Hollman, John H. PT, PhD; Beed, Katherine C. DPT; Buus, Ryan J. DPT; Schleicher, Kenzie L. DPT; Lanzino, Desiree J. PT, PhD
Walking performance is a hallmark of physical function. One's preferred walking speed, for example, reflects quality of life and health status and predicts cognitive decline and remaining life expectancy in older adults.1–5 In addition to preferred walking speed, maximum walking speed may also be important. Maximum walking speed declines more rapidly than preferred walking speed in older adults, due in part to age-related changes in neuromuscular control.6,7 Therefore, assessing maximum walking speed may provide advantages over preferred walking speed for detecting functional decline.
Regardless of whether one measures preferred or maximum walking speed, walking speed is touted as a “vital sign” for function.8 Oftentimes, however, older adults may temporarily be nonambulatory (eg, postsurgical patients) and proxy measures for walking speed may be indicated. Several factors, both nonmodifiable and modifiable, influence walking speed, and therefore, those proxy measures must be considered carefully when one interprets walking speed as a biomarker for quality of life, health status, or physical function. Nonmodifiable factors include age, sex, and height.9,10 Modifiable factors include lower limb strength, cognition, and balance.11–14 It is important to understand which modifiable factors are associated with walking speed, because those can potentially be targeted through intervention.
Less well-understood is how limb coordination, defined as the ability to execute accurate and controlled movements at various speeds,15 influences walking speed. Similar to gait characteristics, normally coordinated limb movements are characterized in part by rhythmic muscle contractions and relaxations that promote easy reversal between opposing muscle groups.16 Limb coordination is assessed in part through tests that examine a person's ability to accurately and quickly perform alternate or reciprocal motions and movement synergies. When assessed quantitatively in healthy individuals, performance on several timed limb coordination tests has been correlated with age, sex, and height,17 similar to relationships observed with walking speed. Since limb coordination tests can be conducted in non–weight-bearing positions, these may serve as a proxy for the measurement of walking speed in individuals who are temporarily nonambulatory. Hollman et al,18 controlling for the effects of sex, age, and height, reported that timed upper limb coordination tests in older adults were more strongly associated with preferred (r = −0.396) and maximum walking speed (r = −0.356) than was a test of lower limb coordination, timed heel-on-shin performance (r = −0.228 and r = −0.288, respectively). It is unclear, however, whether timed upper limb coordination predicts walking speed more than and beyond that which can be accounted for by other variables such as lower limb strength and postural stability, both of which may be difficult to assess in many postsurgical patient populations. The purpose of this study was, therefore, to examine the extent to which timed upper limb coordination performance predicts preferred and maximum walking speed, independent of other variables identified as determinants of walking speed.
Study Design and Participants
We conducted this cross-sectional, exploratory study to examine the relationship between coordination performance and walking speed, controlling for age, height, knee extensor strength, postural stability, and cognitive abilities. Flyers were posted in the community to recruit healthy adults older than 60 years to participate. On the basis of a power analysis, 81 participants were sufficient to detect a change in R2 of 0.10 or higher for any given variable entered into a regression model at α = .05 at a desired statistical power of 0.80.
Eligible participants were English-speaking, community-dwelling individuals older than 60 years who were independent in activities of daily living and community ambulators, which we operationally defined as having a score of 13 or greater on the Rivermead Mobility Index. While the Rivermead Mobility Index was originally developed to assess functional mobility following traumatic brain injury and stroke,19 it has been validated as an index of functional mobility in other patient populations20 and can be used as an index of community ambulation capacity in older adults. Exclusion criteria included any medically diagnosed neurologic pathology or cardiovascular pathology leading to neurologic symptoms, or a history of surgery within 12 months of the test date (eg, hip or knee arthroplasty, back surgery), which may have impacted one's ability to ambulate independently. The Mayo Foundation institutional review board approved the protocol, and all participants provided written informed consent.
Participants who met enrollment criteria and provided informed consent were administered a battery of tests that represent potential determinants of walking speed. On the basis of our interpretation of multiple studies,11–14 the primary determinants of walking speed in healthy older adults appear to be age, sex, height, knee extension strength, cognition, and postural stability. The participants' age, sex, height, and weight were recorded. Investigators then administered sequentially to each participant a survey to quantify comorbidities, a cognitive assessment, a test of isometric knee extension strength, a test of postural stability, a timed upper limb coordination test, and walking tests to determine preferred and maximum walking speeds.
Physical functions like walking can be influenced by the number of comorbid diseases with which a person presents. To examine the associations between multiple clinical tests and walking speed while controlling for comorbidities, we quantified the number of comorbidities with the Functional Comorbidity Index (FCI). The FCI was developed with physical function as the outcome of interest and incorporates a list of 18 diagnoses that correlate with declining function; 1 point is assigned per diagnosis, and the points are summed, yielding a score between 0 and 18.21
We assessed cognition with the Montreal Cognitive Assessment (MoCA), a 30-point test that assesses several cognitive domains and is validated as a screening tool for mild cognitive impairment in older adults.22 The test-retest reliability coefficient of the MoCA is 0.92 and the internal consistency of its items, measured with Cronbach α, is 0.83.22
We quantified isometric knee extension strength of participants' dominant lower extremities using a MicroFET2 handheld dynamometer (Hoggan Health Industries, Salt Lake City, Utah). Participants were seated in a chair with their knees flexed approximately 90°. The dynamometer was stabilized at the anterior distal leg, approximately 5 cm proximal to the ankle joint, with a fixation belt and an examiner's hand. Participants maximally contracted the extensors for 5 seconds. The maximum isometric force produced over 3 repetitions was included in subsequent analyses. A 30-second rest period was provided between repetitions. The method for collecting knee extension strength data was comparable to methods in which the test-retest reliability coefficient exceeded 0.90 in older adults.23
We assessed postural stability with the Functional Reach Test. As per methods described by Duncan et al,24 participants stood adjacent to a wall with an attached meter stick, elevated the arm nearest the wall to 90° of shoulder flexion, and reached as far anteriorly as possible without taking a step, as directed by the examiner. The difference between the ending and starting positions of the tip of the middle finger was recorded as the functional reach distance. The maximum of 3 trials was included in subsequent analyses. The Functional Reach Test correlates highly with laboratory measures of center of pressure excursion, indicating its validity as a test of limits of stability, and is conducted with a test-retest reliability coefficient that exceeds 0.80.24
We assessed upper limb coordination with the timed finger-to-nose test. As per methods described by Lanzino et al,17 the participant alternately touched the tip of an examiner's finger—held at eye level at approximately the participant's arm reach—and the participant's nose as quickly and accurately as possible. After 1 practice trial, the time required to complete 5 cycles of the finger-to-nose movement was measured with a stopwatch and recorded for analysis. Since timed limb coordination performance does not differ between dominant and nondominant limbs,17 we collected data from participants' dominant arms only. Lanzino et al25 reported that timed measurements on the test have an interrater reliability coefficient of 0.92.
Last, preferred and maximum walking speeds were measured over the inner 6 m of a 10-m walkway. Participants were permitted to wear their preferred footwear. In the preferred walking speed trials, participants were instructed to “walk down the path at your normal speed as though you were walking on the sidewalk outside.” In the maximum walking speed trials, they were instructed to “walk down the path as quickly as you can, without running.” Times were recorded with a TracTronix TF 100 infrared dual-beam timing system (TracTronix, Lenexa, Kansas). The mean walking speeds over 2 walks in each condition were included in subsequent analyses. Steffen et al26 reported that comparable methods for measuring walking speed have test-retest reliability coefficients that exceed 0.95.
Descriptive statistics, mean (SD) were calculated for each variable. Pearson product-moment correlation coefficients (r) were used to examine associations among preferred and maximum walking speed with timed finger-to-nose performance, knee extension strength, functional reach, MoCA scores, FCI scores, age, and height. Two stepwise multiple regression models (αentry = .05 and αremoval = .10) along with partial correlation coefficients were used to identify which variables predicted preferred and maximum walking speed, controlling for all other variables. In the first regression model, all 7 predictor variables were entered. In the second model, knee extension strength and functional reach—those variables that either require full weight bearing to complete or may be difficult or contraindicated to measure in postsurgical patient populations—were removed. IBM SPSS (IBM Corporation, Armonk, NY) 21.0 software was used for all analyses.
Assumptions for conducting multiple regression analyses were examined. With the exception of MoCA scores, all other variables were normally distributed (Kolmogorov-Smirnov tests with P > .05). Each predictor variable was linearly associated with preferred and maximum walking speed. Plots of standardized residuals and standardized predicted values indicated the assumptions of homoscedasticity were not violated. Variance inflation factors were less than 2.0 for each variable included in the regressions, suggesting that multicollinearity was not problematic.
Participant Characteristics and Descriptive Data
Demographic data are provided in Table 1. The sample included 84 participants, 60 women (71%) and 24 men (29%), who ranged in age from 60 to 92 years, 75 (9) years, and were predominantly white/Caucasian (99%). Most participants had no history of falls (74%), and their body mass index, 25.8 (4.3) kg·m−2, indicated that most were in the healthy weight to overweight categories.
Descriptive data for each variable are provided in Table 2. Preferred walking speed was 129 (24) cm·s−1, and maximum walking speed was 176 (37) cm·s−1. Preferred walking speed (Figure 1) was positively correlated with functional reach performance (r = 0.529), knee extension strength (r = 0.496), and MoCA scores (r = 0.478) and negatively correlated with FCI scores (r = −0.432), age (r = −0.554), and timed finger-to-nose performance (r = −0.403). Maximum walking speed (Figure 2) was positively correlated with knee extension strength (r = 0.635), functional reach performance (r = 0.533), MoCA scores (r = 0.477), and height (r = 0.199) and negatively correlated with age (r = −0.575), FCI scores (r = −0.433), and timed finger-to-nose performance (r = −0.429).
Regression on Preferred Walking Speed
The first regression model (Table 3) indicated that variance in 5 variables—age, FCI scores, functional reach, knee extension strength, and height—accounted for 55.4% of the variance in preferred walking speed (R2 = 0.554, P < .001). Controlling for all other variables, preferred walking speed was positively correlated with functional reach (partial r = 0.316, P = .004) and with knee extension strength (partial r = 0.246, P = .028) and negatively correlated with age (partial r = −0.379, P = .001), FCI scores (partial r = −0.364, P = .001), and height (r = −0.238, P = .034). The bivariate correlation between preferred walking speed and finger-to-nose performance (Figure 1; r = −0.403) was attenuated (partial r = −0.031, P = .786). On the basis of these findings, preferred walking speed was best predicted with the following regression equation:
Preferred walking speed (cm·s−1) = 234 − 0.9 (age [y]) − 3.9 (FCI score) + 0.9 (functional reach [cm]) + 0.06 (knee extension strength [N]) − 0.4 (height [cm]).
Removing the functional reach and knee extension strength variables, the second regression model (Table 3) indicated that variance in age, FCI scores, and MoCA performance accounted for 48.7% of the variance in preferred walking speed (R2 = 0.487, P < .001). The bivariate correlation between preferred walking speed and finger-to-nose performance (r = −.403) remained attenuated in the second regression model (partial r = −0.179, P = .110). In the absence of valid functional reach scores or knee extension strength measurements, preferred walking speed may alternatively be predicted by
Preferred walking speed (cm·s−1) = 195 − 1.2 (age [y]) − 4.9 (FCI score) + 1.33 (MoCA score).
Regression on Maximum Walking Speed
For maximum walking speed, the first regression model (Table 4) indicated that 5 variables accounted for 63.5% of the variance in maximum walking speed (R2 = 0.635, P < .001). Controlling for all other variables, maximum walking speed was positively correlated with knee extension strength (partial r = 0.464, P < .001), MoCA performance (partial r = 0.204, P = .070), and functional reach (partial r = 0.221, P = .049) and negatively correlated with age (partial r = −0.317, P = .004) and FCI scores (partial r = −0.291, P = .009). The bivariate correlation between maximum walking speed and finger-to-nose performance (r = −0.429) was attenuated (partial r = −0.075, P = .513). According to our findings, maximum walking speed was best predicted with the following regression equation:
Maximum walking speed (cm·s−1) = 163 + 0.17 (knee extension strength [N]) + 1.5 (MoCA score) + 0.9 (functional reach [cm]) − 1.1 (age [y]) − 4.2 (FCI score).
In the absence of valid knee extension strength measurements or functional reach scores, the second regression model (Table 4) indicated that variance in age, FCI scores, MoCA performance, height, and finger-to-nose performance accounted for 55.9% of the variance in maximum walking speed (R2 = 0.559, P < .001). Maximum walking speed may alternatively be predicted by the following equation:
Maximum walking speed (cm·s−1) = 176 − 1.6 (age [y]) − 6.5 (FCI score) + 1.8 (MoCA score) + 0.7 (height [cm]) − 5.5 (finger-to-nose performance [s]).
The purpose of this study was to examine the extent to which timed finger-to-nose coordination performance predicts preferred and maximum walking speed in older adults, independent of other determinants of walking speed. Although the scatter plots (Figures 1 and 2) indicate that coordination performance correlated moderately with both preferred (r = −0.403) and maximum (r = −0.429) walking speed, those relationships were attenuated when other variables in the study were controlled. Timed finger-to-nose coordination performance was not a statistically significant predictor of preferred walking speed, even when knee extension strength and functional reach performance were removed from the analysis (Table 3). Timed finger-to-nose performance contributed to the prediction of maximum walking speed in the absence of knee extension strength and functional reach performance but accounted for only 2.9% of the variance in maximum walking speed more than and beyond the variance accounted for by age, the number of functional comorbidities, MoCA performance, and height (change in R2 = 0.029; Table 4). These findings imply that upper limb coordination assessed via timed finger-to-nose performance is not a strong proxy measure for walking speed in older adults.
According to our findings, the strongest predictors of preferred and maximum walking speed were knee extension strength, functional reach performance, age, and FCI scores. These findings are largely in agreement with those of other studies examining variables that predict walking speed. Among the earliest of those studies, Bohannon11 reported that hip abduction strength, sex, height, and weight explained 13% of the variance in preferred walking speed and that knee extension strength, age, height, and weight explained 41% of the variance in maximum walking speed. In a population-based study, Bohannon later reported that age, height, sex, knee extension force, and waist circumference explained approximately 50% of the variance in preferred walking speed.12 Beyond those variables, other potentially modifiable variables also predict walking speed. Holtzer et al,14 for example, reported that 3 factors related to cognition—verbal intelligence, executive attention, and memory—predict walking speed in healthy older adults. Tiedemann et al13 reported that a composite measure of lower limb strength (sum of knee extension, knee flexion, and ankle dorsiflexion strength) was the strongest predictor of walking speed in older adults, but that other variables, including balance (limits of stability), were independent predictors. Our findings that knee extension strength, functional reach performance, and age are strong predictors of walking speed are consistent with those other findings. In addition, we found that FCI scores—the number of comorbid conditions that affect physical function—also predicted walking speed independently of age, knee extension strength, and other variables included in the study. On the basis of the regression equations, each additional comorbid condition that a person older than 60 years presents with is associated with a 3.9 to 4.9 cm·s−1 reduction in preferred walking speed and with a 4.2 to 6.5 cm·s−1 reduction in maximum walking speed.
In previous work, Hollman et al18 reported that timed upper limb coordination tests were associated with preferred and maximum walking speed in older adults. Of the 5 tests examined (finger-to-nose, pronation-supination, mass grasp, finger opposition, heel-on-shin), performance on the timed finger-to-nose test correlated with both preferred and maximum walking speed (r = −0.512 and −0.513, respectively). Perhaps surprisingly, timed heel-on-shin performance had lower correlations with preferred and maximum walking speed (r = −0.319 and −0.403, respectively) despite being a lower limb measure that, like walking, hinges on lower extremity performance. For that reason, we did not examine lower limb coordination in the present study. Regardless, the magnitudes of the bivariate correlations between timed finger-to-nose performance and preferred, and maximum walking speed (r = −0.403 and −0.429, respectively) obtained in the present study were comparable to those prior results. While the magnitudes of the bivariate correlations between timed finger-to-nose performance and walking speed were comparable between the studies, the major difference is that we included and controlled for additional variables in the present study (knee extension strength, limits of stability, and comorbidities), which enabled us to better examine the predictive relationships among the variables.
The bivariate correlations indicate that there is clearly a negative relationship between timed limb coordination performance and walking speed. Older persons with longer coordination performance times tend to ambulate at slower speeds. Since both tasks are rhythmic in nature, similarities in performance may be because of shared properties between their pattern-generating neural networks.27 Further evidence of a relationship between timed coordination and gait performance is shown in decrements of performa nce in both tasks among persons with neurodegenerative diseases such as dementia.28 Nevertheless, our findings indicate that the relationship between coordination performance and walking speed was attenuated when other measures of function (leg strength, postural stability, and cognition) were obtained. These results may have been tempered by our use of a distal target, external to the participant, for the finger-to-nose test. Swaine et al29 reported that timed finger-to-nose performance did not differ when the test was completed with and without a distal target (examiner's finger). Their participants, however, were aged between 15 and 34 years, whereas in the present study, participants were older, 75 (9) years, and examiners reported that some were more concerned with accuracy at the distal target, potentially leading to the detriment of the speed of their performance. Had a distal target external to the participant not been used for the finger-to-nose measure and had participants focused on speed rather than accuracy, our results may have differed. Another option would have been to assess pronation-supination performance, which also correlates with preferred and maximum walking speeds (r = −0.553 and −0.495, respectively)18 and is as reliable (interclass correlation = 0.92)25 as the finger-to-nose test, with the distinction of not having a distal target.
Some limitations of this study may constrain the generalizability of our findings. The primary limitation is that its observational (cross-sectional) design limits conclusions regarding cause and effect. While several of the variables associated with preferred and maximum walking speed are potentially modifiable, we cannot ascertain whether changes in those variables will induce changes in walking speed. Second, the sample was limited to community-ambulating older adults. The findings may not generalize to clinical populations in whom it may be more relevant to measure walking speed or other proxy measures as a vital sign for physical function. Last, while the sample size was sufficient to analyze bivariate relationships between 7 independent variables and walking speed, the sample was not sufficient to analyze interactions among variables. It is possible that interactions may have accounted for additional variance in walking speed that we did not examine. Prospective studies with more generalizable patient samples, perhaps using coordination tests without external targets, are recommended to examine how changes in coordination or potentially other modifiable bodily functions like strength and postural stability may affect walking speed and, ultimately, the quality of life,1 health status,2,3 and even life expectancy5 in older adults.
Despite the study's limitations, the methods were sufficient to support our conclusions. Specifically, upper limb coordination performance assessed with the timed finger-to-nose test was correlated with both preferred and maximum walking speeds in healthy, community-ambulating older adults. Those relationships, however, were attenuated when other determinants of walking speed such as knee extension strength, postural stability, cognition, functional comorbidities, age, and height were considered. Timed upper limb coordination as measured by the finger-to-nose test, using a distal target, would not appear to be a valid proxy for walking speed when weight-bearing clinical examination procedures are contraindicated.
The Department of Physical Medicine & Rehabilitation, Mayo Clinic, Rochester, Minnesota, provided funding for this study.
1. Ferrucci L, Baldasseroni S, Bandinelli S, et al. Disease severity and health-related quality of life across different chronic conditions. J Am Geriatr Soc. 2000;48(11):1490–1495.
2. Cesari M, Kritchevsky SB, Penninx BWHJ, et al. Prognostic value of usual gait speed in well-functioning older people—results from the Health, Aging and Body Composition Study. J Am Geriatr Soc. 2005;53(10):1675–1680.
3. Studenski S, Perera S, Wallace D, et al. Physical performance measures in the clinical setting. J Am Geriatr Soc. 2003;51(3):314–322.
4. Verghese J, Wang C, Lipton RB, Holtzer R, Xue X. Quantitative gait dysfunction and risk of cognitive decline and dementia. J Neurol Neurosurg Psychiatry. 2007;78(9):929–935.
5. Studenski S, Perera S, Patel K, et al. Gait speed and survival in older adults. JAMA. 2011;305(1):50–58.
6. Clark DJ, Manini TM, Fielding RA, Patten C. Neuromuscular determinants of maximum walking speed in well-functioning older adults. Exp Gerontol. 2013;48(3):358–363.
7. Ko SU, Hausdorff JM, Ferrucci L. Age-associated differences in the gait pattern changes of older adults during fast-speed and fatigue conditions: results from the Baltimore longitudinal study of ageing. Age Ageing. 2010;39(6):688–694.
8. Fritz S, Lusardi M. White paper: “Walking Speed: the Sixth Vital Sign”. J Geriatr Phys Ther. 2009;32(2):2–5.
9. Hollman JH, McDade EM, Petersen RC. Normative spatiotemporal gait parameters in older adults. Gait Posture. 2011;34(1):111–118.
10. Oh-Park M, Holtzer R, Xue X, Verghese J. Conventional and robust quantitative gait norms in community-dwelling older adults. J Am Geriatr Soc. 2010;58(8):1512–1518.
11. Bohannon RW. Comfortable and maximum walking speed of adults aged 20–79 years: reference values and determinants. Age Ageing. 1997;26(1):15–19.
12. Bohannon RW. Population representative gait speed and its determinants. J Geriatr Phys Ther. 2008;31(2):49–52.
13. Tiedemann A, Sherrington C, Lord SR. Physiological and psychological predictors of walking speed in older community-dwelling people. Gerontology. 2005;51(6):390–395.
14. Holtzer R, Verghese J, Xue X, Lipton RB. Cognitive processes related to gait velocity: results from the Einstein Aging Study. Neuropsychology. 2006;20(2):215–223.
15. Schmitz TJ. Examination of coordination. In: O'Sullivan SB, Schmitz TJ, eds. Physical Rehabilitation. 5th ed. Philadelphia, PA: FA Davis Company; 2007:193–225.
16. Byl NN. Multisensory control of upper extremity function. Neurol Rep. 2002;26(1):32–43.
17. Lanzino DJ, Conner MN, Goodman KA, Kremer KH, Petkus MT, Hollman JH. Values for timed limb coordination tests in a sample of healthy older adults. Age Ageing. 2012;41(6):803–807.
18. Hollman JH, Conner MN, Goodman KA, Kremer KH, Petkus MT, Lanzino DJ. Timed limb coordination performance is associated with walking speed in healthy older adults: a cross-sectional exploratory study. Gait Posture. 2013;38(2):316–320.
19. Collen FM, Wade DT, Robb GF, Bradshaw CM. The Rivermead Mobility Index: a further development of the Rivermead Motor Assessment. Disabil Rehabil. 1991;13(2):50–54.
20. Franchignoni F, Brunelli S, Orlandini D, Ferriero G, Traballesi M. Is the River-mead Mobility Index a suitable outcome measure in lower limb amputees?—a psychometric validation study. J Rehabil Med. 2003;35(3):141–144.
21. Groll DL, To T, Bombardier C, Wright JG. The development of a comorbidity index with physical function as the outcome. J Clin Epidemiol. 2005;58(6):595–602.
22. Nasreddine ZS, Phillips NA, Bedirian V, et al. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc. 2005;53(4):695–699.
23. Roy MA, Doherty TJ. Reliability of hand-held dynamometry in assessment of knee extensor strength after hip fracture. Am J Phys Med Rehabil. 2004;83(11):813–818.
24. Duncan PW, Weiner DK, Chandler J, Studenski S. Functional reach: a new clinical measure of balance. J Gerontol. 1990;45(6):M192–M197.
25. Lanzino DJ, Rabinstein A, Kinlaw D, et al. Coordination tests in persons with acute central nervous system pathology: assessment of interrater reliability and known-group validity. J Neurol Phys Ther. 2012;36(3):122–130.
26. Steffen TM, Hacker TA, Mollinger L. Age- and sex-related test performance in community-dwelling elderly people: Six-Minute Walk Test, Berg Balance Scale, Timed Up & Go Test, and gait speeds. Phys Ther. 2002;82(2):128–137.
27. Marder E, Calabrese RL. Principles of rhythmic motor pattern generation. Physiol Rev. 1996;76(3):687–717.
28. Franssen EH, Souren LE, Torossian CL, Reisberg B. Equilibrium and limb coordination in mild cognitive impairment and mild Alzheimer's disease. J Am Geriatr Soc. 1999;47(4):463–469.
29. Swaine BR, Desrosiers J, Bourbonnais D, Larochelle J-L. Norms for 15- to 34-year-olds for different versions of the finger-to-nose test. Arch Phys Med Rehabil. 2005;86(8):1665–1669.
aging; coordination impairment; muscle strength; walking