„ KNEE

Primary knee arthroplasty for osteoarthritis

restores patients’ health-­related quality of life
to normal population levels
A PROPENSITY SCORE MATCHED STUDY
L. Z. Yapp,
C. E. H. Scott,
D. J. MacDonald, Aims
C. R. Howie, This study investigates whether primary knee arthroplasty (KA) restores health-­related
A. H. R. W. Simpson, quality of life (HRQoL) to levels expected in the general population.
N. D. Clement
Methods
From Royal Infirmary This retrospective case-­control study compared HRQoL data from two sources: patients
of Edinburgh, undergoing primary KA in a university-­teaching hospital (2013 to 2019), and the Health
Edinburgh, UK Survey for England (HSE; 2010 to 2012). Patient-­level data from the HSE were used to rep-
resent the general population. Propensity score matching was used to balance covariates
and facilitate group comparisons. A propensity score was estimated using logistic regres-
sion based upon the covariates sex, age, and BMI. Two matched cohorts with 3,029 pa-
tients each were obtained for the adjusted analyses (median age 70.3 (interquartile range
(IQR) 64 to 77); number of female patients 3,233 (53.4%); median BMI 29.7 kg/m2 (IQR 26.5
to 33.7)). HRQoL was measured using the three-­level version of the EuroQol five-­dimension
questionnaire (EQ-­5D-­3L), and summarized using the Index and EuroQol visual analogue
scale (EQ-­VAS) scores.

Results
Patients awaiting KA had significantly lower EQ-­5D-­3L Index scores than the general pop-
ulation (median 0.620 (IQR 0.16 to 0.69) vs median 0.796 (IQR 0.69 to 1.00); p < 0.001). By
one year postoperatively, the median EQ-­5D-­3L Index score improved significantly in the
KA cohort (mean change 0.32 (SD 0.33); p < 0.001), and demonstrated no clinically relevant
differences when compared to the general population (median 0.796 (IQR 0.69 to 1.00) vs
median 0.796 (IQR 0.69 to 1.00)). Compared to the general population cohort, the postoper-
ative EQ-­VAS was significantly higher in the KA cohort (p < 0.001). Subgroup comparisons
demonstrated that older age groups had statistically better EQ-­VAS scores than matched
peers in the general population.

Conclusion
Patients awaiting KA for osteoarthritis had significantly poorer HRQoL than the general
population. However, within one year of surgery, primary KA restored HRQoL to levels ex-
pected for the patient’s age-, BMI-, and sex-­matched peers.

Cite this article: Bone Joint J 2023;105-B(4):365–372.

Introduction there is continuing scrutiny regarding the evidence

Correspondence should be
Knee osteoarthritis (OA) is a leading cause of base for primary KA.6
sent to L. Z. Yapp; email: disability, and is estimated to have a global prev- The impact that a treatment has on health-­
liam.yapp@nhs.scot
alence of 22.9% (654.1 million) in people aged related quality of life (HRQoL) is an increas-
© 2023 The British Editorial
Society of Bone & Joint Surgery
over 40 years.1,2 Primary knee arthroplasty (KA) ingly important aspect when determining its
doi:10.1302/0301-620X.105B4. for end-­stage OA is a common procedure which value to patients and to the healthcare system.7
BJJ-2022-0659.R1 $2.00
successfully relieves pain and enhances function The EuroQol five-­dimension questionnaire (EQ-­
Bone Joint J
2023;105-B(4):365–372. in the majority of patients.3-­5 Despite this success, 5D)8 is a common tool for measuring HRQoL and

THE BONE & JOINT JOURNAL 365

366 L. Z. YAPP, C. E. H. SCOTT, C. R. HOWIE, D. J. MACDONALD, A. H. R. W. SIMPSON, N. D. CLEMENT

Table I. Propensity score matched cohorts 1:1.

Covariates General population KA cohort Standardized mean difference
Unmatched Matched Unmatched Matched
Total, n 22,143 3,029 3,029
Median age, yrs (IQR) 49 (35.0 to 64.0) 71.0 (64 to 78) 69.9 (64 to 76) 2.27 -0.074
Male 9,732 1,459 1,366 -0.01 -0.062
Female 12,411 1,570 1,663 0.01 0.062
Median BMI, kg/m2 (IQR) 26.2 (23.8 to 30.2) 29.4 (26.4 to 33.3) 30.1 (26.6 to 34.2) 0.59 0.099
IQR, interquartile range; KA, knee arthroplasty.

is recommended for use by the National Institute for Health
and Care Excellence (NICE) when undertaking cost-­ utility
analyses.9,10
The burden of primary KA is likely to increase over the
coming decades, and it is estimated that one in five patients
awaiting KA are living with negative EQ-­ 5D scores in
health states termed “worse than death”.11,12 Disruption to the
delivery of elective orthopaedic surgery following the global
COVID-­19 pandemic has led to longer waiting times in the
UK and is associated with measurable deterioration in these
patients’ HRQoL.13–16
The majority of studies investigating HRQoL following
KA are primarily interested in the ‘within-­patient’ change.17-­20
However, it is less clear as to how HRQoL differs between KA
patients and their peers in the general population. Consequently,
there is limited context for the associated improvement seen in
HRQoL following primary KA. The aim of this study was to
compare HRQoL in patients undergoing primary KA for OA
with their age- and sex-­matched peers in the general population.
Methods
Study design. This retrospective case-­control study compares
patients undergoing primary KA to the general population.
This study is reported in accordance with the Strengthening
the Reporting of Observational Studies in Epidemiology guide-
lines.21 Demographic and patient-­ reported outcome measure
(PROM) data were obtained from two separate sources: the
Edinburgh Orthopaedic Research Database and the Health
Survey for England (HSE).
Primary KA cohort. The primary KA cohort comprised pa-
tients treated in a university-­teaching hospital during the period
from 1 January 2013 to 31 December 2019. Data were collect-
ed prospectively and stored in an electronic research database.
Preoperatively, patients completed standardized questionnaires
upon attendance at the preadmission clinic. Postoperative
data were collected at one year following surgery via a post-
al questionnaire. Ethical approval to collect PROMs data was
obtained from the Scotland (A) Research Ethics Committee
(20/SS/0125).
Patients were included in the study if they were undergoing
primary KA for OA during the study period. Patients who did
not respond to the questionnaire and/or had incomplete ques-
tionnaires (n = 651 (14.5%)) or who underwent KA for non-­OA
diagnoses (n = 661; 14.7%) were excluded for the purposes
of this study (n = 1,312). A further 144 patients (3.2%) were
excluded due to missing BMI data. During the follow-­ up
period, there were 4,485 patients undergoing primary KA,
of whom 3,029 (67.5%) satisfied the inclusion and exclusion
General population Preop Postop
Postop
Preop
General
population
-0.5 0.0 0.5 1.0
EQ-5D-3L Index score
Fig. 1
Box and whisker with density plot EuroQol five-­dimension three-­
level questionnaire (EQ-­5D-­3L) Index scores: matched cohorts. Preop,
preoperative knee arthroplasty cohort; postop, postoperative knee
arthroplasty cohort.
criteria of the study (median age 69.7 years (interquartile range
(IQR) 64.2 to 76.1); females 1,663 (54.9%), median BMI 30.1
kg/m2 (IQR 26.6 to 34.2)).
General population cohort. The HSE is a national survey per-
formed every year to monitor trends in the health of adults and
children. Data from the HSE were accessed through the UK
Data Service and responses were considered representative of
the general population.22–24 For the purposes of this study, de-­
identified individual level data for the years 2010 to 2012 was
extracted as those years included the EQ-­5D-­3L as part of the
survey. Data access is restricted to individuals based in the UK
and higher education organisations which are part of the UK
Access Management Federation.
HSE respondents were excluded if they had incomplete data
or if they were aged under 16 years. During the period 2010
to 2012, there were 35,062 participants in the HSE. Of those,
9,743 (27.8%) were aged under 16 years and a further 3,176
(9.1%) had incomplete EQ-­5D-­3L data, leaving 22,143 (63.1%)
participants in the control group (median age 49.0 years (IQR
35.0 to 64.0); females 12,411 (56.0%), median BMI 26.7 kg/
m2 (IQR 23.8 to 30.2)). To account for non-­responder bias, the
individual weighting variable published with each HSE dataset
was used.22–24

Follow us @BoneJointJ THE BONE & JOINT JOURNAL

 PRIMARY KNEE ARTHROPLASTY FOR OA RESTORES PATIENTS’ HEALTH-­RELATED QUALITY OF LIFE TO NORMAL POPULATION LEVELS 367

Table II. Summary of EuroQol five-­dimension three-­level Index scores: matched cohorts.
Variable Median EQ-­5D-­3L (IQR)  Difference p-­value† Median Difference p-­value†
 (95% CI)* postoperativeEQ-­5D-­ (95% CI)‡
General General Preoperative
3L (IQR)
population population
(unmatched) (matched)
Sex
Male 1.00 (0.796 to 0.796 (0.701 to 0.620 (0.159 to 0.31 (0.30 to < 0.001 0.796 (0.690 to 1.000) 0.00002 (0.00004 to < 0.001
1.000) 1.000) 0.691) 0.31) 0.002)
Female 1.00 (0.727 to 0.796 (0.691 to 0.552 (0.090 to 0.31 (0.30 to < 0.001 0.760 (0.689 to 1.000) 0.00004 (-­0.00005 to 0.106
1.000) 1.000) 0.691) 0.31) 0.00004)
Age, yrs
Under 45 1.000 (0.848 to 0.848 (0.726 to 0.620 (0.159 to 0.38 (0.18 to < 0.001 0.691 (0.620 to 0.760) 0.14 (-­0.002 to 0.24) 0.090
1.000) 1.000) 0.620) 0.64)
45 to 54 1.000 (0.796 to 0.812 (0.725 to 0.518 (0.106 to 0.38 (0.31 to < 0.001 0.760 (0.656 to 1.000) 0.03 (0.00005 to 0.15) < 0.001
1.000) 1.000) 0.691) 0.46)
55 to 64 0.848 (0.725 to 0.796 (0.691 to 0.362 (0.088 to 0.31 (0.31 to < 0.001 0.760 (0.620 to 1.000) 0.03 (0.00006 to 0.05) < 0.001
1.000) 1.000) 0.691) 0.38)
65 to 74 0.796 (0.725 to 0.796 (0.725 to 0.589 (0.159 to 0.31 (0.31 to < 0.001 0.796 (0.691 to 1.000) 0.00007 (0.00002 to 0.02)< 0.001
1.000) 1.000) 0.691) 0.33)
75 to 84 0.779 (0.689 to 0.760 (0.691 to 0.620 (0.159 to 0.23 (0.19 to < 0.001 0.796 (0.691 to 1.000) -0.00003 (-­0.000002 to 0.342
1.000) 1.000) 0.691) 0.31) 0.000005)
Over 85 0.727 (0.620 to 0.760 (0.691 to 0.587 (0.101 to 0.27 (0.19 to < 0.001 0.796 (0.691 to 1.000) -0.00005 (-­0.07 to 0.130
0.850) 0.850) 0.691) 0.31) 0.00004)
Total 1.000 (0.760 to 0.796 (0.691 to 0.620 (0.159 to 0.31 (0.30 to < 0.001 0.796 (0.690 to 1.000) 0.00005 (0.00003 to < 0.001
1.000) 1.000) 0.691) 0.31) 0.00009)
*General population (matched) vs preoperative KA.
†Mann-­Whitney U test.
‡General population (matched) vs postoperative KA.
CI, confidence interval; EQ-­5D-­3L, EuroQol five-­dimension three-­level questionnaire; IQR, interquartile range; KA, knee arthroplasty.

Table III. Summary of EuroQol visual analogue scale scores: matched cohorts.
Variable Median EQ-­VAS (IQR)  Difference p-­value† Median Difference p-­value†
 (95% CI)* postoperative EQ-­ (95% CI)‡
General General Preoperative
VAS (IQR)
population population
(unmatched) (matched)
Sex
Male 80 (70.0 to 90.0) 80.0 (65.0 to 90.0) 79.0 (60.0 to 85.0) 0.50 (0.00003 to 0.001 80.4 (70.0 to 90.0) -3.00 (-­5.00 to -1.00 < 0.001
3.20)
Female 80 (70.0 to 90.0) 79.0 (60.0 to 90.0) 70.2 (57.0 to 84.3) 1.50 (0.00002 to < 0.001 80.0 (69.9 to 90.1) -4.00 (-­5.10 to -2.00) < 0.001
4.90)
Age, yrs
Under 45 80 (70.0 to 90.0) 80 (63.8 to 80.0) 55 (42.5 to 89.8) 2.00 (-­10.0 to 0.675 77 (65.5 to 87.8) -4.16 (-­18.0 to 10.0) 0.537
30.0)
45 to 54 80 (70.0 to 90.0) 80 (64.0 to 90.0) 70 (50.0 to 84.0) 9.00 (2.0 to 10.2) < 0.001 81 (64.0 to 90.3) -2.00 (-­7.0 to 0.00004) 0.106
55 to 64 80 (70.0 to 90.0) 80 (60.0 to 90.0) 71 (52.0 to 85.0) 4.80 (1.00 to 6.00) < 0.001 80 (69.9 to 90.0) -2.00 (-­5.00 to -0.1) < 0.001
65 to 74 80 (70.0 to 90.0) 80 (69.0 to 90.0) 74 (60.0 to 86.0) 4.80 (1.00 to 6.00) < 0.001 80 (70.0 to 90.3) -1.00 (-­3.0 to -0.2) < 0.001
75 to 84 75 (60.0 to 85.0) 75 (60.0 to 85.0) 72 (60.0 to 82.0) 0.00001 (-­ < 0.001 80 (70.0 to 90.1) -5.70 (-­9.0 to -4.0) < 0.001
0.000006 to 0.20)
Over 85 70 (50.0 to 80.0) 70 (51.5 to 85.0) 71 (59.7 to 81.5) -0.00002 (-­4.80 to 0.536 80 (70.0 to 90.0) -6.2 (-­10.0 to -1.0) < 0.001
2.00)
Total 80 (70.0 to 90.0) 80 (60.0 to 90.0) 72 (60.0 to 85.0) 1.0 (0.00003 to < 0.001 80.2 (70.0 to 90.1) -3.3 (-­5.0 to -2.0) < 0.001
3.6)
*General population (matched) vs preoperative KA.
†Mann-­Whitney U test.
‡General population (matched) vs postoperative KA.
CI, confidence interval; EQ-­VAS, EuroQol visual analog scale; IQR, interquartile range; KA, knee arthroplasty.

Primary outcome. The primary outcome of interest was the dif-
ference in HRQoL (as measured by the EQ-­5D-­3L Index Score)
between the KA cohort and the control group. The Index score
is generated from a health profile based upon answers to five
questions related to the dimensions Mobility, Self-­care, Usual
activities, Pain and Discomfort, and Anxiety and Depression.25
Each question has three possible levels indicating the degree of
impairment (“None” = 1, “Some” = 2, and “Extreme” = 3) and
there are 243 potential health states.
The Index score is derived by applying value sets to the
health state data, based upon the relative importance a defined
population places upon different health problems.26 The use
of preference-­based weights enables the general population’s
views on health problems to be accounted for and accounts for

VOL. 105-B, No. 4, APRIL 2023

368 L. Z. YAPP, C. E. H. SCOTT, C. R. HOWIE, D. J. MACDONALD, A. H. R. W. SIMPSON, N. D. CLEMENT

General population Preoperative Postoperative

Mobility Mobility Mobility

Anxiety & Depression Anxiety & Depression Anxiety & Depression

Self-care Self-care Self-care

Usual activities Pain & Disability Usual activities Pain & Disability Usual activities Pain & Disability

Fig. 2

Radar plot EuroQol five-­dimension three-­level questionnaire (EQ-­5D-­3L) domain scores: matched cohorts. Each segment reflects possible domain
scores (1 - Lowest, 2 - Middle, 3 - Highest).

General population Preop Postop

deteriorate with increasing age.28 Furthermore, sex-­based dif-
ferences in the degree of problems reported for each domain
Male Female have previously been observed.28 Finally, patients with elevated
BMI (classed overweight or obese) are known to have lower
100 (poorer) preoperative Index and EQ-­VAS scores than those with
normal BMI.29
Statistical analysis. All data handling, cleaning, and statistical
75 analysis was undertaken using R Studio v. 1.3.959 (USA). The
distribution of continuous variables was plotted to assess appro-
EQ-VAS score

priateness of parametric or non-­parametric tests of differences.

50
25
0 chi-­squared test. A p-­value of less than 0.05 was considered sta-
General Preop Postop General Preop
population population
Fig. 3
Box and whisker plot EuroQol visual analogue scale (EQ-­VAS) scores:
matched cohorts. Preop, preoperative knee arthroplasty cohort; postop,
postoperative knee arthroplasty cohort.
intrinsic variation that can exist between nations. Using popu-
lation weighting, the Index score summarizes each possible
health profile on a numerical scale (-­0.594 to 1.0). A score of
1 indicates full health, and scores of 0 and less than 0 indicate
a state equivalent to being dead and a state worse than death,
respectively. Index scores were calculated using the ‘Time
Trade Off’ value set for the UK.7
Secondary outcome. The EQ visual analogue scale (VAS) is
a secondary element of the EQ-­5D-­3L which allows respond-
ents to rate their overall health from 0 (worst health imaginable)
to 100 (best health imaginable). The EQ-­VAS encompasses an
individual’s health beyond the five dimensions covered in the
health profile.27
Variables and potential confounders. The age profiles of both
groups differed significantly, with a greater proportion of young-
er age groups represented in the HSE cohort. Overall HRQoL,
as measured by the EQ-­5D, has been shown to significantly
Parametric and non-­parametric distributions were summarized
using the mean with standard deviation (SD) or the median
with IQR, respectively. Differences between general population
and KA cohort EQ-­5D-­3L Index scores were compared using
two-­sided unpaired Mann-­Whitney U test. Differences between
categorical variables between cohorts were measured using the
Postop tistically significant.
Propensity score matching. Propensity score matching with-
out replacement was undertaken using the MatchIt program to
balance covariates and facilitate group comparisons.30 Using a
1:1 'nearest neighbour' method, a propensity score was estimat-
ed using logistic regression, based upon the covariates sex, age,
and BMI. Balancing was assessed by estimating standardized
mean differences for each covariate. Two matched cohorts with
3,029 cohorts were obtained for the final analyses (Table I). The
standardized mean differences for each covariate were below
0.1, suggesting adequate balance.31
Results
Crude analyses. Unadjusted analyses demonstrated significant
differences in HRQoL between the general population and KA
cohorts. The general population had significantly higher Index
and EQ-­VAS scores than the KA cohort at both pre- and postop-
erative levels (Tables II and III).
Matched cohorts: general population vs preoperative KA.
Following balancing of covariates, the preoperative KA cohort
had significantly lower (poorer) Index scores than the general
population cohort (median preoperative KA 0.620 (IQR 0.16 to
0.69) vs median general population 0.796 (IQR 0.69 to 1.00); p
< 0.001, Mann-­Whitney U test) (Table II; Figure 1). The gener-
al population reported significantly lower levels of problems in

Follow us @BoneJointJ THE BONE & JOINT JOURNAL

 PRIMARY KNEE ARTHROPLASTY FOR OA RESTORES PATIENTS’ HEALTH-­RELATED QUALITY OF LIFE TO NORMAL POPULATION LEVELS 369

Table IV. Descriptive summary of EuroQol five-­dimension three-­level questionnaire components: matched cohorts.
Component General population, n (%) Preoperative, n (%) p-­value† Postoperative, n (%) p-­value‡
Mobility < 0.001 < 0.001
1 1,898 (62.7) 259 (9.4) 1,690 (55.8)
2 1,129 (37.3) 2,485 (90.5) 1,335 (44.1)
3 2 (0.1) 2 (0.1) 4 (0.1)
Problems* 1,131 (37.4) 2,487 (90.6) 1,339 (44.2)
Self-­care < 0.001 < 0.001
1 2,745 (90.6) 2,166 (78.9) 2,506 (82.7)
2 273 (9.0) 570 (20.8) 502 (16.6)
3 11 (0.4) 10 (0.4) 21 (0.7)
Problems* 284 (9.4) 580 (21.2) 523 (17.3)
Usual activities < 0.001 < 0.001
1 2,117 (69.9) 499 (18.2) 1,47 (51.1)
2 834 (27.5) 1,976 (72.0) 1,379 (45.5)
3 78 (2.6) 271 (9.9) 103 (3.4)
Problems* 912 (30.1) 2,247 (81.9) 1,482 (48.9)
Pain and Discomfort < 0.001 < 0.001
1 1,360 (44.9) 21 (0.8) 1,206 (39.8)
2 1,444 (47.7) 1,649 (60.0) 1,669 (55.1)
3 225 (7.4) 1,076 (39.2) 154 (5.1)
Problems* 1,669 (55.1) 2,725 (99.2) 1,823 (60.2)
Anxiety and Depression < 0.001 < 0.001
1 2,271 (75.0) 1,842 (67.1) 2,416 (79.8)
2 702 (23.2) 826 (30.1) 562 (18.6)
3 56 (1.8) 78 (2.8) 51 (1.7)
Problems* 758 (25.0) 904 (32.9) 613 (20.3)
*Level 2 + 3.
†Chi-­squared test; general population vs preoperative KA.
‡Chi-­squared test; general population vs postoperative KA.
EQ-­5D-­3L, EuroQol five-­dimension three-­level questionnaire; KA, knee arthroplasty.

every domain measured by the EQ-­5D-­3L (Table IV; Figure 2).
The median EQ-­VAS was significantly lower in the preopera-
tive KA group when compared to the general population cohort
(Table III; Figure 3).
Matched cohorts: general population vs postoperative KA.
At one year postoperatively, the overall median EQ-­ 5D-­
3L
Index score had significantly improved in the KA cohort com-
pared to preoperative levels (mean difference 0.322 (SD 0.33);
p < 0.001, Wilcoxon signed-­rank test). There was no clinically
relevant difference in the overall Index scores between post-
operative KA and general population cohorts (median postop-
erative KA cohort 0.796 (IQR 0.69 to 1.00) vs median general
population cohort 0.796 (0.69 to 1.00)) (Figures 1 and 2).
The median EQ-­VAS was significantly higher one-­year
post-­KA (median preoperative 72.0 (IQR 60.0 to 85.0) vs median
postoperative 80.2 (IQR 70.0 to 90.1); p < 0.001, Wilcoxon
signed-­rank test). The postoperative EQ-­VAS was significantly
higher (better) than the general population, however this was
not a clinically important difference (Table III; Figure 3).
Discussion
This study provides important context for both the significant
impact of knee OA and the positive outcomes of primary KA.
Patients with end-­stage knee OA awaiting KA have signifi-
cantly poorer HRQoL than the general population. However,
within one year of surgery, primary KA restores HRQoL to
levels comparable with the age-, sex-, and BMI-­ matched
general population.
The findings of this study are generalizable and are consis-
tent with results demonstrated in other countries across the
world. Teni et al32 reported patients with knee OA had lower
(poorer) EQ-­5D scores than the Swedish population, but they
did not adjust for differences between cohorts. Dechartres
et al33 demonstrated that KA patients had greater levels of
disability than the French population for specific activities,
such as bending forwards and walking distances greater than
500 yards. However, following surgery, overall physical
disability improved to levels similar to the general popula-
tion. Miettinen et al34 reported that the majority of KA patients
failed to achieve similar levels of HRQoL to the Finnish popu-
lation. Although previous authors have compared HRQoL in
patients undergoing KA and the general population before,
we believe that the current study is one of the first to use
matching techniques to overcome age-, sex-, and BMI-­related
differences in each group, which enables more meaningful
comparisons. This study provides context for how end-­stage
knee OA affects HRQoL and highlights the significant detri-
mental impact of this condition. Furthermore, in an era when
elective waiting lists are rising exponentially,13,14 these data
demonstrate the importance of this procedure to healthcare
providers and are useful for counselling patients about the
potential benefits of KA.

VOL. 105-B, No. 4, APRIL 2023

370 L. Z. YAPP, C. E. H. SCOTT, C. R. HOWIE, D. J. MACDONALD, A. H. R. W. SIMPSON, N. D. CLEMENT
Patients undergoing primary KA may have different long-­
term mortality rates than the general population.35 One assump-
tion for this association is that there is an inherent selection
bias: namely, that patients undergoing primary KA are gener-
ally healthier than their age- and sex-­matched peers. However,
the current study demonstrated that prior to surgery, patients
undergoing primary KA had significantly poorer levels of self-­
reported health, as measured by the EQ-­5D, than the general
population. Greater (better) HRQoL has been shown to be
associated with lower mortality risk.36 Therefore, restoration of
HRQoL could potentially influence life expectancy in patients
undergoing primary KA.
At one year, patients undergoing KA in older age groups
obtained greater EQ-­ VAS scores than their matched peers.
However, this finding should be interpreted with caution, as
while the differences achieved statistical significance they were
not above the threshold to be considered clinically important.37
Therefore, the observed between-­group differences in HRQoL
in these age groups may not be clinically relevant.
A limitation of this study is that it is a retrospective analysis,
which restricts the ability to prove causality. Although the data
are observational, the use of propensity score matching helps
to balance covariates and theoretically reduces the influence
of confounding.31 It is not possible to fully account for unmea-
sured confounding and therefore a risk of bias within these
results does still exist. Level 1 studies have demonstrated that
KA leads to sustained improvements in HRQoL.38,39 Skou et al38
performed a randomized control trial (RCT) comparing total
KA with non-­surgical management for end-­stage OA and found
that KA provided significantly greater improvements in the
EQ-­5D scores at 12 months. Furthermore, the TOPKAT RCT
demonstrated that improvements in the EQ-­5D-­3L Index scores
persisted up to five years following KA surgery.39
A second limitation is that we did not consider the influence
of race or measures of socioeconomic deprivation between
cohorts. In adults aged over 55 years, HRQoL is worse for
minority ethnic groups and has been associated with increased
prevalence of comorbidity and social deprivation.40 Abel et
al41 compared levels of socioeconomic inequality between
constituent countries of the UK, identifying that Scotland and
England had similar levels of deprivation. However, Scotland
has higher levels of overall mortality and lower levels of ethnic
diversity than England.41,42 The links between race and health
are complex. Racial health inequalities are often considered an
extension of socioeconomic inequalities, with ethnic minority
groups commonly inhabiting a more disadvantaged socioeco-
nomic profile.43 In Scotland, the majority of non-­white minority
groups are less socioeconomically deprived and are quite
different to those that live in England.44 These differences could
lead to unmeasured bias between both cohorts.
A third limitation is that the HSE data do not include infor-
mation regarding the presence or absence of knee OA and/or
other debilitating musculoskeletal conditions. While it is esti-
mated that up to 45% of adults will be diagnosed with knee
OA,45 the lifetime risk of actually requiring KA surgery in the
UK appears to be significantly lower (~6% to 7%).46 Although
some patients included in the general population cohort may
have significant symptoms related to their knee, we suspect that
Follow us @BoneJointJ
the number of patients with end-­stage OA awaiting KA is likely
to be low and will not meaningfully detract from the current
study’s findings. In addition, it is important to be clear that
the current study is not trying to compare patients undergoing
primary KA with disease-­free individuals. The purpose of this
study was to examine how the health profiles of patients under-
going primary KA compare with their age-, sex-, and BMI-­
matched peers in the general population.
Approximately 15% (n = 651) of patients who underwent
primary KA for OA were lost to follow-­up, which may indi-
cate a degree of reporting bias. However, Ross et al47 have
previously analyzed non-­responders to PROMs questionnaires
following primary hip and KA in our institution. These authors
demonstrated that there were no significant differences in the
demographic details (age, sex, BMI) or levels of satisfaction
in ‘ultimate responder’ and ‘persistent non-­responder’ cohorts.
This would suggest that responders and non-­responders from
our institution do not demonstrate systematic differences, and
consequently any missing data related to this should be consid-
ered as missing completely at random.48
Finally, the KA cohort is from a single centre and may not
be representative of the ‘average’ patient in the UK. However,
the demographic data of patients undergoing primary KA in
the 19th annual report from the National Joint Registry, which
collects data from England, Wales, Northern Ireland, the Isle
of Man, and Guernsey, are similar to those of the KA cohort
in this study (median age 70 years (IQR 63 to 76); female n =
810,433 (56.2%); mean BMI 30.9 kg/m2) (Table I).49 Overall
comorbidity and case complexity, two factors which can influ-
ence outcome after primary KA, were not able to be compared
between groups.50 This is relevant because patients considered
suitable for elective surgery may have fewer medical comor-
bidities than their peers. However, HRQoL has not been shown
to vary in patients with or without comorbidities following
primary KA.51
In conclusion, this study provides context for both the nega-
tive impact of knee OA and the positive outcome of primary KA
on patients’ HRQoL. Patients awaiting knee arthroplasty for
knee OA have significantly poorer HRQoL than their matched
peers in the general population. However, within one year of
surgery, primary knee arthroplasty restored HRQoL to levels
comparable with the age- and sex-­matched general population.
‍ ‍Take home message
-  Patients awaiting primary knee arthroplasty for knee
osteoarthritis have significantly poorer health-­related quality of
life (HRQoL) than their matched peers in the general population.
-  However, within one year of surgery, primary knee arthroplasty
restored HRQoL to levels comparable with the age- and sex-­matched
general population.
Twitter
Follow L. Z. Yapp @lzyapp
Follow C. E. H. Scott @EdinburghKnee
References
1. Safiri S, Kolahi A-­A, Smith E, et al. Global, regional and national burden of
osteoarthritis 1990-­2017: A systematic analysis of the Global Burden of Disease
Study 2017. Ann Rheum Dis. 2020;79(6):819–828.
THE BONE & JOINT JOURNAL
 PRIMARY KNEE ARTHROPLASTY FOR OA RESTORES PATIENTS’ HEALTH-­RELATED QUALITY OF LIFE TO NORMAL POPULATION LEVELS
2. Cui A, Li H, Wang D, Zhong J, Chen Y, Lu H. Global, regional prevalence, incidence
and risk factors of knee osteoarthritis in population-­based studies. EClinicalMedicine.
2020;29–30:100587.
3. Hamilton DF, Lane JV, Gaston P, et al. What determines patient satisfaction
with surgery? A prospective cohort study of 4709 patients following total joint
replacement. BMJ Open. 2013;3(4):1–7.
4. Scott CEH, Howie CR, MacDonald D, Biant LC. Predicting dissatisfaction
following total knee replacement: A prospective study of 1217 patients. J Bone Joint
Surg Br. 2010;92-­B(9):1253–1258.
5. Dainty JR, Smith TO, Clark EM, Whitehouse MR, Price AJ, MacGregor
AJ. Trajectories of pain and function in the first five years after total hip and knee
arthroplasty : an analysis of patient reported outcome data from the National Joint
Registry. Bone Joint J. 2021;103-­B(6):1111–1118.
6. Blom AW, Donovan RL, Beswick AD, Whitehouse MR, Kunutsor SK. Common
elective orthopaedic procedures and their clinical effectiveness: umbrella review of
level 1 evidence. BMJ. 2021;374(1):1511.
7. Whitehead SJ, Ali S. Health outcomes in economic evaluation: the QALY and
utilities. Br Med Bull. 2010;96(1):5–21.
8. EuroQol Group. EuroQol--a new facility for the measurement of health-­related
quality of life. Health Policy. 1990;16(3):199–208.
9. No authors listed. Guide to the methods of technology appraisal 2013. National
Institute for Health and Care Excellence. 2013. www.nice.org.uk/process/pmg9/​
chapter/foreword (date last accessed 20 February 2023).
10. Longworth L, Yang Y, Young T, et al. Use of generic and condition-­specific
measures of health-­related quality of life in NICE decision-­making: A systematic
review, statistical modelling and survey. Health Technol Assess. 2014;18(9):1–224.
11. Ackerman IN, Bohensky MA, Zomer E, et al. The projected burden of primary
total knee and hip replacement for osteoarthritis in Australia to the year 2030. BMC
Musculoskelet Disord. 2019;20(1):90.
12. Scott CEH, MacDonald DJ, Howie CR. “Worse than death” and waiting for a joint
arthroplasty. Bone Joint J. 2019;101-­B(8):941–950.
13. Oussedik S, MacIntyre S, Gray J, McMeekin P, Clement ND, Deehan DJ.
Elective orthopaedic cancellations due to the COVID-­19 pandemic: where are we
now, and where are we heading? Bone Jt Open. 2021;2(2):103–110.
14. Yapp LZ, Clarke JV, Moran M, Simpson AHRW, Scott CEH. National operating
volume for primary hip and knee arthroplasty in the COVID-­19 era: A study utilizing
the Scottish arthroplasty project dataset. Bone Jt Open. 2021;2(3):203–210.
15. Clement ND, Scott CEH, Murray JRD, Howie CR, Deehan DJ, IMPACT-­Restart
Collaboration. The number of patients “worse than death” while waiting for a hip
or knee arthroplasty has nearly doubled during the COVID-­19 pandemic. Bone Joint J.
2021;103-­B(4):672–680.
16. Clement ND, Wickramasinghe NR, Bayram JM, et al. Significant deterioration
in quality of life and increased frailty in patients waiting more than six months for
total hip or knee arthroplasty : a cross-­sectional multicentre study. Bone Joint J.
2022;104-­B(11):1215–1224.
17. Schatz C, Klein N, Marx A, Buschner P. Preoperative predictors of health-­related
quality of life changes (EQ-­5D and EQ VAS) after total hip and knee replacement: a
systematic review. BMC Musculoskelet Disord. 2022;23(1):58.
18. Liddle AD, Pandit H, Judge A, Murray DW. Patient-­ reported outcomes
after total and unicompartmental knee arthroplasty: A study of 14,076 matched
patients from the National Joint Registry for England and Wales. Bone Joint J.
2015;97-­B(6):793–801.
19. Mohammad HR, Judge A, Murray DW. A matched comparison of the patient-­
reported outcome measures of 38,716 total and unicompartmental knee replacements:
an analysis of linked data from the National Joint Registry of England, Northern
Ireland and Isle of Man and England’s National PROM collection programme. Acta
Orthop. 2021;92(6):701–708.
20. Jenkins PJ, Clement ND, Hamilton DF, Gaston P, Patton JT, Howie CR.
Predicting the cost-­ effectiveness of total hip and knee replacement: A health
economic analysis. Bone Joint J. 2013;95-­B(1):115–121.
21. von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of
Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting
observational studies. Lancet. 2007;370(9596):1453–1457.
22. No authors listed. Health Survey for England 2010. NatCen Social Research
Department of Epidemiology and Public Health University College London. 2015.
http://doi.org/10.5255/UKDA-SN-6986-3 (date last accessed 9 February 2022).
23. NatCen Social Research Department of Epidemiology and Public Health
University College London. Health Survey for England 2011. 2013. http://doi.org/​
10.5255/UKDA-SN-7260-1 (date last accessed 9 February 2022).
VOL. 105-B, No. 4, APRIL 2023
371
24. NatCen Social Research Department of Epidemiology and Public Health
University College London. Health Survey for England 2012. 2014. http://doi.org/​
10.5255/UKDA-SN-7480-1 (date last accessed 9 February 2022).
25. EuroQol Research Foundation. EQ-­5D-­3L User Guide., 2018: 1–34. euroqol.org/​
publications/user-guides
26. Devlin N, Parkin D, Janssen B. Methods for Analysing and Reporting EQ-­5D Data.
Cham, 2020.
27. Feng Y, Parkin D, Devlin NJ. Assessing the performance of the EQ-­VAS in the NHS
PROMs programme. Qual Life Res. 2014;23(3):977–989.
28. Kind P, Dolan P, Gudex C, Williams A. Variations in population health
status: results from a United Kingdom national questionnaire survey. BMJ.
1998;316(7133):736–741.
29. Baker P, Petheram T, Jameson S, Reed M, Gregg P, Deehan D. The association
between body mass index and the outcomes of total knee arthroplasty. J Bone Joint
Surg Am. 2012;94-­A(16):1501–1508.
30. Ho DE, King G, Stuart EA. MatchIt: nonparametric preprocessing for parametric
causal inference. J Stat Softw. 2011;42(8):1–28.
31. Austin PC. An introduction to propensity score methods for reducing the effects of
confounding in observational studies. Multivariate Behav Res. 2011;46(3):399–424.
32. Teni FS, Rolfson O, Devlin N, et al. Longitudinal study of patients’ health-­related
quality of life using EQ-­5D-­3L in 11 Swedish National Quality Registers. BMJ Open.
2022;12(1):e048176.
33. Dechartres A, Boutron I, Nizard R, et al. Knee arthroplasty: disabilities in
comparison to the general population and to hip arthroplasty using a French national
longitudinal survey. PLoS One. 2008;3(7):e2561.
34. Miettinen HJA, Mäkirinne-­Kallio N, Kröger H, Miettinen SSA. Health-­
related quality of life after hip and knee arthroplasty operations. Scand J Surg.
2021;110(3):427–433.
35. Yapp LZ, Clement ND, Moran M, Clarke JV, Simpson A, Scott CEH. Long-­
term mortality rates and associated risk factors following primary and revision knee
arthroplasty. Bone Joint J. 2022;104-­B(1):45–52.
36. Phyo AZZ, Freak-­Poli R, Craig H, et al. Quality of life and mortality in the
general population: a systematic review and meta-­analysis. BMC Public Health.
2020;20(1):1596.
37. Kang S. Assessing responsiveness of the EQ-­5D-­3L, the Oxford Hip Score, and the
Oxford Knee Score in the NHS patient-­reported outcome measures. J Orthop Surg
Res. 2021;16(1):18.
38. Skou ST, Roos EM, Laursen MB, et al. A randomized, controlled trial of total knee
replacement. N Engl J Med. 2015;373(17):1597–1606.
39. Beard DJ, Davies LJ, Cook JA, et al. The clinical and cost-­ effectiveness
of total versus partial knee replacement in patients with medial compartment
osteoarthritis (TOPKAT): 5-­year outcomes of a randomised controlled trial. Lancet.
2019;394(10200):746–756.
40. Watkinson RE, Sutton M, Turner AJ. Ethnic inequalities in health-­related quality
of life among older adults in England: secondary analysis of a national cross-­sectional
survey. Lancet Public Health. 2021;6(3):e145–e154.
41. Abel GA, Barclay ME, Payne RA. Adjusted indices of multiple deprivation to
enable comparisons within and between constituent countries of the UK including an
illustration using mortality rates. BMJ Open. 2016;6(11):e012750.
42. Schofield L, Walsh D, Feng Z, et al. Does ethnic diversity explain intra-­UK variation
in mortality? A longitudinal cohort study. BMJ Open. 2019;9(3):e024563.
43. Tunstall H, Mitchell R, Gibbs J, Platt S, Dorling D. Socio-­demographic diversity
and unexplained variation in death rates among the most deprived parliamentary
constituencies in Britain. J Public Health (Oxf). 2012;34(2):296–304.
44. Walsh D, Buchanan D, Douglas A, et al. Increasingly diverse: the changing ethnic
profiles of Scotland and Glasgow and the implications for population health. Appl
Spatial Analysis. 2019;12(4):983–1009.
45. Murphy L, Schwartz TA, Helmick CG, et al. Lifetime risk of symptomatic knee
osteoarthritis. Arthritis Rheum. 2008;59(9):1207–1213.
46. Culliford DJ, Maskell J, Kiran A, et al. The lifetime risk of total hip and knee
arthroplasty: results from the UK general practice research database. Osteoarthritis
Cartilage. 2012;20(6):519–524.
47. Ross LA, O’Rourke SC, Toland G, MacDonald DJ, Clement ND, Scott CEH.
Loss to patient-­reported outcome measure follow-­up after hip arthroplasty and knee
arthroplasty: patient satisfaction, associations with non-­response, and maximizing
returns. Bone Jt Open. 2022;3(4):275–283.
48. Bhaskaran K, Smeeth L. What is the difference between missing completely at
random and missing at random? Int J Epidemiol. 2014;43(4):1336–1339.
49. No authors listed. 19th Annual Report 2022: National Joint Registry. National
Joint Registry. 2022. https://reports.njrcentre.org.uk/Portals/0/PDFdownloads/​
372 L. Z. YAPP, C. E. H. SCOTT, C. R. HOWIE, D. J. MACDONALD, A. H. R. W. SIMPSON, N. D. CLEMENT
NJR%2019th%20Annual%20Report%202022.pdf (date last accessed 20 February
2023).
50. Podmore B, Hutchings A, van der Meulen J, Aggarwal A, Konan S. Impact of
comorbid conditions on outcomes of hip and knee replacement surgery: A systematic
review and meta-­analysis. BMJ Open. 2018;8(7):e021784.
51. Podmore B, Hutchings A, Skinner JA, MacGregor AJ, van der Meulen J.
Impact of comorbidities on the safety and effectiveness of hip and knee arthroplasty
surgery. Bone Joint J. 2021;103-­B(1):56–64.
Author information:
L. Z. Yapp, BMSc (Hons), MBChB, MRCSEd, Specialty Registrar in
Orthopaedics
C. E. H. Scott, MD, MSc, BSc, FRCSEd (Tr&Orth), MFSTEd, Consultant
Orthopaedic Surgeon and NRS Clinical Research Fellow
D. J. MacDonald, BA, MIQA, CQP, Research Associate
C. R. Howie, FRCSEd (Tr&Orth), FRCS (Glas), Consultant Orthopaedic
Surgeon
N. D. Clement, PhD, MD, FRCS (Tr&Orth), Consultant Orthopaedic Surgeon
Department of Orthopaedics, Division of Clinical and Surgical Sciences,
University of Edinburgh, Edinburgh, UK; Department of Trauma &
Orthopaedic Surgery, Royal Infirmary of Edinburgh, Edinburgh, UK.
A. H. R. W. Simpson, MA, BCh, DM, FRCS (Ed&Eng), George Harrison-­
Law Professor of Orthopaedics and Honorary Consultant Orthopaedic
Surgeon, Editor-­in-­Chief, Department of Orthopaedics, Division of Clinical
and Surgical Sciences, University of Edinburgh, Edinburgh, UK; Department
of Trauma & Orthopaedic Surgery, Royal Infirmary of Edinburgh, Edinburgh,
UK; Bone & Joint Research, London, UK.
Author contributions:
L. Z. Yapp: Conceptualization, Methodology, Investigation, Formal analysis,
Visualization, Writing – original draft, Writing – review & editing.
Follow us @BoneJointJ
C. E. H. Scott: Conceptualization, Methodology, Investigation, Formal
analysis, Supervision, Visualization, Writing – review & editing.
D. J. MacDonald: Project administration, Resources, Data curation, Writing
– review & editing.
C. R. Howie: Project administration, Resources, Data curation, Writing –
review & editing.
A. H. R. W. Simpson: Project administration, Resources, Supervision, Data
curation, Writing – review & editing.
N. D. Clement: Conceptualization, Methodology, Supervision, Investigation,
Formal analysis, Visualization, Writing –review & editing.
Funding statement:
The authors received no financial or material support for the research,
authorship, and/or publication of this article.
ICMJE COI statement:
N. D. Clement is a member of the editorial board of Bone & Joint Open,
Bone & Joint 360, and The Bone & Joint Journal. C. E. H. Scott reports
an institutional research grant from Stryker, and consulting fees from
Stryker and Smith & Nephew, all of which are unrelated to this study. C.
E. H. Scott is also a member of the editorial boards of The Bone & Joint
Journal and Bone & Joint Research, and is on an advisory board for Pfizer.
A. H. R. W. Simpson is the Editor-­in-­Chief of Bone & Joint Research, and
reports multiple grants from RCUK, Charities, and Stryker, unrelated to
this study.
Ethical review statement:
Ethical approval to collect patient-­reported outcome measures (PROMs)
data as part of the Edinburgh Orthopaedic Research Database (EORD)
was obtained from the Scotland (A) Research Ethics Committee (20/
SS/0125). Approval to analyze PROMs data as part of this project was
obtained from the EORD Data Controller.
This article was primary edited by A. Wood.
THE BONE & JOINT JOURNAL