Background
Total knee arthroplasty (TKA) is a common and effective procedure for the treatment of end-stage osteoarthritis of the knee. Patients undergoing TKA are primarily geriatric and in the US 75 % of TKA were performed on Medicare beneficiaries [
1]. Cram et al. found that annual TKA volume among Medicare enrollees increased 161 % and per capita utilization of TKA increased 99.2 % between 1991 and 2010 [
2]. This increase reflects the observed improvements in pain, function, and independence in an elderly population and increasing acceptance of invasive treatment [
3,
4]. Four of five patients undergoing total knee arthroplasty are satisfied with their knee replacement postoperatively [
5].
However, both risks and costs may be increased in geriatric patients with comorbidities, which may lead to fewer elective orthopedic procedures. Patients aged 85 years or older are 41 % less likely to receive TKA than their younger counterparts [
1]. Patients who did not undergo surgery were older and almost half of these patients were not offered TKA [
6]. Prior investigations into the effects of age on various surgeries have found mixed associations [
7‐
11], and the impact of age on candidacy for TKA is undefined. A better understanding of the risks and benefits of TKAs in this population would allow for more appropriate patient selection.
In order to clarify the effect of age on total knee arthroplasty, we systematically reviewed and evaluated studies from 1990–2015 that investigated the association of age on surgical outcomes.
Methods
Search strategy
This systematic literature review was conducted according to the PRISMA checklists [
12]. As a review of previously published data, it was exempt from review by the local Institutional Review Board. We included all research studies that assessed clinically important outcomes of mortality, venous thromboembolism (pulmonary embolism and/or deep vein thrombosis), myocardial infarction, length of stay, and functional outcomes on geriatric patients undergoing primary total knee arthroplasty. Functional outcomes included all studies with either quantified changes in either self-reported or performance-based pain, mobility, or other symptoms.
A health sciences librarian (GX) conducted a structured literature review using the PubMed database, Scopus database, and the Cochrane database of systematic reviews for articles published from January 1, 1990 to December 16, 2015 with keywords and medical subject heading (MeSH) terms seen in Additional file
1. We reviewed the reference lists of retrieved articles to identify studies that were not obtained from the primary literature searches. Additional studies were identified through Clinicaltrials.gov.
Inclusion and exclusion criteria
Studies were included if they met the following criteria: published in English and grouped patients by age compared to a reference younger group in patients who underwent primary total knee arthroplasty (TKA). We excluded studies if they were case reports, commentaries, guidelines, meta-analyses, systematic literature reviews, editorials, animal studies, studies that did not separate outcomes for TKA and total hip arthroplasty, pediatric studies, reviews, and those that did not contain the outcomes of interest. We excluded patients undergoing simultaneous bilateral knee replacements, as these patients are typically more carefully selected by surgical providers and there is evidence of increased complication rates [
13]. Knee revision arthroplasties were also excluded as a greater fraction of these procedures are urgent and the complication rate differs significantly from primary knee procedures [
2].
Data extraction and risk of bias assessment
Two authors independently screened the titles and abstracts from 3851 citations to assess whether they met the inclusion or exclusion criteria. If the screeners did not agree whether the study was relevant, the third author reviewed the article in detail. After this initial screening, at least two of the three reviewers independently evaluated full-text articles in detail (PJ, EK, MF) and abstracted data for each article. The first author reviewed all inconsistent assessments and the reviewers resolved their disagreements by consensus.
The reviewers abstracted data from the study papers on year of publication, study design, older age group age range, younger age group age range, mortality among the older and younger group, incidence of myocardial infarction in the older and younger group, incidence of venous thromboembolism in the older and younger group, length of stay, and functional outcomes. The investigators used the Newcastle-Ottawa risk of bias tool to assess the quality of each article [
14].
Statistical analysis
Meta-analysis was performed using unadjusted data extracted from each article. We employed random-effect models to obtain pooled odds ratio estimates, using Microsoft Excel 2007 and the Cochrane Review Manager (RevMan) version 5.3 [
15]. To assess heterogeneity, we used the Cochran Q statistic and the
I
2
statistic.
For length of stay and functional data qualitative terms were used to describe either a relative increase or decrease in number of days stayed in the hospital of older patients relative to younger patients and improved, worse, or no difference in functional outcomes of older patients relative to younger patients. P values <0.05 were determined a priori to be statistically significant. Due to the heterogeneous assumptions used in multivariate analyses, we report only univariate analyses.
Discussion
The currently available literature provides limited information on the potential safety and efficacy of knee replacement in the geriatric population. This information is limited by the ability to uniformly define either our population of interest or the optimal endpoints to measure. Although morbidity may be increased in this population, the absolute incidence of each endpoint is small while benefits appear to be uniform across the age groups. Thus, elective knee arthroplasty could be beneficial in select geriatric individuals.
There was variation in methodology, including the definition of age groups and the duration of follow-up, however the weight of available evidence suggests increased mortality for geriatric patients in both the immediate post-operative period as well as several years following TKA. Although mortality is a principle indicator of patient safety, it has limited application as a marker for total knee arthroplasty. Short-term mortality after TKA is less than 0.5 % [
19,
38]. Mahomed, et al. in a study of Medicare recipients in 2000 found that there was a 90-day mortality rate of 0.6; (95 % confidence interval, 0.6–0.7) which was fewer deaths than would be expected in the unselected Medicare population who did not undergo total joint replacement [
39]. Geriatric patients were enrolled for elective surgery in all identified studies, which introduces the potential for substantial selection bias in sampling. For knee replacements in the general population, mortality returns to baseline levels approximately one month after surgery [
40]. It is not clear that this trend would apply to the geriatric population, but over several years it seems likely that baseline mortality trends would dominate. The life expectancy for an 85-year-old patient in the US is approximately 6 years [
41]. Examining mortality rates 10 years postoperatively in this population does not make clinical sense.
Second, there is limited evidence of an increased complication rate for TKA performed in the geriatric population. In addition to the meta-analysis above, two large studies that did not meet inclusion criteria with over 200,000 patients reported statistically significant increased rates of acute MI in short term follow-up periods of six weeks (≥80 years [HR 8.20 (2.38–28.22)], 60–79years [HR 2.55 (0.77–8.42)] [
42] and ninety days [RR 5.0 (2.8–8.6) >90 years relative to 65–69 years [
39]. In contrast, there was no significant increase in DVT or PE in older patients despite increased incidence of common risk factors such as cancer, congestive heart failure, and impaired mobility [
43]. This may have been due to the low incidence of postoperative DVT, as low as 0.4 % in Medicare patients receiving TKA using prophylactic strategies [
2]. Other potential explanations include more aggressive perioperative DVT prophylaxis in geriatric patients with comorbidities, as well as selection bias towards patients with minimal comorbidities in elective TKA. DVT prophylaxis techniques were not mentioned in the majority of studies. Hilton [
21] stated that all patients used antithrombotic stockings and 43 % used heparin subcutaneously twice daily. Biau [
23] stated that “prophylaxis was given to everyone” and Clement [
24] stated that DVT prophylaxis was given based according to the Scottish Intercollegiate Guidelines. Kennedy mentioned that the elderly patients may have had more thorough venous thromboembolism prophylaxis rates [
28].
Both geriatric and control patients had similarly high satisfaction with their joints ranking similar scores for overall satisfaction and improvement in functional status. Compared to the population norm for a particular age group, patients aged 75 years and over improved significantly, becoming similar to population norms for this age [
44]. Studies have shown that geriatric have similar to worse functional scores at baseline compared to younger patients [
2] implying that geriatric patients may delay surgery longer or they are being denied surgery that younger patients would have received. Despite this, patients who underwent total joint arthroplasty had significantly improved WOMAC scores compared with those who did not undergo surgery, however, 45 % of those who did not have surgery were not offered a surgical option [
6]. The other 55 % felt that pain and complications of surgery and needing someone to care for them after surgery were hurdles to surgery [
6]. Although the direct cost of the operation may be increased due to the greater LOS, this cost may need to be weighed not against younger controls but against the increased caregiver needs and loss of independence in those patients treated without invasive therapy. Further studies may be necessary to better define the cost-effectiveness of the procedure in the geriatric population.
Limitations
This review was limited due to the nature of the extant available literature. Definitions varied, not only the definition of “geriatric” but also of mortality, myocardial infarction, DVT, functional recovery, which complicate efforts to synthesize the data into a single answer. The vast majority of studies reviewed for this article compared a selected geriatric population to matched, younger controls. There is limited data describing how these patients were selected, and the results may not be applicable to unselected geriatric patients presenting for initial evaluation. Many of the studies were done in the 1990’s making it difficult to compare to contemporary surgeries, as there have been significant changes in intraoperative and perioperative management since that time.
Conclusions
In conclusion, the available data suggests that selected geriatric patients face similar risk to younger patients undergoing elective TKA, with comparable improvements in functional status. In patients with good functional status and minimal medical comorbidities, age alone should not contraindicate knee arthroplasty. Future, dedicated investigations are need improve selection criteria.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
EK, PJ and MF performed the initial conceptualization and identification of qualifying and analysis of articles. MS performed meta-analysis calculations. GX developed and performed the database searches. All authors with article writing and read and approved the final manuscript.