The Relationship Between Pain, Neuropsychological Performance, and Physical Function in Community-Dwelling Older Adults with Chronic Low Back Pain

ABSTRACT

Objective. Chronic pain and cognitive impairment are prevalent and disabling in older adults (OA), but their interrelationship has not been rigorously tested. We did so in OA with chronic low back pain (CLBP).

Design. A total of 323 OA (160 pain-free, 163 CLBP; mean age 73.5 years, 45% female) had neuropsychological (NP) testing with the Repeatable Battery for the Assessment of Neuropsychological Status, Trail Making Test, and the Grooved Pegboard Test. Pain intensity was measured with the McGill Pain Questionnaire Short Form. Physical performance (gait speed, functional reach, chair rise, trunk rotation, and static/dynamic lifting), psychosocial disruption (Geriatric Depression Scale, the Short Form-36 Mental Health and Role Limitations-Emotional Composite scale), and self-reported disability (Functional Status Index, the Short From-36 Physical Functioning/Role-Physical Composite scale) were also measured.

Setting. Outpatient research laboratory.

Results. There were no group differences in age, gender, or educational level, but significant differences in NP scores (P = 0.01) were found. Five scales accounted for the differences: immediate memory (P = 0.002), language (P = 0.004), delayed memory (P = 0.04), mental flexibility (Trails B [P = 0.02]), and Grooved Pegboard (P = 0.05). NP scores were significantly correlated with physical performance (R² = 0.30, P < 0.001), but not self-reported disability (R² = 0.04, P = 0.52) or psychosocial disruption (R² = 0.05, P = 0.46). NP function was correlated with pain intensity (R² = 0.17, P < 0.001), and NP function mediated the relationship between pain and physical performance.

Conclusions. OA with CLBP demonstrated impaired NP performance as compared with pain-free OA. Further, pain severity was inversely correlated with NP performance, and NP performance mediated the relationship between pain and physical performance. Future research should examine whether cognitive function and impaired physical performance can be improved with pain reduction.

Low Back Pain, Chronic Pain, Cognitive Function, Neuropsychological Performance, Physical Performance

Introduction

Physical disability is one of the most feared consequences of aging, and can be contributed to by a wide range of factors, including cognitive impairment [1,2], psychosocial disruption [3,4], and physical disorders that directly impact function, such as cerebrovascular accidents, hip fractures, and arthritis [1]. Because pathology in each of these domains (i.e., cognitive, psychosocial, and physical) can independently lead to functional decline, disorders that impact all three raise an especially prominent red flag in the minds of geriatrics practitioners. Chronic nonmalignant pain, estimated to exist in at least 50% of community-dwelling older adults (OA) [5], is one such disorder. Numerous studies have demonstrated an association between chronic nonmalignant pain and physical disability [6–13], as well as depression [9,14–18]. While medications used to treat chronic pain can result in cognitive disruption (e.g., opioid-induced delirium [19]), the relationship between chronic pain itself and cognitive function has been less well studied.

A number of studies have indicated that acute pain disrupts cognitive function. Experimentally induced acutely painful stimuli have been shown to impair attention [20,21]. Clinical studies examining the relationship between pain and cognitive function in OA have been performed in the postoperative setting. While a number of postoperative factors have been shown to contribute to delirium, such as infection, electrolyte disturbances, and hypoxia [22], multiple studies have demonstrated the importance of pain itself as a cause of postoperative cognitive dysfunction. Morrison and colleagues demonstrated a cause and effect relationship between pain and delirium in older surgical patients [23] and between higher doses of morphine and improved neuropsychological (NP) function in these individuals [23]. Duggleby and Lander suggested that pain following total hip replacement in patients aged 50–80 years was a strong predictor of mental status decline in the postoperative period [24]. Lynch and colleagues, in a study of 361 patients (mean age 67 years) undergoing elective noncardiac surgeries, found that pain was an independent risk factor for the development of delirium [25]. Heyer and colleagues have recently demonstrated that postoperative pain, not duration of surgery or dose/type of anesthetic, predicted impaired NP performance in postoperative spinal surgery patients over the age of 60 years [26].

Studies examining the relationship between chronic pain and NP function have been performed primarily in younger adults. Eccleston has conducted several experimental studies in chronic pain patients that indicate high pain intensity interferes with performance of complex cognitive tasks via mechanisms of attention [27–29]. Others have shown that patients with chronic nonmalignant pain treated with long-term opioids experience improved NP performance [30,31]. Pain in cancer patients also has been shown to deteriorate NP performance more than opioids used in the course of treatment [32]. More recently, investigators have demonstrated that in younger individuals, chronic low back pain (CLBP) is associated with impaired performance on a complex emotional decision-making task, suggesting abnormal prefrontal processing [33]. Abnormal brain chemistry [34] and thalamocortical atrophy has also been uncovered in these patients [35].

Neither the relationship between NP performance and chronic pain in community dwelling OA, nor the role that NP function plays in mediating the relationship between chronic pain and physical function, has been investigated. The purpose of this study was to do so in independent, community-dwelling OA with the most common of regional musculoskeletal pain syndromes, CLBP. Specifically, we hypothesized that: 1) OA with CLBP would demonstrate impaired performance on NP tests of attention and concentration as compared with pain-free controls; 2) greater pain intensity would be associated with lower performance on these tests; and 3) NP performance would mediate the relationship between pain and impaired physical performance.

Methods

Subjects

Subjects were 323 English-speaking community-dwelling OA (age 65–84 years) with CLBP, defined as pain of at least moderate intensity (measured with the pain thermometer [36]), every day or almost every day, for at least the past 3 months (mean pain duration = 14.2 years, N = 163), or they were pain-free (N = 160). Pain-free was defined as no pain or pain occurring less than once per week of little intensity as measured with the pain thermometer. All subjects were cognitively intact and signed informed consent prior to their participation. Subjects were recruited via newspaper advertisements and tear-off fliers placed in the community and in primary care clinics. They were screened in two phases: first over the telephone (N = 2007), then on site by one of the investigators (D.K.W.) with a structured history and physical examination (N = 610). Exclusion criteria included cognitive impairment (Folstein Mini-Mental State Examination [MMSE] <21 adjusted for age and education) (N = 24), severe visual or hearing impairment, acute illness or pain, and medical conditions that could make the lifting task (see below) potentially unsafe (e.g., postural instability, severe cardiac or respiratory disease).

Subject demographics are shown in Table 1. As displayed in Table 1, pain-free and CLBP subjects were not significantly different with respect to age, gender, educational level, ethnicity, or current living situation. However, compared with pain-free subjects, CLBP subjects were found to have a significantly higher number of comorbidities (Cumulative Illness Rating Scale), lower Folstein scores, and significantly higher depression scores.

Table 1

Subject demographics

Variable	Group		P Value
	Pain-Free	CLBP
Sample size (N)	160	163	—
Age (years)
Mean	73.5	73.6	0.87
SD	4.8	5.2
Gender (N)
Male	94	83	0.16
Female	66	80
Education (%)
High school graduate	17.3	25.2	0.14
Some college (or trade school)	19.9	22.5
College graduate	62.8	52.3
Ethnicity (N)
White	142	141	0.82
African American	15	18
Hispanic	3	4
Current living situation (N)
Live alone	54	48	0.17
Live with spouse	100	99
Live with other family members	4	9
Live with others (nonfamily)	2	7
Modified Cumulative Illness Rating Scale
Mean	2.3	2.9	<0.001
SD	1.5	1.8
Folstein Mini Mental State Examination
Mean	28.7	28.3	0.01
SD	1.3	1.3
Geriatric Depression Scale
Mean	1.7	4.8	<0.001
SD	2.5	4.9
Duration of pain (years)
Mean	—	14.2	—
SD		14.6
McGill Pain Questionnaire Short Form, total score
Mean		12.2
SD		8.3

CLBP = chronic low back pain; SD = standard deviation.

Procedures

All subjects underwent the following assessments.

Pain intensity was measured with the McGill Pain Questionnaire Short Form [37] that has been validated in community-dwelling OA with CLBP [38,39].

NP function was evaluated in two phases. Subjects first underwent screening with the MMSE [40]. Those that scored <21 were excluded from further participation. Subjects included following this initial screen underwent additional testing with the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) that measures immediate and delayed memory, visuospatial skills, language, and attention [41]. The RBANS was selected because it is a broad-based measure of NP performance with relatively low respondent burden. Additionally, there is also increasing evidence of its reliability and utility with OA, and several large sample studies recently have been published that provide detailed normative information for community dwelling OA [42,43]. Mental flexibility was assessed with Part B of the Trail Making Test [44], motor coordination was assessed with the Grooved Pegboard Test (dominant hand) [45], and general intelligence was screened with the National Adult Reading Test [46]. Subjects with evidence of dementia based on the RBANS (N = 1) were excluded from further study participation, and their data were not included in the final analyses.

Opioid use data were collected at the time of the baseline-structured history and physical examination. Because only nine subjects were using opioids, these data were not analyzed further.

Psychosocial disruption was assessed with the Geriatric Depression Scale [47] and the Short Form-36 Mental Health and Role Limitations-Emotional Composite Scale [48,49].

General medical comorbidity was assessed using the Cumulative Illness Rating Scale [50], collected during the structured history and physical examination.

Physical Function/Disability Measures

Because self-reported and performance-based (i.e., observed) measures of physical function have been shown to tap distinct constructs, both were included as part of the assessment [51]. A broad range of reliable and ecologically valid performance-based measures was included, some of which were axially specific, and others were measures of general physical performance.

Observed Measures

1. A lifting task that has been validated in young and older patients with CLBP was performed on a Work Simulator (Baltimore Therapeutic Equipment [BTE] Company, Baltimore, MD, USA), with the subject standing on a force platform (AMI OR-6, MA, USA) [6,52,53]. Subjects lifted a resistive load attached to a 12-inch handle from knee to waist level, then returned the handle to the holder. The resistive load was equivalent to 40% of the subject's mean voluntary static lifting strength that was measured by calculating the mean of three trials of a static bilateral leg lift at knee level using a Chatillon Muscle Strength Dynamometer (Sammons Preston, Bolingbrook, IL, USA). The BTE Work Simulator applied resistance during the up-phase of the lift. Subjects performed the task for a maximum of 15 minutes, with a 15-second rest interval between lifts. Subjects were instructed to lift until they felt physically unable to continue or were told by the experimenter to stop. The task was terminated if: 1) maximum heart rate was reached; 2) the subject demonstrated unsafe body mechanics; 3) the subject was unable to perform at the designed pace; or 4) the time limit was reached. Physical performance measures of the lifting task included mean static lifting strength, number of lifts completed, and a work index. The work index was calculated as the weight lifted times the number of lifts performed during the lifting task.

2. Gait speed was included as a general measure of physical function because of its predictive validity for disability in OA. Twenty-five foot gait speed was assessed based on the methods of Bohannon [54].

3. Functional reach was included as a measure of balance performance [55–57].

4. Chair rise was included because of its ergonomic validity OA with CLBP. Subjects were asked to sit in a lightly padded hard-backed chair, place their arms across their chest, and stand. If successful, the subject is then asked to return to sitting and after a brief rest, repeat the sit-to-stand five times for a timed score.

5. Trunk rotation was included as a measure of spinal mobility and endurance that also has ergonomic validity because of frequent encounters with trunk rotation in the course of daily activities (e.g., retrieving objects from drawers or cabinets, dressing, and reaching items while seated in a car). This test was administered in a standardized format of 10 rotations for a timed score [58].

Self-Reported Measures

Functional Status was measured with the Difficulty subscale of the Functional Status Index, which defines the amount of difficulty experienced in performing specific activities of daily living [59], and the Short Form-36 Physical Functioning/Role-Physical Composite scale [48,49].

Statistical Analyses

Prior to conducting the primary statistical analyses, the statistical models used were evaluated for violations of heteroscedasticity of errors and nonlinearity using standard graphical methods. No measures needed to be transformed as a result of these analyses. To provide better control for experimentwise errors rates for hypotheses that evaluated group differences between pain-free and CLBP subjects on continuous measures (e.g., NP test results), a multivariate analysis of variance (manova) approach, based on the unweighted general linear model, was used. When the manovas showed significant effects, univariate anovas and canonical variates were examined to determine which variables were most important in the obtained significant difference [60,61]. For regression analyses that evaluated the association between NP test findings and measures of subjects' physical performance, self-reported disability, psychosocial disruption, and pain severity ratings, Cohen's set correlation was used [62,63]. Set correlation provides a single general framework of the study of association. It is the realization of the multivariate general linear model and therefore is a natural generalization of simple and multiple correlation. In contrast to canonical correlation, it yields a partitioning of variance in terms of the original variables, rather than their canonical transformations. The building blocks of set correlation are sets of variables, which may be categorical or quantitative. They may also comprise interactions or products of measured variables. Set correlation also allows statistically partialing sets of variables. Partialing (residualizing) the contributions of NP findings on the association between pain severity and physical performance measures was used to evaluate the mediational role of NP function, while still maintaining our multivariate analytical strategy. Chi-square analyses were used to evaluate group differences on dichotomous and ordinal measures. The systat Version 11 statistical package was used to compute all analyses [64]. P ≤ 0.05 was used to indicate statistical significance.

Results

The means and standard deviations by experimental group for the eight NP function measures are displayed in Table 2. A manova, with subject group as the independent variable and the eight NP test scores as the dependent measures, indicated significant differences existed between the groups on one or more of these NP measures (F[8,314]= 2.57, P = 0.01). Although the groups differed on measures of depression and comorbidities (Table 1), these measures were not used as covariates in the manova analysis because neither was significantly correlated with NP scores (for depression, R² = 0.02, P = 0.46; for comorbidities, R² = 0.04, P = 0.10). Follow-up univariate anovas demonstrated significant group differences occurred for the following five NP measures: RBANS immediate memory, language, delayed memory, Part B of the Trail Making Test, and Grooved Pegboard Test (Table 2). To better understand these significant differences, effect sizes were computed. As can be seen in Table 2, these effect sizes ranged from a high of 0.34 for RBANS immediate memory to a low of 0.22 for RBANS delayed memory and Grooved Pegboard. Thus, although these effect sizes are of modest magnitude, these findings confirm our hypothesis that CLBP subjects would demonstrate impaired performance on NP tests.

Table 2

Means (standard deviations) by experimental group for neuropsychological test results

Measure	Pain-Free (N = 160)	CLBP (N = 163)	anovaP Value	Effect Size
RBANS—Immediate memory	103.56 (13.99)	98.53 (15.50)	0.002	0.34
RBANS—Visuospatial	96.48 (17.57)	95.67 (16.78)	0.671
RBANS—Language	102.87 (12.59)	99.14 (10.45)	0.004	0.32
RBANS—Attention	105.96 (15.53)	105.34 (14.53)	0.712
RBANS—Delayed memory	97.91 (15.31)	94.41 (16.11)	0.046	0.22
Trails B (T score)	53.57 (11.36)	50.73 (10.22)	0.019	0.26
Grooved Pegboard (dominant, T score)	45.04 (9.39)	42.76 (11.02)	0.047	0.22
NART—VIQ	98.32 (14.77)	98.16 (13.29)	0.919

For all measures, higher scores indicate better per formance.

RBANS = Repeatable Battery for the Assessment of Neuropsychological Status; CLBP = chronic low back pain; NART—VIQ = National Adult Reading Test––verbal IQ.

A set correlation analysis with the eight NP measures as the predictor set and the two self-reported psychosocial disruption variables as the criterion set was found to be nonsignificant (R² = 0.05, RAO F[16,622] = 0.99, P = 0.46), indicating that subjects' reports of psychosocial disruption were not associated with NP functioning. Similarly, a set correlation analysis between NP measures and measures of self-reported disability was nonsignificant (R² = 0.04, RAO F[16,622]= 0.94, P = 0.52). However, the set correlation analysis between NP measures and physical performance measures was statistically significant (R² = 0.29, RAO F[48,1475] = 2.21, P < 0.001). To evaluate whether this significant association was confounded by comorbidities, a second set correlation analysis was computed that partialed scores on Cumulative Illness Rating Scale from both NP and physical performance measures. This partial set correlation model was still statistically significant (R² = 0.30, RAO F[48,1470] = 2.28, P < 0.001).

Further interpretation of this partial set correlation indicated that the first (χ²[48] = 104.1, P < 0.001), but not the second (χ²[35] = 42.5, P = 0.18), canonical correlation was statistically significant. To further interpret which physical performance measures were most related to the NP tests, canonical loadings for the first significant canonical variate were used and are presented in Table 3. Canonical loadings are correlations that represent how much variance a specific measure shares with the discriminant function. In the present application, these loadings are preferred to canonical or regression coefficients because both the dependent set (physical performance measures) and the independent set (NP measures), as expected, had substantial intercorrelations among themselves, and these intercorrelations can create suppressor effects that reduce or distort these coefficients [65].

Table 3

Set correlation results, canonical loadings (correlations) between neuropsychological test results and measures of physical performance

Construct Set/Measure	Canonical Loading	Multiple Regression Results
Neuropsychological set
RBANS—Immediate memory	−0.24
RBANS—Visuospatial	−0.80
RBANS—Language	0.01
RBANS—Attention	−0.46
RBANS—Delayed memory	−0.49
Trails B	−0.59
Grooved Pegboard (dominant)	−0.64
NART—VIQ	0.03
Physical performance set
Mean static lift	−0.76	R 2 = 0.12, P < 0.001
Work index	−0.58	R 2 = 0.09, P < 0.001
Gait speed	0.65	R 2 = 0.10, P < 0.001
Functional reach	−0.48	R 2 = 0.07, P = 0.007
Chair rise	0.12	R 2 = 0.03, P = 0.220
Trunk rotation	0.21	R 2 = 0.03, P = 0.275

RBANS = Repeatable Battery for the Assessment of Neuropsychological Status; NART—VIQ = National Adult Reading Test––verbal IQ.

As displayed in Table 3, examination of the canonical loadings indicated that the physical performance variables most strongly associated with the NP variables were mean static lift, the dynamic work index, gait speed, and functional reach. Separate multiple regressions for these four physical performance measures indicated they were all significantly predicted by the NP measures, but chair rise and trunk rotation were not (Table 3). As can be seen in Table 3, lower mean static lifting strength, lower work indices, longer gait times, and shorter functional reaches were associated with lower NP test scores. Inspection of the canonical loadings for the NP tests indicated that scores on the RBANS immediate memory and language test, and the National Adult Reading Test—verbal IQ were not as strongly associated with physical performance scores compared with the other five NP tests, that is, RBANS visuospatial, attention, and delayed memory, Part B of the Trail Making Test, and Grooved Pegboard Test (dominant).

To evaluate our hypotheses related to pain intensity ratings, several set correlation analyses were conducted using only the data of CLBP subjects. A set correlation analysis indicated a significant association between their MPQ pain scores and their NP scores (R² = 0.17, RAO F[8,154]= 3.88, P < 0.001). Inspection of the canonical loadings indicated that RBANS attention and visuospatial scores and Part B of the Trail Making Test and Grooved Pegboard scores were most strongly correlated with MPQ pain severity scores. These associations were inversely related, that is, higher pain intensity scores were predictive of lower NP scores.

Pain intensity scores were also found to be significantly associated with physical performance scores (R² = 0.08, RAO F[6,156] = 2.25, P = 0.041). Thus, as NP findings were significantly correlated with both pain intensity scores and physical performance scores, NP scores may serve as a mediator between pain intensity and physical performance. To test whether NP functioning mediated the association between pain severity and physical performance scores in CLBP subjects, a set correlation that partialed NP scores from both MPQ and physical performance measures was computed. This analysis indicated that pain scores were no longer significantly associated with physical performance measures (R² = 0.04, RAO F[6,148] = 1.53, P = 0.172). Thus, these findings support our hypothesis that NP performance would mediate the relationship between pain and physical performance.

Finally, to evaluate whether the association between NP functioning and physical performance was independent of pain severity, two set correlation analyses were computed: one that compared NP and physical performance scores, and the other that partialed pain severity scores on MPQ from both NP and physical performance measures. The un-partialed association between NP and physical performance measures was R² = 0.31 (P < 0.001), and the partialed association was R² = 0.29 (P < 0.001), indicating that pain severity did not significantly modify the association between NP and physical performance scores in CLBP subjects.

Discussion

This is the first study performed exclusively in independent community-dwelling OA demonstrating an association between chronic nonmalignant pain and decrements in NP performance. We also demonstrated that lower NP performance scores were associated with poorer physical performance. Further, NP performance mediated the relationship between pain and physical performance.

The pattern of NP performance changes was dependent on pain severity. Group differences in NP performance were apparent for immediate and delayed memory, language, mental flexibility (Part B of the Trail Making Test), and manual dexterity (Grooved Pegboard). Memory deficits in chronic pain patients have been demonstrated by other investigators [66], although deficits in language, mental flexibility, and manual dexterity have not. Examination of the relationship between pain severity and NP performance in the subjects with CLBP revealed decrements in performance on tasks of attention, visuospatial skills, mental flexibility, and manual dexterity that were associated with increased pain severity. Other experimental studies of chronic pain patients also have found attentional deficits in association with high pain severity [27–29], but our study is the first to demonstrate a relationship between pain severity and other NP tasks. Clinical and experimental studies of acute pain have demonstrated an inverse relationship between pain intensity and attention [20,21,24,26]. A possible mechanism accounting for this relationship is thought to be the occupation of working memory capacity by pain that leads to less available attentional capacity for other tasks [67]. Additional studies are needed to determine the mechanism(s) underlying the relationship between pain and the other NP performance deficits found by our research.

Other investigators have suggested that psychosocial disruption mediates the relationship between pain and cognitive function [68–71]. While there was a statistically significant difference in number of depressive symptoms reported in our pain-free and CLBP subjects, the number was well below the clinically significant cutoff for the Geriatric Depression Scale [47]. Thus, an alternate mechanism for the NP performance differences in these individuals should be hypothesized. Recent data that demonstrated the presence of brain atrophy in patients with CLBP suggest a pattern of atrophy distinct from that in patients with chronic depression or anxiety [72–74]. Thus, while psychosocial disruption, including depression, anxiety, dysfunctional coping, and fear, has been shown to be more common in patients that suffer from chronic pain [9,14–18], it appears that there are unique mechanisms underlying the direct effect of chronic pain on the central nervous system.

From a clinical perspective, it is noteworthy that the frequency of opioid prescription in our participants was quite low. In order to rigorously examine the relationship between opioids and cognitive function, the relationship between morphine equivalents and NP performance should be tested. There was an insufficient number of participants taking opioids in our study sample to examine this relationship. Others have shown that patients taking opioids may suffer decline in cognitive function [75], but there is substantial interindividual and inter-opioid variability in this phenomenon [19,76,77]. Our study results combined with that of others should prompt practitioners to prescribe opioids for patients with chronic nonmalignant pain carefully and methodically. An initial adverse NP response to a particular opioid should not erect a barrier to proceeding down this therapeutic path for the patient in whom such treatment is warranted and for whom pain reduction may ultimately result in improved cognitive function.

The relationship between NP performance and physical performance has not, to our knowledge, been previously demonstrated in OA with chronic pain. Our findings corroborate a similar relationship previously demonstrated by Binder and colleagues in 125 cognitively intact OA with unknown pain status that were enrolled in a randomized trial of exercise or hormone replacement therapy [78]. Cognitive processing speed was found to be a significant predictor of physical frailty as measured by a modified physical performance test. Ble and colleagues have recently demonstrated a relationship between executive function and gait velocity on an obstacle course in older adult participants of the InCHIANTI study [79]. The limitation of these studies as well as ours is their cross-sectional design. While the relationship between cognitive processing speed and physical performance in Binder's study, between executive function and obstacle course gait velocity in Ble's study, and between pain and NP performance in our study was strong, causation models could not be tested. Tabbarah and colleagues examined 7-year longitudinal data in community-dwelling OA and found an association between declines in cognitive performance and declines in performance of both routine and more challenging physical tasks [80].

The mediating role that NP performance played in the relationship between pain and physical performance should be further explored in future research. Recent data demonstrate thalamocortical atrophy in patients with CLBP of heterogeneous age [35]. These preliminary data need to be corroborated, as does the relationship between structural brain changes and changes in NP performance, as well as the degree to which such changes can be prevented or reversed. Traditional models of CLBP-related disability emphasize the direct effects of pain on musculoskeletal function, such as flexibility, strength, and endurance, or on psychosocial function. Our data challenge these traditional notions, and suggest that the central nervous system plays a key role in mediating the relationship between chronic pain and functional compromise.

It should also be highlighted that some of the physical performance measures thought to be axially specific (i.e., chair rise, trunk rotation) bore no significant relationship to NP performance, but gait speed, functional reach, and the static/isoinertial lifting tasks did. Isoinertial lifting is a well-validated, axially specific measure [52,53]. It is a demanding test of physical capacity, thus ceiling effects are rarely encountered with this parameter. It is not surprising therefore that the lifting task was identified as one of the physical performance parameters that was associated with NP performance, while the less demanding chair rise, functional reach, and trunk rotation tasks were not. The association between NP performance and gait velocity adds an interesting element to the body of literature, demonstrating that gait velocity is a powerful predictor of functional decline and disability [81–83]. It appears, based on our study findings, that gait velocity not only is a marker of general physical health and function, but also may be a marker of the vigor of central nervous system processing. Functional reach is a clinical measure of balance [55–57,84]. Our subjects were excluded if they were demented or had a history of recent unexplained falls. Thus, the demonstrated link in these individuals between subtle deficits in NP performance and balance performance extends the well-established relationship between dementia and falls [85].

Our study participants were all independent OA without dementia and with very low levels of depressive symptomatology. The prevalence of opioid use, as noted earlier, was also quite low. The medical, psychosocial and functional profiles of these OA were therefore not typical of chronic pain patients. Even so, these individuals had subclinical deficits in NP performance and in physical performance with demanding tasks. These findings have interesting implications for OA with less cognitive and physical reserves, in whom pain may more overtly lead to deterioration, such as cognitively impaired OA in whom pain has been shown to be associated with agitated behavior [86,87].

The findings from this study corroborate and extend existing work that demonstrates a relationship between pain and cognitive performance, and between cognitive performance and physical performance. NP performance should be included as both an outcome measure and a mediating variable in future studies designed to test chronic pain interventions for OA. Whether the cognitive deficits associated with CLBP are reversible is unknown. Whether geriatricians, psychiatrists, and other practitioners can add chronic pain to the list of modifiable factors that contribute to dementia, and whether pain amelioration can forestall the physical disability that results from cognitive deterioration should be questions addressed by future research.

Acknowledgment

This work was supported by a grant from the National Institutes of Health, R01 AGQ18299.

Presented at the 11th World Congress on Pain, Sydney, Australia, August 2005.