Total hip or knee replacement for osteoarthritis: Mid- and long-term quality of life

Authors

  • Anne-Christine Rat,

    Corresponding author
    1. INSERM, CIC-EC CIE6, Centre Hospitalier Universitaire Nancy, Epidémiologie et Evaluation Cliniques, Nancy-University, Paris Descartes, Metz P Verlaine, Research unit Approches Psychologiques et Epidémiologiques des Maladies Chroniques, EA 4360, and Centre Hospitalier Universitaire Nancy, Service de Rhumatologie, Nancy, France
    • Centre d'Epidémiologie Clinique, CIC-INSERM CIE6, Service d'Epidémiologie et Évaluation Cliniques, Hôpital Marin-Centre Hospitalier Universitaire de Nancy, 2 Avenue du Maréchal de Lattre de Tassigny, CO 34, 54035 Nancy Cédex, France
    Search for more papers by this author
  • Francis Guillemin,

    1. INSERM, CIC-EC CIE6, Centre Hospitalier Universitaire Nancy, Epidémiologie et Evaluation Cliniques, Nancy-University, Paris Descartes, Metz P Verlaine, Research unit Approches Psychologiques et Epidémiologiques des Maladies Chroniques, EA 4360, Nancy, France
    Search for more papers by this author
  • Georges Osnowycz,

    1. CH Neufchateau, Service de Chirurgie, Neufchateau, France
    Search for more papers by this author
  • Jean-Pierre Delagoutte,

    1. Centre Hospitalier Universitaire Nancy, Service de Chirurgie Orthopedique, Nancy, France
    Search for more papers by this author
  • Christian Cuny,

    1. CHR Metz, Service de Chirurgie Orthopedique, Metz, France
    Search for more papers by this author
  • Didier Mainard,

    1. Centre Hospitalier Universitaire Nancy, Service de Chirurgie Orthopedique, Nancy, France
    Search for more papers by this author
  • Cédric Baumann

    1. INSERM, CIC-EC CIE6, Centre Hospitalier Universitaire Nancy, Epidémiologie et Evaluation Cliniques, Nancy-University, Paris Descartes, Metz P Verlaine, Research unit Approches Psychologiques et Epidémiologiques des Maladies Chroniques, EA 4360, Nancy, France
    Search for more papers by this author

Abstract

Objective

To compare quality of life (QOL) scores 3 and 10 years after total hip arthroplasty (THA) or total knee arthroplasty (TKA) for osteoarthritis with QOL scores in a general population, and to determine factors associated with QOL after surgery.

Methods

Data were obtained from 2 multicenter cohorts of patients with THA or TKA: 232 patients were recruited during 2003 (3-year cohort) and 221 patients were recruited during 1994 (10-year cohort). Preoperative data (QOL, radiograph results) and followup data (demographics, comorbidities, pain locations, environmental factors, and QOL) were collected. QOL data for the general population were obtained from a 2003 population-based survey.

Results

A total of 195 and 89 patients for the 3- and 10-year cohorts, respectively, were followed up; the mean age at followup was 73 years. For both of the cohorts, physical functioning and role-physical or role-emotional QOL scores were lower than those for a general population with comparable age. Scores for pain, mental health, and social dimensions were lower than those for the reference population for only the 10-year cohort. For both cohorts, increased number of comorbidities, painful locations other than THA or TKA location, and unfavorable environmental factors were associated with impaired QOL. Low preoperative QOL scores were predictive of impaired QOL at followup for only the 3-year cohort.

Conclusion

THA or TKA can improve QOL, but the benefits may be time limited. Addressing environmental factors and treating comorbidities and pain in locations other than the arthroplasty location could have mid- and long-term effects on the QOL of patients with THA or TKA.

INTRODUCTION

Hip or knee osteoarthritis (OA) is a major cause of pain and disability. For OA that does not respond to medical treatment, total hip arthroplasty (THA) or total knee arthroplasty (TKA) is effective in the short term for alleviating pain and restoring function (1–5). In 2000 in Canada, 44.9 per 1,000 women and 20.8 per 1,000 men were estimated to require THA or TKA (6). In the UK, the rate of hip replacements is estimated to increase by 40% between 1996 and 2026 because of demographic changes alone (7). The proportion of people with THA or TKA is increasing because of the broadening of indications and the aging of the population. Long-term survival of prostheses has also increased; more than 90% of patients with THA do not need revision of their prosthesis 10 years after surgery (8), and 80% of patients with TKA do not need revision after 15 years (9).

Most studies of TKA and THA have followed patients for only 1 to 2 years after surgery, and evidence for longer-term outcomes is less extensive (10). Results for the long term have focused on functional status or activities of daily life (11, 12), and rarely on quality of life (QOL) (13, 14). Several studies have combined different diagnoses (e.g., rheumatoid arthritis and OA), but the QOL probably differs with the conditions.

THA or TKA can decrease disability and improve QOL. In several studies, patients' QOL one year after surgery was close to that of control groups (1, 15). However, the sustainability of the benefits of arthroplasty over time is not guaranteed. A few recent studies, especially long-term studies, showed that physical abilities are impaired after THA or TKA (11, 12, 16–19). Except for QOL or functional ability at the time of surgery, factors associated with QOL after the procedures are not well studied. Painful arthroplasty, pain in other joints, comorbidities, and environmental factors could be major determinants of QOL. Further information on determinants of QOL after THA or TKA might help satisfy patient needs, particularly if environmental factors or comorbidities are associated with QOL. Specific care could improve QOL. In addition, data on determinants of long-term QOL might help in selecting patients for surgery.

The objectives of this study were to 1) compare QOL scores 3 and 10 years after THA or TKA for OA with age- and sex-adjusted QOL scores in a general population with comparable age, and 2) determine factors associated with QOL after surgery.

PATIENTS AND METHODS

Design.

The study involved 2 multicenter cohorts of patients who underwent THA or TKA. Patients for the 3-year cohort (n = 232) (20) were included between April 2002 and June 2004, and patients for the 10-year cohort (n = 221) (3) were recruited during 1994. Patients' vital statuses were checked in the national identity directory for France (Répertoire national d'identification des personnes physiques). Patients were then mailed questionnaires (see below) and were followed up by 2 phone call reminders (1 and 2 months after the first mailing).

The study protocol was approved by the regional ethics committee (Comité Consultatif sur le Traitement de l'Information en matière de Recherche dans le domaine de la Santé, 05546) and the Comité National Informatique et Liberté (02–1181), which ensures the confidentiality of information.

Study sample.

Patients were included if they had OA according to American College of Rheumatology criteria (21, 22) and if they were scheduled to receive THA or TKA within the next 3 months. Consecutive patients followed in 4 outpatient clinics were asked to participate. Arthroplasty was usually unilateral, or if bilateral, a delay of more than 1 year between the 2 surgeries was mandatory. Patients were not included if the surgery was a revision of a previous procedure, if they did not speak French, or if they had another disabling disorder.

Both cohorts were recruited from the same surgical centers in the Lorraine region of eastern France, with one specialized teaching center (Nancy University Hospital) and 3 local centers (Nancy, Metz, and Neufchateau).

Data collected.

Baseline data.

Demographic and clinical data for the 2 cohorts were collected at inclusion, before surgery (Figure 1). Age, sex, OA joint, and years of schooling were recorded. The severity of OA was measured in terms of pain (on a 0–100-mm visual analog scale [VAS]), walking distance (in meters), Kellgren/Lawrence staging of radiograph results (23), and pain, function, and clinical data by the Harris score for hip OA (24) and by the Index of Severity for Knee OA (ISK) (25). For the 3-year cohort, QOL data were collected by a generic questionnaire, the Medical Outcomes Survey Short Form 36 (SF-36), and a specific questionnaire, the Osteoarthritis Knee and Hip Quality of Life questionnaire (OAKHQOL). For the 10-year cohort, QOL data were collected by a generic questionnaire, the Nottingham Health Profile (NHP), and a specific questionnaire, the Arthritis Impact Measurement Scales 2 (26, 27).

Figure 1.

Data collected. * = outcome criteria: dependant variable; THA = total hip arthroplasty; TKA = total knee arthroplasty; ISK = Index of Severity for Knee osteoarthritis; VAS = visual analog scale; QOL = quality of life; SF-36: Medical Outcomes Survey Short Form 36; OAKHQOL = Osteoarthritis Knee and Hip Quality of Life questionnaire; NHP = Nottingham Health Profile; AIMS2 = Arthritis Impact Measurement Scales 2; WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index; FCI = Functional Comorbidity Index; WHOQOL-BREF = World Health Organization Quality of Life BREF.

We also collected data on postoperative complications over a 1-year period. These complications included fractures, heart complications, phlebitis, dislocation, infection, and hematoma. Success of surgery was defined as an improvement of >50% in the Harris hip score or ISK, and a >50% decrease in VAS score for pain, 1 year after surgery. Patients completed QOL questionnaires 6 months and 1 year after surgery.

Followup data.

Followup data were collected in 2006 and 2007, 3 and 10 years after inclusion for the 3- and 10-year cohorts, respectively. Data were collected on age, occupational activity, marital status, residence, body mass index, number of THAs or TKAs, number of painful locations, comorbidities (Functional Comorbidity Index [FCI]), environmental factors (World Health Organization Quality of Life questionnaire-brief version [WHOQOL-BREF]), QOL (SF-36 and OAKHQOL), functional status (Western Ontario and McMaster Universities Osteoarthritis Index [WOMAC]), and pain (VAS).

Measures.

Generic QOL questionnaire.

The SF-36 (28, 29) is the most common generic QOL instrument used for patients with OA, particularly in the context of THA or TKA. It contains 36 items for 8 dimensions: physical functioning, physical role, mental health, emotional role, bodily pain, social functioning, vitality, and general health. We chose not to use summary component scores because recent results suggested that for patients with hip or knee OA, aggregate scores as they are currently defined may not be optimal (30).

The NHP (31) is intended to provide a brief indication of a patient's perceived emotional, social, and physical health problems. Part I contains 38 questions in 6 subareas (energy level, pain, emotional reaction, sleep, social isolation, and physical abilities). Each question is assigned a weighted value; the sum of the weighted values for each subarea totals 100.

Knee- and hip-specific OA QOL questionnaire.

The OAKHQOL (version 2.3) (32) is a specific QOL questionnaire for patients with knee or hip OA. It is self-administered and comprises 43 items in 5 dimensions: physical activity, mental health, pain, social support, and social activities, plus 3 independent items.

The Arthritis Impact Measurement Scales 2 (26, 27) contains 78 items and measures 5 health status components: physical, pain, role, social, and affects. Dimension scores of the 4 instruments were standardized from 0 (worst QOL) to 100 (best QOL).

Pain and functional status.

Pain was measured by a VAS and functional status was measured by the WOMAC (33), the Harris score for hip OA (24), and the ISK (25). Scores for function (latter 3 scales) were standardized from 0 (worst function) to 100 (best function).

Environmental factors and comorbidities.

To study environmental factors, patients completed the environmental dimension of the WHOQOL-BREF (34), which contains items asking about the physical environment, such as satisfaction with conditions of the living place, satisfaction with transport and health services, and availability of leisure activities.

Presence of comorbidities was quantified by the FCI (35). This index was developed to predict physical function in the general population. It contains 18 diagnoses scored by adding the number of “yes” answers, a score of 0 indicating no comorbid illness, and a score of 18 indicating the highest number of comorbid illnesses.

The number of painful locations is a marker of diffuse disease. Patients were asked to draw their painful areas on a schema. The number of painful areas (hands, elbows, shoulders, hips, knees, back, and neck) was totaled to create 2 variables: presence of painful THA or TKA location and presence of other painful locations.

Population data for QOL.

To compare patients' SF-36 scores with scores for the French general population, we used results from a population-based survey of perceived health status and use of health care services conducted by the French National Institute of Statistics and Economic Studies between October 2002 and September 2003 (36). During the survey, 25,931 people were consulted, the mean age was 47 years (range 18–100 years), and 53% were women. A total of 1,038 people were between ages 70 and 75 years and completed QOL questionnaires.

Statistical analysis.

Baseline characteristics and followup data were summarized with descriptive statistics. Data for the 2 cohorts were compared by use of the Student's t-test for quantitative variables and the chi-square or Fisher's exact tests for qualitative variables. QOL scores were described at patient inclusion and 1 and 3 years after surgery for the 3-year cohort. An analysis of variance was used to compare QOL scores for patients and the reference population, with adjustment for age and sex.

Student's t-test or Pearson's correlation testing was used to analyze the association of preoperative or followup data and end point QOL scores. Multivariate linear regression models were used to examine factors predicting QOL, with separate regression models constructed for each QOL dimension. Models were adjusted for age at surgery, sex, and OA joint. Preoperative QOL scores for each corresponding dimension, comorbidities, environmental factors, number of painful locations other than THA or TKA location, and variables found to be associated with QOL on bivariate analysis (P < 0.1) were used as candidate variables in the multivariate analysis. To avoid overadjustment, the Harris hip score or ISK was not entered in the models with QOL physical activity or pain dimensions.

Statistical analysis involved the use of SAS, version 9.1, for Windows (SAS Institute, Cary, NC). A P value less than 0.05 was considered statistically significant.

RESULTS

The response rate for the study was 84% (195 of 232) for the 3-year cohort and 40% (89 of 221); 69% (89 of 129) for those actually contacted) for the 10-year cohort (Figure 2).

Figure 2.

Study flow of patients in the 3- and 10-year cohorts who underwent total hip arthroplasty or total knee arthroplasty.

Difference between responders and nonresponders.

For the 3-year cohort, responders (n = 195) and nonresponders (n = 37) did not differ in sociodemographic data, clinical data, patient-reported outcome measures, presence of complications within 1 year after surgery, or success of surgery.

For the 10-year cohort, as compared with nonresponders (or those lost to followup; n = 132), responders (n = 89) were, on average, younger at surgery (mean ± SD age 61.9 ± 11.8 versus 66.5 ± 1.0 years; P = 0.005), less frequently lived alone (19% versus 37%; P = 0.005), and had lower VAS scores for pain before surgery (mean ± SD 63.7 ± 20.2 versus 70.1 ± 17.7 mm; P = 0.03) and higher Harris hip score or ISK (less impaired function, mean ± SD 47.5 ± 12.0 versus 42.8 ± 15.0; P = 0.01) and physical activity scores before surgery (better QOL, mean ± SD 71.8 ± 14.1 versus 67.4 ± 17.3; P = 0.048).

Patient characteristics at baseline.

The main demographic and clinical characteristics of the patients are shown in Table 1. Before surgery, the mean ± SD VAS scores for pain were 63.7 ± 17.2 and 63.8 ± 20.2 for the 3- and 10-year cohorts, respectively. For the 3-year cohort, 18.9% (n = 34), 66.7% (n = 120), and 14.4% (n = 26) of the patients had Kellgren/Lawrence scale grades of 2, 3, and 4, respectively, and for the 10-year cohort, the percentages were 7.8% (n = 5), 51.6% (n = 33), and 40.6% (n = 26), respectively. The success of the procedure was 94.3% (116 of 123) for the 3-year cohort and 98.2% (53 of 54) for the 10-year cohort.

Table 1. Demographic, social, clinical, and health characteristics of patients in the 3- and 10-year cohorts who underwent total hip arthroplasty (THA) or total knee arthroplasty (TKA)*
 3-year cohort (n = 195)10-year cohort (n = 89)P
  • *

    Values are the mean ± SD unless otherwise indicated. SF-36 = Medical Outcomes Survey Short Form 36; NHP = Nottingham Health Profile; ISK = Index of Severity for Knee osteoarthritis; FCI = Functional Comorbidity Index; WHOQOL-BREF = World Health Organization Quality of Life questionnaire-brief version; OAKHQOL = Osteoarthritis Knee and Hip Quality of Life questionnaire; WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index.

  • Scores range from 0 (worst environment) to 100 (best environment).

  • Scores range from 0 (worst quality of life) to 100 (best quality of life).

  • §

    Scores range from 0 (worse functional ability) to 100 (best functional ability).

Baseline data   
 Women, no. (%)112 (57.4)50 (56.2)0.84
 Joint, no. (%)  0.09
  Hip125 (64.1)66 (74.2) 
  Knee70 (35.9)23 (25.8) 
 Years of schooling, no. (%)  0.69
  Primary137 (73.7)68 (77.3) 
  Secondary42 (22.6)16 (18.2) 
  University7 (3.8)4 (4.6) 
 SF-36   
  Physical functioning37.1 ± 22.2  
  Mental health55.3 ± 19.3  
  Pain34.0 ± 15.3  
  Social functioning62.7 ± 23.2  
 NHP   
  Physical abilities 48.8 ± 20.8 
  Emotional reaction 73.9 ± 27.8 
  Pain 29.1 ± 25.9 
  Social isolation 83.7 ± 25.1 
 Harris hip score41.4 ± 13.247.9 ± 11.30.002
 ISK50.6 ± 13.346.4 ± 14.20.22
Followup data   
 Age, years72.4 ± 8.373.3 ± 11.90.54
 Body mass index, kg/m228.8 ± 4.930.4 ± 14.70.38
 Rural residence, no. (%)120 (71.9)43 (55.8)0.01
 Living as a couple, no. (%)131 (73.6)51 (60.7)0.03
 Occupational activity (yes), no. (%)4 (2.1)7 (8.2)0.04
 Time since surgery, years2.7 ± 0.711.9 ± 0.6 
 No. of THAs or TKAs, no. (%)  0.05
  1151 (77.4)59 (66.3) 
  241 (21.0)23 (25.8) 
  3 or 43 (1.5)7 (7.9) 
 Comorbidities (FCI; range 0–18)2.5 ± 2.03.2 ± 2.40.02
 Back pain, no. (%)91 (54.2)46 (59.0)0.48
 Visual impairment, no. (%)51 (31.1)23 (31.9)0.90
 Cardiac insufficiency, no. (%)29 (17.9)21 (29.6)0.05
 Environment (WHOQOL-BREF)70.6 ± 20.664.9 ± 22.20.04
 No. of painful THAs or TKAs, no. (%)  0.05
  0141 (72.3)53 (59.5) 
  143 (22.1)25 (28.1) 
  211 (5.6)11 (12.4) 
 No. of painful locations other than THA or TKA location, no. (%)  0.08
  086 (44.1)32 (36.0) 
  144 (22.6)15 (16.9) 
  >165 (33.3)42 (47.2) 
 Walking distance, meters1,928 ± 2,1801,346 ± 1,4890.02
 Outdoor walks every day152 ± 84.956 ± 71.80.01
 Domestic assistance32 ± 18.927 ± 32.50.02
 Use of a stick54 ± 29.338 ± 45.20.01
 OAKHQOL   
  Physical activities57 ± 2444 ± 26< 0.0001
  Mental health74 ± 2261 ± 26< 0.0001
  Pain66 ± 2554 ± 320.003
  Social support58 ± 2865 ± 280.08
  Social activities64 ± 3059 ± 290.22
 WOMAC§   
  Function66 ± 2152 ± 26< 0.0001
  Pain70 ± 2157 ± 26< 0.0001

Patient characteristics at followup.

The mean ± SD age at followup was 73 ± 9.5 years and was similar in the 2 cohorts; 57% of the patients were women.

Patients in the 10-year cohort more commonly used domestic help and a technical device to walk than did those in the 3-year cohort. For the 2 cohorts, pain without precise localization was <30 on a 0–100 VAS scale; 32% had residual pain in an operated joint. Among patients with a painful location, 43% and 53% of the 3- and 10-year cohorts, respectively, had THA or TKA pain. Twenty-two patients had 2 painful THA or TKA locations.

As compared with the 10-year cohort, patients in the 3-year cohort more frequently lived in a rural area, were in a relationship, and had fewer comorbidities, and had a more favorable perception of environmental factors, better QOL or function, and better pain scores.

Evolution of QOL scores.

For the 3-year cohort, the QOL scores of the SF-36 improved as early as 6 months postoperatively and up to 1 year, and then reached a plateau at 3 years.

Comparison of QOL for patients with THA or TKA and the reference population ages 70–75 years.

QOL scores for the 10-year cohort were significantly lower than those for the reference population after adjusting for age and sex (P < 0.05) (Figure 3). Differences in scores were 5 points or more, which is considered to be greater than the minimally important clinical difference for SF-36 scores (37). For the 3-year cohort, physical functioning, role-physical, and role-emotional dimension scores were significantly lower than those for the reference population by at least 5 points.

Figure 3.

Mean Medical Outcomes Survey Short Form 36 (SF-36) scores for the 3- and 10-year cohorts who underwent total hip arthroplasty or total knee arthroplasty and comparison with the general population after adjusting for age and sex. Norms are SF-36 mean scores for the general population ages 70–75 years. * = P < 0.05.

Factors associated with QOL.

On bivariate analysis, low QOL at followup was associated with high mean age at surgery, living alone, low QOL scores before surgery, a high number of comorbidities, unfavorable environmental factors, and the presence of painful locations other than the THA or TKA location. Having undergone THA was associated with high QOL scores only for the 10-year cohort, and a low Harris score or ISK before surgery was associated with low QOL only for the 3-year cohort. Painful THA or TKA location was associated with a decreased SF-36 bodily pain score and OAKHQOL physical activity score for the 3-year cohort only.

Results of the multivariate analysis are shown in Tables 2 and 3. Factors associated with current status, such as high number of comorbidities, increased number of unfavorable environmental factors, and painful location other than THA or TKA location, were associated with low QOL at followup. Except for the mental health dimension, preoperative low QOL scores were predictive of low QOL scores 3 years but not 10 years after surgery. The variance explained by the different models was fair to high (range 22–68%).

Table 2. Factors associated with quality of life (QOL) 3 years after total hip arthroplasty (THA) or total knee arthroplasty (TKA)*
 SF-36 physical functioningSF-36 bodily painSF-36 mental healthSF-36 social functioningOAKHQOL physical activityOAKHQOL painOAKHQOL mental healthOAKHQOL social activities
βPβPβPβPβPβPβPβP
  • *

    Multivariate models were all adjusted for sex and joint. SF-36 = Medical Outcomes Survey Short Form 36; OAKHQOL = Osteoarthritis Knee and Hip Quality of Life questionnaire; FCI = Functional Comorbidity Index; R2 = variance explained by the models.

  • Preoperative QOL score for each corresponding dimension (e.g., preoperative physical functioning score of the SF-36 for the regression model with physical functioning).

Age at surgery, years−0.550.0020.080.650.020.860.140.14−0.300.10−0.040.84−0.020.89−0.550.04
Marital status (couple vs. alone)−6.080.11−1.080.77−0.540.85−4.360.250.080.981.890.63−2.370.4193.760.50
QOL at baseline (range 0–100)0.28< 0.0010.100.290.31< 0.0010.170.010.38< 0.0010.28< 0.0010.34< 0.0010.280.002
Painful location other than THA or TKA location (no/yes)8.090.0116.91< 0.0018.420.0016.670.039.680.00312.93< 0.0012.830.253.190.48
Comorbidity (FCI; range 0–18)−2.180.009−2.330.004−1.570.01−2.94< 0.001−1.910.02−2.370.006−2.84< 0.001−2.610.03
Environmental factors (range 0–100)0.40< 0.0010.34< 0.0010.27< 0.0010.270.0010.32< 0.0010.33< 0.0010.31< 0.0010.45< 0.001
R20.43 0.42 0.48 0.32 0.38 0.43 0.50 0.29 
Table 3. Factors associated with quality of life (QOL) 10 years after total hip arthroplasty (THA) or total knee arthroplasty (TKA)*
 SF-36 physical functioningSF-36 bodily painSF-36 mental healthSF-36 social functioningOAKHQOL physical activityOAKHQOL painOAKHQOL mental healthOAKHQOL social activities
βPβPβPβPβPβPβPβP
  • *

    Multivariate models were all adjusted for sex and joint. SF-36 = Medical Outcomes Survey Short Form 36; OAKHQOL = Osteoarthritis Knee and Hip Quality of Life questionnaire; FCI = Functional Comorbidity Index; R2 = variance explained by the models.

  • Preoperative QOL score for each corresponding dimension (e.g., preoperative physical abilities score of the Nottingham Health Profile for the regression model with physical functioning).

Age at surgery, years−0.420.140.120.650.230.250.340.19−0.370.100.440.070.150.32−0.410.08
Marital status (couple vs. alone)12.130.127.910.242.690.617.440.28−1.480.792.630.695.310.1814.480.03
QOL at baseline (range 0–100)−0.060.700.200.090.240.010.220.110.210.220.180.240.38< 0.0010.080.64
Painful location other than THA or TKA location (no/yes)8.650.2115.480.014.890.314.450.4610.280.0521.110.00112.110.0022.000.75
Comorbidity (FCI; range 0–18)−2.070.19−3.220.02−2.860.01−3.000.03−4.13< 0.001−3.360.02−3.17< 0.0012.470.06
Environmental factors (range 0–100)0.140.360.140.300.380.0010.250.060.280.010.390.0050.47< 0.0010.70< 0.001
R20.22 0.39 0.54 0.37 0.46 0.47 0.68 0.36 

DISCUSSION

Because long-term QOL after THA or TKA for OA has been little studied, we aimed to compare QOL scores 3 and 10 years after surgery with QOL scores in the general population and determine factors associated with QOL after surgery. Three years after THA or TKA, physical functioning and role dimensions of QOL remained limited for patients as compared with values for the aged-matched general population. At 10 years after surgery, scores for all of the QOL dimensions were lower than those for the reference population. Comorbidities, environmental factors, and the presence of a painful location other than the THA or TKA location contributed to low QOL. Preoperative QOL scores were predictive of QOL at 3 years but not 10 years after surgery.

Our results complement those of recent studies describing activity limitations for patients with THA or TKA. In a population-based survey, physical activities, self-care activities, and housekeeping tasks were impaired in 815 patients with THA or TKA (11, 12, 18). The SF-36 physical functioning score was lower than that for age-, sex-, and general practitioner–matched controls 8 years after surgery. In contrast, mental health did not decrease and instead improved with time (13, 14). In another study, 10 to 20 years after THA, scores for pain, physical mobility, emotional reactions, and social life were lower than those for control subjects (19). Our study confirmed that functional abilities are impaired several years after THA or TKA and that all other QOL dimensions are also impaired. The strength of our study lies in its use of generic and specific instruments as compared with the use of only generic questionnaires in previous studies. In addition, our results can be generalized to people who had THA or TKA but are not living in a nursing home.

Previous data showed that pain was ameliorated and physical function improved as early as 6 months after arthroplasty; scores improved slowly up to 1 year later, then reached a plateau at 3 years (20, 38). In addition, in another study, 1 year after THA, SF-36 scores reached reference population scores, but pain and physical function scores remained poor for patients with TKA (39). However, one study found that patients with THA and the general population did not differ in the evolution of functional ability (11, 12). We found no difference in patients' QOL at 1 and 3 years. QOL probably declined later and was particularly low 10 years after surgery in all of the dimensions, although physical function and role limitations were still low for the 3-year cohort.

Sociodemographic factors and the type of joint of the arthroplasty were not associated with QOL after surgery. In previous studies, with the exception of the joint (outcome is better with THA), sociodemographic and clinical characteristics were not consistently associated with better outcome (2, 5, 17, 38). High body mass index had a negative impact on functional status in one study (40) but not in others (17, 41). Older age was not associated with poor outcomes (41, 42).

This study highlights the importance of comorbidities in QOL. One study found a number of comorbid conditions not associated with functional status after arthroplasty (17), but in another study, impaired mental health, obesity, and the presence of at least one geriatric disease (visual impairment, falls, incontinence, impaired balance) was associated with low WOMAC function scores (40). These inconsistent results highlight the importance of using validated instruments to measure comorbidities. We used a validated questionnaire (the FCI) (35) specifically developed to explain and predict physical functioning. Moreover, other QOL dimensions were associated with comorbidities. Musculoskeletal pain in locations other than the TKA or THA location has a strong influence on QOL after arthroplasty. Hip or knee OA is often associated with other OA locations. Interestingly, we could differentiate the effect of THA or TKA regional pain and other painful locations. THA or TKA regional pain was prevalent but had no clear impact on QOL, whereas pain in other locations had negative consequences for QOL.

Environmental factors are well-known determinants of QOL in chronic diseases (43). In one study, patients with THA or TKA reported better housing accessibility and more assistive devices and helpers than did controls, but did not report more participation restrictions in terms of economic situation, housing, social relationships, and holidays (44). In our study, physical environment factors, but not marital status, were strongly associated with QOL after surgery. In a few other studies, social support was associated with THA or TKA outcome 1 or 2 years after surgery (2, 5, 17, 38). Social support may be more important just after surgery but not later; an instrument dedicated to social support is needed to capture this effect.

Preoperative QOL, clinical scores, and pain intensity predicted QOL for patients 3 years but not 10 years after THA or TKA. Short- and medium-term studies have consistently shown a low level of pain and good function before surgery associated with low pain and good function after surgery (2, 5, 17, 38). However, patients with poor baseline pain or physical function are more likely to experience greater improvement (14, 45, 46). The influence of preoperative QOL on long-term QOL is unknown. Our finding of no association of preoperative clinical status and QOL 10 years after surgery suggests that QOL over the long term is influenced more by general health state than by local disease.

Surgery is not successful in all cases. In previous studies, QOL or functional state was not satisfactory for 5–10% of patients with THA and 10–15% of patients with TKA (4, 40, 41). In one study, at 3.6 years after surgery, for 31% of patients, the WOMAC scores improved by less than 10 units on a 100-point scale (17). However, in our study, surgery success did not predict QOL at 3 or 10 years after arthroplasty. It would also be interesting in future studies to include data on postoperative complications due to surgery in the short and long term.

Our study has limitations. First, at inclusion, the 2 cohorts completed different QOL questionnaires. Indeed, both cohorts were recruited at different periods, and the availability of questionnaires and standards of use differed. The SF-36 is now the most commonly used generic QOL questionnaire, and reference scores for the general population in numerous countries allow for comparisons. Unfortunately, 10 years ago, this questionnaire was not adapted for French, and the NHP was the most popular generic questionnaire. In addition, no specific QOL questionnaire existed for OA, and the OAKHQOL was developed later (32). However, for both cohorts, QOL was measured by a generic and a specific questionnaire. Second, data for TKA and THA were combined to avoid a lack of power. Several studies concluded less and slower improvement after TKA than after THA (4, 47, 48), whereas other results claimed that the benefits of both were the same (2, 12, 49). Our analysis did not show any differences in QOL between procedures, although all of the statistical models were adjusted for the joint. Third, the followup of patients, particularly for the 10-year cohort, was incomplete, mainly because of death (20%) or migration (17%) of the patients. Some nonresponders probably lived in a nursing home because of decreased ability for self-care and maintenance of activities of daily living. However, in this case, QOL was probably more impaired, which would reinforce our results. Fourth, the sample size of the 10-year cohort was low (n = 89). With a power of 80% and an alpha risk of 0.05, at least 85 patients were needed for a correlation coefficient of 0.3 to be statistically significant. Finally, the 2 cohorts differed in characteristics and indications for arthroplasty have recently changed, with older patients and patients with less impaired QOL and a less severe Kellgren/Lawrence OA grade now frequently undergoing arthroplasty.

Our findings have several implications for patient care. Precise descriptions of disability after total arthroplasty should help surgeons better inform patients about the outcome of arthroplasty so that patients have realistic expectations after the procedure. This will help patients balance delaying surgery and make decisions about whether or not to undergo surgery. Moreover, information on determinants of QOL after THA or TKA might help satisfy patient needs. Sociodemographic characteristics should not prevent patients from undergoing surgery. The timing of the surgery is important and the procedure should not be unduly delayed. Addressing environmental factors and treating comorbidities and pain in locations other than the knee or hip could have a great impact on QOL.

In conclusion, after THA or TKA, impaired QOL persists over time despite substantial improvement in condition. Comorbidities, environmental factors, and the presence of painful locations other than the THA or TKA location are the main factors associated with postoperative QOL. Preoperative QOL is predictive of QOL at 3 years but not 10 years after surgery.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Rat had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Rat, Guillemin, Mainard, Baumann.

Acquisition of data. Rat, Osnowycz, Delagoutte, Cuny, Mainard.

Analysis and interpretation of data. Rat, Guillemin, Baumann.

Ancillary