Reference Equations for the Six-Minute Walk

in Healthy Adults
PAUL L. ENRIGHT and DUANE L. SHERRILL
Respiratory Sciences Center, University of Arizona, Tucson, Arizona

In order to establish reference equations for prediction of the total distance walked during six min-
utes (6MWD) for healthy adults, we administered the standardized 6-min walk test to 117 healthy
men and 173 healthy women, aged 40 to 80 yr. Oxygen saturation (SaO2), pulse rate, and the degree
of dyspnea (Borg scale) were determined before and at the end of the walk. The median distance
walked was 576 m for men and 494 m for women. The 6MWD was significantly less for men and
women who were older and heavier, and for shorter men. The resulting gender-specific regression
equations explained about 40% of the variance in the distance walked for healthy adults: for men,
6MWD 5 (7.57 3 heightcm) 2 (5.02 3 age) 2 (1.76 3 weightkg) 2 309 m, and for women, 6MWD 5
(2.11 3 heightcm) 2 (2.29 3 weightkg) 2 (5.78 3 age) 1 667 m. These reference equations may be
used to compute the percent predicted 6MWD for individual adult patients performing the test for
the first time, when using the standardized protocol. Enright PL, Sherrill DL. Reference equations
for the six-minute walk in healthy adults. AM J RESPIR CRIT CARE MED 1998;158:1384–1387.

The ability to walk for a distance is a quick and inexpensive data collection have been previously detailed (3). Briefly, the popula-
measure of physical function, and an important component of tion in this study was a randomly chosen, stratified cluster of persons
quality of life, since it reflects the capacity to undertake day- in Tucson, Arizona, consisting of 3,805 individuals in 1,655 households
to-day activities. The 6-min walk can be performed by many who were enrolled in 1972 and 1973, with new enrollees added by
marriage and birth.
elderly, frail, and severely limited patients who cannot be tested
with standard (and more expensive) maximal cycle ergome- Six-Minute Walk Test
ter or treadmill exercise tests. The distance walked in 6 min
The 6-min walk test was conducted according to a standardized proto-
(6MWD) is reduced by several types of diseases, including ob- col (2). Subjects were instructed to walk from one end to the other of
structive lung disease, heart failure, arthritis, and neuromuscu- a 100-ft. hallway at their own pace, while attempting to cover as much
lar disease (1, 2). ground as possible in the allotted 6 min. Technicians encouraged sub-
Previous studies using the 6-min walk were conducted in jects with the standardized statements “You’re doing well” or “Keep
groups of patients with a disease associated with physical dys- up the good work,” but were asked not to use other phrases. Subjects
function, but did not include relatively healthy persons. We were allowed to stop and rest during the test, but were instructed to
desired a standard measure of physical function in a longitudi- resume walking as soon as they felt able to do so. Dyspnea, as mea-
nal, population-based study, and included the 6-min walk as sured with the modified Borg dyspnea scale (4), oxygen saturation
(SaO2), and pulse rate were assessed at the start and end of the 6-min
part of a comprehensive follow-up examination of the adults
walk test. Subjects were also asked at the end of the walk whether
in the cohort. The results allowed determination of the predic- they had experienced any of the following symptoms: dyspnea, chest
tors (based on height, weight, body mass index [BMI], age, pain, lightheadedness, or leg pain.
and gender) of the 6MWD in the healthy subset of study par-
ticipants. Cardiovascular Assessments
Prior to the 6-min walk test, we measured each subject’s ankle-arm in-
METHODS dex (AAI), a sensitive subclinical measure of reduced blood flow to
the legs (5), performed an electrocardiogram (ECG), recorded the
Study Population
subject’s systolic blood pressure and self-reported experience of high
The subjects were participants in the Tucson Epidemiological Study blood pressure, and took a history of stroke, any heart problem, or
of Airways Obstructive Disease. The study design and methods of any kind of heart or lung surgery since the time of the subject’s com-
pletion of the last previous questionnaire. Current use of diuretic, an-
tihypertensive, antiarrthymic, and anticoagulant medications was also
(Received in original form October 27, 1997 and in revised form May 20, 1998) self-reported.
Supported by Specialized Centers of Research Grant HL14136 from the National
Heart, Lung and Blood Institute. Pulmonary Assessments
Correspondence and requests for reprints should be addressed to Paul Enright, Spirometry was performed according to American Thoracic Society
M.D., UAZ/UMC Room 2342, 1501 N. Campbell Ave., Tucson, AZ 85724. E-mail: criteria (6), with reference values previously obtained from healthy
[email protected] members of our cohort (7). Subjects were classified as being never-
Am J Respir Crit Care Med Vol 158. pp 1384–1387, 1998 smokers, ex-smokers, or current smokers on the basis of their self-
Internet address: www.atsjournals.org administered questionnaire responses. Other pulmonary data collected
Enright and Sherrill: Six-Minute Walk 1385

TABLE 1 ated with 6MWD in the entire cohort. If the factor was a continuous
EXCLUSIONARY FACTORS FOR THE HEALTHY SUBSET variable, participants who fell beyond the 95th percentile toward ab-
normality were to be excluded from the healthy group. Therefore, the
Factor No. Excluded first step in our procedure was to build a linear regression model to
determine the correlates of 6MWD. For the remaining healthy subset
Age . 80 yr 35
BMI . 35 15
of participants, we chose the anthropomorphic factors that were sig-
AAI , 0.9 39 nificantly independently associated with 6MWD to enter into a model
History of stroke 8 predicting 6MWD. The lower limit of the normal range was defined as
Use of diuretics 42 the 5th percentile.
FEV1 , 70% predicted 48
Current smoker 52 RESULTS
Definition of abbreviations: AAI 5 ankle-arm blood pressure index; BMI 5 body mass Two hundred and five males and 253 females completed the
index.
Two hundred and nineteen participants were excluded (not the total of 239 given
6-min walk. The initial regression model (stepwise) deter-
here) because some participants had more than one factor that excluded them from mined the factors independently associated with a reduced
the healthy subset. 6MWD. Table 1 lists the exclusionary factors resulting from
this model and the number of participants excluded for each
factor. Exclusionary factors included age . 80 yr, BMI . 35
included breathing medications, a history of lung cancer, and any se- (indicating obesity), AAI , 0.90 (indicating peripheral vascu-
rious respiratory illness, lung surgery, or limited activity related to lar disease), a history of stroke, use of diuretics (suggesting hy-
lung disease. The subject’s standing height in stocking feet was mea- pertension or congestive heart failure), an FEV1 less than 70%
sured to the nearest centimeter, using a stadiometer, and the subject’s predicted (suggesting lung disease), and current smoking. The
weight was measured with a balance beam scale that was recalibrated only anthropomorphic variables that were significant in the
every month. model were age and weight, both of which had negative coeffi-
Statistical Methods cients. Height was significantly correlated with 6MWD in men
only. Figures 1 and 2 are scattergrams showing the bivariate
We decided a priori to define the healthy group by excluding subjects
association of age with 6MWD in men and women, including
with disease-related factors that were significantly negatively associ-

Figure 1. (A) Scattergram of the relationship between 6MWD and Figure 2. (A) Scattergram of the relationship between 6MWD and
age in 117 healthy adult men. The linear regression line is super- age in 173 healthy adult women. The linear regression line is su-
imposed, surrounded with the 95% confidence interval (CI) lines perimposed, surrounded with the 95% confidence interval (CI)
for the regression line, or (B), by the CI lines for the individual pre- lines for the regression line, or (B), by the CI lines for the individual
dicted values. predicted values.
1386 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 158 1998

TABLE 2 metric factors of age, weight, and height (for men) were inde-
CHARACTERISTICS OF THE HEALTHY ADULT PARTICIPANTS pendently associated with the distance walked. The gradual
reduction of skeletal muscle mass and strength that generally
Median (5th, 95th percentiles)
occurs with aging (14), and the increased prevalence of debili-
Men Women tating diseases, which we did not objectively measure, are
Characteristic (n 5 117) (n 5 173) probably responsible for the shorter distance walked in our
Age 59.5 (43.1, 77.0) 62.0 (45.0, 79.0) more elderly study participants. A taller height is associated
Height, cm 176 (164, 185) 162 (151, 173) with a longer stride, which makes walking more efficient,
BMI, kg/m2 27.8 (22.1, 33.9) 25.5 (20.6, 32.4) probably resulting in a longer distance walked by taller men.
Baseline pulse 74 (58, 100) 80 (60, 105) Obesity increases the workload for a given amount of exer-
Change in pulse 125 (28, 179) 120 (26, 158) cise, probably resulting in the shorter distanced walked by
Baseline SaO2 96% (92, 98) 96% (93, 99)
women with a higher body weight or BMI. A small study of
Change in SaO2 0.0 (28.0, 12.0) 0.0 (216.0, 12.0)
FEV1, %pred 96% (75, 118) 106% (81, 125) elderly men with COPD found that correcting for body weight
6MWD, m 576 (399, 778) 494 (310, 664) improved the correlation of 6MWD with maximal oxygen
consumption (12). Our results suggest that when 6MWDs are
Definition of abbreviations: BMI 5 body mass index; 6MWD 5 6-min walk distance.
reported in future studies, they should be corrected for age,
height, weight, and gender.
About 60% of the variance in 6MWD remains unexplained
the fitted regression lines. Either weight or BMI were signifi- by our models. Future population-based studies of relatively
cant when entered into the models, and resulted in nearly healthy persons that include the 6-min walk test may be able
equivalent values for R2. to provide models that explain more of the variance and
After the exclusionary factors were applied, 117 men and thereby narrow the normal range (and increase the lower limit
173 women remained in the healthy subset. Table 2 gives their of the normal range provided by our reference equations). For
characteristics. Table 3 gives the gender-specific equations for instance, knowledge of exercise habits and cardiopulmonary
predicting 6MWD. These models explained 42% and 38% of conditioning, and of the presence of musculoskeletal problems
the variation in 6MWD for men and women, respectively (R2 (such as arthritis), might improve future models.
values). After the predicted 6WMD value from Table 3 for an The 6MWD was an excellent independent predictor of
individual patient is computed, the lower limit of the normal morbidity and mortality after 1 yr in 898 patients with heart
range (LLN) for that patient may be obtained by subtracting failure (15). The mean distance walked by these patients was
153 m from the 6MWD value for men or by subtracting 139 m 374 m (SD 5 117 m). It remains to be seen whether or not the
for women. 6MWD will be a strong independent predictor of morbidity and
mortality in population samples or in patients with COPD.
DISCUSSION Caution should be exercised when applying our regression
equations to patients who have characteristics that fall outside
The 12-min walking test was introduced in 1968 as a guide to
of our cohort, including non-Causians and those younger than
physical fitness (8), and was later applied to patients with
40 yr or older than 80 yr. Larger values may be expected from
chronic obstructive pulmonary disease (COPD) (9). It was
persons who have previously performed 6-min walk tests and
then found that decreasing the time of the walk to 6 min did
those to whom nonstandardized encouragement is given.
not significantly reduce the utility of the test (2). The 6-min
walk test has been validated by high correlation with work-
loads, heart rate, and SaO2, the dyspnea responses when com- References
pared with standard bicycle ergometry and treadmill exercise
1. Guyatt, G. H., M. J. Sullivan, P. J. Thompson, E. L. Fallen, S. O. Pugsley,
tests (10–12). There is a learning effect when the test is per- D. W. Taylor, and L. B. Berman. 1985. The six-minute walk: a new
formed on two successive days, with a mean 15% improve- measure of exercise capacity in patients with chronic heart failure.
ment in distance walked (13), but this effect is not important Can. Med. Assoc. J. 132:919–923.
when determining cross-sectional correlations, or when using 2. Butland, R. J. A., J. Pang, E. R. Gross, A. A. Woodcock, and D. M. Ged-
the results as a baseline predictor of later events. Our results des. 1982. Two, six, and 12 minute walking tests in respiratory disease.
apply to the first (initial) 6-min walk performed by a patient. B.M.J. 284:1607–1608.
3. DiPede, C., G. Viegi, J. J. Quackenboss, P. Boyer-Pfersdorf, and M. D.
In the gender-specific models, we found that the anthropo- Lebowitz. 1991. Respiratory symptoms and risk factors in an Arizona
population samples of Anglo- and Mexican-American whites. Chest
99:916–922.
TABLE 3 4. Wilson, R. C., and P. W. Jones. 1989. A comparison of the visual ana-
logue scale and modified Borg scale for the measurement of dyspnea
REFERENCE EQUATIONS FOR 6-MIN WALK DISTANCE
during exercise. Clin. Sci. 76:277–282.
IN HEALTHY ADULTS
5. Hiatt, W. R., J. A. Marshall, J. Baxter, R. Sandoval, W. Hilderbrant,
Men: L. R. Kahn, and R. F. Hamman. 1990. Diagnostic methods for periph-
6MWD 5 (7.57 3 heightcm) 2 (5.02 3 age) 2 (1.76 3 weightkg) 2 309 m. eral arterial disease in the San Luis Valley diabetes study. J. Clin. Epi-
Alternate equation using BMI*: demiol. 43:597–606.
6MWD 5 1,140 m 2 (5.61 3 BMI) 2 (6.94 3 age) 6. American Thoracic Society. 1987. Standardization of spirometry: 1987
When using either equation, subtract 153 m for the LLN update. Am. Rev. Respir. Dis. 136:1285–1298.
7. Knudson, R. J., M. D. Lebowitz, C. J. Holberg, and B. Burrows. 1983.
Women:
Changes in the normal maximal expiratory flow-volume curve with
6MWD 5 (2.11 3 heightcm) 2 (2.29 3 weightkg) 2 (5.78 3 age) 1 667 m.
growth and aging. Am. Rev. Respir. Dis. 127:725–734.
Alternate equation using BMI:
8. Cooper, K. H. 1968. A means of assessing maximal oxygen intake.
6MWD 5 1,017 m 2 (6.24 3 BMI) 2 (5.83 3 age)
J.A.M.A. 203:201–204.
When using either equation, subtract 139 m for the LLN
9. McGavin, C. R., M. Artvinli, H. Naoe, and G. J. R. McHardy. 1978. Dys-
Definition of abreviations: BMI 5 body mass index; 6MWD 5 6-min walk distance; pnea, disability, and distance walked: comparison of estimates of exer-
LLN 5 lower limit of normal. cise performance in respiratory disease. B.M.J. 2:241–243.
* BMI in kg/m2. 10. Langenfeld, H., B. Schneider, W. Grimm, M. Beer, M. Knoche, G. Rieg-
Enright and Sherrill: Six-Minute Walk 1387

ger, and K. Kochsiek. 1990. The six minute walk test: an adequate ex- and N. T. Bateman. 1992. Portable liquid oxygen and exercise ability
ercise test for pacemaker patients? PACE 13:1761–1765. in severe respiratory disability. Thorax 47:781–789.
11. Spence, D. P. S., J. G. Hay, J. Carter, M. G. Pearson, and P. M. A. Cal- 14. Tolep, K., and S. G. Kelsen. 1993. Effect of aging on respiratory skeletal
verley. 1993. Oxygen desaturation and breathlessness during corridor muscles. Clin. Chest Med. 3:363–378.
walking in COPD: effect of oxitropium bromide. Thorax 48:1145–1150. 15. Bittner, V., D. H. Weiner, S. Yusuf, W. J. Rogers, K. M. McIntyre, S. I.
12. Bernstein, M. L., J. A. Despars, N. P. Singh, K. Avalos, D. W. Stansbury, Bangdiwala, M. W. Kronenberg, J. B. Kostis, R. M. Kohn, M. Guil-
and R. W. Light. 1994. Reanalysis of the 12 minute walk in patients lotte, B. Greenberg, P. A. Woods, and M. G. Bourassa. 1993. Predic-
with COPD. Chest 105:163–167. tion of mortality and morbidity with a six minute walk test in patients
13. Leach, R. M., A. C. Davidson, S. Chinn, C. H. C. Twort, I. R. Cameron, with left ventricular dysfunction. J.A.M.A. 270:1702–1707.