Journal of Public Health Advance Access originally published online on June 29, 2006
Journal of Public Health 2006 28(3):267-273; doi:10.1093/pubmed/fdl020
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Pain and overall health status in older people with hip and knee replacement: a population perspective
Louise Linsell, Medical Statistician1
Jill Dawson, Senior Research Scientist2
Krina Zondervan, MRC Fellow and Epidemiologist3
Peter Rose, General Practitioner and University Lecturer4
Andrew Carr, Nuffield Professor of Orthopaedic Surgery5
Tony Randall, General Practitioner and Senior Research Fellow6
Ray Fitzpatrick, Professor7
1 Centre for Statistics in Medicine, Wolfson College Annexe, University of Oxford, Oxford OX2 6UD, UK
2 Division of Public Health and Primary Health Care, University of Oxford, Old Road, Oxford OX3 7LF, UK
3 Wellcome Trust Centre for Human Genetics, University of Oxford, UK
4 Division of Public Health and Primary Health Care, University of Oxford, Old Road, Oxford OX3 7LF, UK
5 Nuffield Department of Orthopaedic Surgery, University of Oxford, Nuffield Orthopaedic Centre, Oxford OX3 7LD, UK
6 Division of Public Health and Primary Health Care, Old Road, Oxford OX3 7LF, UK
7 Division of Public Health and Primary Health Care, University of Oxford, UK
Address correspondence to Jill Dawson, E-mail: jill.dawson{at}dphpc.ox.ac.uk
Objective To investigate the health-related quality of life and presence of hip or knee pain according to whether or not people had had previous hip or knee arthroplasty.
Study design and setting Cross-sectional survey representing randomly selected sample of 5500 elderly (65+) people. Pain prevalence rates obtained from standard screening questions. Standard pain severity ratings obtained for each hip and knee.
Results People with a past arthroplasty had worse health status compared to other people (p < 0.001 for all but two SF-36 dimensions). Hip or knee pain was more prevalent amongst people with past hip or knee replacement than amongst those without (62.5% versus 36.5% respectively; following adjustment for age and sex: Mantel-Haenszel combined odds ratio = 2.90, 95% CI 2.30–3.68, p < 0.001). More replaced knee joints were symptomatic than replaced hip joints (OR = 1.62, p = 0.022).
Conclusions Elderly people with a past hip or knee arthroplasty have significantly greater health and social care needs than other people — especially those related to pain and mobility. This may reflect the generalised nature of the underlying disease process.
Keywords: hip, knee, pain, prevalence, survey
 |
Introduction |
|---|
 TOP Introduction Material and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Hip and knee replacement surgery (arthroplasty) is widely performed, particularly for the treatment of severe osteoarthritis. The prevalence of arthritis increases with age, and an ageing population structure means that demand for these forms of surgery will continue to rise. Given that even the most successful arthroplasty surgery ultimately has a limited life, the outcomes of such surgery have major implications for health and social care planning and provision.1
Several studies in the UK have investigated the prevalence of hip or knee pain in the population, generally with the intention of defining population requirements for arthroplasty.2–4 There are nevertheless few population-based data regarding the association of overall health status with hip or knee arthroplasty. In addition, little is known of the extent to which hip and knee problems in elderly people might be directly related to having a past hip or knee arthroplasty. However, some evidence was presented from a French community-based study,5 which reported that people with a history of hip or knee replacement had significantly greater activity limitations than other people. This study was nevertheless unable to examine whether the limitations were directly related to the replaced joints as opposed to another (new) joint problem. They also made no comparison between hips and knees while noting that those with a knee arthroplasty could have been at higher risk of persistent problems relative to those with hip arthroplasty.5,6
Using data from a cross-sectional survey of elderly (>65 years) people, we aim to (i) compare the overall health status in people who have had a previous hip or knee arthroplasty with people who have never had a hip or knee replaced and (ii) examine the extent to which self-reported pain in a hip or knee is associated with having a past arthroplasty of that joint.
|
Material and methods |
|---|
|
TOPIntroductionMaterial and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Local research ethics committee approval was obtained for the study [Applied and Qualitative Research Ethics Committee (AQREC) reference A01.060].
A random sample of 5500 residents of one health authority in the southeast of England, aged 65 years and above, was obtained from the Health Authority register representing January 2002. A postal questionnaire and covering letter was sent out to everyone within a 2-week period during April 2002. Following two postal reminders (including a second copy of the questionnaire) and procedures to check living status and address, a response rate of 66.3% (3341 of 5039 eligible people) was obtained. Additional details of this study have been reported elsewhere.7
Questionnaire
The questionnaire was divided into a general section, a hip section and a knee section. The general section contained a few demographic items and the anglicized version of the SF-36 general health questionnaire.8 The SF-36 contains 36 items and is widely used as a generic health status instrument. It provides scores on eight dimensions: physical functioning, role limitations due to physical problems, bodily pain, social functioning, general mental health, role limitations due to emotional problems, energy/vitality and general health perceptions representing the last 4 weeks. Scores for each dimension range from 0 (poor health) to 100 (good health).
The hip section began with a screening question that has also been used in other studies.2,3,9 ‘During the past 12 months, have you had pain in or around either of your hips on most days for one month or longer?’ Items concerning which hip was symptomatic and details of any previous hip replacement surgery then followed, with standard response categories offered separately for the left and right hips. Where the respondent reported having a hip replaced, the next question was whether this had occurred during the previous 12 months, or longer ago. These questions were asked of all respondents. Patients reporting a symptomatic hip were also asked to rate pain severity in each hip during the last 4 weeks on a scale ranging from ‘none’ to ‘very severe’. Questions about knee in the knee section were identical to those asked about hips in the hip section (the word knee substituting the word hip). The order of the hip and knee sections was reversed in half of the questionnaires.
Statistical analysis
In the first stage of the analysis, we defined two comparison groups: people who had ever had a hip or knee joint replaced (arthroplasty group) and people with no history of replacement (non-arthroplasty group). We compared the two groups with respect to (i) socio-demographic characteristics, (ii) overall health status and (iii) prevalence of hip and knee pain.
We used Pearson’s 
2 test to examine differences in the socio-demographic characteristics. The overall health status of each group was assessed using the eight dimensions of the SF-36 and analysed using ordered logit estimation, a non-parametric method appropriate for ordinal response variables.10,11 The estimated odds ratios compared the SF-36 scores in the arthroplasty group with the non-arthroplasty group, so an odds ratio <1 implied worse state of health in the arthroplasty group. Estimates for each score were adjusted for age and sex, and analyses were run under the assumption of proportional odds, that is, the effect of each covariate is homogeneous across each ordinal category within each SF-36 dimension. Finally, we examined differences in the prevalence of current hip or knee pain using the Mantel–Henszel statistic, stratified by age and sex.
In the second stage of the analysis, we focussed on the arthroplasty group. The aims here were to (i) investigate the association of pain with replaced joints in more detail and (ii) make comparisons between hip and knee replacements. Examining these issues on a person level was problematic, because some individuals had more than one painful joint and/or more than one hip or knee replacement; so, we performed a joint-specific analysis, treating each hip and knee joint as the unit of analysis.
Using replaced joints (not persons) as the denominator, we calculated the prevalence of current symptoms in replaced hip joints versus replaced knee joints. The replaced joints were grouped according to whether they had been replaced during the previous 12 months versus longer ago to determine how much of the reported pain could probably represent post-operative pain from a recent replacement. Generalized estimating equations (GEEs) were used to estimate the odds ratio of replaced knee joints to hip joints for pain prevalence in (i) all replaced joints and (ii) joints replaced >12 months ago. GEE provides population-averaged estimates that take account of the lack of independence between measures originating from the same individual (i.e. two hip joints and two knee joints).12 The model assumed an equal-correlation structure between hip and knee joints and was adjusted for age and sex.
Joint-specific pain severity ratings were reported for symptomatic hip and knee replacements, once again using replaced joints (not persons) as the denominator. Individual joints were allocated to three levels of severity: a ‘mild category’ consisting of any joints where usual pain in the last 4 weeks had been rated as ‘none’ (people who reported a symptomatic hip or knee in the previous 12 months occasionally had no symptoms within the last 4 weeks.), ‘very mild’ or ‘mild’; a ‘moderate category’ comprising any joints that had been rated as ‘moderate’ and a ‘severe category’ containing any joints that had been rated as either ‘severe’ or ‘very severe’. All analyses were conducted using STATA 8.0.
|
Results |
|---|
|
TOPIntroductionMaterial and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Socio-demographic characteristics of arthroplasty and non-arthroplasty group
Amongst the 3341 respondents, we identified 331/3027 (10.9%) people with a hip or knee replacement and 2696/3027 (89.1%) people with no history of replacement; 314 people did not respond to the question about past joint replacement, hence the lesser denominator. Of the people reporting previous replacement, 210 (63.4%) had had a hip replacement only, 97 (29.3%) a knee replacement only and 24 (7.3%) both a hip and a knee replacement.
Table 1 shows the socio-demographic characteristics of respondents according to whether they had ever/never had a hip or knee replaced. We found a higher rate of replacement in the older age groups and also amongst women. Having a past hip or knee replacement was also associated with living alone, and, amongst females, there was an association between past hip or knee replacement and obesity.
|
Overall health status of the arthroplasty and non-arthroplasty group
Group scores for the eight dimensions of the SF-36 are summarized in Table 2. The scores are summarized as medians because of the skewed nature of the data, but means are also provided for comparison to other literature. People with previous replacement had consistently lower scores than people with no replacement, particularly in the three dimensions: physical function, role limitation (physical) and bodily pain. The differences remained for most of the dimensions after adjusting for age and sex; all the estimated odds ratios were <1, implying that the arthroplasty group had significantly worse health than non-arthroplasty group. These differences were significant in all dimensions except for emotional role limitation and mental health function.
|
Prevalence of hip and knee pain in the arthroplasty and non-arthroplasty group
A higher proportion of people reported having hip or knee pain in the arthroplasty group compared to the non-arthroplasty group (hip or knee pain/past hip or knee replacement: 207/331, 62.5%; hip or knee pain/no replacement: 983/2696, 36.5%). Following adjustment for age and sex, there was evidence that people with previous replacement were likely to be currently symptomatic compared to people with no previous replacement (Mantel–Haenszel combined odds ratio = 2.90, 95% CI 2.30–3.68, P < 0.00). Of the people with hip or knee pain and previous replacement, in 70/206 (34.0%) people, the source of the pain was reported as coming from a replaced joint only; in 66/206 (32.0%), it was from a non-replaced joint and in 70/206 (34.0%), pain was from both a replaced and a non-replaced joint.
Individual joint-level analysis
Amongst the people in the arthroplasty group, there were 301 replaced hip joints (originating from 163 unilateral and 69 bilateral replacements) and 167 replaced knee joints (from 73 unilateral and 47 bilateral replacements) for which the side of replacement was reported. Table 3 shows prevalence of pain in the total sample of 468 replaced joints. A considerable proportion of replaced hip and knee joints were currently symptomatic (47.5% knees versus 34.6% hips), particularly amongst joints replaced longer than 12 months ago, where pain was unlikely to represent post-operative pain or complications (44.0% knees versus 28.0% hips). A greater proportion of replaced knee joints were currently symptomatic compared with replaced hip joints (GEE odds ratio = 1.62, P = 0.022). This difference was more pronounced in joints replaced longer than 12 months ago (GEE odds ratio = 1.85, P = 0.009).
|
Table 4 summarizes hip and knee pain severity levels in replaced joints that were currently symptomatic. The pain experienced in most replaced joints in the last 4 weeks was rated as mild to moderate, although in a substantial proportion the pain was described as severe.
|
|
Discussion |
|---|
|
TOPIntroductionMaterial and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Main findings of this study
The overall health status of people with previous hip or knee joint replacement was found to be considerably worse than that of people with no replacement. The differences were more evident in the physical dimensions of health such as physical function, physical role limitation and bodily pain. However, there was little difference in psychological well-being; thus, the scores for mental health and role limitation (emotional) were fairly comparable. We also found evidence that people with previous hip or knee replacement were more likely to report hip or knee pain than people with no replacement.
On examining the source of the hip and knee pain amongst people with replacement, using a joint-level analysis, we discovered that a substantial proportion of replaced joints were currently symptomatic (48% knees versus 35% hips). The numbers reduced slightly when we excluded recently replaced joints, where symptoms were quite likely to be related to temporary post-operative factors, but the proportion remained high. Replaced knee joints had a consistently higher prevalence of reported pain compared to replaced hip joints, even following adjustment for the age and sex of the person.
What is already known on this topic
To date, the main source of evidence for outcomes following arthroplasty comes from prospective observational studies, which tend to either take the form of survivorship studies of particular types of joint prostheses (these studies may also include clinical assessments)13,14 or from joint registries (particularly from Scandinavia).15,16 The outcome measure that has most often been used in such studies is ‘the occurrence of revision surgery’. Outcomes of arthroplasty need to be monitored for many reasons,17 and they need to be assessed using appropriate measures. An appraisal of effectiveness that is based purely on revision surgery as a measure is limited—because many people will not undergo revision surgery (for a variety of reasons) even if their arthroplasty has become painful or dysfunctional. The proportion of people alive, fit or willing to proceed to revision surgery may well decrease as people age.
What this study adds
Hip and knee replacement surgery is widely performed, particularly on elderly people, and the outcomes of such surgery have major implications for health and social care provision; yet, few population-based data exist on which planning decisions for such provision might be based, and only few studies offer any substantive evidence on outcomes.5,18,19 We could find none from the UK.
This paper complements, but extends, evidence reported from a French community-based study5 which found that, following adjustment for age and sex, people with a history of hip or knee arthroplasty had significantly greater activity limitations than other people. This was particularly the case for activities that involved hip and knee function such as climbing stairs, walking and housekeeping activities. However, unlike this study, the French study could not reveal whether the limitations were directly related to the replaced joints as opposed to another (new) joint problem. They also combined information concerning people with past hip or knee replacements, whereas this study has been able to present data specific to each joint.
An additional strength of this study is that our cross-sectional survey of a random sample of older people obtained a very satisfactory response rate, and our study sample is broadly representative of people aged 65 years and over in the southeast of England,7 although we cannot claim that it represents the UK population as a whole.
Since the early 1990s, total hip and knee replacements have been considered similarly effective,20 but this has not always been the case. Our finding that a higher rate of symptoms affecting replaced knees by comparison with replaced hips could therefore be because of knee replacement procedures that had occurred many years ago when knee arthroplasty was less successful. Nevertheless, it is also possible that our study findings represent evidence that arthroplasty continues to be a somewhat less successful procedure for knees than for hips. Certainly, this subject warrants further research.
Limitations of this study
The importance of any conclusions that might be drawn from the individual hip/knee comparisons in this study are limited by the fact that we had minimal information on the length of time that individual prostheses had been in situ. A ‘rule of thumb’ exists that no more than 10% of hip or knee prostheses need to be revised within the first 10 years following implantation,21–23 which suggests that 90% of hip and knee replacements are ‘successful’ for at least 10 years with most lasting between 15 and 20 years. Given that all prostheses have a limited life (with aseptic prosthetic loosening the main cause of ‘failure’ and symptoms1), it is more likely that a proportion of the symptomatic replaced joints in our sample were reaching this limit. Nevertheless, it is not possible for us to put an exact interpretation on whether those replaced joints that were currently symptomatic represented ‘failed joints’ or simply ‘less than perfect results’. We also could not say whether joints reported as being painful were more or less painful than they were before the arthroplasty took place.
Finally, we emphasize that the findings presented in this paper do not challenge the accepted view that hip and knee arthroplasty has a major impact on pain and disability for people who most usually have end-stage Osteoarthritis (OA). The benefits of arthroplasty may be time limited, but they do generally last for many years, for all but few people.24 However successful an individual joint replacement might be, because OA tends to affect more than one joint,7 arthroplasty recipients will nevertheless invariably continue to have symptoms emanating from other joints, with attendant consequences for their quality of life.
|
Acknowledgements |
|---|
|
TOPIntroductionMaterial and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Financial support was provided by grant from the NHS Executive (Southeast Region).
|
References |
|---|
|
TOPIntroductionMaterial and methodsResultsDiscussionAcknowledgementsReferences |
|---|
- DawsonJ, Fitzpatrick R, Fletcher J, Wilson R. Osteoarthritis affecting the hip and knee. In: Andrew Stevens, James Raftery, Jonathon Mant, Sue Simpson (eds). Health Care Needs Assessment—the Epidemiologically Based Needs Assessment Reviews (First Series), Vol. 1. Ch 8, Oxford: Radcliffe Publishing, 2004.
- FrankelS, Eachus J, Pearson N et al. Population requirement for primary hip-replacement surgery: a cross-sectional study. Lancet 1999;353:1304–9.[CrossRef][ISI][Medline]
- JuniP, Dieppe P, Donovan J et al. Population requirement for primary knee replacement surgery: a cross-sectional study. Rheumatology 2003;42:526–1.
- TennantA, Fear J, Pickering A et al. Prevalence of knee problems in the population aged 55 years and over: identifying the need for knee arthroplasty. BMJ 1995;310:1291–3.
[Abstract/Free Full Text] - BoutronI, Poiraudeau S, Ravaud J-F et al. Disability in adults with hip and knee arthroplasty: a French national community based survey. Ann Rheum Dis 2003;62:748–54.
[Abstract/Free Full Text] - JinksC, Lewis M, Croft P. Health status after hip or knee arthroplasty. Ann Rheum Dis 2003;62:700–1.
[Free Full Text] - DawsonJ, Linsell L, Zondervan K et al. Epidemiology of hip and knee pain and its impact on overall health status in older adults. Rheumatology 2004;43:497–504.
[Abstract/Free Full Text] - BrazierJE, Harper R, Jones NM et al. Validating the SF-36 health survey questionnaire: new outcome measure for primary care. BMJ 1992;305:160–4.[ISI][Medline]
- O’ReillySC, Muir KR, Doherty M. Knee pain and disability in the Nottingham community: association with poor health status and psychological distress. Br J Rheumatol 1998;37:870–3.
[Abstract/Free Full Text] - McCullaghP.Regression models for ordinal data (with discussion). J R Statist Soc 1980;42:109–42.
- ZavoinaW, McKelvey RD. A statistical model for the analysis of ordinal level dependent variables. J Math Sociol 1975;4:103–20.
- ZegerSL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986;42:121–30.[CrossRef][ISI][Medline]
- RitterMA, Campbell ED. Long-term comparison of the Charnley, Muller, and Trapezoidal-28 total hip prostheses. A survival analysis. J Arthroplasty 1987;2:299–308.[Medline]
- GillGS, Mills DM. Long-term follow-up evaluation of 1000 consecutive cemented total knee arthroplasties. Clin Orthop 1991;273:66–76.
- HavelinLI, Engesaeter LB, Espehaug B et al. The Norwegian Arthroplasty Register: 11 years and 73,000 arthroplasties. Acta Orthop Scand 2000;71:337–453.[CrossRef][ISI][Medline]
- PaavolainenP, Pukkala E, Pulkkinen P, Visuri T. Causes of death after total hip arthroplasty: a nationwide cohort study with 24,638 patients. J Arthroplasty 2002;17:274–81.[CrossRef][ISI][Medline]
- SwiontkowskiMF.Why we should collect outcomes data. J Bone Joint Surg Am 2002;85:S14–S15.
- HawkerG, Wright J, Coyte P et al. Health-related quality of life after knee replacement. J Bone Joint Surg Am 1998;80:163–73.
[Abstract/Free Full Text] - MohamedNN, Barrett JA, Katz JN et al. Rates and outcomes of primary and revision total hip replacement in the United States Medicare population. J Bone Joint Surg Am 2003;85:27–32.
[Abstract/Free Full Text] - MoranCG, Horton TC. Total knee replacement: the joint of the decade. BMJ 2000;320:820.
- NIH Consensus Development Panel. Total hip replacement. NIH Consensus Development Panel on total hip replacement. JAMA 1995;273:1950–6.[Abstract]
- NobleJ, Hilton RC. Total knee replacement [editorial]. BMJ 1991;303:262.
- ScuderiGR, Insall JN, Windsor RE, Moran MC. Survivorship of cemented knee replacements. J Bone Joint Surg Br1989;71:798–803.
- FitzpatrickR, Shortall E, Sculpher M et al. Primary total hip replacement surgery: a systematic review of outcomes and modelling of cost-effectiveness associated with different prostheses. Health Technol Assess 1998;2:1–64.[Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||