; 70–84 is considered a good results; 85–100 is considered an excellent result.

Table 8. Meta analysis for HSS

Table 8. Meta analysis for HSS

(0–2 years)
	Citation	N1	N2	NTotal	Effect	Lower	Upper
	Ververeli et. al., 199552	41	41	82	1.56	1.06	2.07
	Ververeli et. al., 199552	42	42	84	1.82	1.30	2.34
	Worland et. al., 199853	37	37	74	6.01	4.90	7.13
	Worland et. al., 199853	43	43	86	4.56	3.74	5.39
Fixed	Combined (4)	163	163	326	2.47	2.15	2.78
Random	Combined (4)	163	163	326	3.43	1.66	5.21
(2–5 years)
	Citation	N1	N2	NTotal	Effect	Lower	Upper
	Hasegawa et. al., 200254	9	9	18	3.32	1.66	4.98
	Hasegawa et. al., 200254	131	131	262	4.17	3.73	4.60
	Hsu et. al., 199855	113	113	226	2.20	1.86	2.53
	Larson et. al., 200156	82	82	164	3.24	2.77	3.71
	Liu & Chen, 1998 57	64	64	128	5.51	4.74	6.29
	Liu & Chen, 199857	24	24	48	4.21	3.13	5.28
	Moskal & Diduch, 199858	488	514	1002	2.15	2.00	2.31
	Rand & Gustilo, 199660	182	182	364	3.20	2.88	3.51
Fixed	Combined (8)	1093	1119	2212	2.63	2.51	2.74
Random	Combined (8)	1093	1119	2212	3.45	2.74	4.16
(5 + years)
	Citation	N1	N2	NTotal	Effect	Lower	Upper
	Diduch et. al., 199761	80	88	168	4.10	3.56	4.65
	Healy et. al., 200263	103	103	206	2.16	1.81	2.51
	Healy et. al., 200263	56	56	112	2.68	2.16	3.20
	Ikejiani et. al., 200064	45	45	90	3.21	2.56	3.85
	Ikejiani et. al., 200064	140	140	280	3.02	2.68	3.37
	Malkani et. al., 199565	84	118	202	2.39	2.02	2.75
	O'Rourke et. al., 200266	92	106	198	2.56	2.18	2.94
	O'Rourke et. al., 200266	22	28	50	2.05	1.33	2.76
	Regner et. al., 199767	103	120	223	3.99	3.53	4.45
	Schroder et. al., 200168	52	102	154	3.59	3.06	4.11
Fixed	Combined (10)	777	906	1683	2.87	2.73	3.01
Random	Combined (10)	777	906	1683	2.97	2.53	3.40

Table 9. Meta analysis for KS

Table 9. Meta analysis for KS

(0–2 years)
	Citation	N1	N2	NTotal	Effect	Lower	Upper
	Bert et. al., 2000, 200169	90	264	354	1.48	1.22	1.74
	Bert et. al., 2000, 2001 69	90	264	354	2.88	2.56	3.20
	Bourne et. al., 199570	50	50	100	2.91	2.34	3.49
	Bourne et. al., 199570	50	50	100	4.00	3.31	4.70
	Bourne et. al., 199570	19	50	69	1.12	.55	1.69
	Bourne et. al., 199570	26	50	76	.83	.33	1.33
	Deshmukh et. al., 200227	130	177	307	1.03	.79	1.28
	Deshmukh et. al., 200227	130	177	307	3.22	2.88	3.57
	Heck et. al., 199872	268	291	559	1.24	1.05	1.42
	Heck et. al., 199872	268	291	559	1.56	1.37	1.75
	Lin et. al., 200273	36	53	89	4.94	4.07	5.81
	Lin et. al., 200273	42	69	111	3.92	3.26	4.58
	Lin et. al., 200273	36	53	89	2.69	2.10	3.29
	Lin et. al., 200273	42	69	111	2.99	2.43	3.55
	Matsueda & Gustilo, 200074	148	148	296	1.60	1.33	1.86
	Matsueda & Gustilo, 200074	143	143	286	2.55	2.23	2.86
	Matsueda & Gustilo, 200074	148	148	296	2.78	2.46	3.11
	Matsueda & Gustilo, 200074	143	143	286	1.49	1.22	1.75
Fixed	Combined (18)	1859	2490	4349	1.85	1.77	1.92
Random	Combined (18)	1859	2490	4349	2.35	1.95	2.76
(2–5 years)
	Citation	N1	N2	NTotal	Effect	Lower	Upper
	Bullens et. al., 200175	86	108	194	3.41	2.96	3.86
	Jenny & Jenny, 199877	32	32	64	3.35	2.56	4.14
	Jenny & Jenny, 199877	32	32	64	2.28	1.63	2.93
	Jenny & Jenny, 199877	93	93	186	1.85	1.51	2.20
	Jenny & Jenny, 199877	93	93	186	3.55	3.09	4.02
	Ranawat et. al., 199779	96	118	214	3.66	3.21	4.10
	Ranawat et. al., 199779	96	118	214	1.79	1.47	2.11
	Rand & Gustilo, 199660	182	195	377	1.89	1.64	2.13
	Rand & Gustilo, 199660	182	195	377	3.70	3.36	4.03
	Yang et. al., 200181	86	86	172	1.32	.99	1.66
	Yang et. al., 200181	86	86	172	3.40	2.92	3.87
Fixed	Combined (11)	1064	1156	2220	2.47	2.35	2.58
Random	Combined (11)	1064	1156	2220	2.73	2.16	3.30
(5 + years)
	Citation	N1	N2	NTotal	Effect	Lower	Upper
	Brown et. al., 200182	18	18	36	2.97	1.96	3.99
	Brown et. al., 200182	250	250	500	3.25	2.98	3.51
	Cloutier et. al., 200183	89	130	219	6.14	5.49	6.78
	Cloutier et. al., 200183	89	130	219	2.08	1.75	2.42
	Duffy et. al., 199884	46	46	92	3.27	2.63	3.91
	Duffy et. al., 199884	47	47	94	1.11	.67	1.55
	Duffy et. al., 199884	47	47	94	4.27	3.52	5.03
	Duffy et. al., 199884	46	46	92	.55	.13	.97
	Gill et al., 1999108	223	223	446	.62	.43	.81
	Gill et al., 1999108	223	223	446	2.38	2.14	2.63
	Healy et. al., 200263	103	103	206	1.11	.81	1.40
	Healy et. al., 200263	103	103	206	2.26	1.91	2.61
	Healy et. al., 200263	56	56	112	1.29	.88	1.71
	Healy et. al., 200263	56	56	112	3.22	2.64	3.79
	Indelli et. al., 200287	85	91	176	3.51	3.03	3.99
	Indelli et. al., 200287	85	91	176	1.45	1.11	1.78
	Martin et. al., 199788	231	231	462	4.11	3.79	4.43
	Miyasaka et. al., 199789	46	83	129	1.57	1.16	1.98
	Miyasaka et. al., 199789	46	83	129	4.51	3.84	5.18
	Mokris et. al., 199790	90	90	180	2.73	2.32	3.15
	Mokris et. al., 199790	90	90	180	3.52	3.05	4.00
	Mont et. al, 199991	101	101	202	1.23	.93	1.54
	Mont et. al., 199991	101	102	203	3.74	3.28	4.20
	O'Rourke et. al., 200266	92	106	198	3.65	3.19	4.11
	O'Rourke et. al., 200266	92	106	198	1.65	1.32	1.98
	O'Rourke et. al., 200266	22	28	50	4.15	3.11	5.19
	O'Rourke et. al., 200266	22	28	50	1.04	.43	1.65
	Sextro et. al., 200193	50	61	111	3.61	2.99	4.23
	Sextro et. al., 200193	50	66	116	.09	-.28	.47
Fixed	Combined (29)	2599	2835	5434	2.07	2.00	2.14
Random	Combined (29)	2599	2835	5434	2.57	2.08	3.05

Table 10. Meta analysis for WOMAC

Table 10. Meta analysis for WOMAC

(0–2 years)
	Citation	N1	N2	NTotal	Effect	Lower	Upper
	Beaupre et. al., 200195	34	38	72	1.98	1.40	2.56
	Beaupre et. al., 200195	34	39	73	1.79	1.23	2.35
	Beaupre et. al., 200195	32	40	72	1.10	.59	1.61
	Beaupre et. al., 200195	32	40	72	2.77	2.10	3.44
	Beaupre et. al., 200195	34	38	72	1.15	.64	1.66
	Beaupre et. al., 200195	34	39	73	1.07	.57	1.57
	Beaupre et. al., 200195	32	40	72	2.84	2.16	3.52
	Beaupre et. al., 200195	34	39	73	1.44	.91	1.97
	Beaupre et. al., 200195	34	38	72	1.57	1.03	2.11
	Jones et. al., 200196	221	221	442	1.16	.96	1.36
	Jones et. al., 200196	221	221	442	1.83	1.61	2.06
	Jones et. al., 200196	35	35	70	1.88	1.30	2.46
	Jones et. al., 200196	221	221	442	1.61	1.39	1.82
	Jones et. al., 200196	35	35	70	1.75	1.18	2.31
	Jones et. al., 200196	35	35	70	.99	.48	1.49
Fixed	Combined (15)	1068	1119	2187	1.54	1.44	1.64
Random	Combined (15)	1068	1119	2187	1.62	1.39	1.86

Table 11. Meta analysis for SF-36

Table 11. Meta analysis for SF-36

(0–2 years)
	Citation	N1	N2	NTotal	Effect	Lower	Upper
	Bachmeier et. al., 200194	45	108	153	1.22	.84	1.60
	Beaupre et. al., 200195	32	40	72	1.02	.51	1.52
	Beaupre et. al., 200195	36	39	75	.84	.36	1.33
	Beaupre et. al., 200195	34	40	74	.97	.48	1.47
	Bert et. al., 2000, 200169,99	158	254	412	1.37	1.15	1.59
	Fortin et. al., 199928	47	47	94	1.52	1.05	1.99
	Fortin et. al., 199928	59	59	118	1.10	.70	1.49
	Heck et. al., 199872	268	291	559	1.24	1.06	1.42
	Jones et. al., 200196	35	35	70	.88	.38	1.38
	Jones et. al., 200196	221	221	442	1.19	.99	1.39
Fixed	Combined (10)	935	1134	2069	1.21	1.11	1.30
Random	Combined (10)	935	1134	2069	1.20	1.10	1.30

Tables 8–11 display the effect size (defined as the number of standard deviations of change) for this same data. The functional scores after TKA are consistently higher. The mean effect size for the HSS studies is 3.91 for those with followup up to two years, 3.01 for those 2–5 years, and 2.97 for those studies with more than five years of followup. For the studies using KS scores the mean effect size is 2.35 for those 0–2 years, 2.73 for those 2–5 years, and 2.67 for those 5+ years. For WOMAC studies the mean effect size for 0–2 years of followup is 1.62. The more generic SF-36 scores had the smallest mean effect size; for the studies with 0–2 years of followup it was 1.27 (though this is still considered a “large effect size”). The effect size is considerably higher for those studies where the clinician reports the results compared to those where patient reports are used.

Revisions and Complications

Table 12. Revision rates after primary TKAs

Table 12. Revision rates after primary TKAs

	Based on Knees		Based on Patients
Followup	Revisions	Revisions Other Procedures	Revisions	Revisions Other Procedures
0–2 years	0	0	< 1%	< 1%
2.1-5 years	2.0%	3.5%	1.6%	2.9%
5+ years	2.0%	3.1%	2.1%	3.5%

Revision rates were calculated in several ways. The basic data are shown as an evidence table in Appendix F, Evidence Table 2. Table 12 summarizes the revision rates for primary TKAs. The results are organized to show the rates at different followup intervals and are grouped by both knees and patients. The revision rates are further subdivided into operations specified as revisions and all procedures performed on the knees in question. The revision rate through five or more years is 2.0 percent of knees and 2.1 percent of patients.

The data base used to calculate perioperative complication rates (defined as occurring within six months of the TKA) is shown in Appendix F, Evidence Table 3. Complications were defined by each investigator. The vast majority were “knee related” or deep venous thrombosis. When the unit of analysis was numbers of knees operated on, the complication rate was 5.4 percent; when the denominator was numbers of patients, the rate was 7.6 percent. There were essentially no cardiopulmonary complications reported. Given the number of elderly subjects undergoing a major surgical procedure this suggests that these adverse effects were not addressed in the literature.

Table 13. Assessment of TKA prostheses and surgical (more...)

Table 13. Assessment of TKA prostheses and surgical procedures

Study	Prosthesis Type	Measure(s) and Baseline Score	Followup Length and Score	Notes
Prothesis
Baldwin & Rubinstein, 1996101	Intermedics Natural Knee TKA, (1) Subjects with excellent/good bone quality (GB) vs. (2) Subjects with fair/bad bone quality (BB)	Hospital of Special Surgery (HSS)	Followup = 4 years	Study concludes bone quality had little effect on the four-year outcome of this ingrowth TKA.
		HSS	HSS
		GB: 55	GB: 92
		BB: 48	BB: 90
Bert et al., 2000, 200169,99	Total condylar TKA. Low demand patients were randomized to receive either (1) All-polyethylene or (2) metal-backed implant type. Not reported for medium/high demand subjects	Knee Society Knee Score (KS)	Followup = 1 year	Study hypothesis that prosthetic choice should be determined by peroperative activity level (demand matching) was not validated.
		AP, low demand: 41	KS
		MB, low demand: 38	AP, low demand: 82
		Function score (KSF)	MB, low demand: 87
		AP, low demand: 41	KSF
		MB, low demand: 43	AP, low demand: 72
			MB, low demand: 54
Cloutier et al., 200183	Total condylar, posterior cruciate-retaining	KS: 33	Followup = 10 years	After TKA with PCR both anterior and posterior cruciate ligaments (even degenerate) remain functional after an average of 10 years. Survival at 10 years with end point being revision was 94.8%.
		KSF: 44	KS: 90.7
			KSF: 82
Evanich et al., 199762	Cementless Intermedics Natural Knee TKA using metal-backed, porous-coated patellar component	HSS: 58	Followup = 6–10 years	Overall patellar survivorship was 96%. Study concludes comparatively good results from the use of a metal-backed patellar component if component design, surgical technique and patellar alignment are properly addressed.
			HSS: 98
Ewald et al., 199985	Kinematic nonconstrained TKA, posterior cruciate-retaining	KS: 42	Followup = 10–14 years	Overall revision rate was 6.5%. Data from study suggests patella replacement is not appropriate with this design
		KSF: 37	KS: 82
			KSF: 68
Gill & Joshi, 200186	Cemented posterior cruciate ligament-retaining TKA. Total Condylar Knee (54%) and Kinematic Condylar (46%).	KS: 39	Followup = 16.8 years	Study finds the long-term results of cemented posterior cruciate ligament-retaining TKA excellent in terms of improved function and pain relief.
		KSF: 44	KS: 90
			KSF: 58
Gill et al., 1999108	Total Condylar Knee, posterior cruciate-retaining	KS: 40.3	Followup = 16–21 years	Prosthetic survivorship at 20 years was 96% for revision. Total Condylar with retention of the posterior cruciate produces results comparable to the original Total Condylar Knee with cruciate-sacrifice.
			KS: 88.4
Gioe & Bowman, 2000103	Press-Fit Condylar, (1) All-polyethylene (APT) vs. (2) Metal-backed tibial (MBT) components.	KS	Follow up = 3 years	Study reports TKA with all-polyethylene components functions equivalently to metal-backed tibial components, and is less costly.
		APT: 38.1	KS
		MBT: 35.4	APT: 84.3
		KSF	MBT: 85.4
		APT: 55.9	KSF
		MBT: 57.2	APT: 74.4
			MBT: 72.1
Hsu et al., 199855	Hybrid Miller-Galante I (MGI) TKA using uncemented femoral components with cemented tibial and patellar components	HSS: 64	Followup = 4.8 years	Study does not recommend MGI TKA due to high rate of patellar complications but may be a useful alternative fixation mode in TKA procedures.
			HSS: 90
Indelli et al., 200287	Insall-Burstein II	KS: 41	Followup = 7.5 years	Survivorship analysis using worst-case scenario showed a success rate of 91%.
			KS: 94
Jordan et al., 1997105	Mobile meniscal bearing TKA	KS: 29	Followup = 8 years	Kaplan-Meier survivor analysis, using revision surgery for any mechanical reason, showed a survivorship of 94.6%.
		KSF: 34	KS: 93
			KSF: 94
Larson et al., 200156	Insall-Burstein II posterior-stabilized TKA	HSS: 58	Followup = 4 years	80% and 17% of the knees were rated excellent and good, respectively. Using the patellar resurfacing technique used in this study, patellofemoral complications were only 4.2%.
			HSS: 89
Liu & Chen, 199857	Four different implants used.			Not possible to test effect of prosthesis.
Malakani et al., 199565	Kinematic Condylar prosthesis, posterior cruciate-retaining	HSS: 55	Followup = 10 years	Using revision as end point, rate of survival was 96%. Study found knee scores, rate of survival of implants were similar to reported previously subjects who had a total condylar TKA with sacrifice of the posterior cruciate ligament. Loosening of patellar components was noted to be a major problem.
		KS: 33	HSS: 81
		KSF: 46	KS: 80
			KSF: 64
Meding et al., 200178	Posterior cruciate-retaining TKA (98%)			Not possible to test effect of prosthesis.
	Insall-Burstein II posterior stabilized TKA (2%)
Miyasaka et al., 199789	Total Condylar, posterior cruciate-sacrificing	KS: 28.1	Followup = 14 years	Survival of retention of the prosthesis was 91% at 13 years.
		KSF: 30.2	KS: 88.7
			KSF: 69.2
Mokris et al., 199790	Genesis TKA system, conversion module allowing for posterior cruciate-sacrifice	KS: 50	Followup = 6.5 years	Clinically, results were excellent in 95% of knees, good in 4%.
		KSF: 41	KS: 97
			KSF: 88
Mont et al., 199991	Duracon TKA system, posterior cruciate-retaining	KS: 52	Followup = 5 years	At final follow up 96% of knees had good or excellent results. Almost complete absence of patellofemoral complications was noted.
		KSF: 42	KS: 94
			KSF: 70
O'Rourke et al., 200266	Insall-Burstein II, (1) All-polyethylene (APT) vs. (2) Cemented metal-backed tibial (MBT) components.	KS	Followup = 6.4 years	Modular Insall-Burstein II TKAs were found to function well at followup although the authors noted that the high prevalence of osteolysis in subjects with good/excellent clinical scores was worrisome. Routine followup radiographs after TKA to detect asymptomatic osteolytic changes was recommended.
		APT: 34	KS
		MBT: 30	APT: 87
		KSF	MBT: 85
		APT: 64	KSF
		MBT: 50	APT: 79
		HSS	MBT: 79
		APT: 71	HSS
		MBT: 59	APT: 87
			MBT: 87
Regner et al., 199767	Freeman-Samuelson TKA with three different types of tibial components fixed with macrointerlocking pegs: (1) High density polyethylene without stem (Group 1); (2) Metal-backed tibial without stem (Group 2); (3) Metal-backed tibial with stem (Group 3)	HSS: 42	Followup = 6.8 years	Using revision as end point, rate of survival was 79% at 10 years. Investigators found cementless fixation of this design using the macrointerlocking pegs and no other stabilization resulted in poor fixation and a high revision rate and cannot be recommended.
			HSS: 82
Rinta-Kiikka et al., 199692	Cementless Synatomic TKA, posterior cruciate-retaining	KS: 48.5	Followup = 5–7 years	Clinical survival rate, based on aseptic loosening, was 88.6%.
		KSF: 42.6	KS: 76.9
			KSF: 64.2
Ritter et al., 1995109	Anatomic Graduated Components TKA, posterior cruciate-retaining		Followup = 10.7 years	Clinical survival rate, based on revision, was 98.86% at 15 years.
			KS: 81
Rodriguez et al., 199680	Total Condylar TKA	KSF: 28	Followup = 12.7 years	At the 15-year followup period, survivorship analysis suggested a 91% probability of survival for the prosthesis. Cemented Total Condylar TKA in severe rheumatoid arthritis provided durable pain relief and restoration in function.
			KSF: 55
Schroder et al., 200168	Cementless porous-coated Anatomic Graduated Components TKA	HSS: 52	Followup = 10 years	At followup, 92% of the patients were satisfied or very satisfied with their TKA. Cumulative prosthesis survival after 10–11 years was 97%.
			HSS: 91
Sextro et al., 200193	Kinematic I condylar TKA, posterior cruciate-retaining	KS: 32.8	Followup = 15.7 years	At the 15-year followup period, survivorship was 88.7%, using revision as the endpoint. Study shows good function and survivorship of the Kinematic I condylar TKA.
		KSF: 48.7	KS: 87.9
			KSF: 51.3
Title et al., 2001107	(1) Total Condylar TKA, posterior cruciate-sacrificing (TCP) vs. (2) Press-Fit Condylar, posterior cruciate-substituting (PFC)	KS	Followup = 4 and 4.5 years	Both designs showed comparable pain relief and walking ability.
		TCP: 43.4	KS
		PFC: 44	TCP: 95.4
		KSF	PFC: 96.7
		TCP: 31	KSF
		PFC: 30.4	TCP: 85.5
			PFC: 92.2
Yang et al., 200181	Total condylar-type design with or without posterior cruciate-retention			Not possible to test effect of prosthesis.
Procedures
Bourne et al., 199570	All subjects received single type featuring an anatomic patellofemoral joint, (1) Patella resurfaced group (PR) vs. (2) Patella not resurfaced group (PNR)	KS	Followup = 2 years	The not resurfaced group had significantly less pain at two-year followup. A required longer followup suggested.
		PR: 37	KS
		PNR: 41	PR: 81
		KSF	PNR: 87
		PR: 41	KSF
		PNR: 44	PR: 67
			PNR: 76
Brown et al., 200182	Non reported identical prosthesis type, reports on component asymmetry. (1) Asymmetric TKA (AS) vs. (2) Symmetric TKA (S)	KS	Followup = 6.4 years	No statistical differences in knee scores were noted between right and left TKAs performed with asymetrically sized components.
		AS: 54	KS
		S: 51	AS: 91
			S: 90
Bullens et al., 200175	Press-Fit Condylar TKA, posterior cruciate-retaining (PCR) in 95%	KS: 32.9	Followup = 4.9 years	Five-year survival with revision as end point being revision 99% (best-case scenario), but decreased to 69% with revision, pain scale (visual analog -VAS) >20, satisfaction VAS <80, or lost to follow up as endpoint (worst-case scenario).
		KSF: 29.1	KS: 83.5
			KSF: 51.5
Clark et al., 200198	(1) Posterior cruciate-sacrificed (PCS) vs. (2) Posterior cruciate-retaining (PCR) TKAs	KS	Followup = 2 years	No notable differences between groups at years two and three of followup.
		PCS: 98.8	KS
		PCR: 100.6	PCS: 157.1
		WOMAC	PCR: 156.5
		PCS: 50.4	WOMAC
		PCR: 47.2	PCS: 22.8
			PCR: 18.5
Cohen et al., 199771	AMK, a condylar cruciate-sparing implant, (1) Bilateral, and (2) Unilateral TKA	KS	Followup = 0.5 years	Study concludes simultaneous bilateral TKA does not result in any significant increase in patient morbidity or effect post-op function compared to unilateral TKA.
		B: 53	KS
		U: 55	B: 89
			U: 87
Deshmukh et al., 200227	Cemented Kinemax, patella retained. Role of body weight investigated	KS: 23	Followup = 0.5 years	Study found body weight did not adversely affect the outcome of TKA in the short-term.
		KSF: 42	KS: 79
			KSF: 63
Diduch et al., 199761	Posterior stabilized, posterior cruciate-substituting	HSS: 55	Followup = 8 years	Survival at 18 years with end point being revision was 94%.
			HSS: 92
Duffy et al., 199884	Press-Fit Condylar, (1) Uncemented (UC) vs. (2) Cemented (C) (Press-Fit Condylar)	KS	Followup = 10 years	Survival at end point being revision or aseptic loosening was 72% in the uncemented group and 94% in the cemented group.
		UC: 33	KS
		C: 32	UC: 87.8
		KSF	C: 92.4
		UC: 52.3	KSF
		C: 45.4	UC: 66.3
			C: 72.4
Elke et al., 199576	Unconstrained posterior cruciate ligament-retaining TKA. 424 cemented, 100 uncemented TKA	KS not broken down by cemented vs. uncemented	Followup = 4.8–9.8 years	Cemented TKA can be recommended for patients with RA.
Griffin et al., 1998110	Posterior stabilized, cemented with metal-backed tibial components and patella resurfacing in obese patients	HSS	Followup = 10 years	HSS scores comparable between groups and revision rates were not higher in the obese group at followup.
		Obese: 47.7	HSS
		Not Obese: 55	Obese: 88.3
			Not Obese: 90.3
Harwin, 1998102	Kinemax cemented posterior cruciate ligament-retaining condylar with a symmetrical femoral component articulating with a medially offset symmetrical dome patella component	KS: 38	Followup = 5.1 years	Study suggests cemented TKA with symmetrical patellofemoral resurfacing with an offset patella dome and posterior cruciate ligament-retention yields low patellofemoral complications and reoperations
		KSF: 47	KS: 91
			KSF: 86
Hasegawa et al., 200254	Cruciate-retaining (cementless) and posterior stabilized (cemented)			Not possible to test effect of prosthesis.
Hube et al., 2002104	Midvastus approach for TKA	KS: 52.3	Followup = 3 years	95% of the patients had excellent or good functional result.
			KS: 90.6
Ilkejiani et al., 200064	Genesis knee system, (1) Patella resurfaced group (PR) vs. (2) Patella not resurfaced group (PNR)	HSS	Followup = 2 years	No significant difference between groups with regard to pain, HSS scores, and complications.
		PR: 54.8	HSS
		PNR: 56.0	PR: 89.1
			PNR: 91
Jenny & Jenny, 199877	Search total knee prosthesis which allows retention or replacement of the anterior cruciate ligament (ACL). (1) ACL-retaining group (AR) vs, (2) ACL-replacing (ARP)	KS	Followup = 2–3 years	Results showed clinical and functional outcomes were neither improved nor worsened with the ACL-retaining prosthesis.
		AR: 50	KS
		ARP: 41	AR: 89
		KSF	ARP: 90
		AR: 41	KSF
		ARP: 38	AR: 80
			ARP: 79
Konig et al., 1997, 1998, 200030,106,111	Posterior cruciate-retaining, press-fit condylar TKA using uncemented femoral components with cemented tibial and patellar components	Data from Konig et al., 199785	Followup = 3.2 years	Study showed hybrid TKA provides good results comparable to cemented TKA.
		KS: 28.7	KS: 82.3
		KSF: 45.5	KSF: 71.9
Lombardi Jr et al., 2001112	(1) Maxim posterior cruciate-retaining (PCR) vs. (2) Maxim posterior cruciate-sacrificing (PCS)	KS	Followup = 5 years	No significant differences in outcome between the groups were observed.
		PCR: 118.04	KS - Total
		PCS: 112.90	PCR: 162.16
		KSF	PCS: 158.05
		PCR: 54.77	KSF
		PCS: 47.91	PCR: 71.22
		KS-Pain	PCS: 66.77
		PCR: 16.67	KSF - Pain
		PCS: 13.63	PCR: 44.23
		Converted from HSS	PCS: 44.10
Martin et al., 199788	Press-Fit Condylar TKA.	KS: 28	Followup = 6.5 years	Study reports good results with the Press-Fit Condylar, 95% of patients were pain free on level walking and were satisfied with their functional result.
		KSF: 49	KS: 88
			KSF: 72
Matsueda & Gustilo, 200074	Genesis TKA system	KS	Follow up = 0.5 years	There were no significant differences in the KS score.
		S: 51	KS
		MP: 52	S: 90
		KSF	MP: 90
		S: 47	KSF
		MP: 46	S: 75
			MP: 74
Moskal & Diduch, 199858	Several designs			Not possible to test effect of prosthesis.
Pereira et al., 199859	Kinemax, (1) Posterior cruciate-retaining (PCR) vs. (2) Posterior cruciate-sacrificing (PCS)	HSS	Followup = 3 years	Data revealed no difference in clinical outcome between PCR and PCS.
		PCR: 56.08	HSS
		PCS: 51.12	PCR: 90.2
			PCS: 92.16
Ranawat et al., 199779	Press-Fit Condylar modular TKA, posterior cruciate-substituting	KS: 44	Followup = 4.8 years	Study found Press-Fit Condylar modular TKA resulted in excellent relief of pain and restoration of function with a low prevalence of patellofemoral problems. Survival of the implant at 6 years was 97%.
		KSF: 40	KS: 93
			KSF: 78
Rand & Gustilo, 199660	Genesis TKA system, (1) Resurfacing patellar component (RSC) vs. (2) Inset Biconvex patellar component (BPC)	KS	Follow up = 2.3 years	At followup, KS score was higher in the RSC group. The inset BPC appeared to provide better radiographic alignment than the RSC, but it had a higher incidence of radiolucent lines.
		RSC: 37	KS
		BPC: 42	RSC: 92
		KSF	BPC: 86
		RSC: 48	KSF
		BPC: 44	RSC: 81
		HSS	BPC: 82
		RSC: 60	HSS
		BPC: 57	RSC: 88
			BPC: 88

Although the sampling approach was not specifically designed to search for all outcomes associated with using different types of prostheses or different surgical approaches, we did analyze the studies that fell within the search parameters. In some cases it was difficult to classify a study as primarily addressing either the use of a specific type of prosthesis or testing a specific surgical procedure or technique. Several studies reported prostheses that were used in specific types of procedures. Table 13 is arranged to attempt to classify the emphasis of studies by procedure or prosthesis, but some overlap is inevitable. A number of the studies of prostheses were case series that reported generally good results. A few tested the use of a prosthesis with a specific group of patients. The studies of procedures were a mixture of case studies and comparative studies.

TKA studies assessing prophylaxis for postoperative deep venous thrombosis (DVT) or infection were identified by searching the 611 references meeting and not meeting inclusion criteria. The Cochrane Library was also searched back to 1994. The investigators decided a priori to include only randomized controlled trials (RCTs) with the exception of large cohort studies. Fourteen studies were identified and extracted; nine DVT, three infection, and two tourniquet studies. All included studies were randomized controlled trials with the exception of one large cohort study.²⁴ One trial was identified through The Cochrane Library.²⁵

Table 14. Assessment of TKA procedures/programs

Table 14. Assessment of TKA procedures/programs

Study	Procedure Type	Measure(s) and Baseline Scores	Followup Length and Scores	Notes
Beaupre et al., 200195	(1) Standard exercise and continuous passive motion (CPM) vs. (2) Standard exercise and Slider Board (SB) vs. (3) Standard exercise alone (SE)	Western Ontario and McMaster Osteoarthritis Index (WOMAC) and SF-36, Physical Functioning (SF-36 PF)	Followup = 0.5 years	No differences between groups in WOMAC and SF-36 scores at any measurement interval. When postop rehabilitation regimens that focus on early mobilization are used, adjunct CPM or SB that are added to SE are not required.
		WOMAC	WOMAC
		Pain, CPM: 47	Pain, CPM: 76
		Stiffness, CPM: 44	Stiffness, CPM: 65
		Function, CPM: 51	Function, CPM: 74
		Pain, SB: 46	Pain, SB: 85
		Stiffness, SB: 50	Stiffness, SB: 73
		Function, SB: 41	Function, SB: 81
		Pain, SE: 51	Pain, SE: 79
		Stiffness, SE: 49	Stiffness, SE: 69
		Function, SE: 53	Function, SE: 77
		SF-36, PF	SF-36 PF
		CPM: 31	CPM: 46
		SB: 31	SB: 53
		SE: 31	SE: 55
Healy et al., 200263	(1) Clinical pathway group (CP) vs. (2) No clinical pathway group (NCP)	KS	CP Followup = 5 years	Both groups had excellent relief of pain and improvement in function. The CP program reduced resource utilization and cost.
		CP: 51.58	NCP Followup = 8 years
		NCP: 43.61	KS
		KSF	CP: 92.11
		CP: 49.90	NCP: 90.75
		NCP: 45.18	KSF
		HSS	CP: 75.11
		CP: 60.64	NCP: 74.69
		NCP: 57.68	HSS
			CP: 88.06
			NCP: 86.92
Lin et al., 200273	(1) No (or Pre) clinical pathway group (NCP) vs. (2) Clinical pathway group (CP)	KS	Followup = 2 years	No significant differences were found between groups in the knee scores. Clinical pathway is an effective management tool for TKA.
		CP: 40.6	KS
		NCP: 43.0	CP: 93.6
		KSF	NCP: 93.5
		CP: 46.7	KSF
		NCP: 34.1	CP: 84.2
			NCP:84.7
Ververeli et al., 199552	(1) Continuous passive motion (CPM) vs. (2) no CPM (NCPM)	HSS	Followup = 2 years	No clinical differences in knee scores at follow up. CPM is efficacious in increasing short-term flexion and decreasing need for knee manipulation without increasing costs.
		CPM: 63.5	HSS
		NCPM: 65	CPM: 84.5
			NCPM: 81.3
Worlund et al., 199853	(1) Continuous passive motion (CPM) vs. (2) Professional physical therapy (PT)	HSS	Followup = 0.5 years	Study concludes CPM is an adequate rehabilitation alternative with lower costs and no difference in clinical results.
		CPM: 62.9	HSS:
		PT: 61.7	CPM: 95.3
			PT: 95.7

Several other procedures, which involved primarily non-surgical elements of care, were also described. These are summarized in Table 14. Three of these addressed the use of continuous passive motion as a rehabilitative approach; two studies were positive. The other two studies tested different clinical pathways and showed mixed results.

Table 15. Complications: Prevention of Venous Thrombosis (more...)

Table 15. Complications: Prevention of Venous Thrombosis (VT)/Pulmonary Embolism (PE) studies

Study	Study Type: Intervention; Control	Results	Conclusions
Blanchard et al., 1999113	RCT: (1) Nadroparin calcium, a LMWH, adjusted to body weight., subcutaneously 12 hours before and after surgery then once a day for 10–12 days vs. (2) Continuous intermittment pneumatic compression device (CIPC) of the foot.	DVT: 31/48 (65%, 95% CI 49.5–77.8) for CIPC group and 16/60 (27%, 95% CI 16.1–39.7) for the nadroparin group (p<0.001). Only one patient in the nadroparin group had severe bleeding.	Authors conclude that a once daily fixed, weight-adjusted dose of nadroparin is superior to CIPC of the foot.
Colwell et al., 1995114	RCT, open-label: (1) Postoperative Enoxaparin 30 mg subcutaneously bid vs. (2) Unfractionated heparin 5000 IU subcutaneously three times daily. MDT = 7 (up to 14 days)	77/225 (34%) of heparin subjects and 56/228 (25%) of enoxaparin subjects had an incidence of DVT (p=0.02). Two subjects receiving heparin had a PE, one fatal. three major hemorrhagic episodes in each group.	Study found postoperative enoxaparin more effective and as safe as unfractionated heparin in preventing DVT in patients having elective TKA.
Francis et al., 1996115	RCT: (1) “Two-step” warfarin, administered 10–14 days preoperatively then postoperatively vs. (2) warfarin, one dose night before TKA. Postoperatively, dose adjusted to target INR 2.2. Treatment up to nine days.	Occurrence of DVT nearly identical, 39% in the two-step regimen vs. 38% for the night before group. Occurrence of proximal VT was 5% vs. 7%, respectively (p ns). Patients in two-step regimen received 1.3 transfusions vs. 0.95 of the night before regimen (p<0.05).	Authors conclude night before warfarin regimen is more convenient and may be associated with less bleeding than the two-step warfarin regimen.
Heit et al., 2000116	Double-blind (DB), placebo-controlled, randomized controlled trial (RCT), TKA and THA. (1) Postoperative Ardeparin, a low molecular-weight heparin (LMWH), 50 anti Xa, IU/kg body weight subcutaneously bid vs. (2) Placebo. Mean duration of treatment (MDT) = 7 days postoperative.	Results for TKA only.	Study found the cumulative incidence of symptomatic DVT or death was not significantly reduced by ardeparin prophylaxis. Authors conclude extended ardeparin use is not clinically important for most patients and future research should identify high-risk patients who would benefit most from extended prophylaxis.
		DVT, PE, or death: 5 (1.4%) for ardeparin group, 6 (1.7%) for placebo, OR = 0.8 (95% CI 0.2–2.7).
		DVT: 1 (0.3%) for ardeparin group, 3 (0.8%) for placebo, OR = 0.3 (95% CI 0.03–3.1).
Leclerc et al., 1996117	DB, RCT: (1) Postoperative Enoxaparin 30 mg subcutaneously bid vs. (2) Postoperative Warfarin, dose adjusted to INR 2–3. MDT = 9 (up to 14 days).	109 of 211 (51%) warfarin subjects had an incidence of DVT vs. 76/206 (37%) of enoxaparin subjects (p=0.003). 22 (10.4%) warfarin and 24 (11.7%) had proximal VT (p>0.2). 6 (1.8%) warfarin and 7 (2.1%) had major bleeding (p>0.2).	Study found postoperative fixed-dose enoxaparin more effective than adjusted-dose warfarin in preventing DVT after TKA. No differences were observed for incidence of proximal VT or clinically overt hemorrhage.
Leclerc et al., 199824	Cohort study, TKA and THA:	Results for TKA cohort only.	Postoperative use of enoxaparin for a mean of nine days is associated with a clinically acceptable rate of symptomatic VT and major hemorrhage. Authors conclude predischarge compression ultrasonography cannot be justified.
	Postoperative Enoxaparin, a LMWH, 30 mg subcutaneously bid. MDT = 9 (up to 14 days.	VT: 33/842 (3.9%, 95% CI 2.6–5.2)
		Symptomatic proximal VT or PE: 23/842 (2.7%)
		Fatal PE: 3/842 (0.4%)
		Major hemorrhage: 24/842 (2.9%)
Perhoniemi et al., 1996118	DB, RCT, TKA, and THA: (1) Postoperative Enoxaparin 40 mg subcutaneously once a day vs. (2) Dihydroergotamine 0.5 mg + Heparin 5000 IU (HDHE) subcutaneously bid. One dose from each group prior to surgery. Treatment for seven days.	Results for TKA and THA combined. Overall incidence of thromboembolic events was low (3%). One DVT seen in the Enoxaparin group and two PE in HDHE group.	Study found the two regimens showed comparable efficacy and overall safety in preventing DVT.
Robinson et al., 1997119	DB, RCT, TKA, and THA. All subjects receiving postoperative warfarin adjusted to target International Normalized Ratio (INR) 2–3 for a mean of 9.8 days were randomized to (1) Bilateral Compression Ultrasonography vs. (2) Sham.	Results for TKA only. In the screening group, one subject developed symptomatic proximal DVT and one had a nonfatal PE. One subject in the sham group had a symptomatic proximal DVT. Asymptomatic DVT was detected in six subjects in the screening group.	Use of warfarin prophylaxis during hospitalization results in very low rates of symptomatic DVT or PE after discharge. Authors conclude use of compression ultrasonography is not justified in this setting.
Westrich & Sculco, 1996120	RCT: (1) Postoperative Aspirin 325 mg bid (also given night of operation) vs. (2) Postoperative Aspirin + Pulsatile pneumatic plantar-compression device (PPPC) for 5 days. Warfarin administered if either treatment was judged to have failed.	22/81 (27%) of PPPC subjects had an incidence of DVT vs. 49/83 (59%) of aspirin alone subjects (p<0.001).	Study confirms safety and efficacy of PPPC with aspirin compared with aspirin alone and supports use of mechanical compression for prophylaxis against DVT.

The review of randomized trials addressing prevention of venous thrombosis and pulmonary embolus uncovered several studies that tested various approaches to anticoagulation and other preventive techniques. These studies are summarized in Table 15. Two studies suggest that compression ultrasonography is not justified. Two find drug therapy better than mechanical approaches. Several studies compared anticoagulant drugs and drug regimens.

Table 16. Complications: Prevention of infection studies (more...)

Table 16. Complications: Prevention of infection studies

Study	Study Type: Intervention; Control	Results	Conclusions
Chiu et al., 2002121	Randomized controlled trial (RCT): (1) Cefuroxime 2 grams -impregnated cement vs. (2) Cement without cefuroxime.	No knees developed deep infections in the Cefuroxime-impregnated cement group vs. 5 knees (3.1%, p=0.02) in the group receiving cement without Cefuroxime.	Cefuroxime-impregnated cement was demonstrated to be effective in the prevention of early to late deep infection after TKA.
Mauerhan et al., 199325	Double blind (DB), RCT, TKA and THA:	Results for TKA cohort only.	Study found no significant differences in the prevalence of wound infections between the two antibiotic regimens.
	(1) Cefuroxime 1.5 grams followed by 750 mg x 2 doses for a total of one day of antibiotic treatment vs. (2) Cefazolin 1 gram three times daily for three days.	Rate of deep wound infection was 0.6% (1/178) for subjects receiving Cefuroxime vs. 1.4% (3/207) for subjects receiving Cefazolin.
Periti et al., 1999122	RCT, TKA, and THA: (1) Teicoplannin 400 mg IV x 1 dose at induction of anesthesia vs. (2) Cefazolin, 5 doses over 24-hour period (2 grams at induction of anesthesia and 1gram daily IV).	Results for TKA cohort only.	Study concludes a single preoperative dose of Teicoplannin ensures adequate surgical antisepsis compared with a standard, multiple-dose regimen of Cefazolin.
		6 (1.5%) subjects in the Teicoplannin group and 7 (1.7%) subjects in the Cefazolin group developed a surgical wound infection during postoperative hospital stay (p ns).

Table 16 summarizes three randomized trial that address infection prevention. Each compares alternative antibiotic regimens.

Table 17. Complications: Tourniquet studies

Table 17. Complications: Tourniquet studies

Study	Study type: Intervention; Control	Results	Conclusions
Abdel-Salam & Eyres, 1995123	RCT: (1) Undergo TKA surgery with tourniquet vs. (2) Surgery without tourniquet.	4/40 (10%) subjects in tourniquet group had DVT within 8–21 days after surgery vs. none (0/40) for the no tourniquet group.	Study concludes TKA can be safely performed without the use of the tourniquet.
		5/40 (12.5%) subjects in tourniquet group had a wound infection within 10 days of surgery vs. none (0/40) for the no tourniquet group.
Wakankar et al., 1999124	Randomized controlled trial (RCT): (1) Undergo TKA surgery with tourniquet vs. (2) Surgery without tourniquet.	One subject in tourniquet group had an asymptomatic DVT on post-op ultrasonography.	Study found no siginificant differences in the incidence of DVTs or wound complications. Use of tourniquet is safe for TKA.

Table 17 shows two randomized trials that tested the use of tourniquets in performing TKAs. One concluded tourniquets were safe and the other that they did not reduce surgical complications.