- Top of page
- SUBJECTS AND METHODS
- AUTHOR CONTRIBUTIONS
Chronic disease may represent one of a number of age-related catabolic processes that contribute to the progression of sarcopenia (1). Given that common factors may influence changes to different components of the musculoskeletal system (2), it is possible that osteoarthritis (OA) is predictive of muscle deterioration during aging. Cross-sectional studies have demonstrated that knee and hip OA are associated with decreased muscle mass (3, 4) and muscle strength (4–6) in older adults. This association may be explained by arthrogenous muscle inhibition, where changes in afferent input from the affected joint result in reduced efferent motor neuron stimulation of the attendant skeletal muscles (7). While it has been demonstrated that quadriceps weakness may be a risk factor for the development of OA (8), it is also possible that muscle atrophy and attendant strength decline following the development of OA. This may be a result of joint pain, which may lead to the avoidance of activity that stimulates skeletal muscle synthesis (9).
Aging is associated with an increased risk of falling that may be a consequence of muscle weakness (10). OA has also been associated with self-reported recurrent falls (11), and the relationship may be partly explained by reduced muscle strength. Older adults with self-reported OA have been observed to exhibit poorer postural stability (12). A cross-sectional analysis of the Tasmanian Older Adult Cohort Study (TASOAC) also demonstrated that self-reported pain and dysfunction have a modest but significant association with objectively measured falls risk (13). If OA is predictive of sarcopenia progression and falls in older adults, a change will need to be considered in the manner in which researchers and clinicians address OA-associated disability; however, longitudinal studies are required to confirm this. The aim of this prospective study was to examine whether self-reported joint pain, stiffness, and dysfunction, and/or radiographic OA (ROA), are predictive of changes in lean mass, strength, and falls risk in community-dwelling older adults.
Significance & Innovations
Greater declines in lower extremity muscle strength and quality and greater increases in risk of falls over nearly 3 years were generally observed in older women, but not men, who reported knee or hip pain at baseline.
Changes in muscle strength and quality did not differ for those with baseline knee or hip radiographic osteoarthritis (ROA) compared to those without ROA.
ROA may, however, contribute to sarcopenia progression and falls by inducing pain and stiffness that lead to dysfunction at the knee joint.
SUBJECTS AND METHODS
- Top of page
- SUBJECTS AND METHODS
- AUTHOR CONTRIBUTIONS
This study was conducted within the TASOAC study, an ongoing, prospective, population-based study primarily aimed at identifying factors associated with the development and progression of OA and osteoporosis in community-dwelling older adults. Ethics approval was provided by the Southern Tasmanian Health and Medical Human Research Ethics Committee. Written informed consent was obtained from all of the participants.
The cohort consisted of 1,100 men and women ages 50–79 years, selected from electoral rolls in Southern Tasmania (population 229,000) using stratified simple random sampling by sex without replacement (response rate 57%). Institutionalized older adults were excluded from participating in the study and participants were also excluded due to contraindication for magnetic resonance imaging (MRI), as MRI tests were required to examine OA progression.
Enrolled participants attended a clinic at the Menzies Research Institute between March 2002 and September 2004. Participants were invited to attend a followup clinic approximately 2 to 3 years later. Previous diagnosis of cardiovascular disease (CVD), diabetes mellitus, and rheumatoid arthritis (RA), as well as smoking history, were self-reported by questionnaire at baseline.
Height was measured to the nearest 0.1 cm (with shoes, socks, and headwear removed) using a Leicester stadiometer (Invicta). Weight was measured to the nearest 0.1 kg (with shoes, socks, and bulky clothing removed) using electronic scales (Heine) that were calibrated at the beginning of each clinic. Body mass index (BMI; kg/m2) was calculated from these measurements.
Participants underwent a whole body scan by dual x-ray absorptiometry (DXA), using a Hologic Delphi densitometer from which soft tissue composition was determined. A variable of leg lean mass was calculated as the sum of lean mass in both legs.
Knee extension strength.
Knee extension strength of the dominant leg was measured using a seated isometric contraction of the knee extensors. A strap was positioned 10 cm above the lateral malleolus. The strap was attached to a 100-kg pocket balance, which was connected to the back of the chair. Participants were instructed to push against the strap by extending the lower leg from the knee joint until maximum contraction force was achieved. The best score from 2 attempts was recorded.
Leg strength and muscle quality.
Whole leg strength was measured to the nearest kilogram in both legs simultaneously, using a dynamometer (TTM Muscular Meter). This test examines isometric strength, predominantly of the quadriceps and hip extensors, and has been described in detail previously (14).
Two trials were recorded, and the mean score was taken as the criterion value for whole leg strength. Intraclass correlation coefficients (ICCs) demonstrated high reproducibility between trials 1 and 2 at both baseline (ICC 0.95; 95% confidence interval [95% CI] 0.94, 0.96) and followup (ICC 0.96; 95% CI 0.95, 0.97) in the present study.
We also examined leg muscle quality (LMQ; an estimate of specific force), which has been shown to decrease with age, and is commonly used in studies of aging muscle (15). LMQ was calculated as the magnitude of leg strength from the whole leg strength test divided by the combined lean mass of the legs from the DXA scans using the following formula:
Falls risk assessment.
The Physiological Profile Assessment (PPA; Prince of Wales Medical Research Institute) assessed participant falls risk at baseline and followup. This protocol has been described previously (16). The PPA examines 5 physiologic domains (vision, reaction time, proprioception, strength, and balance) and provides a standardized falls risk score. Falls risk scores less than 0 indicate a low risk of falls, between 0 and 1 indicate a mild risk, between 1 and 2 indicate a moderate risk, and greater than 2 indicate a high risk.
Physical activity was measured over 7 consecutive days following the initial clinic using an Omron (Omron HJ-003 & HJ-102) or Yamax (Yamax SW-200) pedometer. Participants were also required to complete a pedometer diary recording details, including daily step counts and duration of pedometer use. The pedometer protocol and data handling techniques have been described previously (14).
Knee pain, stiffness, and dysfunction were assessed by the validated Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) (17) at baseline. The questionnaire was self-administered and each dimension was assessed by a series of questions answered using a scale from 1 (no pain, stiffness, or dysfunction) to 10 (severe pain, stiffness, or dysfunction). Dichotomous variables were created to represent “severe” pain, stiffness, or dysfunction, equivalent to a WOMAC index score in the highest quartile for each respective measure.
The prevalence of “any neck pain,” “any shoulder pain,” “any hand pain,” “any back pain,” “any hip pain,” “any knee pain,” and “any foot pain” was also assessed by questionnaire at baseline with a “yes” or “no” answer to the following question: “Do you have pain at any of these sites today?”
Knee ROA was examined by radiographs at baseline using a standing anteroposterior semiflexed view of both knees, assessed using the Altman atlas (18). This protocol has been described previously (19). Each of the following was assessed on a scale of 0–3 (where 0 = no disease and 3 = most severe disease): medial joint space narrowing (JSN), lateral JSN, medial femoral osteophytes, medial tibial osteophytes, lateral femoral osteophytes, and lateral tibial osteophytes. Each score was arrived at by consensus with 2 readers who simultaneously assessed the radiographs with immediate reference to the Altman atlas. Intraobserver repeatability was assessed in 40 subjects with ICCs of 0.65–0.85.
Hip ROA was examined using weight-bearing anteroposterior radiographs with both feet in 10° internal rotation, and assessed in the same manner. Features assessed included axial JSN, superior JSN, and osteophytes. Again, each score was arrived at by 2 readers and intraobserver repeatability was assessed in 40 subjects with ICCs of 0.60–0.87.
Participants were classified using dichotomous variables for the presence or absence of JSN or osteophytes in the hip or knee at baseline. Dichotomous variables of knee or hip ROA were created to represent a score of >0 for either JSN or osteophytes, and continuous variables were also created using the total score for JSN and osteophytes of the knee and hip, respectively.
We initially examined differences in baseline descriptive characteristics between those who reported any knee pain at baseline and those who did not. Knee pain was chosen as a stratifying variable because we hypothesized that pain at this site would have the strongest associations with dependent variables of lower extremity lean mass and strength. Independent t-tests assessed differences in continuous variable descriptive characteristics and chi-square tests examined differences for dichotomous variables. In order to examine the effect of self-reported pain at more than 1 site on lower extremity muscle mass and strength, analysis of variance tests also compared LMQ at baseline across sex-stratified groups classified as reporting no pain, pain at 1 site only (neck, shoulders, hands, back, hips, knees, or feet), and pain at 2 or more sites. LMQ was the only outcome variable assessed, as it is a composite measure of muscle mass and strength and therefore reduced the requirement for multiple comparisons.
In order to more completely elucidate associations of OA with subsequent changes in muscle parameters, multivariable regression analyses examined associations of absolute change in outcome factors over 2.6 years with both baseline pain at any site and ROA of the knees and hips. These analyses were stratified by sex and were adjusted for age, CVD, diabetes mellitus, RA, smoking history, and the value of the outcome variable at baseline. Each analysis was also adjusted for change in body fat and physical activity levels, as well as time from baseline assessment to followup assessment.
In order to further examine the potential pathways for associations between ROA, pain, and muscle parameters, a path analysis was conducted. Path analysis is complimentary to regression analysis but separates direct and indirect effects through a medium variable, allowing interpretation of causal relationships between multiple variables (20). It was hypothesized that the association between knee ROA and change in LMQ is mediated by pain, stiffness, and dysfunction. As such, knee ROA score was used as the exogenous variable, and WOMAC scores for knee pain, stiffness, and dysfunction were used as intermediate endogenous variables. Successive multiple regressions were performed to estimate path coefficients. Knee pain and ROA measures were selected for this analysis as they were expected to be more strongly associated with lower extremity muscle mass and strength, for which LMQ was used as a composite variable.
For all statistical analyses, a P value of less than 0.05 (2-tailed) or a 95% CI not including the null point was considered statistically significant. All statistical tests were performed using Intercooled Stata 10.1 for Windows.
- Top of page
- SUBJECTS AND METHODS
- AUTHOR CONTRIBUTIONS
Of the 1,099 participants who attended the baseline clinic, 224 (20%) were lost to followup due to reasons such as death, disability, institutionalization, relocation interstate or overseas, joint replacement, or withdrawal of consent. Of 875 participants who provided both baseline and followup data, 18 did not have complete DXA data, 79 failed to complete the muscle strength tests, 4 did not complete the falls risk assessment, and 65 were not assessed for ROA. Therefore, 709 participants (50% women, mean ± SD age 62 ± 7 years) were included in all data analyses. Included participants were younger (mean ± SD age 61.9 ± 7.1 versus 63.1 ± 7.6 years; P = 0.031), had higher knee extension strength (mean ± SD 31.5 ± 10.8 versus 28.2 ± 11.7 kg; P < 0.001), and were less likely to report any knee pain (43% versus 51%; P = 0.042) than excluded participants.
Table 1 shows results from analyses of baseline descriptive characteristics. These analyses revealed that for both men and women, those who reported any knee pain had greater mean body weight and BMI, as previously reported (6), and among women only, had greater mean body fat percentages and lower mean physical activity. Those reporting knee pain were also more likely to report neck, shoulder, hand, back, hip, or foot pain, but only men were more likely to demonstrate knee ROA, and no differences were observed in the proportions of hip ROA for either sex. Men with any knee pain demonstrated greater mean leg lean mass, whereas women with any knee pain had lower mean knee extension strength and whole leg strength. For both sexes, those who reported any knee pain had lower mean LMQ. Figure 1 shows LMQ at baseline categorized by groups who reported no pain, pain at 1 site only, or pain at 2 or more anatomic sites. This graph demonstrates that for women there was a significant trend for decreasing LMQ with increasing number of pain sites (P = 0.01), whereas no association was observed for men (P = 0.30). These comparisons were unadjusted for covariates.
Table 1. Values for descriptive characteristics among Tasmanian Older Adult Cohort Study participants at baseline, stratified by sex and self-reported knee pain*
| ||Men (n = 352)||Women (n = 357)|
|No knee pain (n = 197)||Knee pain (n = 155)||P||No knee pain (n = 201)||Knee pain (n = 156)||P|
|Age, years||63.0 ± 7.3||62.0 ± 7.2||0.192||62.0 ± 7.0||61.7 ± 7.5||0.638|
|Height, cm||173.8 ± 6.3||174.4 ± 6.3||0.355||160.8 ± 6.0||161.3 ± 6.4||0.403|
|Weight, kg||81.6 ± 11.9||85.8 ± 12.5||< 0.001†||69.7 ± 11.6||73.2 ± 14.2||0.008†|
|BMI, kg/m2||27.0 ± 3.5||28.2 ± 3.8||0.001†||27.0 ± 4.4||28.2 ± 5.6||0.018†|
|Body fat, %||27.2 ± 4.4||28.0 ± 5.2||0.073||39.0 ± 5.0||40.1 ± 5.5||0.046†|
|PA, steps × 10−3/day||8.9 ± 3.7||9.3 ± 3.7||0.338||9.1 ± 3.6||8.2 ± 3.4||0.018†|
|Neck pain, %||32||50||< 0.001‡||40||63||< 0.001‡|
|Shoulder pain, %||29||47||< 0.001‡||34||60||< 0.001‡|
|Hand pain, %||22||54||< 0.001‡||33||62||< 0.001‡|
|Back pain, %||48||72||< 0.001‡||52||76||< 0.001‡|
|Hip pain, %||25||47||< 0.001‡||29||60||< 0.001‡|
|Foot pain, %||22||45||< 0.001‡||23||55||< 0.001‡|
|Knee ROA, %||59||73||0.003‡||69||76||0.148|
|Hip ROA, %||49||46||0.537||49||55||0.284|
|Leg lean mass, kg||18.6 ± 2.7||19.3 ± 2.7||0.009†||13.8 ± 2.6||14.1 ± 2.7||0.307|
|Knee extension strength, kg||37.8 ± 8.8||36.2 ± 9.9||0.086||25.5 ± 8.2||22.8 ± 9.2||0.003†|
|Whole leg strength, kg||131.2 ± 39.7||127.7 ± 45.0||0.430||63.7 ± 25.7||55.3 ± 24.9||0.002†|
|LMQ, kg/kg||7.0 ± 1.9||6.6 ± 2.1||0.037†||4.7 ± 1.9||4.0 ± 1.9||0.001†|
|Falls risk score||0.3 ± 0.8||0.3 ± 0.8||0.954||0.2 ± 0.7||0.3 ± 1.0||0.126|
Figure 1. Mean ± SE leg muscle quality (LMQ) for men and women classified with no pain, pain at 1 site, and pain at 2 or more sites at baseline. * = significant trend within sex.
Download figure to PowerPoint
Differences in change in muscle parameters over 2.6 years according to baseline pain status are shown in Table 2. The reported data are multivariable regression coefficients (95% CIs). Coefficients represent the difference in mean change of the outcome variable for those who reported pain, stiffness, or dysfunction, compared to those who did not. In women, any knee pain predicted a greater decline in knee extension strength, whole leg strength, LMQ, and falls risk. Also, for women only, severe knee pain, stiffness, and dysfunction predicted greater declines in knee extension strength, and severe knee stiffness and dysfunction predicted greater declines in falls risk score. Hip pain predicted a greater decline in knee extension strength in women, but no differences were observed for changes in muscle parameters with baseline pain at any other anatomic site (data not shown). Additionally, no differences in change in muscle parameters were observed for men who reported pain at any anatomic site, or stiffness or dysfunction in the knees, compared to those who did not (data not shown).
Table 2. Multivariable regression coefficients expressing sex-stratified differences in change in lower extremity lean mass, function, and falls risk over 2.6 years according to baseline joint pain, stiffness, and dysfunction status*
| ||Leg lean mass, kg||Knee extension strength, kg||Whole leg strength, kg||LMQ, kg/kg||Falls risk score|
|Men|| || || || || |
| Any knee pain||0.20 (−0.03, 0.42)||1.13 (−0.31, 2.57)||2.46 (−3.70, 8.60)||−0.02 (−0.37, 0.33)||−0.02 (−0.16, 0.11)|
| Any hip pain||−0.05 (−0.29, 0.18)||0.72 (−0.80, 2.24)||2.52 (−4.00, 9.04)||0.17 (−0.20, 0.53)||−0.14 (−0.28, 0.03)|
| Severe knee pain†||0.15 (−0.12, 0.43)||−0.57 (−2.36, 1.22)||2.08 (−5.83, 9.98)||−0.08 (−0.53, 0.37)||−0.09 (−0.26, 0.08)|
| Severe knee stiffness†||0.18 (−0.09, 0.46)||−0.66 (−2.46, 1.14)||0.60 (−7.27, 8.47)||−0.09 (−0.54, 0.35)||−0.02 (−0.19, 0.15)|
| Severe knee dysfunction†||0.12 (−0.16, 0.39)||0.31 (−1.79, 1.85)||5.65 (−2.08, 13.38)||0.16 (−0.27, 0.60)||−0.03 (−0.20, 0.14)|
|Women|| || || || || |
| Any knee pain||0.01 (−0.16, 0.26)||−1.94 (−3.37, −0.51)‡||−4.76 (−9.43, −0.12)‡||−0.41 (−0.76, −0.06)‡||0.18 (0.03, 0.34)‡|
| Any hip pain||0.06 (−0.15, 0.27)||−1.53 (−2.95, −0.11)‡||−0.85 (−5.47, 3.77)||−0.12 (−0.47, 0.23)||0.08 (−0.08, 0.23)|
| Severe knee pain†||−0.02 (−0.26, 0.23)||−2.47 (−4.13, −0.81)‡||−3.13 (−8.79, 2.53)||−0.23 (−0.66, 0.20)||0.09 (−0.10, 0.27)|
| Severe knee stiffness†||0.21 (−0.06, 0.46)||−2.20 (−3.97, −0.43)‡||−1.29 (−7.38, 4.79)||−0.24 (−0.71, 0.22)||0.26 (0.07, 0.45)‡|
| Severe knee dysfunction†||0.08 (−0.17, 0.32)||−2.85 (−4.53, −1.17)‡||−1.78 (−7.47, 3.91)||−0.27 (−0.70, 0.17)||0.20 (0.02, 0.38)‡|
Multivariable regression analyses adjusting for confounders revealed a higher total knee ROA score at baseline predicted, and surprisingly, a greater increase in leg lean mass over 2.6 years in both sexes (men: 0.06 kg per unit increase in knee ROA score; P = 0.004 and women: 0.048 kg per unit increase in knee ROA score; P = 0.017). No other significant associations were observed between baseline ROA and change in muscle parameters (data not shown).
Figure 2 demonstrates the hypothesized links between knee ROA and change in LMQ in women, and provides standardized beta coefficients estimated by path analysis. As no associations were observed between pain and muscle parameters in men, this analysis was conducted only among women. Path analysis revealed that knee ROA scores had direct significant positive effects on knee pain, stiffness, and dysfunction assessed by the WOMAC. Knee pain and stiffness also had direct significant positive effects on dysfunction, but no significant direct effect on change in LMQ. Knee dysfunction alone had a significant direct negative effect on change in LMQ in older women.
Figure 2. Path analysis examining the hypothesized links between knee radiographic osteoarthritis (ROA) and reduced lower extremity muscle function in older women. Bold coefficients are significant. LMQ = leg muscle quality.
Download figure to PowerPoint
- Top of page
- SUBJECTS AND METHODS
- AUTHOR CONTRIBUTIONS
The primary finding of this study is that any knee and hip pain, and more severe knee pain, stiffness, and dysfunction, reported by older women predicts greater declines in lower extremity muscle strength and quality, and greater increases in falls risk over 2.6 years. No associations were observed for pain measures in older men, however.
To the best of our knowledge, this is the first prospective cohort study of community-dwelling older adults that directly examines the influence of pain and ROA on muscle mass, strength, and quality and falls risk. The musculoskeletal conditions sarcopenia and OA have both been linked with increased risk of disability (21, 22) and falls (10, 11) in the aging population. The results from the present study indicate that pain may contribute further to disability and falls risk in women by exacerbating age-related muscle strength and quality declines.
The primary findings of this study suggest that self-reported pain, rather than structural change associated with OA, predicts muscle decline and falls risk in older adults, and this is supported by previous research. Disability assessed by the WOMAC index in approximately 70 hospital outpatients with symptomatic knee OA was not associated with radiographic change, but was strongly related to pain severity (23). Also, quadriceps strength, knee pain, and age were more important determinants of disability measured using the Stanford Health Assessment Questionnaire than the severity of knee ROA in approximately 160 adults over the age of 55 years (24). In more than 5,000 elderly women from the Study of Osteoporotic Fractures followed for 7 years, those with self-reported physician-diagnosed OA had an increased risk of falls, yet those with ROA actually demonstrated a reduced risk of recurrent falls in the first year (25). A cross-sectional analysis of 850 TASOAC participants has also demonstrated that falls risk is associated with WOMAC-assessed pain and dysfunction, but not ROA (13). The present study adds to this area by demonstrating that over nearly 3 years in women, lower extremity joint pain, stiffness, and dysfunction (but not ROA) are associated with declines in muscle parameters and increases in objectively-measured falls risk scores in older adults.
McAlindon et al (24) have suggested that pain may be a mechanism by which OA leads to functional decline, and our findings support this premise. Our path analysis performed in women revealed that knee ROA had a direct effect on self-reported knee pain, stiffness, and dysfunction, and knee dysfunction was subsequently associated with a deleterious change in LMQ over 2.6 years. These results may therefore suggest that self-reported knee pain, stiffness, and dysfunction mediate the relationship between OA and poor muscle strength. This conclusion is supported by a cross-sectional analysis of more than 500 men and women with a mean age of approximately 55 years, which demonstrated that significant associations of hand ROA with grip strength and function were mediated by pain (26), and the Observational Arthritis Study in Seniors study also identified that knee pain intensity mediated the association between knee ROA and change in performance over 30 months (27). It is worth noting that in the present study, women with knee pain were no more likely than those without knee pain to have knee ROA, suggesting that other causes of joint pain may contribute to muscle declines.
It is unclear why joint pain may contribute to reductions in muscle strength. This may be a result of reluctance on the behalf of the participant to perform a maximal voluntary contraction due to expectation of pain experience. However, it has also been suggested that arthrogenous muscle inhibition occurs because changes in afferent input from the affected joint result in reduced efferent motor neuron stimulation of the attendant skeletal muscles (7). A case–control study has demonstrated that patients with symptomatic knee OA are more likely to have quadriceps activation failure than those without OA (28). This explanation is supported to some extent by the fact that change in leg lean mass did not differ for those with and without pain, and was actually greater for those with higher total knee ROA scores at baseline.
It was unexpected that leg lean mass would increase more for those with a higher baseline knee ROA score in this cohort. Due to the multiple comparisons performed in this study, it is possible that this may represent a chance finding. However, this observation may be linked to an infiltration of fatty acids into the skeletal muscles in those with higher total fat mass (29), which is likely in those with ROA; DXA does not entirely exclude intramuscular fat from measures of lean mass (30), and so it is possible that apparent increases in lean mass were in fact a result of infiltration of fat into the lower extremity muscle groups. Alternatively, this association may reflect that increased joint damage may be present in those who are more active (31), and these older adults are likely to demonstrate muscle hypertrophy compared to less active counterparts. Given that this finding was based on the total knee ROA score, it is also possible that reductions in lean mass occur in the early stages of structural joint changes, with increases in lean mass following adaptations to pain and stiffness. Regardless of the mechanism, the apparent improved leg lean mass in those with higher knee ROA scores was not associated with any improvement in muscle strength.
Nevertheless, the results from this study highlight the substantial discordance between observed ROA and self-reported pain (32, 33), and this may be related in some part to the inclusion of JSN in ROA assessments. The fact that cartilage is aneural would suggest that JSN is unlikely to contribute to pain development in OA, and further research must examine whether different criteria for diagnosing ROA may be more predictive of future pain development and declines in muscle parameters.
Associations between OA, muscle parameter, and falls risk changes differ for men and women. One explanation may be that women with knee OA have been shown to report higher levels of self-reported pain and physical disability than men with similar OA severity (34), and as such our findings may reflect different responses to the pain experience. Future research on sex differences in the experience of pain may open pathways for preventing OA-associated muscle declines. Furthermore, associations of lower extremity muscle strength with pain and stiffness at different sites should be investigated. Baseline knee pain was a stronger predictor of muscle parameter declines than hip pain, probably due to the major role of the quadriceps muscles in our assessments, whereas severe knee stiffness, but not knee pain, predicted a greater increase in falls risk. It is possible that knee stiffness contributes more to postural instability than knee pain.
This study has several limitations. First, the response rate for the TASOAC study was somewhat low (57%), and although the retention of participants at followup was high (82%), those lost to followup were significantly older, had lower knee extension strength, and were more likely to report knee pain. Therefore, the associations between pain, ROA, muscle parameters, and falls risk may differ for older adults who demonstrate poorer health. Second, only limited data were collected regarding therapies for OA, which may have played a role in improvements in physical function. Third, the reproducibility of radiograph reading was good rather than excellent, and this may explain the lack of associations observed between ROA and change in muscle parameters and falls risk. Fourth, there is currently no “gold standard” of measurement for LMQ, so the DXA method used in this study has not been assessed against such a standard. Finally, these results may not be generalized to institutionalized older adults or those over the age of 80 years.
In conclusion, self-reported knee and hip pain, and knee stiffness and dysfunction, are associated with declines in lower extremity muscle strength and quality and increases in falls risk over 2.6 years in older women. Further research is required to examine whether these associations exist for men, and to determine whether ROA predicts future changes in muscle parameters and risk of falls.