Demographic, pain, and quality of life factors in unilateral and bilateral total knee arthroplasty: A population-based retrospective cohort study of 23,033 patients
Abstract
Background
The Alberta Bone and Joint Health Institute (ABJHI) has a comprehensive registry capturing clinical, demographic, and patient-reported outcome measures on all knee arthroplasties performed in the province of Alberta, Canada. This large database provides an opportunity to study the outcomes of total knee arthroplasties (TKA) in patients who underwent unilateral or bilateral procedures.
Methods
Pain and quality of life outcomes of 29,157 TKAs performed in 23,033 patients for knee osteoarthritis performed in Alberta, Canada from January 2013 to February 2020 were examined. This study: (1) investigates whether having both knees replaced results in better pain and quality of life outcomes than replacing only one knee (using multivariate analysis of variance) and (2) identifies if pain and quality of life outcomes impact the likelihood of needing surgery in the contralateral knee after a first TKA (using univariate and multivariate Cox proportional hazards models). Patient-reported outcomes were measured using questionnaires: the Western Ontario and McMaster Universities Arthritis Index (WOMAC) to assess knee pain, and the EuroQol-5D-5L (EQ-5D) to assess quality of life.
Results
Patients who had surgery on both knees had approximately 5% better WOMAC (pain) and EQ-5D (quality of life) questionnaire scores 12 months post-surgery than those who had surgery on only one knee. Additionally, for every 1-point higher pre-operative WOMAC pain score (i.e., less pain) there was a 1% decreased likelihood of undergoing a contralateral TKA compared to a 46% decreased likelihood per 1-point improvement in quality of life. Dementia, back pain, and hospital readmission were associated with a 68%, 17%, and 44% decreased likelihood of contralateral TKA respectively. By contrast, patients with obesity were 36% more likely to undergo bilateral TKA (p < 0.05).
Conclusions
Preoperative knee pain and quality of life play a statistically significant role in determining which patients undergo bilateral TKA. Additionally, bilateral TKA resulted in better patient-reported outcomes compared to unilateral TKA. Though clinical significance cannot be drawn from the low response rates for the patient-reported outcomes, these findings provide a basis for future research on patient-reported outcomes within the Canadian healthcare system.
Background
Total knee arthroplasty (TKA) is the final treatment option to manage knee pain from osteoarthritis [1]. While nearly 90% of patients with knee osteoarthritis have radiographic evidence of osteoarthritis in both knees [2,3], only one-third of these patients go on to have both knees replaced [3-5]. It is unclear why some people choose to get one knee replaced and not the other when osteoarthritis in both knees is common.
To date, research investigating the factors associated with having both knees replaced has not yet captured the variety and combinations of comorbidities that explain progression or contraindication to a contralateral TKA. Further, it is not clear whether pre-operative pain or the level of pain recovery after an index TKA can predict successful pain relief after TKA on the contralateral knee [6,7]. Thus, this study examines the outcomes of TKAs performed in Alberta, Canada to investigate whether 1) bilateral TKA results in better pain and quality of life outcomes than unilateral TKA, and 2) pain and quality of life outcomes after a first TKA affect the likelihood of needing surgery in the contralateral knee.
Methods
Data Source
Data was collected by the Alberta Bone and Joint Health Institute (ABJHI) between January 2013 and February 2020. The ABJHI registry captures administrative, clinical, demographic and patient-reported outcomes measures on all hip and knee arthroplasties in the province of Alberta, Canada. Ethics approval was obtained from the Conjoint Health Research Ethics Board at the University of Calgary (REB20-0070), and permission to use the data for this retrospective review was granted by the ABJHI.
Study Design
All patients (29,180) who underwent TKA in Alberta during this time period were captured in the registry. Alberta uses a centralized triage and health care system, and thus criteria for TKA are consistent province-wide [8]. We compared the results between patients who had surgery on one knee (unilateral) and those who had surgery on both knees (bilateral). The cohort was divided into the following groups:
- Unilateral: patient having TKA on one side at a time (i.e., only one surgery that day)
- Two-staged bilateral: patients having TKA on one side, but have already had TKA on the other side (i.e., both knees underwent TKA in two separate operations on different days)
- One-staged bilateral: patients having bilateral surgery under same anesthesia (same day)
Exclusion Criteria
For the purposes of this analysis, diagnoses unlikely to be due to primary osteoarthritis, including inflammatory arthritis, fracture, bone malignancies, osteonecrosis, and trauma, as well as patients under 30 years of age were excluded.
Variables and outcomes
Patient-reported outcome scores were compared before and after surgery (at 3- and 12-months post-surgery) using two questionnaires: the Western Ontario and McMaster Universities Arthritis Index (WOMAC) to assess knee pain [9], and the EuroQol-5D-5L (EQ-5D) to assess quality of life. WOMAC scores were transformed to report a range from 0 (worst) to 100 (best) whereas EQ-5D scores range from 0 (worst) to 1 (best), using the Canadian formula [10,11]. The questionnaires were administered over the phone, on paper or electronically prior to or at a clinic visit. To determine the clinical relevance in patient-reported outcomes, patient acceptable symptom states (PASS), a dichotomous patient-reported outcome threshold used to identify cut-off points in acceptable numerical patient-reported outcome scores, were used. PASS were defined as 75 for WOMAC pain and 0.78 for EQ-5D quality of life, where values above these thresholds represented an acceptable symptom state [12,13].
The objective variables of interest were age, sex, surgical date, readmission within 30 days post-TKA, revision surgery conducted within 365 days, and presence of comorbidities. Comorbidities and chronic conditions were extracted by ABJHI from documented pre-admission comorbidities in the Discharge Abstract Database (DAD), the National Ambulatory Care Reporting System (NACRS), and physician claims data either prior to TKA hospital admission, or up to 5 years prior to the TKA using Canadian Institute for Health Information (CIHI) Population Grouper data. These included the presence of obesity, a rheumatoid condition, diabetes, back pain, stroke, depression, chronic mental health concerns, any type of cancer, chronic cardiac disease, chronic pulmonary disease, chronic renal disease, deep vein thrombosis, history of drug and/or alcohol abuse, chronic hepatic disease, asthma, human immunodeficiency virus (HIV), and/or pneumonia. Additionally, a total risk was computed as the sum of the presence of any of the comorbidities, to a maximum of 3 (i.e., a value of 3 captures patients with 3 or more comorbidities).
Statistical analysis
Pre-surgical characteristics of the whole cohort, as well as unilateral and bilateral groups, were calculated using proportions and means as appropriate. Age and sex distributions for each group are presented as histograms for visual comparison. The cumulative incidence of bilateral TKA after index TKA was calculated using the Kaplan-Meier method. Time between TKAs was reported as median interquartile range (IQR) without accounting for censoring because of the low percentage (< 50%) of the cohort who proceeded to bilateral TKAs. Surgical outcomes were compared between the groups using Fisher’s exact test as the incidence of adverse events was small [14]. Response rates and characterization of missing values were reported according to guidelines provided by The International Society of Arthroplasty Registries PROMS Working group [15].
WOMAC and EQ-5D scores were compared between the groups using multivariate analysis of variance (MANOVA) with group as a between-subjects factor and time (pre-operative, 3 months after first TKA, 12 months after last TKA).
Each risk factor’s contribution to a unilateral TKA patient’s likelihood of undergoing bilateral TKA was assessed using Cox proportional hazards models, with associations presented as a hazard ratio (HR) and 95% confidence intervals (CI) [low, high]. Significant variables from a univariate model were then entered into the following three multivariate models in a stepwise manner:
- Model 1: included objective factors alone (age, sex, obesity, diabetes, depression, stroke, dementia, mental health, back pain, readmission within 30 days, and revision with 1 year)
- Model 2: included patient-reported outcomes alone (pain and quality of life assessed using the WOMAC and EQ-5D questionnaires respectively)
- Model 3: included all variables to understand the relative contribution to patients at risk for a bilateral TKA
Two-sided p-values were used for all analyses, and p < 0.05 was considered significant. Multiple comparisons were adjusted with Bonferroni correction factor as appropriate. Statistical analysis was performed with SPSS (v28.0), R (v4.0.3), and RStudio (v2022.07.2+576).
Results
Descriptive Statistics
37,336 knees from 29,180 patients underwent elective, primary TKA in Alberta between January 2013 and February 2020. Of the 23,033 patients that fit the inclusion criteria, 71.8% underwent unilateral TKA, 26.6% had two-stage bilateral TKAs, and 1.6% had one-stage bilateral TKAs (Figure 1). One-stage bilateral TKA patients were younger than two-stage bilateral and unilateral TKA patients (mean 63 years vs. 66 and 67 years, p < 0.001) and were more likely to be male (51% vs. 37% and 40%, p < 0.001) (Table 1). While the one-stage bilateral group appeared to have a greater proportion of patients with no comorbid conditions (27% vs. 23% and 23%), this difference was not statistically significant (p = 0.11). The age distributions by sex for the whole cohort, unilateral, as well as the two-stage and one-stage bilateral TKA cohorts are shown in Figure 2.
Figure 1.
| Whole Cohort n = 23033 | Unilateral TKA n = 16539 | Two-Stage Bilateral TKA n = 6124 | One-Stage Bilateral TKA n=370 | |||
|---|---|---|---|---|---|---|
| Age in years , mean (SD) | 66.7 (9.0) | 67.1 (9.2) | 65.8 (8.3) | 63.4 (7.9) | ||
| Female, n (%) | 14040 (61.0) | 9980 (60.3) | 3879 (63.3) | 181 (48.9) | ||
| Risk Factors, n (%) | ||||||
| Asthma | 1906 (8.3) | 1355 (8.2) | 529 (8.6) | 22 (5.9) | ||
| Back Pain | 5997 (26.0) | 5459 (27.0) | 1457 (23.8) | 81 (21.9) | ||
| Cancer | 3156 (13.7) | 2355 (14.2) | 756 (12.3) | 45 (12.2) | ||
| Pulmonary | 2223 (9.7) | 1692 (10.2) | 505 (8.2) | 26 (7.0) | ||
| Renal | 589 (2.6) | 452 (2.7) | 134 (2.2) | 3 (0.8) | ||
| Deep Vein Thrombosis | 888 (3.9) | 644 (3.9) | 237 (3.9) | 7 (1.9) | ||
| Dementia | 215 (0.9) | 189 (1.1) | 25 (0.4) | 1 (0.3) | ||
| Depression | 3337 (16.4) | 2713 (16.4) | 1012 (16.5) | 53 (14.3) | ||
| Diabetes | 4708 (20.4) | 3389 (20.5) | 1254 (20.5) | 65 (17.6) | ||
| Drug/Alcohol Abuse | 862 (3.7) | 657 (4.0) | 187 (3.0) | 18 (4.9) | ||
| Obesity | 5645 (24.5) | 3704 (22.4) | 1863 (30.4) | 78 (21.1) | ||
| Stroke | 313 (1.4) | 238 (1.4) | 73 (1.2) | 2 (0.5) | ||
| Mental Health | 844 (3.7) | 631 (3.8) | 209 (3.4) | 4 (1.1) | ||
| Hepatic | 288 (1.3) | 221 (1.3) | 66 (1.1) | 1 (0.3) | ||
| HIV | 3 (<0.1) | 2 (<0.1) | 0 (0.0) | 0 (0.0) | ||
| Pneumonia | 2158 (9.4) | 1616 (9.8) | 507 (83) | 35 (9.5) | ||
| Total Risk, (%) | ||||||
| 0 Risk Factors | 22.9 | 22.9 | 22.6 | 27.3 | ||
| 1 Risk Factor | 31.1 | 30.6 | 32.1 | 35.1 | ||
| 2 Risk Factors | 23.6 | 23.6 | 23.6 | 21.6 | ||
| > 3 Risk Factors | 22.5 | 22.9 | 21.7 | 15.9 | ||
| Pre-TKA Questionnaire Scores, mean (SD) ** | ||||||
| WOMAC Pain* (n = 16439) | 43.2 (17.9) | 44.4 (17.8) | 40.0 (17.4) | 38.3 (21.1) | ||
| EQ5D5L Index (n = 15887) | 0.5 (0.2) | 0.5 (0.2) | 0.5 (0.2) | 0.4 (0.3) | ||
| *Transformed Scores | ||||||
| **n varies depending on questionnaire completion | ||||||
Figure 2.Age and sex distributions for the A) whole cohort, B) unilateral, C) two-stage bilateral, and D) one-stage bilateral TKA cohorts.
After the first TKA, two-stage bilateral patients had fewer revision surgeries(0.6%) performed within 365 days than the one-stage bilateral, and unilateral groups (1.0% and 1.1%, p = 0.03). However, when accounting for the second TKA, the two-stage bilateral patients had more revision surgeries (1.5%, p = 0.004). More unilateral TKA patients (2.7%) were readmitted to hospital within 30 days compared to 1.4% and 1.6% of one-stage and two-stage bilateral TKA patients after the first TKA, respectively (p < 0.001) (Table 2). Similarly, after accounting for the second TKA, two-stage bilateral patients had more hospital readmissions (3.3%). The percentage of patients who died within 365 days of their last TKA was less for the two-stage bilateral group (0.1%) compared to the unilateral and one-stage bilateral (0.9% and 1.1%, p < 0.001). This percentage remained lower (0.5%) for the two-stage bilateral group after accounting for the 2nd TKA (p < 0.001). 275 patients underwent a contralateral TKA within 3-months of the first and were thus excluded from analyses using 3-month post-operative outcomes,summarized inTable 2. Based on the Kaplan-Meier method, the cumulative incidence of two-stage bilateral TKA was 16.4% (95% CI 15.9 to 16.9%) after 1 year, 28.4% (95% CI 27.8 to 29.1%) after 3 years, and 33.1% (95% CI 32.4 to 33.8%) after 5 years. In patients who underwent a two-stage bilateral TKA, the median time between TKAs was 11 months (0.9 years, IQR 0.6 to 1.3). The median timefrom TKA to data extractionin the unilateral cohort was approximately 3 years (3.3 years, IQR 1.6 to 5.1).
| Whole Cohort n = 23033 | Unilateral TKA n = 16539 | Two-Stage Bilateral TKA n = 6124 | One-Stage Bilateral TKA n = 370 | p -value | |
|---|---|---|---|---|---|
| Revision within 365 days | |||||
| First TKA, n (%) | 200 (0.9) | 158 (1.0) | 38 (0.6) | 4 (1.1) | 0.033 |
| All TKAs, n (%) | 90 (1.5) | 0.0045 | |||
| Readmission within 30 days | |||||
| First TKA, n (%) | 538 (2.3) | 438 (2.6) | 95 (1.6) | 5 (1.4) | <0.001 |
| All TKAs, n (%) | 205 (3.3) | 0.0045 | |||
| Death within 365 days, n (%) | |||||
| First TKA, n (%) | 158 (0.7) | 147 (0.9) | 7 (0.1) | 4 (1.1) | <0.001 |
| All TKAs, n (%) | 30 (0.5) | <0.001 | |||
| Years to bilateral TKA, median (IQR) | 0.9 (0.6 to 1.3) | ||||
| Time to data extraction, median (IQR) | 3.3 (1.6 to 5.1) | ||||
Because of the low response rates (approximately 10% of patients completed questionnaires at all 3 timepoints) yet large sample sizes, responses were deemed useful to report and thus werereported according to recommendations by Orr et al [15]. Patient-reported outcomes are reported in Supplementary Table 1 for WOMAC pain and Supplementary Table 2 for EQ-5D with transparency of loss to follow up and its impact of clinical relevance using the clinical relevance ratio (the number of patients that achieved a clinically significant improvement >PASS, over the number of patients at baseline).Because there is a range of PASS thresholds that have been defined for each patient-reported outcome measure, Figure 3 was used to visually see the suitability of the chosen PASS thresholds in this population.
Figure 3.Scatter plots of all responders for WOMAC Pain (blue/ left column) and EQ-5D (green/right column) for A) & B) unilateral C) & D) two-stage bilateral, and E) & F) one-stage bilateral at baseline, 3 months post-TKA, 12 months post-TKA. The grey transparent band indicates scores above the PASS threshold.
To demonstrate the clinical relevance of patients that achieved PASS, the clinical relevance ratio (number of patients that achieved a clinically significant improvement over the number of patients at baseline) was compared to the proportion of patients with a clinically significant outcome (number of patients that achieved a clinically significant improvement over the number of responders at that timepoint), as shown in Figure 4.
Figure 4.Bar charts indicating the mean and standard deviation of all responders for WOMAC Pain (blue/ left column) and EQ-5D (green/right column) for A) & B) unilateral C) & D) two-stage bilateral, and E) & F) one-stage bilateral at baseline, 3 months post-TKA, 12 months post-TKA. The dashed grey line indicates the proportion of above the PASS threshold (clinically significant outcome, calculated as the number of patients who achieved a clinically significant improvement over the number of responders at that timepoint) and the solid grey line indicates the clinical relevance ratio (number of patients that achieved a clinically significant improvement over the number of patients at baseline).
Comparison between unilateral and bilateral TKA patient - reported outcomes
Only 7% of patients completed questionnaires pre-surgery, 3-months and 12-months post TKA, leaving 1,602 patients for comparison of patient-reported outcomes (1,196 unilateral TKA, 406 two-stage bilateral TKA). One-stage bilateral was omitted from this analysis due to the small patient-reported outcome sample size (n = 20). Patients with two-stage bilateral TKAs had poorer pre-operative WOMAC pain levels (41.40 ± 0.87, p <0.001) than unilateral TKA patients (45.71 ± 0.51, p < 0.001) and EQ-5D levels (0.48 ± 0.01 vs. 0.54 ± 0.01, p <0.001) (Figure 5). Trends were similar for pre-operative quality of life (0.48 ± 0.01 vs 0.53 ± 0.01). Both groups reached similar outcomes 3 months following the first TKA with no significant differences (p > 0.05) in pain (-0.80 ± 1.01) or EQ-5D (-0.01 ± 0.80) between groups. However, the two-stage bilateral TKA group achieved greater improvements in pain (3.69 ± 1.02) and EQ-5D (0.03 ± 0.01) 12 months following the last TKA compared to the unilateral TKA group (p < 0.001).
Figure 5.A) Changes in WOMAC pain and B) Changes in EQ-5D-5L quality of life outcomes, from (1) pre-surgery to (2) 3 months following the first TKA and (3) 12 months following the last TKA (i.e., only TKA for the unilateral group and the second TKA for the bilateral group). *** p <0.001 **p <0.01 *p<0.05 for the difference between groups at a given time.A) Changes in WOMAC pain and B) Changes in EQ-5D-5L quality of life outcomes, from (1) pre-surgery to (2) 3 months following the first TKA and (3) 12 months following the last TKA (i.e., only TKA for the unilateral group and the second TKA for the bilateral group). *** p <0.001 **p <0.01 *p<0.05 for the difference between groups at a given time.
Risk factors associated with progression to contralateral TKA
Univariate Cox proportional hazard models (n=22,663) indicated that obesity was associated with increased incidence of contralateral TKA whereas increased age, male sex, stroke, dementia, back pain, hospital readmission within 30 days and TKA revision within 1 year were individually associated with decreased incidence of contralateral TKA (Table 3). Patient-reported outcomes associated with increased incidence of contralateral TKA included poorerpre-surgical WOMAC painandEQ-5D scores (Table 3). Diabetes, depression, stroke, and mental health were not significantly associated with contralateral TKA.
| Risk Factor | HR | 95% CI | p-value |
|---|---|---|---|
| Objective (n = 22663) | |||
| Age | 0.99 | 0.99, 0.99 | < 0.001 |
| Sex (male) | 0.91 | 0.86, 0.96 | < 0.001 |
| Obesity | 1.36 | 1.29, 1.44 | < 0.001 |
| Diabetes | 1.00 | 0.94, 1.07 | > 0.9 |
| Depression | 1.00 | 0.94, 1.07 | > 0.9 |
| Stroke | 0.81 | 0.65, 1.03 | 0.082 |
| Dementia | 0.39 | 0.26, 0.58 | < 0.001 |
| Mental Health | 0.90 | 0.78, 1.03 | 0.14 |
| Back Pain | 0.83 | 0.83, 0.93 | < 0.001 |
| Total Comorbidities | |||
| 1 | 1.05 | 0.98, 1.13 | 0.14 |
| 2 | 1.01 | 0.94, 1.09 | 0.7 |
| >3 | 0.95 | 0.89, 1.03 | 0.2 |
| Hospital readmission (within 30 days) | 0.59 | 0.49, 0.73 | < 0.001 |
| Revision TKA (within 365 days) | 0.57 | 0.42, 0.79 | < 0.001 |
| Death (within 365 days) | 0.14 | 0.06, 0.28 | < 0.001 |
| Patient-Reported (n = 7433) | |||
| Pre-surgical WOMAC Pain | 0.99 | 0.98, 0.99 | < 0.001 |
| Pre-surgical EQ-5D | 0.37 | 0.30, 0.44 | < 0.001 |
In multivariate analyses (n=7,433), only patients with 3-month patient-reported outcomes data were included so all multivariate models were comparable. Model 1 (objective factors only) indicated that obesity was associated with higher probability of a contralateral TKAwhereas olderage, male sex, dementia, back pain, hospital readmission within 30 days, and revision within 1 year were associated with lower probability of a contralateral TKA (???). Model 2, (patient-reported outcomes only), indicated that pre-operative quality of life was associated with higher probability of a contralateral TKA whereas pre-operative WOMAC pain had smaller effects.In Model 3, when accounting for all factors, age, sex and TKA were no longer significantfactors predicting probability of contralateral TKA. Whereas, obesity (HR 1.36 [1.24, 1.50]) was associated with an increased likelihood of needing surgery in the contralateral knee, while dementia (HR 0.32 [0.13, 0.77]), back pain (HR 0.83 [0.75, 0.92]), hospital readmission within 30 days of TKA (HR 0.56 [0.40, 0.79]), poorer pre-surgical WOMAC pain (HR 0.99 [0.99, 0.99]), and worse pre-surgical quality of life score (HR 0.54 [0.43, 0.69]) were associated with a decreased likelihood of having surgery on the contralateral knee (p < 0.05).
| Models | HR | 95% CI | p-value |
|---|---|---|---|
| Model 1: Objective Risk Factors (n = 22663) | |||
| Age | 0.99 | 0.99, 0.99 | < 0.001 |
| Sex (male) | 0.93 | 0.88, 0.98 | 0.008 |
| Obesity | 1.33 | 1.26, 1.41 | < 0.001 |
| Dementia | 0.46 | 0.31, 0.69 | < 0.001 |
| Back Pain | 0.88 | 0.83, 0.93 | < 0.001 |
| Hospital readmission (within 30 days) | 0.64 | 0.52, 0.78 | < 0.001 |
| Revision TKA (within 365 days) | 0.58 | 0.42, 0.80 | < 0.001 |
| Death (within 365 days) | 0.15 | 0.07, 0.32 | < 0.001 |
| Model 2: Patient-Reported Outcomes (n = 7433) | |||
| Pre-surgical WOMAC Pain | 0.99 | 0.99, 0.99 | < 0.001 |
| Pre-surgical EQ-5D | 0.56 | 0.44, 0.71 | < 0.001 |
| Model 3: Combined Model (n = 7433) | |||
| Obesity | 1.37 | 1.24, 1.50 | < 0.001 |
| Dementia | 0.32 | 0.13, 0.77 | 0.011 |
| Back Pain | 0.83 | 0.75, 0.92 | < 0.001 |
| Hospital readmission (within 30 days) | 0.56 | 0.40, 0.80 | 0.001 |
| Death (within 365 days) | 0.36 | 0.12, 1.13 | 0.080 |
| Pre-surgical WOMAC Pain | 0.99 | 0.99, 0.99 | < 0.001 |
| Pre-surgical EQ-5D | 0.54 | 0.43, 0.69 | < 0.001 |
Discussion
Does having both knees replaced result in better patient-reported outcomes than replacing only one knee?
Statistically, bilateral TKA patients had significantly better reports of knee pain and quality of life after 12 months compared to the unilateral cohort. However, we could not determine whether this difference between unilateral and bilateral cohorts is clinically relevant (clinical relevance ratio at 12 months was 0.10 for both pain and quality of life in the unilateral cohort and 0.16 and 0.19 for pain and quality of life in the bilateral cohort) due to the low response rate in patient-reported outcomes within this study. Thus, the 5-10% difference between the unilateral and bilateral cohorts may not be a meaningful improvement. Furthermore, the bilateral cohort had poorer pre-operative patient-reported outcomes but achieved greater improvements after 12 months compared to unilateral cohort. This may be due to the first TKA typically being performed on the more symptomatic knee [6]. Therefore, a greater change in pain is expected for the first TKA. Alternatively, this may suggest that those who report poor pre-operative pain and quality of life outcomes may be more motivated to have a contralateral TKA regardless of improvements after the index TKA.
Do patient-reported outcomes impact the likelihood of needing surgery in the contralateral knee after a first TKA?
Patient-reported outcomes, especially quality of life, have a statistically significant impact on who undergoes bilateral TKA. Female sex, younger age, and the need for revision TKA are also associated with higher probability of contralateral TKA, but these effects were no longer significant after accounting for patient-reported outcomes. Other studies also report small effects of age and sex [16-18], however, these effects were not significant in all studies, likely due to sample sizes. Females with knee osteoarthritis and younger individuals seeking TKA tend to have worse knee pain [16,17,19,20], explaining how pain may account for these variables. Overall, it is likely that knee-related symptoms and quality of life are driving the need for contralateral TKA in these groups.
Predictors of increased likelihood of progression to contralateral TKA included greater pre-surgical pain, poor pre-surgical quality of life, and obesity. In contrast, dementia, back pain, and hospital readmission after the first TKA reduced the likelihood of progression to contralateral TKA. The finding that a contralateral TKA is often not performed when other major medical diagnoses are present may indicate they interfere with accurate assessments of pain, the prioritization of resolving pain, or that the conditions contraindicate a second surgery [21]. Further research is needed to determine if patients with comorbidities would benefit from a contralateral TKA.
Data Limitations
Several measures, such as socioeconomic status and radiographic data, could have provided more insight on the population’s access and eligibility for TKA but were not collected by this registry. Literature shows that while nearly 90% of osteoarthritis patients present with radiographic evidence of osteoarthritis in both knees [2], only a third of these patients undergo bilateral TKAs [3-5], and approximately 37% refused contralateral TKA [22]. However, without radiographic data or record of symptoms in the contralateral knee, it is difficult to determine how representative these statistics are in our cohort. Additionally, given the nature of retrospective data extraction, we likely did not capture all of the contralateral TKAs. However, the long-time interval from TKA to data extraction (3.3 years) relative to the median time to contralateral TKA (11 months) suggests that most progressions from unilateral to bilateral TKA were captured in the dataset.
While the demographic and clinical data were available for all 23,033 TKA patients in Alberta who met our inclusion criteria, only 1,602 were included in analyses requiring patient-reported outcome data at all 3 timepoints (i.e., baseline, 3- and 12-months) and 7,433 were included in the survival analyses due to low completion rates. Those who completed patient-reported outcome measures had higher pre-surgical quality of life, were more likely to be female, and required slightly fewer TKA revisions (data not shown). There remains much debate about the utility and appropriate reporting of patient-reported outcomes. The percentage of any given population under study who complete patient-reported outcome measures is critical and some analysis of whether that group is representative of the whole is important. It is difficult to know if the patients that did not complete questionnaires are missing at random or if there are biases inherent in the group which completed the questionnaires. However, given the large size of this study, the results were deemed useful in providing a basis for future research. Specifically, this study established the importance of comparing the difference between clinical relevance ratio (not found to be clinically significant) and proportion of patients achieving a clinically significant outcome (found to be statistically significant).
Conclusion
In summary, we used population-based data to demonstrate that knee pain and quality of life play a significant role in determining which TKA patients will undergo bilateral TKA. Specifically, patients who underwent surgery on both knees reported statistically better outcomes than those who had surgery on only one knee. Additionally, better preoperative patient-reported pain and quality of life outcomes are significantly linked with a decreased likelihood of undergoing contralateral TKA. While the descriptive statistics and characterization of patient-reported outcomes within the Canadian healthcare system may be useful specifically for researchers, the low response rate for the patient-reported outcomes may not yield clinically impactful conclusions. Thus, further research is required to provide stronger evidence and establish validity and reliability for these observed relationships.
List of Abbreviations
Alberta Bone and Joint Health Institute (ABJHI); Canadian Institute for Health Information (CIHI); confidence intervals (CI); Discharge Abstract Database (DAD); EuroQol-5D-5L (EQ-5D); hazard ratio (HR); human immunodeficiency virus (HIV); interquartile range (IQR); multivariate analysis of variance (MANOVA); National Ambulatory Care Reporting System (NACRS); Patient Acceptable Symptom State (PASS); Total knee arthroplasty (TKA); Western Ontario and McMaster Universities Arthritis Index (WOMAC)
Declarations
Ethics approval and consent to participate
Ethics approval was obtained from the Conjoint Health Research Ethics Board at the University of Calgary (REB20-0070), and permission to use the data for retrospective review was granted by the Alberta Bone and Joint Health Institute. Research was performed in accordance with the Tri-Council Policy and the Declaration of Helsinki. The need for informed consent was waived by the Conjoint Health Research Ethics Board at the University of Calgary.
Availability of data and materials
The datasets analysed for this study are available from the Alberta Bone and Joint Health Institute based in Calgary, Alberta but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of the Alberta Bone and Joint Health Institute.
Competing interests
The authors declare that they have no competing interests.
Funding
Study funding was provided by the Hotchkiss Brain Institute and the University of Calgary. ADH holds a Canada Research Chair in MR Spectroscopy in Brain Injury. SAL (PhD Salary Award 22-0000000127) and SLM (Stars Career Development Award STAR-18-0189) were supported by the Arthritis Society Canada.
Authors’ contributions
SAL was responsible for data conceptualization, formal analysis, data interpretation, data visualization, and wrote the main manuscript. AK and SE were involved in data cleaning, preliminary data analysis, and interpretation. GS, RN, JW provided medical opinions for data interpretation. ADH was involved in the conceptualization, data interpretation, and drafting of the main manuscript. SLM was responsible for conceptualization, data acquisition, data editing, formal analysis, and drafting of the main manuscript. All authors were involved in the review and editing of the manuscript.
Acknowledgements
Thank you to Sofia Azeez, Akintunde Bashorun, Lindsay Pittman, Katelyn Reczek and Christopher Smith from the Alberta Bone and Joint Health Institute for their assistance with data procurement.
References
- Varacello M, Luo TD, Johanson NA. Total Knee Arthroplasty Techniques. StatPearls. 2022.
- Gunther K P, Sturmer T, Sauerland S, Zeissig I, Sun Y, Kessler S, Scharf H P, Brenner H, Puhl W. Prevalence of generalised osteoarthritis in patients with advanced hip and knee osteoarthritis: The Ulm Osteoarthritis Study. Annals of the Rheumatic Diseases. 1998; 57(12)DOI
- McMahon M, Block JA. The Risk of Contralateral Total Knee Arthroplasty After Knee Replacement for Osteoarthritis.. J Rheumatol. 2003; 30:1822-4.
- Santana Daniel C., Anis Hiba K., Mont Michael A., Higuera Carlos A., Piuzzi Nicolas S.. What is the Likelihood of Subsequent Arthroplasties after Primary TKA or THA? Data from the Osteoarthritis Initiative. Clinical Orthopaedics & Related Research. 2019; 478(1)DOI
- Bohm Eric R, Molodianovitsh Katy, Dragan Alina, Zhu Naisu, Webster Greg, Masri Bas, Schemitsch Emil, Dunbar Michael. Outcomes of unilateral and bilateral total knee arthroplasty in 238,373 patients. Acta Orthopaedica. 2016; 87(sup1)DOI
- Qutob Majdi, Winemaker Mitch, Petruccelli Danielle, de Beer Justin. Staged Bilateral Total Knee Arthroplasty. The Journal of Arthroplasty. 2013; 28(7)DOI
- Gabr A, Withers D, Pope J, Santini A. Functional outcome of staged bilateral knee replacements. The Annals of The Royal College of Surgeons of England. 2011; 93(7)DOI
- Marshall Deborah A., Jin Xuejing, Pittman Lindsay B., Smith Christopher J.. The use of patient-reported outcome measures in hip and knee arthroplasty in Alberta. Journal of Patient-Reported Outcomes. 2021; 5(S2)DOI
- Bellamy J, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW. Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. J Rheumatol. 1988; 15(12):1833-40. PubMed
- Herdman M., Gudex C., Lloyd A., Janssen MF., Kind P., Parkin D., Bonsel G., Badia X.. Development and preliminary testing of the new five-level version of EQ-5D (EQ-5D-5L). Quality of Life Research. 2011; 20(10)DOI
- Xie Feng, Pullenayegum Eleanor, Gaebel Kathy, Bansback Nick, Bryan Stirling, Ohinmaa Arto, Poissant Lise, Johnson Jeffrey A.. A Time Trade-off-derived Value Set of the EQ-5D-5L for Canada. Medical Care. 2016; 54(1)DOI
- MacKay C., Clements N., Wong R., Davis A.M.. A systematic review of estimates of the minimal clinically important difference and patient acceptable symptom state of the Western Ontario and McMaster Universities Osteoarthritis Index in patients who underwent total hip and total knee replacement. Osteoarthritis and Cartilage. 2019; 27(10)DOI
- Hawker Gillian A., Conner‐Spady Barbara L., Bohm Eric, Dunbar Michael J., Jones C. Allyson, Ravi Bheeshma, Noseworthy Tom, Dick Donald, Powell James, Paul Paulose, Marshall Deborah A.. Patients’ Preoperative Expectations of Total Knee Arthroplasty and Satisfaction With Outcomes at One Year: A Prospective Cohort Study. Arthritis & Rheumatology. 2020; 73(2)DOI
- Kim Hae-Young. Statistical notes for clinical researchers: Chi-squared test and Fisher's exact test. Restorative Dentistry & Endodontics. 2017; 42(2)DOI
- Orr Melissa N., Klika Alison K., Gagnier Joel J., Bhandari Mohit, Piuzzi Nicolas S.. A Call for a Standardized Approach to Reporting Patient-Reported Outcome Measures. Journal of Bone and Joint Surgery. 2021; 103(22)DOI
- Lamplot Joseph D., Bansal Anchal, Nguyen Joseph T., Brophy Robert H.. Risk of Subsequent Joint Arthroplasty in Contralateral or Different Joint After Index Shoulder, Hip, or Knee Arthroplasty. Journal of Bone and Joint Surgery. 2018; 100(20)DOI
- Sayeed Siraj A., Johnson Aaron J., Jaffe David E., Mont Michael A.. Incidence of Contralateral THA After Index THA for Osteoarthritis. Clinical Orthopaedics & Related Research. 2012; 470(2)DOI
- Sanders Thomas L., Maradit Kremers Hilal, Schleck Cathy D., Larson Dirk R., Berry Daniel J.. Subsequent Total Joint Arthroplasty After Primary Total Knee or Hip Arthroplasty. Journal of Bone and Joint Surgery. 2017; 99(5)DOI
- Hawker Gillian A., Bohm Eric, Dunbar Michael J., Jones C. Allyson, Noseworthy Tom, Marshall Deborah A.. The Effect of Patient Age and Surgical Appropriateness and Their Influence on Surgeon Recommendations for Primary TKA. Journal of Bone and Joint Surgery. 2022; 104(8)DOI
- Petterson S.C., Raisis L., Bodenstab A., Snyder-Mackler L.. Disease-Specific Gender Differences Among Total Knee Arthroplasty Candidates. The Journal of Bone & Joint Surgery. 2007; 89(11)DOI
- Memtsoudis Stavros G., Liu Spencer S.. Bilateral total knee arthroplasties: a call for practice guidelines. Journal of Clinical Anesthesia. 2013; 25(3)DOI
- Sesen Hakan, Demirkale Ismail, Karaduman Mert, Vural Celal Alp, Okkaoglu Mustafa Caner, Altay Murat. Why two-thirds of patients accepted the second session in staged bilateral total knee arthroplasty: a prospective analysis of 111 patients. Knee Surgery, Sports Traumatology, Arthroscopy. 2014; 23(12)DOI