To test the hypothesis that the number of areas of musculoskeletal pain reported is related to incident disability.
To test the hypothesis that the number of areas of musculoskeletal pain reported is related to incident disability.
Subjects included 898 older persons from the Rush Memory and Aging Project without dementia, stroke, or Parkinson's disease at baseline. All participants underwent detailed baseline evaluation of self-reported pain in the neck or back, hands, hips, knees, or feet, as well as annual self-reported assessments of instrumental activities of daily living (IADLs), basic activities of daily living (ADLs), and mobility disability. Mobility disability was also assessed using a performance-based measure.
The average followup was 5.6 years. Using a series of proportional hazards models that controlled for age, sex, and education, the risk of IADL disability increased by ∼10% for each additional painful area reported (hazard ratio [HR] 1.10, 95% confidence interval [95% CI] 1.01–1.20) and the risk of ADL disability increased by ∼20% for each additional painful area (HR 1.20, 95% CI 1.11–1.31). The association with self-report mobility disability did not reach significance (HR 1.09, 95% CI 0.99–1.20). However, the risk of mobility disability based on gait speed performance increased by ∼13% for each additional painful area (HR 1.13, 95% CI 1.04–1.22). These associations did not vary by age, sex, or education and were unchanged after controlling for several potential confounding variables including body mass index, physical activity, cognition, depressive symptoms, vascular risk factors, and vascular diseases.
Among nondisabled community-dwelling older adults, the risk of disability increases with the number of areas reported with musculoskeletal pain.
Musculoskeletal disorders are common and among the major causes of chronic pain in older adults (1). People age ≥65 years represent the fastest growing segment of the US population, which underscores the importance of determining the association of musculoskeletal pain with the development of disability in the elderly (2). Both cross-sectional and longitudinal studies have reported a link between musculoskeletal pain and disability in the elderly (1, 3, 4). However, few longitudinal studies have examined whether the likelihood of developing disability increases as more areas with musculoskeletal pain are reported.In this study, we used data from 898 older participants of the Rush Memory and Aging Project, a longitudinal study of common chronic conditions of aging (5), to test the hypothesis that the risk of developing disability increases as more areas with musculoskeletal pain are reported.
All participants were from the Rush Memory and Aging Project, a longitudinal clinical-pathologic investigation of chronic conditions of old age (5), and all signed an informed consent agreeing to an annual clinical evaluation. In addition, study participation required that participants sign an anatomic gift act, donating, at the time of death, their entire brain and spinal cord, as well as selected nerves and muscles, to Rush investigators. The study was in accordance with the latest version of the Declaration of Helsinki and was approved by the Rush University Medical Center Institutional Review Board.
At the time of these analyses, 1,221 participants had enrolled and completed a baseline evaluation. Eligibility for these analyses required the following: 1) a valid baseline assessment of musculoskeletal pain and disability, 2) the absence of a diagnosis of dementia, stroke, or Parkinson's disease at baseline, and 3) at least 1 followup disability assessment in order to assess incident disability. There were 1,185 persons with a valid musculoskeletal pain and disability assessment at baseline. We excluded 72 persons who met the criteria for dementia at baseline, 120 with previous stroke, and 14 with Parkinson's disease. Also excluded were 28 persons who had completed a baseline evaluation but died before their first followup examination, as well as 30 persons who had not been in the study long enough for followup evaluation. Of the 921 participants who were eligible for followup, there were 898 (97.5%) with complete followup data who were included in these analyses.
Clinical diagnoses were made using a multistep process (5). First, subjects underwent detailed annual cognitive function testing that included 19 cognitive performance tests (see below). Second, the cognitive test data were reviewed by an experienced neuropsychologist who determined if cognitive impairment was present. Next, participants were evaluated in person by an experienced clinician who diagnosed dementia, stroke, or Parkinson's disease using previously published criteria (5).
Musculoskeletal pain was assessed at baseline by asking the participants if they had experienced pain or aching in any of their joints on most days for at least 1 month during the prior year. Individuals who answered affirmatively about having had musculoskeletal pain were then questioned whether they had pain or aching on most days for at least 1 month during the past year in 5 areas including the following: back or neck, hands, hips, knees, or feet. In these analyses, we used the number of areas reported to be painful.
Disability was assessed annually through 3 self-report instruments (5). Instrumental activities of daily living (IADLs) were assessed using items adapted from the Duke Older Americans Resources and Services project, which assesses 8 activities: telephone use, meal preparation, money management, medication management, light and heavy housekeeping, shopping, and local travel. Basic activities of daily living (ADLs) were assessed using a modified version of the Katz Basic ADL scale, which assesses 6 activities: feeding, bathing, dressing, toileting, transferring, and walking across a small room. Mobility disability was assessed using the Rosow-Breslau Functional Health scale, which assesses 3 activities: walking up and down a flight of stairs, walking one-half of a mile, and doing heavy housework like washing windows, walls, or floors. Participants were given the following response choices with regard to their ability to perform each of the above activities: no help, help, unable to do. For these analyses, participants who reported needing help with or an inability to perform 1 or more tasks on each of the 3 scales were classified as being disabled.
Based on the participant's annual gait speed when requested to walk 8 feet, mobility disability was considered present if walking speed was <0.55 meter/second, as previously described in this cohort (6).
Demographic information including date of birth, sex, and years of education were collected through a participant interview. Body mass index (BMI) was determined by dividing measured weight represented in kilograms with the square of measured height represented in meters. Physical activity was assessed using questions adapted from the 1985 National Health Interview Survey. The minutes spent engaged in each of 5 activities were summed and expressed as hours of activity per week, as previously described (5).
A composite measure of cognitive function based on 19 tests was used in these analyses: immediate and delayed recall of story A from Logical Memory (7), immediate and delayed recall of the East Boston Story (8, 9), Word List Memory, Word List Recall, Word List Recognition (10), a 15-item version of the Boston Naming Test (11), Verbal Fluency (10), a 15-item reading test (9), Digit Span Forward, Digit Span Backward (7), Digit Ordering (12), Symbol Digit Modalities Test (13), Number Comparison (14), two indices from a modified version of the Stroop Neuropsychological Screening Test (15), a 15-item version of Judgment of Line Orientation (16), and a 16-item version of Standard Progressive Matrices (17). One additional test, Complex Ideational Material (18), was used only for diagnostic classification purposes. To compute the composite measure of cognitive function, raw scores on each of the individual tests were converted to Z scores using the baseline mean and SD of the entire cohort, and the Z scores of all 19 tests were averaged. Psychometric information on this composite measure is contained in previous publications (5).
Depressive symptoms were assessed with a 10-item version of the Center for Epidemiologic Studies Depression scale (5). We summarized the number of 3 vascular risk factors: hypertension, diabetes mellitus, and smoking, as well as the number of 3 vascular diseases: myocardial infarction, congestive heart failure, and claudication, as previously described (5).
Spearman's rank correlations were used to examine the bivariate associations between the number of areas of reported musculoskeletal pain and demographic variables and other covariates at baseline. Two sample t-tests (with Satterthwaite's method adjustments for unequal variances, when appropriate) and nonparametric tests (as appropriate) were used to compare the baseline characteristics of participants with and without musculoskeletal pain. Next we examined the associations between the 5 areas that were assessed for pain. Since 1 or more latent variables might underlie pain in multiple areas, we used tetrachoric correlations to estimate the correlations of latent normally distributed variables from the two-way tables for each pair of joints. The core analysis employed a Cox proportional hazards model to examine whether the number of painful areas was related to incident self-report disability. All models controlled for age, sex, and education. In subsequent analyses, we added interaction terms for age, sex, and education to examine whether the associations of musculoskeletal pain with risk of disability varied with demographic characteristics. We then added terms for a number of potential confounders that might affect the association of musculoskeletal pain and disability. We employed both linear and quadratic terms for BMI, since both high and low BMI may be associated with adverse health outcomes. In a final model, we examined whether the number of painful areas predicted incident mobility disability based on gait speed performance (6). A priori level of statistical significance was 0.05. All models were validated graphically and analytically. Analyses were programmed in SAS, version 9.1.3 for LINUX (SAS Institute) (19).
There were 898 persons in these analyses with a mean ± SD followup period of 5.6 ± 2.4 years and their baseline characteristics are summarized in Table 1. The mean tetrachoric correlation for all pairs of the 5 areas assessed for pain was 0.65. The mean ± SD number of areas with pain was 0.959 ± 1.41 (range 0–5), indicating that most participants had no pain or only 1 painful area. Approximately 25% of the participants reported multiple painful areas (Figure 1). The presence of musculoskeletal pain in one area was associated with an increased risk of pain in other areas (mean odds ratio 9.0, range 6.8–13.1). For example, reporting pain in the back increased the odds of also reporting pain in the hip by 13-fold (Table 2).
|Variable||No pain (n = 528)||Pain (n = 370)||P†|
|Age, mean ± SD years||79.7 ± 7.4||79.7 ± 6.9||0.985|
|Male, no. (%)||172 (33)||59 (16)||<0.001|
|Education, mean ± SD years||14.7 ± 3.1||14.2 ± 3.1||0.019|
|Mini-Mental State examination, mean ± SD‡||27.9 ± 2.1||28.0 ± 1.9||0.441|
|Body mass index, mean ± SD kg/m2||26.3 ± 4.5||28.3 ± 6.2||<0.001|
|Physical activity, mean ± SD hours/week§||3.4 ± 3.9||2.8 ± 2.7||0.006|
|Global cognition, mean ± SD¶||0.08 ± 0.53||0.08 ± 0.51||0.896|
|Depressive symptoms, mean ± SD#||0.9 ± 1.5||1.8 ± 2.0||<0.001|
|No. vascular risk, mean ± SD||1.1 ± 0.8||1.2 ± 0.8||0.062|
|No. vascular diseases, mean ± SD||0.20 ± 0.5||0.26 ± 0.5||0.068|
|Congestive heart failure||3||7||0.024|
The number of reported areas of musculoskeletal pain was not associated with age (rS = −0.02, P = 0.487) but was inversely related to education (rS = −0.08, P = 0.013) and women reported more areas of musculoskeletal pain (rS = −0.20, P < 0.001). Musculoskeletal pain was associated with a higher BMI (rS = 0.16, P < 0.001), less physical activity (rS = −0.07, P = 0.037), and more depressive symptoms (rS = 0.25, P < 0.001). There were trends for vascular risk factors (rS = 0.06, P = 0.07) and vascular diseases (rS = 0.06, P = 0.07), but pain was not related to cognition (rS = 0.002, P = 0.946).
In the first analysis testing the hypothesis that musculoskeletal pain is associated with the risk of developing disability in IADLs, we restricted the analysis to the 485 of 898 persons (54%) who reported no disability in IADLs at baseline. Over a mean ± SD followup period of 5.8 ± 2.5 years, 295 of 485 persons (60.8%) reported impairment in IADLs. In a proportional hazards model which controlled for age, sex, and education, and examining the hazard ratio (HR) for this model, each additional area of musculoskeletal pain was associated with ∼10% greater risk of developing IADL disability. The percentage of participants developing disability increased with increasing areas of pain, but the percentage was similar for 1 or 2 areas of pain and 3 or more areas (Table 3).
|0||56.2 (181/322)||30.0 (147/490)||54.4 (190/349)||51.8 (192/371)|
|1 or 2||69.8 (74/106)||37.17 (75/202)||61.1 (69/113)||62.5 (95/148)|
|3 or more||70.2 (40/57)||43.8 (53/121)||57.1 (32/56)||61.3 (49/80)|
Alternatively, since the baseline age in this model was also associated with incident IADL disability (HR 1.10, 95% CI 1.01–1.20), we can compare the size of the effect of the increased risk of IADL disability associated with each additional area of musculoskeletal pain by comparing this increased risk with a more familiar metric: the risk of IADL disability associated with increasing baseline age. For example, each additional area of musculoskeletal pain at baseline was associated with an equivalent risk of IADL disability associated with a participant being ∼1.3 years older at baseline (areas of musculoskeletal pain, estimate, 0.093/age, estimate, 0.069 = 1.3 years). The association of more areas of musculoskeletal pain and incident IADL disability did not vary by age, sex, or education (data not shown). The association of musculoskeletal pain and incident IADL disability were unchanged after adjustment for several covariates (Table 4).
|Age, sex, education||1.10 (1.01–1.20)||1.20 (1.09–1.31)||1.09 (0.99–1.20)||1.13 (1.04–1.22)|
|Body mass index, kg/m2†||1.08 (0.99–1.18)||1.17 (1.07–1.27)||1.05 (0.95–1.15)||1.10 (1.02–1.20)|
|Physical activity‡||1.09 (1.00–1.19)||1.19 (1.10–1.30)||1.08 (0.99–1.19)||1.12 (1.03–1.21)|
|Cognition§||1.09 (1.00–1.19)||1.21 (1.11–1.31)||1.08 (0.99–1.19)||1.13 (1.05–1.22)|
|Depressive symptoms¶||1.07 (0.98–1.18)||1.19 (1.09–1.29)||1.06 (0.97–1.17)||1.09 (1.01–1.19)|
|No. vascular diseases#||1.09 (1.00–1.19)||1.20 (1.11–1.31)||1.08 (0.98–1.18)||1.12 (1.03–1.21)|
|No. vascular risk factors**||1.09 (1.00–1.19)||1.20 (1.11–1.31)||1.08 (0.98–1.19)||1.13 (1.05–1.22)|
This analysis was restricted to the 813 of 898 persons (91%) who reported no ADL impairment at baseline. Over a mean ± SD followup period of 5.6 ± 2.4 years, 275 of 813 persons (33.8%) reported ADL impairment. In a proportional hazards model which controlled for age, sex, and education, each additional area of musculoskeletal pain was associated with ∼20% greater risk of developing ADL disability, which was equivalent to the risk of developing Katz disability associated with a participant being ∼2.0 years older at baseline. The percentage of participants developing disability increased with increasing areas of pain as shown in Table 3 and illustrated in Figure 2. Next, we repeated the core model described above with additional terms for possible interactions with demographic variables. The association of musculoskeletal pain and ADL disability did not vary by age, sex, or education (data not shown). The increased risk of ADL disability with more areas of musculoskeletal pain persisted after adjustment for several covariates (Table 4).
Next, we examined the association of musculoskeletal pain with the risk of developing mobility disability. This analysis was restricted to the 518 of 898 persons (57.6%) who reported no mobility disability at baseline. Over a mean ± SD followup period of 5.8 ± 2.5 years, 291 persons (56.2% of 518) reported mobility disability. There was a trend for an increased risk of self-report mobility disability for each additional painful area reported (Table 4). The association of more areas of musculoskeletal pain with incident mobility disability was essentially unchanged even after adjustment for several covariates (Table 4).
We next examined the association of musculoskeletal pain with risk of mobility disability based on physical-performance measures, which may be more sensitive compared with self-report measures. There were 599 of 898 participants (66.7%) without mobility disability at baseline based on a gait speed of >0.55 meter/second. Over a mean followup period of 5.8 years, 336 of 599 persons (56.1%) developed mobility disability. Each additional area of musculoskeletal pain was associated with ∼13% greater risk of developing mobility disability, which was equivalent to the risk of developing mobility disability associated with a participant being ∼2.2 years older at baseline. The percentage of participants developing disability increased with the increasing areas of pain, but the percentage developing disability was similar for both 1 or 2 areas of pain and 3 or more areas of pain (Table 3). The estimate for this model was similar to the estimate from the model based on self-reported mobility disability (0.120 for performance-based versus 0.082 for self-reported). The increased risk of mobility disability with more areas of musculoskeletal pain persisted even after adjustment for several covariates (Table 4).
In this cohort of ∼900 community-dwelling older adults without dementia, stroke, or Parkinson's disease, the risk of developing incident IADL, ADL, and mobility disability increased as more areas with musculoskeletal pain were reported. These associations did not vary by age, sex, or education, and they remained significant after controlling for several possible confounders including BMI, physical activity, cognition, depressive symptoms, vascular risk factors, and vascular diseases. These results extend the prior literature and suggest that musculoskeletal pain may be a modifiable risk factor for decreasing the burden of disability in community-dwelling older adults. Further, they suggest that the number of musculoskeletal areas affected by pain may have prognostic significance.
There has been an increased awareness and understanding of the growing burden of musculoskeletal conditions both in the elderly and to society. Studies that have focused on pain in a particular body area, such as low back pain or lower extremity pain, report that both are associated with a subsequent decline in physical performance (20–22). There have been few longitudinal studies that have examined the contribution of musculoskeletal pain to the development of disability in community-dwelling older adults (23). A recent longitudinal study reported that the number of locations and severity of chronic musculoskeletal pain is associated with risks of falling (24). Pain in multiple areas has been reported to be associated with an increased risk of progression from mild to severe ADL disability and mobility disability in women (3, 25). The current study extends these prior studies by showing that more widespread musculoskeletal pain is associated with an increased risk of developing both IADL and Katz disability. Although the association of musculoskeletal pain and the development of self-reported mobility disability showed a trend for significance, the effect size for this association was similar to the association obtained in a model for incident performance-based mobility disability corroborating this association. A similar percentage of participants with 1 or 2 areas of pain and 3 or more areas of pain subsequently developed IADL and mobility disability (Table 3). Further work is needed to examine whether there may be a threshold effect for pain and the subsequent development of different disabilities.
The current study suggests that the presence of musculoskeletal pain may be an important clinical marker for identifying older adults at increased risk for developing a wide range of disabilities. The results of the current study have important translational implications since they suggest that the public health efforts to encourage lifestyle changes and interventions that would ameliorate pain might increase the efficacy of efforts to decrease the burden of disabilities in our rapidly aging population.
Musculoskeletal conditions vary with regard to their pathophysiology, but they are linked anatomically through the structural changes in the bones, joints, and muscles, leading to long term pain and disability in adults. The basis for the association between musculoskeletal pain and disability is uncertain but is likely multifactorial. Prior studies have shown that radiologic or physical joint changes are not as robustly associated with disability as reports of pain, suggesting that other factors are involved in the association (26). While disability due to musculoskeletal pain is often ascribed to osteoarthritis in a specific joint (e.g., the knee), recent work suggests that pain is multifaceted and complex (27–29). For example, the pain matrix in joint or spine disease constitutes an interaction between structural pathology, neural innervation of the joint (sensory, motor, autonomic), dysfunction of pain processing at spinal and cortical levels, and various environmental and individual determinants (i.e., affective, cognitive). Therefore, pain reflects the integrated result of multiple biologic and psychosocial inputs that are not fully understood (28, 29). Our finding that musculoskeletal pain was associated with the development of a wide range of disabilities suggests that pain may affect changes in the central nervous system pathways that mediate both motor and nonmotor functions. Further work is needed to explicate the biology of the link between pain and disability.
The current study has several limitations. The selected nature of the cohort, which included participants willing to provide organ donation and the low prevalence of musculoskeletal pain, underscores the need for replication of these results in a population-based study. Although the current study provides important longitudinal data about pain and incident disability, it is an observational study so causal inferences are limited. Therefore we cannot determine whether musculoskeletal pain causes disability or whether both share a common underlying pathophysiology. Pain is a multidimensional construct, and the current study assessed the number of areas with musculoskeletal pain during a 1-month period. Further studies are needed to examine the contributions of other dimensions of pain including chronicity, frequency, and severity. Finally, we did not have objective-measures physical examination, radiologic, or pathology findings that would provide more precise characterization and classification of the causes of pain.
Our study is strengthened by having a large cohort in which simultaneous measures of musculoskeletal pain and disability were obtained, by the longitudinal design, and by our ability to control for important covariates, especially detailed cognitive function. Our use of a community-based sample rather than a clinic-based sample reduces one type of selection bias. Further, we excluded persons with common neurologic conditions that cause disability, and we considered several covariates that may affect the relationship of musculoskeletal pain and mobility disability. This study suggests that in ambulatory community-dwelling older adults without disability, the number of regions affected with musculoskeletal pain is related to the subsequent risk of developing disability. Public health programs to encourage lifestyle changes and ameliorate musculoskeletal pain might decrease the burden of disability in our aging population.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Buchman had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Buchman, Shah, Leurgans, Boyle, Wilson, Bennett.
Acquisition of data. Buchman, Shah, Wilson, Bennett.
Analysis and interpretation of data. Buchman, Shah, Leurgans, Boyle, Wilson.
We thank all the participants in the Rush Memory and Aging Project. We also thank Traci Colvin, Sandra McCain, and Tracey Nowakowski for project coordination; Barbara Eubeler, Mary Futrell, Karen Lowe Graham, and Pam A. Smith for participant recruitment; John Gibbons and Greg Klein for data management; Wenqing Fan, MS, for statistical programming; and the staff of the Rush Alzheimer's Disease Center.