Sage Journals: Discover world-class research

Abstract

Objectives: To determine the performance characteristics of gait speed measurements and the physiological cost index (PCI; heart rate change/gait speed) in working adults. Methods: Gait speeds, heart rate changes, andnon–steady state PCIs were calculated in 61 volunteers who worked in our health sciences center. These subjects completed 9 separate 100-foot walk tests in 3 separate sessions. Results: The mean heart rate change after a 100-foot walk was 16.6 ± 8.1 beats per minute. The mean gait speed was 76.1 ± 9.6 meters per minute, and the mean PCI was 0.22 ± 0.11 beats per meter. There were highly significant correlations among all measurements on the 9 separate tests (correlation coefficients 0.41-0.95); gait speed measurements had the highest correlations (0.91-0.95). In a multivariable model hypertension and arthritis were associated with reduced gait speeds. Conclusion: Gait speed, heart rate changes, and non–steady state PCIs have good repeatability when measured over short walks. This information provides a rapid physiological assessment and a method for measuring changes in functional status in healthy subjects and most patients.

Keywords

heart rate gait speed functional capacity physiologic cost index

Functional tests can provide important information about the health status of patients with documented medical disorders and patients receiving routine health care. These tests often include studies that evaluate mobility and balance and measure gait speed. The distances used for gait speed measurement range from 8 feet (2.44 meters) to 400 meters.¹ Shorter tests have more obvious practical applications in a busy clinic and may identify changes in mobility related to aging, medical disorders, medications, and poor conditioning.² Gait speed measurements can also identify patients who are at increased risk for future adverse events, including falls, hospitalizations, and institutionalization.^3,4 These tests do not provide the same physiologic information available from cardiac stress tests or combined cardiorespiratory tests. However, changes in heart rate can be referenced to gait speed, and this calculation provides the physiologic cost index (PCI; heart rate change divided by gait speed).^5,6 The PCI reasonably reflects metabolic demand of walking and does increase in patients with neurological disorders, such as stroke.⁷ Whether or not the PCI provides useful information about adults with chronic medical disorders is unknown. In addition, there is limited information about the repeatability and reproducibility of this test; 2 studies have used young adults to determine reliability.^5,6 We wanted to determine whether or not short walks reproducibly increase the heart rate and whether the gait speed measurement or the PCI calculation was more repeatable (ie, had less within subject variability). We evaluated the repeatability of heart rate change, gait speed, and PCI measurements and made qualitative assessments of gait using the Tinetti scale in a larger group of healthy volunteers with a wider age range. Our results support the recommendation that short walk tests to calculate gait speed provide primary care physicians a rapid method to identify changes in the functional status of clinic patients.

Methods

Study Population

We recruited 61 healthy volunteers from coworkers in the School of Medicine at Texas Tech University Health Sciences Center for this study. The protocol had institutional review board approval, and all participants gave written and informed consent. The study coordinator collected routine demographic information about age, sex, height, weight, and chronic medical conditions, including diabetes, hypertension, congestive heart failure, cerebrovascular disease, chronic obstructive pulmonary disease (COPD), asthma, and arthritis. All medical information was based on self-reports from the participating subjects, and no medical records were reviewed. Inclusion criteria included a willingness to participate and fulltime employment or student status at the health sciences center. Exclusion criteria for the study included any difficulty walking or leg or foot injury within the past 6 months. We did not exclude volunteers with chronic stable health conditions, such as hypertension.

Test Protocol

Participants completed three 100-foot walks in an open corridor on 3 separate occasions. Each walk included a 50-foot walk down the hall, a turn, and a 50-foot walk back to the start point.² Patients and volunteers were told to “walk with their usual pace.” No additional instructions or encouragement were given during the walks. Walk time, O₂ saturations, and pulse rates were recorded. The gait was also qualitatively assessed using the Tinetti gait assessment tool.⁸ This tool evaluates gait characteristics, including gait initiation, step length and height, step continuity, path, trunk swing, and stance; scores can range from 0 to 12. The subject rested 3 minutes before starting the first walk and between each of the walk sessions. Walk trials were scheduled according to the availability of the participant and the study investigator, and trial were completed within 1 week. There was no effort to schedule them at the same time of each day. This protocol met the necessary conditions to establish repeatability: same measurement procedure, same observer, same instrument, same location, and a short time frame for testing.

Physiological Assessment

Physiological cost index was calculated by dividing the heart rate change (final heart rate minus starting heart rate; beats/min) by gait speed (m/min) under nonsteady gait conditions at the end of the 100-foot walk test. The rationale for this distance largely developed from practical considerations of what is reasonably possible in a busy outpatient clinic.² The PCI provides a reasonable estimate of energy expenditure (O₂ consumption) and can help evaluate functional performance.^5,6

Statistical Analysis

Descriptive statistics were used to calculate the means, standard deviations, and frequencies. To determine the bivariate correlations between categorical patient characteristics and the gait speed, PCI, and heart rate change, t tests were used. Multiple linear regression analysis was used to control for the confounding variables. Intraclass correlation was calculated to measure consistency between walks of each subject. The repeatability coefficient was derived from the within-subject variance and standard deviation calculated by 1-way analysis of variance for subjects across all walks and across walks 2 and 3 in each session.⁹The repeatability coefficient is a measure of reliability or variability; its measurement is based on the testing conditions described above. Reproducibly measures the variability in a test under less uniform conditions, such as testing at different sites by different observers. Validity addresses the degree to which a test measures the desired outcomes and was not measured in our study. The standard errors of the measurement (SEM) was derived from the within subject standard deviation for all walks and normalized to the mean to calculate the SEM%.⁷ This calculation represents a coefficient of variation based on analysis of variance. SSPS 17.0.3 (IBM Inc, Armonk, NY) was used for statistical analysis, and P values ≤.05 were considered significant.

Results

This study included 61 subjects with a mean age of 46 ± 12.5 years and a mean body mass index (BMI) of 29.36 ± 7.12 kg/m². Forty-seven (77%) were women, 44 were Caucasian, 14 Hispanic, and 2 were Black. Thirteen patients (21.3%) had hypertension, 9 arthritis, 7 asthma, 4 diabetes, and 2 COPD. The mean heart rate change after a 100-foot walk was 16.6 ± 8.1 beats/min, the mean gait speed was 76.1 ± 9.6 m/min, and the mean PCI was 0.22 ± 0.11 beats/m for all 9 walks (Table 1). The heart rates returned to their base line values after a 3-minute rest, and there were no episodes of desaturation (data not shown). Bivariate analysis of age, sex, BMI, and clinical diagnoses are reported in Table 2. BMI, hypertension, and arthritis had statistically significant effects on gait speed but not on heart rate change or PCI. In a multivariable regression model, both hypertension (β = −0.296; 95% confidence interval = −13.98 to −0.93; P = .026) and arthritis (β = −0.260; 95% confidence interval = −14.29 to −0.35; P = .04) reduced gait speed but not heart rate change or PCI. Medications with the potential to change heart rate responses to exercise are reported in Table 3. All subjects had qualitatively normal gaits with an average Tinetti score of 12 (data not shown).

Table 1.

Differences Between Walks.

	Gait Speed (m/min)	HR Change (beats/min)	Physiological Cost Index (beats/m)
All 9 walks (mean ± SD)	76.12 ± 9.64	16.59 ± 8.1	0.22 ± 0.11
6 walks (mean ± SD)	76.51 ± 9.33	16.66 ± 7.89	0.22 ± 0.11
Difference between second and third walk (mean ± SD)	0.56 ± 5.04	−0.02 ± 8.8	−0.0019 ± 0.11
Repeatability coefficient (9 walks)	17.61	36.48	0.49
Repeatability coefficient (6 walks)	10.83	28.36	0.37
Standard error of the mean (SEM)	1.19	0.72	0.0098
SEM%^a	1.56	4.3	4.45

SEM% = SEM/mean × 100.

Table 2.

Bivariate Analysis.

	Gait Speed	P	HR Change	P	PCI	P
Age in years
≤48 (n = 31)	75.22 ± 7.99	.45	17.68 ± 6.52	.13	0.24 ± 0.09	.11
>48(n = 30)	77.05 ± 10.51		15.47 ± 4.37		0.20 ± 0.06
Gender
Female (47)	75.25 ± 8.80	.18	16.43 ± 6.07	.68	0.22 ± 0.08	.96
Male (14)	79.03 ± 10.6		17.14 ± 3.91		0.22 ± 0.05
BMI in kg/m²
≤30 (n = 35)	79.45 ± 9.77	.001	16.18 ± 4.99	.51	0.20 ± 0.06	.09
>30 (n = 26)	71.64 ± 6.38		17.15 ± 6.46		0.24 ± 0.09
HTN
Yes (n = 13)	71.82 ± 7.69	.04	15.87 ± 4.96	.60	0.22 ± 0.07	.92
No (n = 48)	77.28 ± 9.40		16.80 ± 5.83		0.22 ± 0.08
Arthritis
Yes (n = 9)	70.20 ± 5.36	.04	12.27 ± 2.78	.01	0.17 ± 0.03	.04
No (n = 51)	77.06 ± 9.54		17.45 ± 5.67		0.23 ± 0.08
Asthma
Yes (n = 7)	75.53 ± 6.64	.82	18.09 ± 3.94	.49	0.24 ± 0.06	.43
No (n = 53)	76.38 ± 9.61		16.52 ± 5.80		0.22 ± 0.08
DM
Yes (n = 4)	71.43 ± 10.24	.32	16.35 ± 7.80	.24	0.28 ± 0.10	.14
No (n = 56)	76.30 ± 9.23		19.86 ± 5.51		0.22 ± 0.07
COPD
Yes (n = 2)	66.37 ± 0.08	.06	8.11 ± 2.99	.10	0.12 ± 0.04	.92
No (n = 59)	76.45 ± 9.26		16.88 ± 5.48		0.22 ± 0.08

Abbreviations: HR, heart rate; PCI, physiological cost index; BMI, body mass index; HTN, hypertension; DM, diabetes mellitus; COPD, chronic obstructive pulmonary disease.

Boldfaced numbers represent statistically significant differences at a P≤ .05.

Table 3.

Medications Relevant to Heart Rate Responses.

Medication	Number of patients (%)
Respiratory medications
Albuterol, albuterol/ipratropium	3 (4.9)
Fluticasone/salmeterol	3 (4.9)
Montelukast	1 (1.6)
Cardiac medications
Ca²⁺ blocker	5 (8.2)
ACEI	4 (6.6)
Beta-blocker	3 (4.9)
Diuretic	3 (4.9)
ARB	2 (3.2)
Other medications
Thyroid supplement	4 (6.6)

Abbreviations: ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker.

There were highly significant correlations between heart rate changes, gait speed, and PCI when calculated for all 9 separate measurements or when calculated for the first and second walk, the first and third walk, and the second and third walk for each test day (Table 4). The mean differences for heart rate change, gait speed and PCI, the repeatability coefficients, and the SEM% are reported in Table 1.

Table 4.

Intraclass Correlation Coefficients (ICC).

	ICC	95% Confidence Interval
Gait speed (9 walks)	0.91	0.88-0.94
Gait speed (walks 1 and 2)	0.91	0.88-0.94
Gait speed (walks 1 and 3)	0.90	0.87-0.94
Gait speed (walks 2 and 3)	0.95	0.92-0.96
HR change (9 walks)	0.41	0.32-0.53
HR change (walks 1 and 2)	0.42	0.31-0.54
HR change (walks 1 and 3)	0.49	0.38-0.61
HR change (walks 2 and 3)	0.44	0.33-0.56
PCI (9 walks)	0.43	0.33-0.54
PCI (walks 1 and 2)	0.42	0.31-0.54
PCI (walks 1 and 3)	0.49	0.38-0.61
PCI (walks 2 and 3)	0.47	0.36-0.59

Abbreviations: HR, heart rate; PCI, physiological cost index.

Discussion

Our results demonstrate that a 100-foot walk test reproducibly increases the heart rate in healthy adult volunteers. This allows the calculation of a non–steady state PCI, which in turn links function (gait speed) with physiologic stress (heart rate change). Steady state heart rate responses would likely require walking for 3 minutes, and this time requirement would increase the practical difficulties with testing in a clinic.⁶ The gait speeds and the calculated PCIs in our study are similar to the results reported in other healthy adult cohorts.^5,6,10,11 The test–retest variability in this study had satisfactory intraclass correlations (good, r = 040-0.75 to excellent, r> 0.75). The correlations for gait speed were much better than the correlations for heart rate change and PCI, and the correlations between walks 2 and 3 were better than the correlations between all 3 walks and the correlations between other test pairs. This result suggests that practice improves repeatability. The repeatability coefficient and the SEM% were smaller for the gait speed measurement than the heart rate change and the PCI.^7,9 Finally, the clinical diagnoses of hypertension and arthritis were associated with reduced the gait speed in these relatively healthy workers. This suggests that this test can identify small changes in performance status related to chronic stable medical conditions, at least in a group of subjects.

Gait speed measurements provide a simple method for patient testing that can be used in almost all settings. Lusardi and Fritz¹² have called walking speed the sixth vital sign and concluded that it is “easily measurable, clinically interpretable, and potentially a modifiable risk factor.” Bohannon¹ used the information from the 2001-2002 National Health and Nutrition Examination Survey to determine normal gait speed. He reported that the speed calculated over an 8-foot distance was similar to speed calculated over a 20-foot distance and that important factors explaining gait speed included age, stature, waist circumference, gender, and knee extension force.¹ In a study using community-based subjects, gait speed increased with height, lower extremity strength, and VO₂ max and decreased with age, the presence of depressive symptoms, and poor physical health status.¹³ Gait speed is a good predictor of the development of disability, institutionalization, and hospitalization in studies with older community-based subjects.^3,4 It decreases in some chronic medical conditions, such as hypertension, cerebrovascular disease, diabetes, renal disease, COPD, and aortic stenosis.^14-18

The PCI requires measurement of both heart rate changes and gait speed. Jaiyesimi and Fashakin⁵ and Bailey and Raatcliffe⁶ studied PCI measurements in young healthy adults to determine test–retest reliability. These volunteers walked on a treadmill at their preferred gait speed; the PCI measurements were calculated at non–steady state (at 1 minute), steady state, (at 4 minutes), and postexercise times.⁵ These 2 studies concluded that the PCI was a reasonable estimate of energy expenditure and that it could be used to evaluate locomotor disability, functional performance, and the effect of interventions in people with orthopedic impairment.^5,6 Graham et al¹⁹ measured PCI and O₂ consumption on 2 different tracks. They concluded that the PCI measurement was reliable and that the track configuration influenced the result because it influenced the pace of walking. PCI measurements have been done in patients with stroke, cerebral palsy, spinal cord injury, and rheumatoid arthritis.^20-26Danielssonet al²¹ reported a clear difference between patients with a stroke and healthy controls. They calculated that 53% of the variability in PCI could be explained by age, sex, group designation (stroke vs control), and O₂ consumption.²¹

The PCI in healthy adults depends on age and conditioning, and it depends on neuromuscular function, orthopedic impairment, and cardiorespiratory status in patients with chronic medical problems. Steven et al²⁶ demonstrated that anti-inflammatory drugs (indomethacin and naproxen) increased the preferred walking speed and decreased the PCI in patients with rheumatoid arthritis. This study demonstrates that these measurements provide a simple quantitative method to measure responses to drugs in some patients. However, the interpretation of PCI results in patients is complicated. The most important issue is whether or not a particular patient has a normal heart rate response to a given level of walking stress. Beta-blockers can reduce the heart rate response to exercise and may reduce the PCI without a change in gait speed. Therefore, the interpretation of abnormal PCIs in patients can be difficult because multiple factors influence heart rate change and gait speed. Consequently, this test has the most utility when serial tests are used to evaluate therapeutic interventions, to monitor disease course, and to monitor aging. It also encourages physicians to think about patient impairment using a physiologic analysis.

Our study involved a relatively large number of healthy adults who performed multiple tests (9 replicates total) in 3 separate sessions and analyzed the repeatability in these measurements. However, these results may not be generalizable to older adults and some clinic patients. We used a walk distance that likely produced non–steady state conditions in some subjects and possibly required more motivation and fitness than 8-foot and 10-meter walk tests.^1,27 Some of our patients were on medications, which might influence the heart rate response to walking, but the number of patients on any give class of medication was too small to allow meaningful subgroup comparisons. In addition, testing was done at different times during the day and unmeasured factors, such as medication dosing time, food intake, and fatigue could have influenced the results. However, these variables are probably unavoidable in a clinic setting. In general, the heart rate changes, gait speed, and PCIs in our subjects were similar to subjects in other studies.

In summary, our study demonstrates that gait speed, heart rate change, and PCI measurements are repeatable in healthy working adults. Our results indicate that simple measurements of gait speed will likely provide the most reproducible estimate of functional status and can form the basis for additional decision making in patient care in clinics and rehabilitation facilities. Multiple walk tests can reduce the test variability, but this approach has practical disadvantages and is probably not necessary when serial tests are used to follow the patients. Finally, gait speeds should be interpreted in the context of clinical anchors, when possible.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Rishi Raj is a pulmonary and critical care medicine physician in the Department of Internal Medicine at Northwestern University in Chicago.

Hoda Mojazi Amiri is an Internal Medicine resident at Eisenhower Medical Center in Rancho Mirage, CA.

Helen Wang is a fourth year medical student at Texas Tech University Health Sciences Center in Lubbock, TX.

Kenneth M. Nugent is a pulmonary and critical care medicine physician in the Department of Internal Medicine at Texas Tech University Health Sciences Center in Lubbock, TX.

References

Bohannon

. Population representative gait speed and its determinants.J GeriatrPhysTher. 2008;31:49-52.

Raj

Guerra

Sehli

. One hundred-foot walk test for functional assessment of clinic patients. Am J Med Sci. 2009;338:361-367.

Guralnik

Simonsick

Ferrucci

. A short physical performance battery assessing lower extremity function: association with self-reported disability and prediction of mortality and nursing home admission. J Gerontol. 1994;49:M85-M94.

Reuben

Siu

Kimpau

. The predictive validity of self-report and performance-based measure of function and health.J Gerontol. 1992;47:M106-M110.

Jaiyesimi

Fashakin

. Reliability of physiological cost index measurements. Afr J Med Med Sci. 2007;36:229-234.

Bailey

Ratcliffe

. Reliability of physiological cost index measurements in walking normal subjects using steady-state, non-steady-state and post-exercise heart rate recording. Physiotherapy. 1995;81:618-623.

Faria

Teixeira-Salmela

Neto

Rodrigues-de-Paula

. Performance-based tests in subjects with stroke: outcome score, reliability and measurement errors. Clin Rehabil. 2012;26:460-468.

Tinetti

. Performance-oriented assessment of mobility problems in elderly patients.J Am Geriatr Soc. 1986;34:119-126.

Dolmage

Hill

Evans

Goldstein

. Has my patient responded? Interpreting clinical measurements such as the 6-minute walk test. Am J Respir Crit Care Med. 2011;184:642-646.

10.

Bohannon

. Comfortable and maximum walking speed of adults aged 20-79 years: reference values and determinants. Age Ageing. 1997;26:15-19.

11.

Al-Obaidi

Wall

Al-Yaqoub

Al-Ghanim

. Basic gait parameters: a comparison of reference data for normal subjects 20 to 29 years of age from Kuwait and Scandinavia. J Rehabil Res Dev. 2003;40:361-366.

12.

Fritz

Lusardi

. White paper: “Walking speed: the sixth vital sign”. J GeriatPhysTher. 2009;32:2-5.

13.

Buchner

Cress

Esselman

. Factors associated with changes in gait speed in older adults. J Gerontol A BiolSci Med Sci. 1996;51:M297-M302.

14.

Stenholm

Chia

Simonsick

Ferrucci

. Gait pattern alterations in older adults associated with type 2 diabetes in the absence of peripheral neuropathy—results from the Baltimore Longitudinal Study of Aging. Gait Posture. 2011;34:548-552.

15.

Rosano

Longstreth

Jr Boudreau

. High blood pressure accelerates gait slowing in well-functioning older adults over 18-years of follow-up. J Am Geriatr Soc. 2011;59:390-397.

16.

Ilgin

Ozalevili

Kilinc

Sevinc

Cimrin

Ucan

. Gait speed as a functional capacity indicator in patients with chronic obstructive pulmonary disease. Ann Thorac Med. 2011;6:141-146.

17.

Painter

Kuskowski

. A closer look at frailty in ESRD: getting the measure right. Hemodial Int. 2013;17:41-49.

18.

Green

Woglom

Genereux

. Gait speed and dependence in activities of daily older adults with severe aortic stenosis. Clin Cardiol. 2012;35:307-314.

19.

Graham

Smith

White

. The reliability and validity of the physiological cost index in healthy subjects while walking on 2 different tracks. Arch Phys Med Rehabil. 2005;86:2041-2046.

20.

Fredrickson

Ruff

Daly

. Physiological cost index as a proxy measure for the oxygen cost of gait in stroke patients. Neurorehabil Neural Repair. 2007;21:429-434.

21.

Danielsson

Willen

Sunnerhagen

. Measurement of energy cost by the physiological cost index in walking after stroke. Arch Phys Med Rehabil. 2007;88:1298-1303.

22.

Raja

Joseph

Benjamin

Minocha

Rana

. Physiological cost index in cerebral palsy: its role in evaluating the efficiency of ambulation. J Pediatr Orthop. 2007;27:130-136.

23.

Gylfadottir

Dallimore

Dean

. The relation between walking capacity and clinical correlates in survivors of chronic spinal poliomyelitis. Arch Phys Med Rehabil. 2006;87:944-952.

24.

Van Hedel

Wirz

Dietz

. Assessing walking ability in subjects with spinal cord injury: validity and reliability of 3 walking tests. Arch Phys Med Rehabil. 2005;86:190-196.

25.

Ijzerman

Nene

. Feasibility of the physiological cost index as an outcome measure for the assessment of energy expenditure during walking. Arch Phys Med Rehabil. 2002;83:1777-1782.

26.

Steven

Capell

Sturrock

MacGregor

. The physiological cost of gait (PCG): a new technique for evaluating nonsteroidal anti-inflammatory drugs in rheumatoid arthritis. Br J Rheumatol. 1983;22:141-145.

27.

Peters