knee osteoarthritis

The Relationship Between Prospectively Assessed Body Weight and Physical Activity and Prevalence of Radiological Knee Osteoarthritis in Postmenopausal Women

CASSANDRA SZOEKE, LORRAINE DENNERSTEIN, JANET GUTHRIE, MARGARET CLARK, and FLAVIA CICUTTINI

ABSTRACT.

Objective. To determine the effect of weight and physical activity on the prevalence of radiological knee osteoarthritis (OA) in a cohort of middle-aged women.

Methods. The longitudinal phase of the Melbourne Women's Mid-life Health Project is a population-based prospective study of 438 Australian-born women who have been followed annually over 11 years. Of these women, 257 (59%) remained in longitudinal assessment at eleventh year of followup, and 224 of these women agreed to undergo radiographs of their knees. Radiographs were scored for features of OA [osteophytes and joint space narrowing (JSN)] using a validated scale, by 2 investigators who were blinded to questionnaire results. Data were obtained by both self-administered and face-to-face interview questionnaires.

Results. The average weight increase over the study period of 11 years was 4 kg (range –14 to 25 kg). Of the 224 women evaluated, 65 (29%) had knee joint osteophytes and 95 (42%) had evidence of knee JSN. Current weight and weight at baseline were independent factors associated with a higher prevalence of both osteophytes and JSN in all compartments of the knee. The average amount of physical activity over the 11 years of followup was a significant factor independently associated with an increased prevalence of patellofemoral JSN and approached significance for tibiofemoral osteophytes and total knee JSN.

Conclusion. Our study supports a longterm detrimental effect of weight on the knee joint and suggests the importance of longterm weight maintenance programs in preventing knee OA. The average amount of physical activity was associated with an increased prevalence of some features of knee OA. (First Release Aug 1 2006; J Rheumatol 2006;33:1835–40)

Key Indexing Terms:

RADIOLOGICAL
OSTEOARTHRITIS
MENOPAUSE 
BODY WEIGHT
PHYSICAL ACTIVITY

---

From the Office for Gender and Health, Department of Psychiatry, University of Melbourne; Department of Medicine, The Royal Melbourne Hospital; Department of Epidemiology and Preventive Medicine, Monash University, Melbourne; and the Department of Rheumatology, Alfred Hospital, Prahran, Victoria, Australia.

Radiographs were funded by a grant from the Shepherd Foundation. Dr. Szoeke has received research funding from the Arthritis Foundation of Australia, the University of Melbourne (Viola Edith Scholarship and JA Thompson Prize), and the Royal Australian College of Physicians (Tweedle Fellowship). Data entry and analysis was supported by a grant from the Australian Menopausal Society. The Melbourne Women's Midlife Health Project baseline data collection was funded by the Victorian Health Promotion Foundation.

C.E.I. Szoeke, MBBS, Bsc (Hons), Office for Gender and Health, Department of Psychiatry, University of Melbourne and Department of Medicine, Royal Melbourne Hospital; L. Dennerstein, MBBS, FRANZCP, PhD, Professor; J.R. Guthrie, PhD; M.S. Clark, PhD, Office for Gender and Health, Department of Psychiatry, University of Melbourne; F.M. Cicuttini, MBBS, FRACP, PhD, Associate Professor, Department of Epidemiology and Preventive Medicine, Monash University and Department of Rheumatology, Alfred Hospital.

Address reprint requests to Dr. C.E.I. Szoeke, P.O. Box 2026, The Royal Melbourne Hospital, Parkville, 3050, Victoria, Australia. E-mail: Cassandra.Szoeke@mh.org.au

Accepted for publication May 9, 2006.

---

Osteoarthritis (OA) is the most common musculoskeletal disease, affecting more than one-quarter of the population over 60 years of age^1-3. Pain and limitation of function caused by the symptoms of OA affect many aspects of an individual's health and quality of life⁴. Given that the treatment of OA is largely symptomatic, prevention is likely to play a large part in reducing the prevalence of disease.

Obesity is a major international epidemic⁵. Given that women at midlife have an increased tendency for gaining weight^6,7, the investigation of the relationship between the development of OA and weight is particularly important in this population. Obesity is a strong risk factor for the development of OA and also plays a significant role in worsening the disability of established OA^8-11. The magnitude of the weight increase that significantly increases the risk of OA may be small. In a twin study, twins with OA were on average 3–5 kg heavier than their co-twins¹². If such subtle changes in weight can affect the development of OA, this has significant clinical relevance, as small weight loss may be more easily achieved than dramatic weight loss. A twin study examining the association of body mass index (BMI) and knee OA implied that environmental factors are greater in influence than genetics¹³. Identification of such modifiable risk factors in the development of this disease, which has limited treatments, is of enormous clinical importance.

A number of studies have suggested that intensive physical activity increased the risk of knee OA^14,15. However, it is still unclear whether there is a different effect of physical activity on radiological OA, where the definition depends on osteophytes rather than joint space narrowing (JSN)¹⁶. This is important, since osteophytes may be a consequence of forces applied to joints, but it is JSN that is a surrogate marker of knee cartilage. There is some emerging evidence for a beneficial effect of physical activity on joint cartilage¹⁷.

We examined a cohort of healthy middle-aged women to determine whether weight and physical activity over 11 years followup were independently associated with final knee OA.

MATERIALS AND METHODS

The Melbourne Women's Mid-life Health Project is a population-based prospective study of Australian-born women. The study began in 1991 (baseline) with the use of random digit dialing to interview 2,001 Australian-born women between 45–55 years old and residing in Melbourne. The response rate was 71%. Of these women, 779 were eligible for longitudinal assessment (they had menses in the prior 3 months and were not taking oral contraceptives or hormone therapy)¹⁸, of whom 438 (56%) were recruited for longitudinal assessment. Two hundred fifty-seven (59%) of the longitudinal cohort attended the office for the eleventh year of followup. Of these women, 224 agreed to undergo radiographs of their hands and knees in the following year.

Compared to participants who presented for radiographs (n = 224), the remaining women who entered the longitudinal study (n = 214) were significantly more likely to be older and to have smoked more at baseline. Ethics approval was obtained from the Melbourne Health Research Directorate, the University of Melbourne Human Research Ethics Committee, and Monash University Ethics Committee.

The height and weight of participants were measured by field workers and BMI calculated. Average weight over 11 years of followup was calculated by summation of weight in each year, then divided by the number of years. Physical activity was assessed by the question "How often do you participate in physical activities or sports for fitness or recreational purposes?" with a scale from 0 – 7 (every day, 4 – 6 times a week, 2 – 3 times a week, once a week, a few times a month, less than once a month, never). Data was transformed with frequency described by 4 categories: "Daily" physical activity was defined as activity reported every day or 4 – 6 times per week, "weekly" as once to 2 – 3 times per week, "monthly" as less than once to a few times per month, and "never" as never does sport¹⁹. This form of assessing physical activity was validated as comparable to the longer "Minnesota leisure-time physical activity" questionnaire²⁰. Average physical activity was calculated by summing the score from each year and dividing by the number of years of followup.

Both knees were radiographed in each participant. Radiographs were conducted of the knees, both in a weight-bearing anteroposterior view in full extension and in skyline view in 45° flexion using a perspex positioning wedge. Patellofemoral joint disease was based on the radiological findings on the skyline view²¹.

The tibiofemoral and patellofemoral joint compartments of the knee were graded on a 4-point scale (0 – 3) for both JSN and osteophytes and scored using a standard atlas²¹, by 2 investigators who were blinded to questionnaire results. The variable called "total knee osteophytes" was defined as an osteophyte present at any one of the knee compartments. Total knee narrowing variable was defined as JSN at any one of the knee compartments.

Characteristics for the study population (n = 224) at baseline and eleventh year of followup were compared using independent t-test or chi-squared analysis. Variables selected for the analysis were identified as important by review of the literature. Logistical regression was used to determine the effect of weight and physical activity over the study period on the prevalence of OA. Variables selected for this multivariate analysis were those found to be significant in the bivariate analysis and not highly intercorrelated. Statistical package SPSS 13.1 was used for all analyses.

RESULTS

Questionnaires were completed by 257 participants in the eleventh year of followup and 224 women underwent radiographic assessment.

Table 1 shows the characteristics of the study population at first year of longitudinal followup and eleventh year of followup. At baseline the mean age of participants was 49.7 years (SD 2.5). At the eleventh followup year, participants' mean age was 60.0 years (SD 2.5) and they were heavier (72.7 kg vs 68.6 kg) with an average weight gain of 4 kg (range –14 to 25 kg). For the 159 participants who gained weight over the period, the mean weight gained was 6.25 kg (range 0.5 to 25 kg). Thirty-six women recorded a weight loss over the study period with average weight loss of 4.3 kg (range 0.5 – 14 kg). The mean BMI of the cohort increased from 25.9 (SD 4.8) to 27.7 (SD 5.5).

Table 1. Characteristics of the participants (n = 224) at baseline and at 11th year of followup.

At baseline, 56 (25%) participants reported daily physical activity and 46 (20%) reported doing no physical activity. In 2002, 89 (40%) participants reported daily physical activity and 29 (13%) no activity. Compared to baseline, in 2002 participants were more likely to report exercising frequently (p = 0.005). When radiographed, 65 (29%) participants had knee joint osteophytes and 95 (42%) had evidence of knee JSN (Figure 1). With respect to radiological assessment of OA, there was good agreement of results between observers. Correlation coefficient was 0.992 (p = 0.01) and 0.968 (p = 0.01) for intraobserver and interobserver knee scoring, respectively.

[click, then close, image]

Figure 1. The prevalence of knee osteoarthritis in this cohort of women. Results are reported for various compartments and features of knee OA. No. of patients = 224.

Weight at baseline was a predictor of both osteophytes and narrowing in all compartments of the knee in 2003, independent of current age, current level of physical activity, hormone therapy (HT) use, and smoking (Table 2). Current weight at eleventh year of followup was associated with osteophytes in both tibiofemoral and patellofemoral compartments and the total knee osteophyte score, as well as narrowing in the tibiofemoral but not the patellofemoral joint. This relationship persisted after adjusting for age, current level of physical activity, HT use, and smoking (Table 2) for all variables except tibiofemoral osteophytes.

Table 2. Relationship between radiological knee OA and weight at baseline and 11th year of followup.

Simple multivariate analysis, adjusting for age, showed that average weight (defined as the summation of all weights recorded over the study period divided by the number of years of followup) was associated with osteophytes at all sites and JSN at all sites except the patellofemoral joint. Total knee osteophytes showed association with average weight adjusting for current level of physical activity, HT use, and smoking (Table 3).

Table 3. Relationship between radiological knee OA and average weight over study period.

The mean amount of exercise performed by the individual over the study period was associated with tibiofemoral osteophytes and patellofemoral JSN in simple multivariate analysis and persisted after adjusting for BMI, HT use, current level of physical activity, and smoking status for patellofemoral narrowing (Table 4).

Table 4. Relationship between radiological knee OA and average physical activity over 11 years.

DISCUSSION

Our study supports the role of weight as an important factor associated with osteophytes and JSN in all compartments of the knee. We showed that the weight of women at baseline was associated with the development of OA 11 years later. Our data also suggest that the average amount of physical activity is associated with some features of knee OA.

Few studies have looked at the influence of weight over time associated with subsequent development of OA. Our findings are consistent with similar findings in men²², where weight in youth was predictive of latent knee OA. We also found that the average weight carried over 11 years was predictive of OA after adjusting for HT use, current physical activity, and smoking status. As radiological OA was assessed at one timepoint only, it is possible that the increase in weight may have been due to the presence of knee OA that reduced activity due to pain and limited movement. However, our community-based subjects had mild radiological OA and when we examined those with no current pain, the results were similar (data not shown). In addition, results were adjusted for levels of physical activity.

There is evidence that small changes in body weight may significantly increase the risk of OA¹². Our study suggests that both current weight and past weight are important factors associated with knee OA. This suggests that programs to maintain longterm healthy weight are likely to play an important role in reducing the risk of knee OA, particularly in light of the evidence that environmental modification of BMI can influence OA¹³.

Average physical activity over the 11 years was independently associated with a higher prevalence of tibiofemoral osteophytes and patellofemoral narrowing adjusting for age, although results remained significant only for patellofemoral narrowing after adjusting for HT use, current physical activity, current BMI, and smoking status. There was a trend towards an association with tibiofemoral osteophytes (p = 0.08).

Previous studies on the influence of physical activity on OA have been conflicting. Karlson, et al found that increased physical activity was not associated with the need for hip replacement due to OA¹⁰. In contrast, the Framingham Study showed a 3 times increased risk of knee OA for those in the highest quartile of physical activity²³. In our study there was a trend for more frequent physical activity in women at first year of followup to be associated with tibiofemoral osteophytes and not with JSN. Osteophytes and joint space width represent different aspects of the pathogenesis on knee OA^24,25. Osteophytes may reflect traction forces across joints¹⁶, while joint space width reflects the status of articular cartilage. This may explain the different effect of physical activity on knee OA. It may also explain some of the conflicting findings in the literature, since in the past there has been an overreliance on assessment of osteophytes in defining OA. We also found that physical activity was associated with increased prevalence of JSN, but not osteophytes, at the patellofemoral joint. It may be that physical activity, as measured in our study, results in different biochemical effects at the tibiofemoral and patellofemoral joints that may explain these differences.

A potential limitation of this study is that only 57% of the initial participants had complete data over 11 years of followup. Examination of baseline characteristics between participants and nonparticipants shows those who remained in followup were younger and fewer were smoking at baseline, but they were not significantly different with respect to their weight or reported physical activity, the major risk factors examined in this study. A major limitation is the measurement of physical activity, which asked about frequency of activity or sports for fitness or recreation. This question was chosen at the start of the study as being the most feasible means of obtaining information from a large population sample from whom a large amount of information was to be obtained. Although this question has been shown to be comparable in terms of measuring time spent exercising with a more complex questionnaire²⁰ it does not provide information on the type or intensity of the exercise. In particular as regards our study, it does not provide information on particular joint involvement or on the weight-bearing character of the activities. However, previous data from this cohort²⁶ have shown that the majority of the physical activities participated in by these women involved walking or gardening –– both of which have considerable involvement with the knees.

We do not have data on incident OA, since radiographs were not done at baseline. However, the strength of our study is that we prospectively collected data on weight and physical activity, the 2 risk factors of interest, over 11 years. We also used an objective, validated method for determining the presence and absence of knee OA based on radiographs.

Our findings confirm that weight is an important risk factor for OA in middle-aged women and suggest that the average amount of physical activity is associated with an increased prevalence of some features of knee OA. We found that past weight (11 years ago) and the average weight over the 11 years of followup were both independently associated with a higher prevalence of knee OA. This suggests that weight has a detrimental, longterm effect on joints and highlights the importance of longterm weight maintenance programs in preventing knee OA. Further work is needed to clarify the effect of physical activity on knee health.

REFERENCES

Search PubMed for:

1. Leveille SG. Musculoskeletal aging. Curr Opin Rheumatol 2004;16:114-8. [MEDLINE]

2. WHO. The burden of musculoskeletal conditions at the start of the new millennium. World Health Organ Tech Rep Ser 2003;919:i-x, 1-218, back cover.

3. Reginster JY. The prevalence and burden of arthritis. Rheumatology Oxford 2002;41 Supp 1:3-6.

4. Dias RC, Dias JM, Ramos LR. Impact of an exercise and walking protocol on quality of life for elderly people with OA of the knee. Physiother Res Int 2003;8:121-30. [MEDLINE]

5. WHO. Obesity: preventing and managing the global epidemic. Report of a WHO consultation. World Health Organ Tech Rep Ser 2000;894:i-xii, 1-253.

6. Dubnov G, Brzezinski A, Berry EM. Weight control and the management of obesity after menopause: the role of physical activity. Maturitas 2003;44:89-101. [MEDLINE]

7. Guthrie JR, Dennerstein L, Dudley EC. Weight gain and the menopause: a 5-year prospective study. Climacteric 1999;2:205-11.[MEDLINE]

8. Miller GD, Rejeski WJ, Williamson JD, et al. The Arthritis, Diet and Activity Promotion Trial (ADAPT): design, rationale, and baseline results. Control Clin Trials 2003;24:462-80. [MEDLINE]

9. Yoshimura N, Nishioka S, Kinoshita H, et al. Risk factors for knee osteoarthritis in Japanese women: heavy weight, previous joint injuries, and occupational activities. J Rheumatol 2004;31:157-62.[MEDLINE]

10. Karlson EW, Mandl LA, Aweh GN, et al. Total hip replacement due to osteoarthritis: the importance of age, obesity, and other modifiable risk factors. Am J Med 2003;114:93-8. [MEDLINE]

11. Felson DT, Zhang Y, Anthony JM, et al. Weight loss reduces the risk for symptomatic knee osteoarthritis in women. The Framingham Study. Ann Intern Med 1992;116:535-9. [MEDLINE]

12. Cicuttini FM, Baker JR, Spector TD. The association of obesity with osteoarthritis of the hand and knee in women: a twin study. J Rheumatol 1996;23:1221-6. [MEDLINE]

13. Manek N, Hart D, Spector TD, MacGregor AJ. The association of body mass index and osteoarthritis of the knee joint. Arthritis Rheum 2003;48:1024-9. [MEDLINE]

14. McAlindon TE, Wilson PW, Aliabadi P, et al. Level of physical activity and the risk of radiographic and symptomatic knee osteoarthritis in the elderly: the Framingham study. Am J Med 1999;106:151-7. [MEDLINE]

15. Garrick JG, Requa RK. Sports and fitness activities: the negative consequences. J Am Acad Orthop Surg 2003;11:439-43. [MEDLINE]

16. Hannan MT, Felson DT, Anderson JJ, Naimark A. Habitual physical activity is not associated with knee osteoarthritis: the Framingham Study. J Rheumatol 1993;20:704-9. [MEDLINE]

17. Jones G, Glisson M, Hynes K, Cicuttini F. Gender differences in cartilage development: a possible explanation for variations in knee osteoarthritis in later life. Arthritis Rheum 2000;43:2543-9.[MEDLINE]

18. Dennerstein L, Dudley EC, Hopper JL, et al. A prospective population-based study of menopausal symptoms. Obstet Gynecol 2000;96:351-8. [MEDLINE]

19. Guthrie JR. Physical activity: measurement in mid-life women. Acta Obstet Gynecol Scand 2002;81:595-602. [MEDLINE]

20. Guthrie JR. Physical activity: measurement in mid-life women. Acta Obstet Gynecol Scand 2002;81:595-602. [MEDLINE]

21. Altman RD, Hochberg M, Murphy WA Jr, et al. Atlas of individual radiographic features in osteoarthritis. Osteoarthritis Cartilage 1995;3 Suppl A:3-70.

22. Gelber AC, Hochberg MC, Mead LA, et al. Body mass index in young men and the risk of subsequent knee and hip osteoarthritis. Am J Med 1999;107:542-8. [MEDLINE]

23. Felson DT, Zhang Y, Hannan MT, et al. Risk factors for incident radiographic knee osteoarthritis in the elderly: the Framingham Study. Arthritis Rheum 1997;40:728-33. [MEDLINE]

24. Neame R, Zhang W, Deighton C, et al. Distribution of radiographic osteoarthritis between the right and left hands, hips, and knees. Arthritis Rheum 2004;50:1487-94. [MEDLINE]

25. Cicuttini FM, Spector T, Baker J. Risk factors for osteoarthritis in the tibiofemoral and patellofemoral joints of the knee. J Rheumatol 1997;24:1164-7. [MEDLINE]

26. Guthrie JR, Smith AM, Dennerstein L, Morse C. Physical activity and the menopause experience: a cross-sectional study. Maturitas 1994;20:71-80.[MEDLINE]

PSORIATIC ARTHRITIS

Therapies for Axial Disease in Psoriatic Arthritis. A Systematic Review
PETER NASH

ABSTRACT.

Prevalence rates for axial involvement in psoriatic arthritis (PsA) vary from 40% to 74% depending upon criteria for diagnosis. In the absence of trial evidence to assess axial involvement in PsA, the GRAPPA group, by consensus, has suggested that outcome measures and therapies for axial disease in ankylosing spondylitis (AS) be used. This systematic review addresses the management of axial disease in PsA, and provides treatment recommendations based on the AS literature. (First Release May 15 2006; J Rheumatol 2006;33:1431–4)

Key Indexing Terms:

ANKYLOSING SPONDYLITIS
PSORIATIC ARTHRITIS
AXIAL DISEASE 
SPONDYLOARTHROPATHY
PSORIASIS
ANTIRHEUMATIC THERAPY

---

From the Department of Medicine, University of Queensland, Cotton Tree, Australia.

P. Nash, MBBS (Hons), FRACP.

Address reprint requests to Dr. P. Nash, Department of Medicine, University of Queensland, PO Box 59, Cotton Tree, Australia. E-mail: pnash@tpg.com.au

---

INTRODUCTION

Axial disease in psoriatic arthritis (axPsA) is common; prevalence rates range from 40% of PsA patients if inflammatory spinal pain is used as the diagnostic criteria, to 78% if radiologically defined sacroiliitis is required¹. Despite shared features between axial disease in PsA and ankylosing spondylitis (AS), important distinctions have been described:

• Reduced male preponderance

• Less overall disease severity

• Less severe sacroiliitis

• Less syndesmophyte development (e.g., "spotty" asymmetric distribution of marginal and paramarginal syndesmophytes with random progression)

• Less cervical involvement

• Better preservation of spinal mobility

• Relative absence of ligamentous ossification

• Relative sparing of apophyseal joints

• Reduced association with HLA-B27 positivity

(prevalence rates from 21% to 76% with a strong association when positive with sacroiliitis)^2-4

These observations, however, are based upon studies of small patient populations either assessed retrospectively or followed prospectively for relatively short periods^2-4.

Outcome Measures

Outcome measures used in the assessment of AS have not been validated in PsA, and assumptions of equivalent utility may be false. For example, the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) scores were significantly lower in PsA than in AS and correlated poorly with external indicators of disease activity⁵. In addition, spinal assessments have shown poor sensitivity and specificity for sacroiliac joint measures, including Schober's test and its modifications, chest expansion, cervical range of movement, and sacroiliac joint tenderness. Sensitivity to change and minimal clinically significant differences have not been established⁶.

Outcome measures in axPsA are under active assessment; however, the Assessments in Ankylosing Spondylitis (ASAS) Working Group has developed response criteria for improvement in AS, and by consensus, these will be used in this review⁷. ASAS has validated composite measures measuring disease activity (BASDAI), function (Bath AS Functional Index, BASFI), patient global assessment (Bath AS Global Index, BASGI), and spinal mobility (Bath AS Metrology Index, BASMI). Structural damage is evaluated using validated radiographic instruments (Bath AS Radiology Index/Stoke AS Spine Score, BASRI/SASSS)⁷.

Search Strategy

A literature search of Medline, EMBASE, CINHAL, and the Cochrane Library from 1966 to 2005 was done using the key MeSH terms ankylosing spondylitis, psoriatic arthritis, axial disease, spondyloarthropathy (SpA), psoriasis, spondylitis, and antirheumatic therapy, alone and in combination, and of the following terms with PsA, SpA, or spondylitis: salazopyrine, methotrexate, physiotherapy, pamidronate, gold oral and intramuscular, azathioprine, cyclosporine, hydroxychloroquine, infliximab, etanercept, adalimumab, and tumor necrosis factor (TNF).

The resulting database was culled for articles specifically addressing axial disease and its therapy in psoriasis, PsA, AS, or SpA. Abstracts from annual meetings of the American College of Rheumatology and the European Congress of Rheumatology were admitted as evidence if sufficient detail was available to determine level of evidence.

RESULTS

Physiotherapy — Level of Evidence 1a, Grade A

In a review of 6 trials of physiotherapy in AS (561 participants), supervised group physical therapy significantly improved global health and functioning, pain, and stiffness compared with home-based individual exercise⁸. In 2 trials, individualized home exercise programs had greater effects on spinal mobility and physical function, compared with no intervention. In 3 trials, supervised group physiotherapy produced better patient global assessments, compared with home-exercise programs. In another study, 3 weeks of inpatient spa-exercise therapy followed by 37 weeks of weekly outpatient group physiotherapy (without spa) showed evidence for effects on pain, physical function, and patient global assessment, compared with outpatient group physiotherapy alone. An additional trial of a home-based exercise and disease education package showed benefit for spinal mobility and function in AS⁹.

Nonsteroidal Antiinflammatory Drugs

Nonsteroidal antiinflammatory drugs (NSAID) appear to be rapidly efficacious in relieving signs and symptoms of pain and morning stiffness (level of evidence 1a, grade A) in axial disease. Continuous NSAID usage can reduce radiological progression (level of evidence 1b, grade A), although separate treatments for inflammation and progression of structural damage have been suggested¹⁰.

Coxibs may have fewer gastrointestinal side effects, but cardiovascular issues may nullify that benefit. Equivalent efficacy has been demonstrated for one coxib¹¹, and superior efficacy over 12 months for etoricoxib over a nonselective NSAID (naproxen) has been shown (level of evidence 1b, grade A)¹². Two NSAID at maximum recommended/tolerated and continuous doses are recommended before concluding that a patient is NSAID refractory (level of evidence 4, grade C).

Simple Analgesia

No disease-specific evidence supports paracetamol or other analgesics in axial disease of AS or PsA.

Corticosteroids

Significant benefit has been noted with intraarticular glucocorticoids for active peripheral arthritis (level of evidence 4, grade C) and given under fluoroscopic or computer tomographic guidance into the sacroiliac joints (level of evidence 1b, grade A). A dose-response, double-blind comparison of 1 g versus 375 mg of methylprednisolone given as 3 consecutive daily intravenous (IV) infusions demonstrated no significant differences¹³. Open studies evaluating pulse IV methylprednisolone 1 g for 3 consecutive days have demonstrated prompt improvement lasting 3–21 months (level of evidence 2b, grade B)¹⁴.

Disease Modifying Antirheumatic Drugs

Sulfasalazine (SSZ). Sulfasalazine has been studied in PsA, but only small percentages of patients had axial involvement, and responses were poorly documented. In a review of 12 controlled trials of SSZ, improvement in erythrocyte sedimentation rate (ESR) and morning stiffness favored SSZ over placebo (level of evidence 1a, grade A)¹⁵. In the largest and longest of these trials (221 patients, 36-week course)¹⁶, improvement was significant (p = 0.01) but modest, with 55% and 45% improvements in the PsA response criteria for SSZ and placebo, respectively. The action of SSZ appears to be confined to peripheral arthritis with no evidence of benefit in axial disease (level of evidence 1a, grade A).

A single trial evaluated SSZ in primary outcome analyses of back pain, chest expansion, occiput-to-wall test, and patients' general well-being. Compared with other trials, patients in this study had more peripheral arthritis and the highest level of baseline ESR, but the shortest disease duration. Withdrawals for side effects with SSZ were significant. Spondylitis was not improved, and radiological progression was not retarded with SSZ¹⁶.

Oral formulations of 5-ASA (pentasa and asacol) have been examined in open trials in AS without benefit¹⁷.

Methotrexate (MTX). The efficacy of MTX in PsA was first described in 1964, but benefit has not been demonstrated in axial disease (level of evidence 1a, grade A). A 3 month, double-blind, placebo-controlled study of parenteral MTX in 21 patients who had active skin disease and peripheral arthritis showed significant improvement in joint tenderness and range of motion, extent of skin involvement, and ESR¹⁸.

In a randomized, double-blind, placebo-controlled trial comparing oral low dose pulse MTX with placebo, the only significant response was the physician assessment of arthritis activity¹⁹. In a 24-month study of 38 patients, MTX did not show improvement in radiographic progression compared with matched controls²⁰.

In a randomized, double-blind placebo-controlled trial of 70 patients with AS, MTX showed significant benefit over 24 weeks in physical well-being, BASDAI, BASFI, physician global assessment, patient global assessment, spinal pain, and Health Assessment Questionnaire for Spondyloarthropathies (HAQ-S) (level of evidence 1b, grade A)²¹.

The toxicity, particularly hepatotoxicity, of MTX in psoriasis and PsA has been compared with that in RA. A retrospective study of 104 patients followed over 2 decades did not suggest increased toxicity²². However, analysis is difficult: studies tend to be retrospective, with no controlling for alcohol consumption; older studies include daily therapy; and many do not include folic acid supplementation. No consensus exists as to the indications for liver biopsy either before treatment or at specified intervals during treatment.

Gold salts and antimalarial drugs. Anecdotal case reports have shown no benefit for gold salts and antimalarial drugs in axPsA^23,24.

Azathioprine. Azathioprine was associated with a low response rate and a high withdrawal rate due to adverse effects²⁵.

Thalidomide (level of evidence 3, grade B). A one-year, open-label trial of thalidomide was conducted in 30 men with treatment-refractory AS. Of 26 patients who completed the study, 80% showed > 20% improvement in 3–6 months; 9 patients became pain-free. Peripheral neuropathy and teratogenicity remain major limitations²⁶.

Leflunomide (level of evidence 1b, grade A for non-benefit). A randomized, placebo-controlled, 6-month trial of leflunomide in 45 patients with AS showed no significant benefit, and axial disease was not assessed²⁷.

Cyclosporine (level of evidence 3, grade B). A review of 16 studies of cyclosporine therapy in PsA showed improvements in rash and peripheral arthritis, without evaluating axial disease²⁸. The addition of cyclosporine to MTX nonresponders significantly improved peripheral joint scores, but axial disease was not evaluated²⁹.

Bisphosphonates (level of evidence 1b, grade A). One randomized, controlled, single-center study and 2 open analyses examined IV pamidronate in NSAID-refractory AS patients. In a 6-month, double-blind, dose-response comparison, 60 mg had significant benefit over 10 mg IV pamidronate; 40% of patients had a 50% BASDAI reduction, but no significant reduction in peripheral pain³⁰.

Biological Therapies

Interleukin 1 receptor antagonist anakinra (level of evidence 3, grade B). An open-label, 3-month pilot study of anakinra in 9 patients showed significant improvements in ASAS 20% response (BASFI, BASDAI, AS Quality of Life, C-reactive protein, and ESR), and significant improvement in enthesitis and osteitis lesions as measured by magnetic resonance imaging (MRI)³¹.

Anti-TNF therapy. Trial evidence shows etanercept, infliximab, and adalimumab provide significant improvement in disease activity, range of motion, physical function, and quality of life, both as monotherapy and as add-on therapy to other disease modifying antirheumatic drugs in both psoriasis and PsA^32-40. All 3 agents retard radiologic progression in PsA. Axial disease was not studied, as too few patients had spondylitis. In AS trials, few patients had psoriasis (5–15%), and ASAS 20 responses were inferior in this subset.

Etanercept trials in PsA and AS showed improvements in symptoms and signs of disease, mobility, physical function, and acute phase reactants, with complete resolution or improvement noted in 86% of lesions, documented by MRI, in the AS trial (level of evidence 1a, grade A)^32-34.

Infliximab trials in PsA have shown significant response in NSAID-refractory patients (those with long-standing, severe disease), and benefit was maintained up to 24 months with continued therapy (level of evidence 1a, grade A)^35-37.

Adalimumab therapy in 315 PsA patients treated for 6 months showed significant improvements in joint and skin manifestations, improved quality of life, and retardation of radiological progression (level of evidence 1b, grade A). Only one patient in this study, however, had axial disease^38-40.

TREATMENT RECOMMENDATIONS AND CONCLUSIONS

Validated outcome measures and formal trials in adequate numbers of patients with axPsA are urgently needed to address the safety and efficacy of treatments for axPsA. At present, the ASAS specifications and definitions for diagnosis, assessment of disease, treatment failure, treatment contraindications, and assessment of response are recommended and are summarized as follows:

Treatment Algorithm^41,42

1. Establish diagnosis of PsA.

2. Initiate therapy with physiotherapy and continuous NSAID for clinically symptomatic axial disease.

3. Consider corticosteroid injection to symptomatic sacroiliac joint.

4. If axial disease is persistent, particularly with radiological evidence, commence anti-TNF therapy, with attention to contraindications and appropriate monitoring.

5. Other considerations include methotrexate (with appropriate monitoring) or pamidronate (intravenous infusions).

Until further evidence is available, infliximab, etanercept, and adalimumab are suggested for reduction of signs and symptoms of moderate to severely active axPsA in patients with an inadequate response to at least 2 NSAID. Either SSZ or MTX are suggested in patients with predominantly active peripheral arthritis.

Efficacy, effect size, and number needed to treat have been recently reviewed⁴². As trial evidence becomes available in patients with PsA with specific axial involvement, these recommendations can be reviewed without the necessity of using AS surrogates.

REFERENCES

Search PubMed for:

1. Battistone MJ, Manaster BJ, Reda DJ, Clegg DO. The prevalence of sacroiliitis in psoriatic arthritis: new perspectives from a large, multicenter cohort. A Department of Veterans Affairs Cooperative Study. Skeletal Radiol 1999;28:196-201. [MEDLINE]

2. Hanly JG, Russell ML, Gladman DD. Psoriatic spondyloarthropathy: a long term prospective study. Ann Rheum Dis 1988;47:386-93. [MEDLINE]
3. Helliwell PS, Hickling P, Wright V. Do the radiological changes of classic ankylosing spondylitis differ from the changes found in the spondylitis associated with inflammatory bowel disease, psoriasis, and reactive arthritis? Ann Rheum Dis 1998;57:135-40.[MEDLINE]
4. Gladman DD, Brubacher B, Buskila D, Langevitz P, Farewell VT. Differences in the expression of spondyloarthropathy: a comparison between ankylosing spondylitis and psoriatic arthritis. Clin Invest Med 1993;16:1-7. [MEDLINE]
5. Taylor WJ, Harrison AA. Could the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) be a valid measure of disease activity in patients with psoriatic arthritis? Arthritis Rheum 2004;51:311-5. [MEDLINE]
6. Gladman DD, Cook RJ, Schentag C, et al. The clinical assessment of patients with psoriatic arthritis: results of a reliability study of the Spondyloarthritis Research Consortium of Canada. J Rheumatol 2004;31:1126–31. [MEDLINE]
7. Braun J, Sieper J. Building consensus on nomenclature and disease classification for ankylosing spondylitis: results and discussion of a questionnaire prepared for the International Workshop on New Treatment Strategies in Ankylosing Spondylitis, Berlin, Germany, 18–19 January 2002. Ann Rheum Dis 2002;61 Suppl III:iii61–7.
8. Dagfinrud H, Hagen KB, Kvien TK. Physiotherapy interventions for ankylosing spondylitis. The Cochrane Database of Systematic Reviews 2004; Issue 4: Art. No: CD002822. doi: 10.1002/14651858.CD002822.pub2.
9. Hidding A, van der Linden S, Gielen X, de Witte L, Dijkmans B, Moolenburgh D. Continuation of group physical therapy is necessary in ankylosing spondylitis: results of a randomized controlled trial. Arthritis Care Res 1994;7:90–6. [MEDLINE]
10. Wanders A, van der Heijde D, Landewe R, et al. Nonsteroidal anti-inflammatory drugs reduce radiographic progression in patients with ankylosing spondylitis: a randomized clinical trial. Arthritis Rheum 2005;52:1756–65. [MEDLINE]
11. Dougados M, Behier JM, Jolchine I, et al. Efficacy of celecoxib, a cyclooxygenase 2-specific inhibitor, in the treatment of ankylosing spondylitis: a six-week controlled study with comparison against placebo and against a conventional nonsteroidal anti-inflammatory drug. Arthritis Rheum 2001;44:180-5.[MEDLINE]
12. van der Heijde D, Baraf HS, Ramos-Remus C, et al. Evaluation of the efficacy of etoricoxib in ankylosing spondylitis: results of a fifty-two–week, randomized, controlled study. Arthritis Rheum 2005;52:1205-15. [MEDLINE]
13. Peters ND, Ejstrup L. Intravenous methylprednisolone pulse therapy in ankylosing spondylitis. Scand J Rheumatol 1992;21:134-8. [MEDLINE]
14. Mintz G, Enriquez RD, Mercado U, Robles EJ, Jimenez FJ, Gutierrez G. Intravenous methylprednisolone pulse therapy in severe ankylosing spondylitis. Arthritis Rheum 1981;24:734-6.[MEDLINE]
15. Ferraz MB, Tugwell P, Goldsmith CH, Atra E. Metaanalysis of sulfasalazine in ankylosing spondylitis. J Rheumatol 1990;17:1482-6. [MEDLINE]
16. Clegg DO, Reda DJ, Weisman MH, et al. Comparison of sulfasalazine and placebo in the treatment of ankylosing spondylitis. A Department of Veteran Affairs Cooperative Study. Arthritis Rheum 1996;39:2004-12. [MEDLINE]
17. Dekker-Saeys BJ, Dijkmans BA, Tytgat GN. Treatment of spondyloarthropathy with 5 aminosalicyclic acid (mesalazine): an open trial. J Rheumatol 2000;27:723-6. [MEDLINE]
18. Ostendorf B, Specker C, Schneider M. Methotrexate lacks efficacy in the treatment of severe ankylosing spondylitis compared with rheumatoid and psoriatic arthritis. J Clin Rheumatol 1998;4:129-36.
19. Willkens RF, Williams HJ, Ward JR, et al. Randomized, double-blind, placebo controlled trial of low-dose pulse methotrexate in psoriatic arthritis. Arthritis Rheum 1984;27:376-81. [MEDLINE]
20. Abu-Shakra M, Gladman DD, Thorne JC, Long J, Gough J, Farewell VT. Longterm methotrexate therapy in psoriatic arthritis: clinical and radiological outcome. J Rheumatol 1995;22:241-5.[MEDLINE]
21. Gonzalez-Lopez L, Garcia-Gonzalez A, Vazquez-Del-Mercado M, Munoz-Valle JF, Gamez-Nava JI. Efficacy of methotrexate in ankylosing spondylitis. A randomized, double blind, placebo controlled trial. J Rheumatol 2004;31:1568-74. [MEDLINE]
22. Wollina U, Stander K, Barta U. Toxicity of methotrexate treatment in psoriasis and psoriatic arthritis — short and long term toxicity in 104 patients. Clin Rheumatol 2001;20:406-10.[MEDLINE]
23. Grasedyck K, Schattenkirchner M, Bandilla K. The treatment of ankylosing spondylitis with auranofin (Ridaura). Z Rheumatol 1990;49:98-9. [MEDLINE]
24. Palit J, Hill J, Capell HA, et al. A multicentre double blind comparison of auranofin, intramuscular gold thiomalate and placebo in patients with psoriatic arthritis. Br J Rheumatol 1990;29:280-3. [MEDLINE]
25. Jones G, Crotty M, Brooks P. Psoriatic arthritis: a quantitative overview of therapeutic options. Psoriatic Arthritis Meta-Analysis Study Group. Br J Rheumatol 1997;36:95-9. [MEDLINE]
26. Wei JC, Chan TW, Lin HS, Huang F, Chou CT. Thalidomide for severe refractory ankylosing spondylitis: a 6-month open-label trial. J Rheumatol 2003;30:2627-31. [MEDLINE]
27. Van Denderen JC, Van der Paardt M, Nurmohamed MT, et al. Double-blind, randomised, placebo-controlled study of leflunomide in the treatment of active ankylosing spondylitis. Ann Rheum Dis 2005 [epub ahead of print].
28. Olivieri I, Salvarani C, Cantini F, Macchioni L, Padula A, Niccoli L. Therapy with cyclosporine in psoriatic arthritis. Semin Arthritis Rheum 1997;27:36-43. [MEDLINE]
29. Fraser AD, van Kuijk AW, Westhovens R, et al. A randomised, double blind, placebo controlled, multicentre trial of combination therapy with methotrexate plus cyclosporin in patients with active psoriatic arthritis. Ann Rheum Dis 2005;64:859-64. [MEDLINE]
30. Maksymowych WP, Jhangri GS, Fitzgerald AA, et al. A six-month randomized, controlled, double-blind, dose-response comparison of intravenous pamidronate (60 mg versus 10 mg) in the treatment of nonsteroidal antiinflammatory drug-refractory ankylosing spondylitis. Arthritis Rheum 2002;46:766-73.[MEDLINE]
31. Tan AL, Marzo-Ortega H, O'Connor P, Fraser A, Emery P, McGonagle D. Efficacy of anakinra in active ankylosing spondylitis: a clinical and magnetic resonance imaging study. Ann Rheum Dis 2004;63:1041–5. [MEDLINE]
32. Mease PJ, Goffe BS, Metz J, VanderStoep A, Finck B, Burge DJ. Etanercept in the treatment of psoriatic arthritis and psoriasis: a randomised trial. Lancet 2000;356:385–90. [MEDLINE]
33. Mease PJ, Kivitz AJ, Burch FX, et al. Etanercept treatment of psoriatic arthritis: safety, efficacy, and effect on disease progression. Arthritis Rheum 2004;50:2264-72. [MEDLINE]
34. Davis JC Jr, van der Heijde D, Braun J, et al. Enbrel Ankylosing Spondylitis Study Group. Recombinant human tumor necrosis factor receptor (etanercept) for treating ankylosing spondylitis: a randomized, controlled trial. Arthritis Rheum 2003;48:3230-6. [MEDLINE]
35. Braun J, Brandt J, Listing J, et al. Treatment of active ankylosing spondylitis with infliximab: a randomised controlled multicentre trial. Lancet 2002;359:1187–93. [MEDLINE]
36. van der Heijde D, Dijkmans B, Geusens P, et al. Ankylosing Spondylitis Study for the Evaluation of Recombinant Infliximab Therapy Study Group. Efficacy and safety of infliximab in patients with ankylosing spondylitis. Results of a randomized, placebo-controlled trial (ASSERT). Arthritis Rheum 2005;52:582-91.[MEDLINE]
37. Antoni CE, Kavanaugh A, Kirkham B, et al. Sustained benefits of infliximab therapy for dermatologic and articular manifestations of psoriatic arthritis: results from the Infliximab Multinational Psoriatic Arthritis Controlled Trial (IMPACT). Arthritis Rheum 2005;52:1227-36. [MEDLINE]
38. Mease PJ, Gladman DD, Ritchlin CT, et al. Adalimumab therapy in patients with psoriatic arthritis: 24-week results of a phase III study ADEPT — Adalimumab Effectiveness in Psoriatic Arthritis Trial [abstract]. Arthritis Rheum 2004;51 Suppl.
39. Mease P, Gladman D, Ritchlin C, Ruderman E, Steinfeld S, Choy E. Adalimumab for the treatment of psoriasis with moderately to severely active psoriatic arthritis. Arthritis Rheum 2005;52:3279-89. [MEDLINE]
40. Mease P, Sharp J, Ory P, et al. Adalimumab treatment effects on radiographic progression of joint disease in patients with psoriatic arthritis: results from ADEPT [abstract]. EULAR 2005; F0212.
41. Braun J, Pham T, Sieper J, et al. International ASAS consensus statement for the use of anti-tumour necrosis factor agents in patients with ankylosing spondylitis. Ann Rheum Dis 2003;62:817–24. [MEDLINE]

42. Zochling J, Van der Heijde D, Dougados M, Braun J. Current evidence for the management of ankylosing spondylitis: a systemic literature review for the ASAS/EULAR management recommendations in ankylosing spondylitis. Ann Rheum Dis 2005; [epub ahead of print].

Physical Activity and Risk of Revision Total Knee Arthroplasty in Individuals with Knee Osteoarthritis: A Matched Case-Control Study

DINA L. JONES, JANE A. CAULEY, ANDREA M. KRISKA, STEPHEN R. WISNIEWSKI, JAMES J. IRRGANG, DAVID A. HECK, C. KENT KWOH, and LAWRENCE S. CROSSETT

ABSTRACT. 

ABSTRACT.

Objective. To determine if physical activity was a risk factor for revision arthroplasty after primary total knee arthroplasty (TKA) due to osteoarthritis (OA) within the previous 15 years.

Methods. This was a matched case-control study. The cases had primary TKA followed by revision arthroplasty. Controls had primary TKA and no revision arthroplasty. Cases and controls were matched for age, sex, number of knees replaced, and date of primary TKA. Standardized telephone interviews were conducted to assess historical leisure activity, occupational activity, and instrumental activities of daily living after primary TKA in metabolic equivalent (MET)-hours per week. Conditional logistic regression was performed to identify the variables that predicted the need for revision arthroplasty.

Results. Seventeen female and 9 male pairs, aged 47 to 85 years, participated. Most of the reported activity was of low impact and low or moderate intensity. Cases reported a median of 44.5 (range 0 to 137) MET-hours of total historical physical activity per week compared with controls' 55.1 (range 0 to 278) MET-hours. Total historical physical activity was not associated with the risk of revision arthroplasty (OR 0.99, 95% CI 0.99–1.01). Participants with primary TKA (controls) consistently reported more MET-hours of leisure and occupational activity than those with revision arthroplasty (cases) regardless of the number of knees replaced or whether or not walking was accounted for.

Conclusion. This study quantified and described patterns of physical activity in a population with TKA. Physical activity did not appear to be a risk factor for revision arthroplasty. Our results suggest that individuals undergoing primary TKA should be encouraged to remain active after surgery. (J Rheumatol 2004;31:1384-90)

Key Indexing Terms:

PHYSICAL ACTIVITY
OSTEOARTHRITIS
TOTAL KNEE ARTHROPLASTY
PRIMARY
REVISION
KNEE

---

From the Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania.

Supported by the Arthritis Foundation and the Foundation for Physical Therapy.

D.L. Jones, PhD, PT, Research Associate, Division of Rheumatology and Clinical Immunology, Department of Medicine, and Clinical Assistant Professor, Department of Physical Therapy, School of Health and Rehabilitation Sciences; J.A. Cauley, DrPH, MPH, Associate Professor; A.M. Kriska, PhD, Associate Professor; S.R. Wisniewski, PhD, Assistant Professor; J.J. Irrgang, PhD, PT, ATC, Assistant Professor, Vice-Chairman Clinical Services, Department of Physical Therapy, School of Health and Rehabilitation Sciences; D.A. Heck, MD, Professor, Department of Orthopaedic Surgery, Indiana University; C.K. Kwoh, MD, Professor of Medicine and Epidemiology, Division of Rheumatology and Clinical Immunology, Department of Medicine, University of Pittsburgh, Center for Health Equity Research and Promotion, Veterans Administration Pittsburgh Health Care System; L.S. Crossett, MD, Assistant Professor, Department of Orthopaedic Surgery, University of Pittsburgh.

Address reprint requests to Dr. D.L. Jones, Division of Rheumatology and Clinical Immunology, University of Pittsburgh, 3500 Terrace Street, S727 Biomedical Science Tower, Pittsburgh, PA 15261. E-mail: Dljst4@pitt.edu

Submitted January 20, 2003; revision accepted January 23, 2004.

---

Osteoarthritis (OA) of the knee is more likely to lead to disability than OA in any other joint^1,2. Total knee arthroplasty (TKA) is a frequently utilized treatment option that can improve pain, function, and the quality of life^3-7. Although good to excellent results have been reported in more than 90% of individuals after knee replacement, roughly 1% require revision arthroplasty per year^6-12. The escalating rate of primary TKA, coupled with the increase in longevity, will most likely lead to a greater number of revision procedures in the future.

Few studies have examined risk factors for prosthetic failure. Factors such as male sex, younger age, greater height and weight, longer length of hospital stay, and a higher number of surgical complications have been associated with an increased risk of revision arthroplasty^10,13-15. Physical activity has been suggested as a possible risk factor for prosthetic failure, through aseptic loosening or polyethylene wear of the primary prosthesis as a result of high-impact activities, or wear of the weight-bearing surfaces due to repetitive mechanical loading of the prosthetic knee joint^10,16-18. Understanding the relationship between physical activity and prosthetic failure is necessary because improving activity levels is an important patient expectation after TKA¹⁹.

Most surgeons recommend avoiding intense and/or high-impact activities after TKA; however, little evidence is available to support such advice^17,20. Several studies recommend certain leisure/sports activities after TKA, or report the frequency with which people returned to such activities after surgery^17,20,21. Another study reported an increased risk of polyethylene failure in individuals who were more active, based on their occupation or retirement status¹⁰. Physical activity was not a risk factor for revision arthroplasty in a longterm study of younger individuals (aged 55 yrs or less) with primary TKA who were followed for an average of 8 years⁶.

Although physical activity comprises a spectrum of components that includes leisure activities, sports, occupational activities, and activities of daily living (basic and instrumental), previous studies assessed only limited components of physical activity after TKA (such as walking or sports activities)^18,21-24. Other studies classified activity after TKA with ordinal rating scales that combined the various components of physical activity, without providing information on the frequency, duration, and intensity of activity^6,10.

Our purpose was to determine if physical activity was a risk factor for revision arthroplasty after primary TKA due to OA. Specifically, we quantified the frequency, duration, and intensity of physical activity after TKA across the entire activity spectrum. A matched case-control study design was used to test the hypothesis that individuals with higher physical activity levels after primary TKA have a greater risk of revision arthroplasty than individuals with lower physical activity levels.

MATERIALS AND METHODS

Participants were recruited for the study between October 1999 and September 2000 at one community hospital and 3 tertiary hospitals after approval by the respective institutional review boards. Eligible participants were identified from the medical records of 12 orthopedic surgeons at 4 group practices. Recruitment materials and a consent form were mailed to all eligible participants. All participants provided written informed consent before enrolling in the study.

The cases and controls were aged 25 years or older. The cases had had primary TKA within the past 2 to 15 years and had also undergone revision arthroplasty. The controls had had primary TKA within the same time and no history of revision arthroplasty. Primary TKA was defined as at least one bicompartmental or tricompartmental knee replacement due to OA. Both cases and controls had primary TKA at least 2 years prior to study entry to allow one year for recovery and one year to assess physical activity at a stable level. Cases and controls were matched on 4 criteria: age (within 5 yrs), sex, number of knees replaced (unilateral vs bilateral), and the date of primary TKA (within 3 yrs).

Revision arthroplasty was defined as revision of at least the tibial or femoral component due to aseptic loosening or mechanical failure within the past 5 years, with a minimum of 2 years between primary and revision surgery. The indications for revision arthroplasty were classified based on a review of surgical and radiographic reports. Individuals who experienced prosthetic failure within 2 years after primary TKA were excluded from the study in order to eliminate those who required early revision arthroplasty due to serious postoperative complications. In order to minimize recall bias in the cases due to pain or recent surgery, the cases must have had revision surgery at least 3 months prior to enrolling in the study. Individuals with a history of knee infection, resection arthroplasty, lower extremity arthrodesis, lower extremity amputation, or more than one revision arthroplasty were excluded from the study.

Surgical data related to primary TKA were obtained by a retrospective review of medical records. Information was collected on the number of knee compartments replaced, the method of prosthetic fixation for each component (cemented vs cementless), the anatomic class of prosthesis (based on the posterior cruciate ligament), thickness of the tibial polyethylene spacer, and number of postoperative complications. Inpatient medical records were reviewed to determine the presence or absence of 16 potential postoperative complications according to Heck, et al²⁵. The number of complications was summed to provide the total number of complications during the hospitalization for primary TKA.

Each participant completed a structured telephone interview to obtain information on sociodemographic, clinical, functional, and physical activity variables. The interviewers were blinded to the participants' case/control status. Self-reported height and weight were used to calculate body mass index²⁶. The presence of comorbid conditions was determined using the comorbidity index of the Lower Limb Questionnaire from the American Academy of Orthopaedic Surgeons' Lower Limb Outcomes Data Collection Questionnaires^27,28. The comorbidity index has shown satisfactory reproducibility and validity for predicting health status^29,30. The comorbidity index can range from a score of 0 (no comorbidities) to 100 (highest level of comorbidities)²⁸. Separate questions were used to assess the history of knee injury or knee surgery prior to primary TKA based on participant self-report.

Ambulation was categorized according to whether or not a participant required assistance (an assistive device or assistance from at least one person) to ambulate on level surfaces and stairs. Participants were also asked if their usual physical activity level 2 years after primary TKA was more, less, or the same as it was 2 years before surgery in order to assess the change in activity between pre- and postoperative primary TKA. Two years was chosen as the timeframe because there were differences among the matched pairs in the length of time from primary to revision arthroplasty, and 2 years after primary TKA was a common point in time for all pairs.

The measurement of historical leisure (including sports) and occupational activity in the cases included activity from the second year after primary TKA through either the fifteenth year after primary TKA or the time of revision surgery, whichever occurred first. For the matched control, leisure and occupational activity were assessed from the second year after primary TKA until the time of revision surgery in the matched case.

The leisure activity section of the Modifiable Activity Questionnaire (MAQ) was used to measure historically the average number of hours per week of participation in 39 leisure and sports activities²². The MAQ has been shown to be reliable in adults and adolescents through field testing and comparison with activity monitors (leisure activity r = 0.69–0.85; occupational activity r = 0.88)^31,32. To weight each activity by its relative intensity, the average number of hours per week for each activity was multiplied by the activity's metabolic equivalent (MET) and summed across all activities to provide the average number of MET-hours per week of historical leisure activity^22,33-36.

The MAQ was also used to assess historical occupational activity based on the physical demands and work schedule of each job²². Occupations such as student, homemaker, and volunteer, or being retired, disabled, or unemployed, were also included. The energy demands of each job were categorized as light, moderate, or hard. The number of hours per week in each category was weighted by its relative intensity (i.e., 1.5 MET for light, 4 MET for moderate, and 7 MET for hard) and summed to provide the average number of MET-hours per week of historical occupational activity.

The work, yard work, and caretaking sections of the YALE Physical Activity Survey were used to assess retrospectively the number of hours per week spent performing 15 instrumental activities of daily living (IADL) during the past month²². The 2-week reproducibility of the instrument has been reported (r = 0.42–0.65)³⁷. The number of hours spent on each activity was multiplied by an intensity code and summed over all of the activities to provide the total number of kilocalories per week, which were then converted to MET-hours per week.

Statistical methods. Descriptive statistics were calculated for all variables. The median and range were also calculated for the physical activity variables due to the non-normal distribution of the data.

Univariate analyses for matched data were conducted between the outcome variable (revision/no revision) and each potential predictor variable using paired t tests and signed-rank tests where appropriate. Ordinal and nominal data were analyzed with either the McNemar test or the Bowker test of symmetry. Because the time spent walking for exercise has been found to be unreliable in many populations, the historical leisure activity variable was analyzed both with and without the inclusion of any time spent walking²². All results include walking as a leisure activity except where noted.

The leisure activities were also categorized according to the level of impact at the knee based on the literature to determine if there were differences in the amount of high-impact leisure activity between the cases and controls^17,38,39. The activities were classified as no (e.g., swimming), low/moderate (e.g., walking), or high-impact (e.g., jogging). In addition, the leisure and occupational activities were stratified based on intensity level to examine if the amount of high-intensity activity differed between cases and controls. Activities that required ≥ 6 MET in energy expenditure were classified as high-intensity activities^31,40,41.

Odds ratios (OR, based on a one MET-hour change in physical activity per week) and 95% confidence intervals (CI) were calculated for the leisure, occupational, and IADL variables. The physical activity variables were analyzed separately, as well as with leisure and occupational activity combined as a total historical physical activity variable. The total historical physical activity variable was analyzed as both discrete and continuous data. For discrete data, the total historical physical activity variable was stratified into high and low activity groups based on the sex-specific median of the distribution (men 51.9 MET-hours, women 49.8 MET-hours) to determine if high versus low physical activity was associated with the need for revision arthroplasty.

After the predictor variables were identified, a stepwise conditional logistic regression analysis for matched data was performed to identify the variables that predicted the need for revision arthroplasty. To avoid excluding any potentially important predictor variables, variables that had a univariate significance level (p) ≤ 0.15, were clinically meaningful, or had non-missing data were entered into the model⁴². The total historical physical activity variable was forced to remain in the model. A significance level (p) of 0.10 was used as the entry and stay criteria in the stepwise model⁴². The hypotheses for all of the statistical tests were 2-tailed except for the physical activity analyses. All statistical tests were conducted with a Type I error rate of 0.05 using the SAS software package (version 8.0; SAS Institute, Cary, NC, USA).

RESULTS

One hundred eighty-nine patients met the eligibility criteria, of whom 64 (34%) were identified as cases and 125 (66%) as controls. Thirty-eight of the eligible cases (59%) and 52 of the eligible controls (42%) enrolled in the study. Reasons for nonenrollment included: failure to respond to the mailings (cases 16%, controls 30%), declined participation in the study (cases 19%, controls 15%), no available forwarding address (cases 3%, controls 8%), and deceased (cases 3%, controls 5%). Enrolled and nonenrolled cases and controls were similar with respect to sex, number of knees replaced, number of months since primary TKA, and mean age, except that enrolled controls were significantly younger than the nonenrolled controls [enrolled 71 (SD 7.4) yrs; not enrolled 75 (SD 8.1) yrs; p = 0.001].

Twenty-six case-control pairs were identified and included 17 female (65%) and 9 male (35%) pairs with a mean age of 70.5 (SD 8.9) years (range 47–85 yrs). Twelve (46%) of the pairs had both knees replaced. Primary TKA in the cases and controls was performed an average of 6.4 (SD 2.3) years previously (range 2–11 yrs). The mean time to prosthetic failure from primary to revision arthroplasty in the cases was 5 (SD 2.3) years (range 2–11 yrs). The indications for revision arthroplasty were polyethylene failure (42%), component loosening (38%), patellar instability (12%), arthrofibrosis (4%), and oversize components (4%). Revision arthroplasty in the cases was performed a mean of 2.6 (SD 1.5) years prior to study entry (range 4 mo to 5 yrs).

The sociodemographic, clinical, surgical, and functional characteristics of the case-control pairs are presented in Table 1. Cases reported a history of knee surgery prior to TKA 3 times more frequently than the controls (p = 0.02). Cementless femoral components were implanted during primary TKA more often in the cases than the controls (p = 0.01). After primary TKA, there were no statistically significant differences between the cases and controls in ambulation on level surfaces; however, the cases were more likely than controls to need assistance to ascend and descend stairs (p = 0.02). There was a tendency for the cases to report a reduction in their usual physical activity level after primary TKA more often than the controls, although the differences were not statistically significant (p = 0.07).

Table 1. Sociodemographic, clinical, surgical, and functional characteristics.

Physical activity was assessed historically over an average period of 4 (SD 2.0) years in the matched cases and controls (range 1–10 yrs). The most frequently reported leisure activity was walking in the cases (65%) and gardening/yardwork in the controls (77%). No case and only 2 controls (8%) engaged in high-impact leisure activities, therefore analyses examining the influence of activity type (high vs low-impact) on risk of revision arthroplasty were not conducted. Retirement was the most frequently reported occupational activity (cases 42%, controls 54%) followed by serving as a homemaker (cases 39%, controls 23%).

The controls reported more median MET-hours per week of historical leisure and occupational activity than the cases, and similar median MET-hours of high-intensity leisure and occupational activity (Table 2). There was no association, however, between leisure or occupational activity (historical or high-intensity) and the risk of revision arthroplasty. At the lower-intensity end of the activity spectrum, the controls again reported more median MET-hours per week of IADL than the cases (Table 2). Participation in IADL was not a risk factor for revision arthroplasty.

Table 2. Components of physical activity.

The controls reported more MET-hours of total historical physical activity per week than the cases (Table 2). Total historical physical activity was not associated with the risk of revision arthroplasty. Similar results were obtained when walking was excluded from the leisure activity component of total activity.

The controls reported more median MET-hours of total historical physical activity than the cases regardless of whether one knee [cases 48.9 (range 13.9–125.2), controls 60.8 (range 0–194.4)] or both knees [cases 27.2 (range 0–137), controls 32.8 (range 10.2–278)] were replaced during primary TKA. In addition, men reported more median MET-hours of total historical physical activity per week than women [men 51.9 (range 0–278), women 49.8 (range 0–125.2)]. The risk of revision arthroplasty in participants with high levels of total historical physical activity was 33% lower than in those with low levels of total activity; however, this finding was not statistically significant (OR 0.67, 95% CI 0.67–1.93).

The following variables were entered into the multivariate model along with the total historical physical activity variable: history of knee surgery before primary TKA, fixation of femoral component, ambulation on stairs, ambulation on level surfaces, and change in usual physical activity. In the final model (Table 3), total historical physical activity was not associated with the risk of revision arthroplasty after adjusting for the other 2 variables in the model. The risk of revision arthroplasty was significantly increased in participants who reported a history of knee surgery before undergoing primary TKA and in those who experienced a reduction in usual activity level after surgery.

Table 3. Total historical physical activity and risk of revision arthroplasty.

DISCUSSION

The purpose of this matched case-control study was to test the hypothesis that physical activity was a risk factor for revision arthroplasty after primary TKA due to OA. Our results showed that participants who reported a higher number of MET-hours per week of leisure activity, occupational activity, or IADL after primary TKA did not have an increased risk of revision arthroplasty. In addition, men reported more total historical physical activity than women in a pattern consistent with previous studies^24,43,44.

To our knowledge, this is the first study that quantified activity across the physical activity spectrum in a population with total knee replacement. Our study took the physical activity assessment one step further by quantifying the frequency, duration, and intensity of participation in a wide range of activities. Further, we were able to match the cases and controls on 4 important characteristics (age, sex, number of knees replaced, and date of primary TKA) that could potentially confound the relationship between physical activity and revision arthroplasty. Although surgeon and hospital characteristics have been reported to influence outcomes after TKA^25,45, we chose not to match on these factors because they have not been linked to physical activity levels.

While physical activity was not related to revision arthroplasty, 2 other variables emerged as potential risk factors for revision surgery. Participants who reported a history of surgery in their knee before primary TKA were more likely to undergo revision arthroplasty than those who did not. Thus, prior surgery (e.g., arthrotomy or arthroscopy) may increase the risk of revision; however, this relationship is a complex issue and requires further study.

Participants who reported a reduction in their usual physical activity level after primary TKA were also more likely to undergo revision arthroplasty than those who reported that their activity levels were unchanged or had improved. Activity levels after knee replacement may have been related to the success of the surgery in restoring function⁴⁶. The cases may have curtailed their activity levels after primary TKA due to symptoms or functional limitations associated with their knee. Because this was a retrospective study, data were not available on the preoperative functional status or attainment of postoperative rehabilitation milestones in the participants. Self-reported preoperative pain and functional status were not collected because patient recall of this information postoperatively has been found to have only poor to moderate agreement with information reported prospectively⁴⁷.

In the univariate analyses, participants with cementless femoral components implanted during primary TKA had a greater risk of revision arthroplasty than those with cemented components. Cementless prostheses generally have lower cumulative survival rates than cemented prostheses, and therefore have a greater risk of component loosening due to failure at the bone-implant interface^15,48,49.

The absence of a significant relationship between physical activity and revision arthroplasty may have been due to limited statistical power. Nevertheless, we were able to identify several potential risk factors for revision arthroplasty. More important, we were able to quantify and describe patterns of physical activity in an older population with total knee replacement. Participants with primary TKA (controls) consistently reported more MET-hours of leisure and occupational physical activity than participants with revision arthroplasty (cases) regardless of the number of knees replaced or whether or not walking was accounted for. With respect to the type of physical activities that were performed, there was not enough variability in our sample to examine the influence of high-impact activity on the need for revision arthroplasty.

Physical activity in this study was assessed by self-report, instead of direct evaluation, due to the retrospective design of the study. The physical activity estimates obtained by the questionnaires have been found to be reliable and valid, and allowed us to rank individuals within a group from the least to most active²². Because this was a retrospective study, the historical assessment of physical activity may have been subject to recall bias. The cases, who were interviewed after revision arthroplasty, may have remembered their past physical activity differently than the controls, who did not undergo a second surgery. To minimize any potential recall bias due to pain or recent surgery, we waited at least 3 months after revision arthroplasty before enrolling cases in the study. Recall of physical activity may also have been limited in the controls because the physical activity assessment period was framed by only one specific event (primary TKA), compared to 2 events in the cases (primary and revision TKA).

Based on these results, individuals undergoing primary TKA should be encouraged to remain active after surgery, especially women. Although there are no evidence-based guidelines available for determining which activities are appropriate after surgery, there is general agreement on the recommendation for participation in lower intensity and low-impact activities, with avoidance of high-impact activities^6,17,21,50. Most of the activity reported in this study was of low impact and low or moderate intensity.

Although physical activity was not a significant risk factor for revision arthroplasty in this study, individuals with primary TKA (controls) consistently reported more leisure and occupational activity than those with revision TKA (cases). Further research is required to confirm this observation.

ACKNOWLEDGMENT

We thank Edward J. McClain III, MD; Raj K. Sinha, MD, PhD; Peter Z. Cohen, MD; Anthony M. DiGioia III, MD; Mark A. Goodman, MD; Rodney G. Gordon, MD; Gregory L. Hung, MD; Paul H. Resnick, MD; and Spiro N. Papas, MD, for their assistance in identifying potential participants for this study. We also thank Elizabeth Dames, Keelan R. Enseki, Sheri Anne Hale, Jayme Elizabeth Klein, and Brian M. Klucinec for their assistance with data collection.

REFERENCES

Search PubMed for:

1. Yang SS, Nisonson B. Arthroscopic surgery of the knee in the geriatric patient. Clin Orthop Rel Res 1995;316:50-8. [MEDLINE]

2. Hannan MT, Felson DT, Anderson JJ, Naimark A, Kannel WB. Estrogen use and radiographic osteoarthritis of the knee in women. The Framingham Osteoarthritis Study. Arthritis Rheum 1990;33:525-32. [MEDLINE]

3. Mancuso CA, Ranawat CS, Esdaile JM, Johanson NA, Charlson ME. Indications for total hip and total knee arthroplasties. Results of orthopaedic surveys. J Arthroplasty 1996;11:34-46. [MEDLINE]

4. Dieppe P, Basler HD, Chard J, et al. Knee replacement surgery for osteoarthritis: effectiveness, practice variations, indications, and possible determinants of utilization. Rheumatology 1999;38:73-83.[MEDLINE]

5. Belmar CJ, Barth P, Lonner JH, Lotke PA. Total knee arthroplasty in patients 90 years of age and older. J Arthroplasty 1999;14:911-4. [MEDLINE]

6. Diduch DR, Insall JN, Scott NW, Scuderi GR, Font-Rodriquez D. Total knee replacement in young, active patients: long-term follow-up and functional outcome. J Bone Joint Surg Am 1997;79:575-82. [MEDLINE]

7. Gill GS, Chan KC, Mills DM. 5- to 18-year follow-up study of cemented total knee arthroplasty for patients 55 years old or younger. J Arthroplasty 1997;12:49-54. [MEDLINE]

8. Spector TD, Hart DJ. How serious is knee osteoarthritis. Ann Rheum Dis 1992;51:1105-6. [MEDLINE]

9. Callahan CM, Drake BG, Heck DA, Dittus RS. Patient outcomes following tricompartmental total knee replacement: a meta-analysis. JAMA 1994;271:1349-57. [MEDLINE]

10. Heck DA, Clingman JK, Kettelkamp DG. Gross polyethylene failure in total knee arthroplasty. Orthopedics 1992;15:23-8.[MEDLINE]

11. Lonner JH, Siliski JM, Scott RD. Prodromes of failure in total knee arthroplasty. J Arthroplasty 1999;14:488-92. [MEDLINE]

12. Rathjen KW. Surgical treatment: total knee arthroplasty. Am J Knee Surg 1998;11:58-63. [MEDLINE]

13. Heck DA, Melfi CA, Mamlin LA, et al. Revision rates after knee replacement in the United States. Med Care 1998;36:661-9.[MEDLINE]

14. Johnson GVV, Worland RL, Keenan J, Norambuena N. Patient demographics as a predictor of the ten-year survival rate in primary total knee replacement. J Bone Joint Surg Br 2003;85:52-6.[MEDLINE]

15. Rand JA, Trousdale RT, Ilstrup DM, Harmsen WS. Factors affecting the durability of primary total knee prostheses. J Bone Joint Surg Am 2003;85:259-65. [MEDLINE]

16. Hilding MB, Ryd L, Toksvig-Larsen S, Mann A, Stenström A. Gait affects tibial component fixation. J Arthroplasty 1999;14:589-93. [MEDLINE]

17. McGrory BJ, Stuart MJ, Sim FH. Participation in sports after hip and knee arthroplasty: review of literature and survey of surgeon preferences. Mayo Clin Proc 1995;70:342-8. [MEDLINE]

18. McClung CD, Zahiri CA, Higa JK, Amstutz HC, Schmalzried TP. Relationship between body mass index and activity in hip or knee arthroplasty patients. J Orthop Res 2000;18:35-9.[MEDLINE]

19. Mancuso CA, Sculco TP, Wickiewicz TL, et al. Patients' expectations of knee surgery. J Bone Joint Surg Am 2001;83:1005-12. [MEDLINE]

20. Mallon WJ, Callaghan JJ. Total knee arthroplasty in active golfers. J Arthroplasty 1993;8:299-306. [MEDLINE]

21. Bradbury N, Borton D, Spoo G, Cross MJ. Participation in sports after total knee replacement. Am J Sports Med 1998;26:530-5. [MEDLINE]

22. Pereira MA, Fitzgerald SJ, Gregg EW, et al. A collection of Physical Activity Questionnaires for health-related research. Med Sci Sports Exerc 1997;29 Suppl 6:S1-205.[MEDLINE]

23. Zahiri CA, Schmalzried TP, Szuszczewicz ES, Amstutz HC. Assessing activity in joint replacement patients. J Arthroplasty 1998;13:890-5. [MEDLINE]

24. Schmalzried TP, Szuszczewicz ES, Northfield MR, et al. Quantitative assessment of walking activity after total hip or knee replacement. J Bone Joint Surg Am 1998;80:54-9. [MEDLINE]

25. Heck DA, Robinson RL, Partridge CM, Lubitz RM, Freund DA. Patient outcomes after knee replacement. Clin Orthop Rel Res 1998;356:93-110. [MEDLINE]

26. US Department of Health and Human Services, Public Health Service. Healthy People 2000 National Health Promotion and Disease Prevention Objectives. Washington, DC: US Department of Health and Human Services; 1990.

27. American Academy of Orthopaedic Surgeons, American Association of Hip and Knee Surgeons, American Orthopaedic Society for Sports Medicine, et al. Lower limb questionnaire. Version 2000. [Internet. Cited May 3, 2004] Available from: www3.aaos.org/research/normstdy/lwintr.cfm

28. Hunsaker FG, Cioffi DA, Amadio PC, et al. The American Academy of Orthopaedic Surgeons outcomes instruments: Normative values from the general population. J Bone Joint Surg Am 2002;84:208-15. [MEDLINE]

29. Sangha O, Stucki G, Fossel AH, Katz JN. A simplified method to assess comorbidity in clinical and health services research of rheumatic diseases [abstract]. Arthritis Rheum 1995;38 Suppl:S177.

30. Sangha O, Stucki G, Liang MH, Fossel AH, Katz JN. The Self-Administered Comorbidity Questionnaire: A new method to assess comorbidity for clinical and health services research. Arthritis Rheum 2003;49:156-63. [MEDLINE]

31. Kriska AM, Knowler WC, LaPorte RE, et al. Development of a questionnaire to examine relationship of physical activity and diabetes in Pima Indians. Diabetes Care 1990;13:401-7.[MEDLINE]

32. Kriska AM, Sandler KB, Cauley JA, LaPorte RE, Hom DL, Pambianco G. The assessment of historical physical activity and its relation to adult bone parameters. Am J Epidemiol 1988;127:1053-63. [MEDLINE]

33. Ainsworth BE, Haskell WL, Leon AS, et al. Compendium of physical activities: classification of energy costs of human physical activities. Med Sci Sports Exerc 1993;25:71-80. [MEDLINE]

34. Durnin JVGA, Passmore R. Energy, work, and leisure. London: Heinnemann Educational Books Ltd.; 1967.

35. Passmore R, Durnin JVGA. Human energy expenditure. Physiol Rev 1955;35:801-40.

36. Wilson PWF, Paffenbarger RS Jr, Morris JN, Havlik RJ. Assessment of methods for physical activity and physical fitness in population studies: a report of a NHLBI workshop. Am Heart J 1986;111:1177-93. [MEDLINE]

37. DiPietro L, Caspersen CJ, Ostfeld AM, Nadel ER. A survey for assessing physical activity among older adults. Med Sci Sports Exerc 1993;25:628-42. [MEDLINE]

38. Kilgus DJ, Dorey FJ, Finerman GA, Amstutz HC. Patient activity, sports participation, and impact loading on the durability of cemented total hip replacements. Clin Orthop Rel Res 1991;269:25-31. [MEDLINE]

39. Jakes RW, Khaw KT, Day NE, et al. Patterns of physical activity and ultrasound attenuation by heel bone among Norfolk cohort of European Prospective Investigation of Cancer (EPIC Norfolk): population based study. BMJ 2001;322:140. [MEDLINE]

40. Taylor HL, Jacobs DR Jr, Schucker B, Knudsen J, Leon AS, Debacker G. A questionnaire for the assessment of leisure time physical activities. J Chron Dis 1978;31:741-55.

41. Blair SN, Haskell WL, Ho P, et al. Assessment of habitual physical activity by a seven-day recall in a community survey and controlled experiments. Am J Epidemiol 1985;122:794-804.[MEDLINE]

42. Hosmer DW, Lemeshow S. Applied logistic regression. 2nd ed. New York: John Wiley & Sons; 2000.

43. Caspersen CJ, Christenson GM, Pollard RA. Status of the 1990 physical fitness and exercise objectives — Evidence from NHIS 1985. Publ Health Reports 1986;101:587-92.[MEDLINE]

44. Stephens T, Jacobs DR Jr, White CG. A descriptive epidemiology of leisure-time physical activity. Publ Health Reports 1985;100:147-58.[MEDLINE]

45. Katz JN, Barrett J, Mahomed NN, Baron JA, Wright J, Losina E. Association between volume and perioperative outcomes of primary and revision total knee replacement [abstract]. Arthritis Rheum 2003;48 Suppl:S220.

46. Jones DL, Cauley JA, Kriska AM, et al. The role of physical activity on the need for revision total knee arthroplasty in individuals with osteoarthritis of the knee. Pittsburgh: University of Pittsburgh; 2001.

47. Lingard EA, Wright EA, Sledge CB, Kinemax Outcomes Group. Pitfalls of using patient recall to derive preoperative status in outcome studies of total knee arthroplasty. J Bone Joint Surg Am 2001;83:1149-56. [MEDLINE]

48. Lotke PA, Garino JP. Revision total knee arthroplasty. Philadelphia: Lippincott-Raven Publishers; 1999.

49. Robertsson O, Knutson K, Lewold S, Lidgren L. The Swedish Knee Arthroplasty Register 1975-1997: An update with special emphasis on 41,223 knees operated on in 1988-1997. Acta Orthop Scand 2001;72:503-13.

50. Ritter MA, Eizember LE, Fechtman RW, Keating EM, Faris PM. Revision total knee arthroplasty. A survival analysis. J Arthroplasty 1991;6:351-6. [MEDLINE]