The 6-min Walk Test*

A Quick Measure of Functional Status in Elderly

Adults
Paul L. Enright, MD; Mary Ann McBurnie, PhD; Vera Bittner, MD;
Russell P. Tracy, PhD; Robert McNamara, MD; Alice Arnold, PhD; and
Anne B. Newman, MD; for the Cardiovascular Health Study

Objectives: To determine the correlates of the total 6-min walk distance (6MWD) in a population
sample of adults > 68 years old.
Methods: The standardized 6-min walk test (6MWT) was administered to the Cardiovascular
Health Study cohort during their seventh annual examination.
Results: Of the 3,333 participants with a clinic visit, 2,281 subjects (68%) performed the 6MWT.
There were no untoward events. The mean 6MWD was 344 m (SD, 88 m). Independent general
correlates of a shorter 6MWD in linear regression models in women and men included the
following: older age, higher weight, larger waist, weaker grip strength, symptoms of depression,
and decreased mental status. Independent disease or risk factor correlates of a shorter 6MWD
included the following: a low ankle BP, use of angiotensin-converting enzyme inhibitors, and
arthritis in men and women; higher C-reactive protein, diastolic hypertension, and lower FEV1 in
women; and the use of digitalis in men. Approximately 30% of the variance in 6MWD was
explained by the linear regression models. Newly described bivariate associations of a shorter
6MWD included impaired activities of daily living; self-reported poor health; less education;
nonwhite race; a history of coronary heart disease, transient ischemic attacks, stroke, or diabetes;
and higher levels of C-reactive protein, fibrinogen, or WBC count.
Conclusions: Most community-dwelling elderly persons can quickly and safely perform this
functional status test in the outpatient clinic setting. The test may be used clinically to measure
the impact of multiple comorbidities, including cardiovascular disease, lung disease, arthritis,
diabetes, and cognitive dysfunction and depression, on exercise capacity and endurance in older
adults. Expected values should be adjusted for the patient’s age, gender, height, and weight.
(CHEST 2003; 123:387–398)

Key words: 6-min walk; elderly; exercise; functional status; heart failure

Abbreviations: 6MWD ⫽ 6-min walk distance; 6MWT ⫽ 6-min walk test; AAI ⫽ ankle-arm index;
ACE ⫽ angiotensin-converting enzyme; ADL ⫽ activities of daily living; ATS ⫽ American Thoracic Society;
AVAS ⫽ additive and variance stabilizing transformation; BMI ⫽ body mass index; CHF ⫽ congestive heart failure;
CHS ⫽ Cardiovascular Health Study; CI ⫽ confidence interval; CVD ⫽ cardiovascular disease; LVM ⫽ left ventricular
mass; MI ⫽ myocardial infarction; MMSE ⫽ Mini-Mental State Examination; TIA ⫽ transient ischemic attack

T heinexpensive
ability to walk for a distance is a quick and
performance-based measure, and an
imal cycle ergometry or treadmill exercise tests.2
Walking tests are more reliable than other perfor-
important component of quality of life, since it
reflects the capacity to undertake day-to-day activi- For editorial comment see page 325
ties or, conversely, functional limitation.1 The 6-min
walk test (6MWT) can be performed by many el- mance-based measures in elderly persons, such as
derly, frail, and severely limited patients who cannot timed chair stands and weight lifting.3 The 6-min
be tested using standard (and more expensive) max- walk distance (6MWD) is known to be reduced by

*From the University of Alabama at Birmingham (Dr. Bittner), This research was supported by contracts N01-HC-85079
Birmingham, AL; PAD Clinical Trial Center (Dr. McBurnie), through N01-HC-85086, N01-HC-35129, and N01-HC-15103
Seattle, WA; Pediatrics (Dr. Newman), University of Pittsburgh, from the National Heart, Lung, and Blood Institute.
Pittsburgh, PA; University of Vermont (Dr. Tracy); University of Manuscript received November 6, 2001; revision accepted June
Arizona (Dr. Enright) and private practice (Dr. McNamara), 6, 2002.
Phoenix, AZ; University of Washington Coordinating Center (Dr. Correspondence to: Paul Enright, MD, 4460 East Ina Rd, Tucson,
Arnold), Seattle, WA. AZ 85718; e-mail: [email protected]

www.chestjournal.org CHEST / 123 / 2 / FEBRUARY, 2003 387

several types of diseases, including obstructive lung position and elbow flexed 90°. Grip strength was measured three
disease, heart failure, arthritis, and neuromuscular times for each hand, and the highest value (in kilograms) from the
participant’s dominant hand was used for our analysis. A medi-
disease.4,5 cation inventory was obtained at each examination,9 but the
Previous studies using the 6MWT were limited to specific indication for each medication was not determined.
patients with a given disease, and did not include Participants assessed their general health by answering the
elderly persons from several community samples. question, “Would you say, in general, your health is (excellent,
The Cardiovascular Health Study (CHS) is a pro- very good, good, fair, or poor)?” Limitation of instrumental
activities of daily living (ADL) was defined as trouble performing
spective observational study of a general population any of the following: light or heavy housework, shopping, meal
sample designed to study the epidemiology and risk preparation, money management, and using the telephone.
factors associated with cardiovascular disease (CVD) Symptoms of depression were assessed using the modified
in the elderly. The correlates of four performance- Center for Epidemiologic Studies depression scale of 0 to 30.10
based measures— gait speed, timed chair stands, The categorical variable depression (a mood, not a diagnostic
category) was defined as a score of ⱖ 15. Good social support was
grip strength, and maximal inspiratory pressure— defined as a score of ⱖ 12 using a standard scale.11 Cognitive
obtained during the baseline examination of the function was assessed by trained interviewers using a modified
cohort have previously been described.6 The stan- Mini-Mental Status Examination (MMSE),12 scored on a scale of
dardized 6MWT was included in a follow-up exam- 0 to 100 (including both serial 7s and spelling “world” back-
ination of the cohort, along with many tests of wards). All of the measurements described thus far were concur-
rent with the examination in which the 6MWT was done.
cardiovascular and pulmonary risk factors and dis- From 1992 through 1993, we measured each subject’s systolic
ease. We hypothesized that the 6MWD would be BP in both ankles (tibial arteries) and their right arm (brachial
associated with many of these factors. artery) at rest in supine position, using a hand-held Doppler
transducer, and later calculated the ankle-arm index (AAI), a
sensitive subclinical measure of reduced blood flow to the legs.13
Blood was obtained while fasting, and analyzed for RBC and
Materials and Methods WBC counts, blood chemistry, and lipoprotein levels.14 At base-
line for each cohort, blood was analyzed for C-reactive protein
Study Population levels.15 Echocardiography was performed during visits from
1995 through 1996.16
Participants in the CHS were selected using a Medicare
eligibility list provided by the US Health Care Financing Admin-
CVD Assessment
istration for the four participating communities: Forsyth County,
North Carolina; Pittsburgh, PA; Sacramento County, California; Health status was assessed at baseline through self-report of
and Washington County, Maryland. These communities are physician diagnosis of diseases. Self-report of CVDs were vali-
diverse in proportion of minorities, education and income levels, dated according to standardized criteria by the medications used,
degree of urbanization, death rates, and availability of medical and by examination data such as BP, ECG,17 echocardiography,
care. The initial study cohort of 5,201 participants was recruited and carotid ultrasound.18 Cardiovascular events occurring after
and examined in 1989 through 1990. An additional cohort of 687 baseline and prior to the 6MWT were validated by a review of
African Americans was enrolled in 1992 and 1993 in order to medical records and adjudicated according to standardized cri-
enhance the representation of the study. The 6MWT was done teria.19
just once, during the seventh annual clinic examination of the
original cohort and the fourth annual examination of the added
Pulmonary Assessments
cohort (June 1996 to May 1997).
The following were exclusion factors for study entry: institu- Spirometry was performed according to American Thoracic
tionalized, terminal illness; inability to walk, communicate, or Society (ATS) criteria,20 with reference values previously ob-
give informed consent; or likely to move from the area during the tained from healthy members of our cohort.21 The smoking status
next 3 years. Enrolled CHS participants were younger, more of each participant at each examination was categorized as
educated, and more likely to be married and white than those never-smoker, former smoker, or current smoker, using re-
who refused or were ineligible. The CHS design and recruitment sponses to the standardized ATS DLD-78 respiratory question-
are described in detail elsewhere.7,8 The research protocol was naire.22
reviewed and approved by the institutional review board for Exclusions from the 6MWT included the following: regular use
human studies at the four clinical centers, and informed consent of an ambulatory aid (cane or walker); a resting oxygen saturation
was obtained. ⬍ 90%; inability to walk due to musculoskeletal problems; chest
pain in the previous 4 weeks; a heart attack, angioplasty, or heart
Examinations surgery in the previous 3 months; heart rate ⬍ 50 beats/min at
rest (unless a physician or nurse determined that an AV block or
Study participants completed standardized interviews in both conduction problem was not the cause of the bradycardia); heart
the home and the field center, and extensive examinations at the rate ⬎ 110 beats/min at rest; acute ST-T wave changes on the
field centers at baseline and during the annual examinations. Not ECG; participant refusal; or judgment of the clinic staff that the
all components were repeated at each annual visit. Standing participant would probably not be able to complete the walk
height was measured in stocking feet to the nearest centimeter safely (technician discretion). These exclusions were probably
using a stadiometer, and weight was measured using a balance conservative (excluding many participants who would have ea-
beam scale, recalibrated monthly. Handgrip strength was mea- gerly and safely performed the test) because physicians could not
sured using a Jamar dynamometer set at the second handle be present in the clinics during all of the examinations to assess
position. The participants were seated with their wrist in a neutral and treat symptomatic participants.

388 Clinical Investigations

6MWT steroidal anti-inflammatory drugs for both genders; and estrogen
in women); and “function” variables (grip strength, MMSE score,
The 6MWT was conducted according to a standardized proto- and depression score). To determine entry and removal of
col,23 using an internal hallway with the 100-foot distance marked candidate variables from the model, p values of 0.05 and 0.06
by colored tape on the floor. Participants were told that “the were used, respectively. The piecewise terms were included in
purpose of this test is to see how far you can walk in six minutes.” the model based on partial F tests for simultaneous significance
They were then instructed to “walk from end to end of the of both coefficients. Once the stepwise selection was completed,
hallway at your own pace, in order to cover as much ground as the final models were refit using only the selected variables in
possible.” Each minute, technicians encouraged the participants order to minimize the amount of missing data. Appropriateness
with the standardized statements “You’re doing well” or “Keep up of the functional forms was reaffirmed using AVAS.
the good work,” but were asked not to use other phrases. Adjustment was not made for multiple comparisons. All p
Participants were allowed to stop and rest during the test, but values are presented as relative measures of the strength of the
were instructed to resume walking as soon as they were able to do associations and should not be strictly interpreted because of the
so. The technician used a mechanical lap counter to count the large number of statistical tests performed. Analyses were carried
number of laps completed, and an electronic timer with a buzzer out using SPSS for Windows (Release 9.0; SPSS; Chicago, IL)26
that sounded 6 min after the walk started. Before the walk started and S-Plus.27
and at the end of the 6-min walk, participants were shown a Since the 6MWD was associated with age, gender, race, height,
modified Borg dyspnea scale24 printed on a card and asked to and weight, reference equations for the healthy subset of partici-
“indicate your current degree of shortness of breath” on a scale of pants were determined using these variables as predictors of 6MWD
“0 ⫽ nothing at all” to “10 ⫽ very, very severe.” At the end of the in a linear regression model. Participants with factors associated with
walk, they were asked if they had experienced any of the a shorter 6MWD were excluded from the healthy subset.
following specific symptoms: dyspnea, chest pain, light-headed-
ness, or leg pain, or any other symptoms.

Results
Statistical Methods
6MWT Exclusions and Safety
Preliminary descriptive analyses included frequencies, histo-
grams, and error bar plots to examine bivariate relationships with Of the 3,333 participants who attended the 1996
total distance walked. For bivariate associations, Pearson ␹2 tests through 1997 clinic visit, approximately one third were
were used to evaluate associations between categorical variables, excluded or chose not to try the 6MWT (Table 1).
analysis of variance F tests for associations between continuous
and categorical variables, and t tests for partial correlations
between continuous measures.
Multiple linear regression analysis was performed to determine
relationships between total distance walked and potential predic- Table 1—Participation, Exclusions, and Completion of
tors. Variables were examined for the linearity of their relation- the 6MWT
ship with 6MWD using the additive and variance stabilizing
transformation (AVAS) in S-Plus (StatSci; Seattle, WA).25 AVAS Variables No. (%)
is a nonparametric regression technique that attempts to find
Refused or unable to participate 297 (6)
smooth transformations that approximate an additive model.
Interviewed outside the clinic 1,080 (23)
These transformations can be used to suggest appropriate func-
Year 9 clinic visit done 3,333 (71)
tional forms for standard linear models. Linear piecewise trans-
Total year 9 participants 4,710 (100)
formations were suggested for AAI, diastolic BP (women only),
Of the 3,333 patients with a year 9 clinic visit done
and MMSE, as was a quadratic form for weight that was centered
Excluded from 6MWT due to 766 (23)
to reduce collinearity between the linear and quadratic terms.
ECG alert 31
Median values appeared to be reasonable cut-points for the
Ambulatory aid 331
piecewise transformations. Analyses were stratified by gender
Aortic stenosis 15
based on results of preliminary analyses that suggested differ-
Heart rate ⬍ 50 beats/min or ⬎ 110 beats/min 153
ences in some relationships by gender, but any variable that was
Systolic BP ⬎ 200 mm Hg or diastolic BP ⬎ 110 22
entered into the model for one gender was also entered into the
mm Hg
model for the other gender.
Recent MI, angioplasty, bypass surgery 7
A series of stepwise regressions were fit for successive blocks of
New chest pain, dyspnea, or fainting 84
covariates. Candidate variables were stepped into the model in
Resting oxygen saturation ⬍ 90% 1
the following groups: demographics (age, race, site indicators,
Technician discretion 122
less than high school education); anthropometry (body mass
Did not participate in 6MWT (unknown reason) 158 (5)
index [BMI], waist circumference, weight, weight squared, stand-
6MWT not attempted 128 (4)
ing height, and arm span); clinical and subclinical disease (prev-
Refused at interview 34
alence for diabetes, arthritis, angina, congestive heart failure
Physically unable 2
[CHF], claudication, myocardial infarction [MI], stroke, and
Technician discretion 58
transient ischemic attack [TIA], FEV1, AAI, diastolic and brachial
Other 34
and tibial systolic BP), measures of inflammation (fibrinogen,
6MWT partially completed 164 (5)
WBC count, and C-reactive protein); smoking status (current vs
Refused during 6MWT description 3
never or former); echocardiographic variables (left ventricular
Physically unable 42
mass [LVM], regional wall motion, and percentage of fractional
6MWT stopped by the technician 88
shortening); medications (␤-blockers, diuretics, vasodilators, an-
Other reason, unknown 31
giotensin-converting enzyme [ACE] inhibitors, calcium channel
Successfully completed the 6MWT 2,117 (64)
blockers, digitalis, hypertension, insulin, lipid-lowering, and non-

www.chestjournal.org CHEST / 123 / 2 / FEBRUARY, 2003 389

The 164 participants who started the walk but stopped Table 3—Symptoms Reported at the End of the
walking before 6 min had elapsed (partial completers) 6MWT*
were included in the analyses for this report. Those Partial
who completed the walk (n ⫽ 2,117) were significantly Completers Completers Total
healthier in many respects when compared to the Any symptoms 517/2,117 (24) 64/164 (39) 581/2,281 (25)
partial completers, or to those who did not perform the Chest pain 24 (4) 5 (6) 29 (4)
test (Table 2). Those with prevalent CVD (a history of Lightheadedness 115 (19) 13 (17) 128 (19)
angina, MI, CHF, claudication, TIA, or stroke) were Leg pain 204 (34) 26 (33) 230 (34)
much less likely to have tried or completed the walk. Other symptoms† 254 (43) 34 (44) 288 (43)
Half of the cohort who did not attend the clinic visit *Data are presented as No./total (%) or No. (%).
(and 37% of those who attended the clinic visit but did †The majority of these symptoms involved muscle or joint pain,
discomfort, or fatigue (n ⫽ 184). Other symptoms also included
not try the walk) reported that their general health was
shortness of breath, tightness or pressure in chest, wheezing
only fair or poor, compared to only 16% of those who (n ⫽ 27); general weakness or fatigue (n ⫽ 26); dizziness or balance
completed the walk. problems (n ⫽ 14); nausea (n ⫽ 6); and miscellaneous symptoms
There were no untoward events associated with (n ⫽ 31). Some participants reported more than one symptom at the
the test (no need for emergency evaluation or ther- end of the test.
apy). All participants were asked about symptoms at
the end of the walk, and approximately 75% said that
they had no symptoms (Table 3). The most common The mean distance walked was 362 m (1,188 feet)
symptoms reported by the others included leg pain, for men and 332 m (1,089 feet) for women. The
muscle or joint pain, discomfort, or fatigue. Only 29 6MWD distribution was skewed toward shorter dis-
participants reported chest pain. Thirty-nine percent tances.
of the 164 participants who were partial completers
reported a symptom at that point (compared to 24% Bivariate and Nonlinear Associations With 6MWD
of those who walked for the entire 6 min); however,
the distribution of symptoms did not differ between There was a nonlinear relationship of body weight
the two groups. and BMI with 6MWD (Fig 1). Patients with a low or

Table 2—Comparison of Study Participants Who Completed the 6MWT, Partial Completers, and Those Who Were
Not Tested*

Completed Partial Visit Done, No Clinic

the 6MWT Completers No 6MWT† Visit
Characteristics (n ⫽ 2,117) (n ⫽ 164) (n ⫽ 1,052) (n ⫽ 1,377) p Value‡

Age 77 (4) 78 (5) 79 (5) 81 (6) ⬍ 0.001

Male gender 40.3 32.9 40.3 35.2 0.005
Nonwhite race 13.9 20.7 20.8 17.4 ⬍ 0.001
Less than high school 22.4 22.6 27.0 35.8 ⬍ 0.001
education
Waist size, cm 96 (12) 99 (16) 98 (14) 98 (14) ⬍ 0.001
Weight, lb 159 (30) 162 (38) 161 (35) 151 (36) 0.33
Height, cm 164 (9) 162 (10) 163 (10) 160 (10) ⬍ 0.001
Angina 18.0 23.8 30.4 30.9 ⬍ 0.001
MI 8.5 15.9 15.3 17.4 ⬍ 0.001
CHF 5.3 6.1 15.4 18.0 ⬍ 0.001
Stroke 3.6 5.5 11.0 13.4 ⬍ 0.001
TIA 2.8 5.5 4.9 6.2 ⬍ 0.001
Claudication 2.2 5.5 4.3 5.6 ⬍ 0.001
Diabetes 13.3 14.6 16.8 8.4 ⬍ 0.001
Arthritis 28.5 36.9 42.7 37.4 ⬍ 0.001
Current smoker 6.4 11.8 8.6 9.1 0.008
Echocardiographic LVM 147 (48) 140 (39) 158 (51) 153 (53) ⬍ 0.001
Limited ADL 0.14 (0.44) 0.13 (0.39) 0.61 (1.1) 1.20 (1.8) ⬍ 0.001
Fair or poor health 15.6 27.4 37.1 50.5 ⬍ 0.001
Depression score 5.1 (4.5) 5.9 (4.6) 7.1 (5.3) 7.4 (6.2) ⬍ 0.001
Cognitive function 93 (8) 91 (10) 88 (14) 73 (25) ⬍ 0.001
*Data are presented as mean (SD) or %.
†This category includes participants who came into the clinic but refused the test, or were excluded from the test, or were unable to perform the
test because of physical, cognitive, or equipment problems.
‡p values are unadjusted, and test for a trend across the categories (four rows).

390 Clinical Investigations

Figure 1. The association of BMI by deciles with 6MWD. Figure 3. The relationship of diastolic BP with 6MWD in
Note the shorter distance walked by obese elderly persons elderly women. A breakpoint was found at a diastolic BP of
(BMI ⬎ 30). 69 mm Hg.

diabetes, lower lung function, and weaker grip

a high weight (or BMI) did not walk as far as the strength; and higher fibrinogen, C-reactive protein,
others. The upper threshold for a lower distance was and WBC count. Participants who reported limita-
a BMI ⬎ 30 or a weight ⬎ 180 lb. We did not use tions in ADL, fair to poor general health, more
BMI in subsequent models since it forces a specific symptoms of depression, and those with lower cog-
relationship between weight and height that is less nitive function (lower MMSE score) did not walk as
powerful as a predictor when compared to using far. Although there was a tendency for those who did
both weight and weight squared in the model. not walk as far to have abnormal echocardiographic
There was also a nonlinear association of ankle findings, these trends were not significant.
BP (and AAI) with 6MWD in both men and
women (Fig 2). AVAS analysis suggested a break-
Independent Predictors of 6MWD in Regression
point at an AAI of 1.1 in women and 1.2 in men.
Models
There was also a breakpoint (at 69 mm Hg) in the
relationship of diastolic (brachial) BP and 6MWD Tables 5, 6 show the independent correlates of
in women (Fig 3). 6MWD, using gender-specific, stepwise, linear re-
For those variables that were significantly corre- gression models. Approximately 30% of the variance
lated with 6MWD, Table 4 lists their mean values for in 6MWD was explained by the final models. The
each quintile of 6MWD; for each categorical vari- distances were approximately 7% lower at two of the
able, the percentages are given. Disease and risk clinics (Hagerstown, MD, and Pittsburgh, PA). Age
factors that were more likely in the lowest quintile of and waist size remained strongly associated with
distance walked included a history of any CVD, 6MWD. The nonlinear effects of weight and AAI
also remained as strong predictors of 6MWD. Racial
differences also persisted in the final models, but
high school education became nonsignificant as dis-
ease variables entered the models. Diastolic BP, lung
function, arthritis, and C-reactive protein level re-
mained significant correlates in women.
The only echocardiographic variable that was sig-
nificant in any model was moderate-to-severe left
ventricular wall motion abnormalities in women
(p ⫽ 0.002). Since only 776 women completed this
examination, the overall strength of that model was
decreased, and so we elected not to enter echocar-
diographic variables into the final models. None of
the echocardiographic variables were significant in
any of the models for men, but the use of digitalis
and ACE inhibitors was associated with lower
Figure 2. The relationship between the AAI of systolic BP and
6MWD. A breakpoint was found at AAI of 1.10 for women and 6MWD, and the use of lipid-lowering medications
AAI of 1.20 for men. was associated with a higher 6MWD.

www.chestjournal.org CHEST / 123 / 2 / FEBRUARY, 2003 391

 Table 4 —Bivariate Associations With the Distance Walked During the 6MWT*

Quintile of Total Distance Walked

1 2 3 4 5
Characteristics 3 to 278 m 279 to 332 m 333 to 367 m 368 to 412 m ⱖ 413 m p Value†

Age, yr 78.8 (5.0) 78.1 (4.4) 77.3 (4.2) 76.7 (4.0) 76.2 (3.3) ⬍ 0.001
Male gender 30.1 33.4 38.5 44.7 52.5 ⬍ 0.001
Nonwhite race 19.0 17.1 16.4 11.8 7.4 ⬍ 0.001
Less than high school education 29.8 28.4 23.1 18.3 12.1 ⬍ 0.001
BMI 27.5 (5.4) 27.3 (4.5) 26.8 (4.0) 26.4 (3.8) 25.8 (3.3) ⬍ 0.001
Waist size, cm 99 (15) 98 (13) 96 (12) 96 (12) 94 (11) ⬍ 0.001
Weight, lb 158 (36) 159 (31) 159 (29) 159 (28) 159 (28) 0.65
Height, cm 161 (9) 162 (9) 164 (9) 165 (9) 167 (9) ⬍ 0.001
Angina 21.1 21.2 17.3 15.4 17.3 0.020
MI 13.7 9.6 6.8 7.3 7.6 ⬍ 0.001
CHF 8.7 6.7 3.5 3.8 4.0 ⬍ 0.001
Stroke 6.7 4.9 2.6 3.0 1.3 ⬍ 0.001
TIA 6.7 4.9 2.6 3.0 1.3 0.001
Claudication 5.9 2.0 1.3 2.1 0.7 ⬍ 0.001
Diabetes 18.1 12.5 14.0 13.3 8.7 ⬍ 0.001
Arthritis 36.4 38.9 27.6 19.8 22.8 ⬍ 0.001
Systolic BP, mm Hg 138 (21) 136 (19) 135 (20) 136 (19) 134 (18) 0.003
Diastolic BP, mm Hg 69 (11) 70 (10) 69 (11) 70 (11) 71 (10) 0.065
AAI 1.07 (0.19) 1.10 (0.17) 1.14 (0.15) 1.13 (0.12) 1.16 (0.13) ⬍ 0.001
FEV1, L 1.7 (0.57) 1.8 (0.53) 1.9 (0.57) 2.1 (0.56) 2.3 (0.59) ⬍ 0.001
Grip strength, kg 24.8 (8.6) 26.0 (9.0) 27.8 (9.4) 29.7 (9.5) 32.0 (9.8) ⬍ 0.001
C-reactive protein 3.7 (7.0) 3.2 (6.0) 3.3 (6.3) 2.8 (6.0) 2.0 (3.0) ⬍ 0.001
Fibrinogen 335 (64) 325 (65) 327 (64) 314 (54) 310 (61) ⬍ 0.001
ln (WBC) 1.83 (0.28) 1.81 (0.25) 1.78 (0.28) 1.77 (0.28) 1.74 (0.25) ⬍ 0.001
Current smoker 8.9 6.6 7.3 6.6 4.5 0.019
Echocardiographic LVM‡ 147 (54) 147 (49) 144 (46) 145 (43) 148 (45) 0.97
Echocardiographic percentage 41.6 (9.2) 41.6 (9.0) 41.9 (8.4) 41.6 (7.8) 42.4 (8.6) 0.30
of fractional shortening‡
Moderate-to-severe regional 3.4 2.7 1.9 2.1 1.9 0.15
wall motion abnormality
Moderate-to-severe left 2.7 3.1 1.9 2.1 1.7 0.20
ventricular ejection fraction
decrease
Limited ADL 0.26 (0.67) 0.16 (0.44) 0.11 (0.32) 0.09 (0.31) 0.05 (0.27) ⬍ 0.001
Fair or poor health 29.8 19.2 16.0 10.8 6.0 ⬍ 0.001
Depression score 6.2 (4.8) 5.6 (4.6) 5.4 (4.6) 4.6 (4.1) 4.2 (3.9) ⬍ 0.001
Cognitive, MMSE 89.8 (11) 92.0 (7.7) 93.3 (7.4) 93.9 (6.7) 95.9 (4.8) ⬍ 0.001
*Data are presented as mean (SD) or %. ln ⫽ natural log.
†Unadjusted p values, testing for a trend across the quintiles of total distance walked.
‡Data are missing for these echocardiography variables from many subjects.

In the final step of entering groups of variables channel blockers), or nonsteroidal anti-inflammatory
into the models, three functional variables were drugs, insulin, or estrogen by women.
significant: score on the MMSE, the depression
symptom score, and grip strength. When they en-
Reference Equations for the 6MWD
tered the models, race and education became much
less important. Participants with better cognitive Approximately one third of the 2,115 patients who
levels, fewer symptoms of depression, and a stronger completed the 6MWT remained in the healthy
grip strength walked farther. subset (Table 7). The 437 healthy women walked a
The following factors were not independent cor- mean of 367 m (confidence interval [CI], 249 to
relates of the distance walked (p ⬍ 0.01) in the final 479 m), and the 315 healthy men walked a mean of
models: educational level, height, current smoking, 400 m (CI, 280 to 532 m). The 71 healthy African-
prevalent coronary heart disease (a history of MI, American participants walked an average of 40 m
angina, CHF, stroke, or claudication), or the use of shorter distance than the others. The relationship of
several cardiovascular medications (diuretics, antihy- age with 6MWD was linear. There were no signifi-
pertensives, vasodilators, ␤-blockers, and calcium cant gender or race interaction terms with age,

392 Clinical Investigations

 Table 5—Linear Regression Models Predicting 6MWD for Women (n ⴝ 1,094), With Same Variables in Model for
Men and Women*

Original Model Function Variables Added

Blocks Coefficient (SE) 95% CI p Value Coefficient (SE) 95% CI p Value

Demographics
Age at year 9 visit, yr ⫺ 4.4 (0.58) ⫺ 5.5 to ⫺ 3.3 ⬍ 0.001 ⫺ 3.4 (0.59) ⫺ 4.6 to ⫺ 2.3 ⬍ 0.001
Black race ⫺ 14.6 (7.1) ⫺ 28.5 to ⫺ 0.66 0.040 ⫺ 11.4 (7.2) ⫺ 25.6 to 2.8 0.115
Education less than high school ⫺ 6.2 (5.6) ⫺ 17.1 to 4.8 0.269 5.1 (5.8) ⫺ 6.3 to 16.5 0.381
Clinic ⬍ 0.001 ⬍ 0.001
Wake Forest Referent Referent
University of California, Davis 0.05 (6.6) ⫺ 12.8 to 12.9 ⫺ 6.1 (6.5) ⫺ 18.9 to 6.7
Hagerstown ⫺ 15.7 (7.1) ⫺ 29.5 to ⫺ 1.9 ⫺ 25.2 (7.1) ⫺ 39.0 to ⫺ 11.4
Pittsburgh ⫺ 35.4 (6.6) ⫺ 48.3 to ⫺ 22.5 ⫺ 41.3 (6.5) ⫺ 54.1 to ⫺ 28.5
Anthropometry
Waist circumference, cm ⫺ 1.50 (0.28) ⫺ 2.06 to ⫺ 0.95 ⬍ 0.001 ⫺ 1.23 (0.28) ⫺ 1.79 to ⫺ 0.68 ⬍ 0.001
Weight (centered), lb 0.36 (0.15) 0.07 to 0.66 0.016 0.16 (0.15) ⫺ 0.15 to 0.46 0.313
Weight squared (centered), lb ⫺ 0.006 (0.001) ⫺ 0.009 to ⫺ 0.003 ⬍ 0.001 ⫺ 0.007 (0.002) ⫺ 0.010 to ⫺ 0.003 ⬍ 0.001
Clinical
AAI 0.006 0.006
ⱕ 1.1 78.6 (25.3) 28.9 to 128.3 78.4 (24.6) 30.1 to 126.8
⬎ 1.1 ⫺ 5.93 (30.0) ⫺ 64.8 to 52.9 0.002 ⫺ 13.3 (29.6) ⫺ 71.3 to 44.8 ⬍ 0.001
Diastolic BP, mm Hg
⬍ 69 ⫺ 0.41 to 0.91 0.23 (0.33) ⫺ 0.41 to 0.87
⬎ 69 0.25 (0.34) ⫺ 2.2 to ⫺ 0.62 ⬍ 0.001 ⫺ 1.4 (0.39) ⫺ 2.2 to ⫺ 0.67 ⬍ 0.001
FEV1, L ⫺ 1.4 (0.40) 25.9 to 49.8 ⬍ 0.001 35.9 (6.0) 24.1 to 47.7 0.001
Arthritis 37.8 (6.1) ⫺ 31.5 to ⫺ 12.4 0.008 ⫺ 17.0 (4.9) ⫺ 26.6 to ⫺ 7.4 0.021
Stroke or TIA prevalence ⫺ 21.9 (4.9) ⫺ 46.8 to ⫺ 7.0 0.002 ⫺ 22.9 (9.9) ⫺ 42.3 to ⫺ 3.5 0.002
ln (C-reactive protein) ⫺ 26.9 (10.1) ⫺ 12.4 to ⫺ 2.8 ⫺ 7.4 (2.4) ⫺ 12.2 to ⫺ 2.6
⫺ 7.6 (2.5)
Medications
ACE inhibitors (any reason) ⫺ 29.4 (7.1) ⫺ 43.4 to ⫺ 15.5 ⬍ 0.001 ⫺ 27.1 (6.9) ⫺ 40.7 to ⫺ 13.6 ⬍ 0.001
Digitalis ⫺ 10.0 (9.1) ⫺ 27.9 to 7.8 0.270 ⫺ 9.2 (9.0) ⫺ 26.9 to 8.4 0.305
Lipid lowering 4.5 (6.8) ⫺ 8.9 to 17.9 0.513 3.6 (6.7) ⫺ 9.4 to 16.7 0.586
Function
MMSE score ⬍ 0.001
ⱕ 95 0.89 (0.44) 0.02 to 1.76
⬎ 95 5.6 (1.4) 2.9 to 8.4
Grip strength 1.7 (0.45) 0.78 to 2.54 ⬍ 0.001
Depression score ⫺ 1.4 (0.49) ⫺ 2.3 to ⫺ 0.41 0.005
* See Table 4 for expansion of abbreviation.

height, or weight in the model predicting 6MWD in The 12-min walking test was introduced in 1968 as
the healthy subset. The total amount of variance in a guide to physical fitness,28 and later applied to
6MWD explained by the model (R2) was only 20%. patients with COPD.29 It was then found that de-
On average, the healthy 752 participants walked only creasing the time to 6 min did not significantly
10.5% farther than the entire group who completed reduce the utility of the test.23 The 6MWT has been
the test. validated by high correlation with workloads, heart
rate, oxygen saturation, and dyspnea responses when
compared to standard bicycle ergometry and tread-
Discussion mill exercise tests in middle-aged adults30 –32 and in
elderly persons.2,33
We found that most elderly persons can safely
perform the 6MWT to quickly measure their func-
Anthropometric Correlates
tional status. We may be the first investigators to
describe associations of 6MWD with impaired ADL; Age, weight, and waist size were independently
self-reported health; education; race; a history of associated with the distance walked in this analysis,
coronary heart disease, TIA, stroke, or diabetes; and and these factors were also associated with gait speed
indexes of inflammation: (C-reactive protein, fibrin- and timed chair stands during their baseline exami-
ogen, and WBC count). nation.6 The gradual reduction of skeletal muscle

www.chestjournal.org CHEST / 123 / 2 / FEBRUARY, 2003 393

 Table 6 —Linear Regression Models Predicting 6MWD for Men (n ⴝ 715), With Same Variables in Model for
Men and Women*

Original Model Function Variables Added

Blocks Coefficient (SE) 95% CI p Value Coefficient (SE) 95% CI p Value

Demographics
Age at year 9 visit, yr ⫺ 3.3 (0.70) ⫺ 4.7 to ⫺ 1.9 ⬍ 0.001 ⫺ 2.0 (0.72) ⫺ 3.4 to ⫺ 0.58 0.006
Black race ⫺ 38.8 (9.6) ⫺ 57.5 to ⫺ 20.0 ⬍ 0.001 ⫺ 25.4 (9.7) ⫺ 44.5 to ⫺ 6.3 0.009
Less than high school education ⫺ 17.2 (7.1) ⫺ 31.1 to ⫺ 3.3 0.015 ⫺ 3.3 (7.3) ⫺ 17.6 to 11.0 0.652
Clinic 0.001 0.001
Wake Forest Referent Referent
University of California, Davis 15.0 (8.5) ⫺ 1.8 to 31.7 7.6 (8.4) ⫺ 9.0 to 24.1
Hagerstown ⫺ 11.9 (9.4) ⫺ 30.2 to 6.5 ⫺ 19.8 (9.3) ⫺ 38.1 to ⫺ 1.5
Pittsburgh ⫺ 11.5 (8.6) ⫺ 28.4 to 5.4 ⫺ 17.5 (8.6) ⫺ 34.3 to ⫺ 0.6
Anthropometry
Waist circumference, cm ⫺ 2.8 (0.59) ⫺ 3.99 to ⫺ 1.69 ⬍ 0.001 ⫺ 1.97 (0.59) ⫺ 3.13 to ⫺ 0.81 0.001
Weight (centered), lb 0.73 (0.23) 0.29 to 1.18 0.001 0.36 (0.23) ⫺ 0.09 to 0.81 0.117
Weight squared (centered), lb ⫺ 0.010 (0.003) ⫺ 0.015 to ⫺ 0.004 0.001 ⫺ 0.009 (0.003) ⫺ 0.014 to ⫺ 0.003 0.002
Clinical
AAI ⬍ 0.001 ⬍ 0.001
ⱕ 1.2 119 (24) 72.7 to 165.7 106 (23) 59.7 to 151.8
⬎ 1.2 ⫺ 36 (44) ⫺ 122.0 to 49.1 ⫺ 26 (43) ⫺ 110 to 58
Diastolic BP, mm Hg 0.27 (0.53) ⫺ 0.78 to 1.32 0.271 ⫺ 0.04 (0.52) ⫺ 1.07 to 0.99 0.237
ⱕ 71 0.64 (0.54) ⫺ 0.41 to 1.69 0.83 (0.53) ⫺ 0.21 to 1.87
⬎ 71 8.2 (5.3) ⫺ 2.2 to 18.6 7.3 (5.2) ⫺ 2.9 to 17.6
FEV1, L ⫺ 17.8 (6.8) ⫺ 31.1 to ⫺ 4.5 0.121 ⫺ 13.6 (6.7) ⫺ 26.8 to ⫺ 0.4 0.162
Arthritis ⫺ 27.0 ⫺ 47.7 to ⫺ 6.2 0.009 ⫺ 26.1 (10.4) ⫺ 46.5 to ⫺ 5.7 0.043
Stroke or TIA prevalence 10.6 ⫺ 11.4 to 0.92 0.011 ⫺ 2.0 (3.2) ⫺ 8.2 to 4.3 0.012
ln (C-reactive protein) ⫺ 5.2 (3.1) 0.096 0.536
Medications
ACE inhibitors (any reason) ⫺ 24.5 (7.9) ⫺ 40.0 to ⫺ 9.0 0.002 ⫺ 23.5 (7.7) ⫺ 38.7 to ⫺ 8.3 0.002
Digitalis ⫺ 34.6 (9.2) ⫺ 52.7 to ⫺ 16.6 ⬍ 0.001 ⫺ 31.6 (9.0) ⫺ 49.3 to ⫺ 13.8 0.001
Lipid lowering 27.8 (9.5) 9.2 to 46.5 0.003 24.0 (9.4) 5.6 to 42.4 0.010
Function
MMSE score ⬍ 0.001
ⱕ 95 2.1 (0.51) 1.06 to 3.05
⬎ 95 5.0 (1.9) 1.37 to 8.72
Grip strength 1.2 (0.40) 0.40 to 1.96 0.003
Depression score ⫺ 1.5 (0.72) ⫺ 2.9 to ⫺ 0.10 0.036
* See Table 4 for expansion of abbreviation.

mass and strength that generally occurs with specific functional form weight/height squared
aging34,35 (and debilitating diseases that we did not was not optimal for describing the relationship of
measure) are probably responsible for the shorter weight and height with 6MWD. Waist size and
distance walked by those ⬎ 85 years old. A taller body weight were more strongly associated with
height is associated with a longer stride, which makes 6MWD than was BMI.
walking more efficient, probably resulting in a longer After correcting for age, gender, height, weight,
distance walked in the taller men and women. After and other confounders, elderly African-American
correcting for other factors, including height, elderly women and men walked a shorter distance when
men did not walk farther than elderly women (in a compared to white men and women in our study.
model that included men and women together; data Investigators in Japan recently reported that the
not shown). mean 6MWD of healthy elderly Japanese men and
Obesity increases the workload for a given amount women36 was similar to that reported for whites.37,38
of exercise, probably resulting in the shorter distance
walked in participants with a higher body weight or
Cardiovascular Correlates
BMI. On average, participants who were obese, as
defined by a BMI ⬎ 30, walked approximately 85% Participants who reported a history of any type of
of the distance completed by those of average body CVD (a history of angina, MI, heart failure, TIA, or
weight. Although BMI is a clinically useful index stroke) did not walk as far as the others (Table 4).
of obesity, exploratory analyses suggested that the This was also true in the Tucson study.37 Men

394 Clinical Investigations

Table 7—Reference Equations for the 6MWD From the ciated with substantially shorter 6MWD (Table 4). In
Healthy Subset of 437 Women and 315 Men the multivariate models, we used two terms for AAI
For the total distance walked in meters for women: (piecewise) since we found a breakpoint in the relation-
493 ⫹ (2.2 ⫻ height) ⫺ (0.93 ⫻ weight) ⫺ (5.3 ⫻ age), with ship between AAI and the 6MWD. The highest
height in centimeters and weight in kilograms. 6MWD was seen when the AAI was approximately 1.1
For men, add 17 m. Subtract 100 m for the lower limit of the (Fig 3). 6MWD decreases as the AAI was higher or
normal range.
lower than 1.1. Previous studies of AAI suggest that a
For the total distance walked in yards for women:
539 ⫹ (6.1 ⫻ height) ⫺ (0.46 ⫻ weight) ⫺ (5.8 ⫻ age), with low AAI is associated with peripheral vascular disease,33
height in inches and weight in pounds. while a high AAI may be due to loss of arterial
For men, add 18 yards. Subtract 109 yards for the lower limit of compliance.43
the normal range.
Pulmonary Correlates
Criteria applied sequentially to exclude participants from the
healthy subgroup: We found that a lower FEV1 was a strong, inde-
Arthritis in hips and knees, excluded 576
pendent predictor of a lower 6MWD in women. A
FEV1 ⬍ 70% predicted, excluded 406
AAI low (⬍ 0.90) or high (⬎ 1.5), excluded 108 low FEV1 is most commonly due to obstructive lung
History of stroke, TIA, or claudication, excluded 84 diseases such as COPD (due to decades of cigarette
Diabetics receiving medications, excluded 74 smoking) and asthma, but is also reduced in diseases
Weight or waist size ⬎ 95th percentile*, excluded 55 that restrict lung volumes. Previous investigators
Cognitive impairment (MMSE score ⬍ 80), excluded 39
have used the 6MWD as a measure of the severity of
Diastolic hypertension (⬎ 90 mm Hg), excluded 21
COPD and an outcome measure in COPD treat-
*Gender-specific cutpoints for obesity were ⬎ 90 kg for women ment studies.23,31,44,45 We found that the FEV1 was
and 101 kg for men, and for high waist size 118.5 cm for women
and 116 cm for men.
also associated with gait speed and grip strength in
the elderly women of this cohort.6 Current smoking
was bivariately associated with a significantly shorter
distance walked in both studies, but in the gender-
receiving ACE inhibitors and those receiving digi-
specific models, smoking status was replaced by
talis walked approximately 90% as far as others. This
indexes of subclinical diseases that are known to be
is probably because they were receiving these med-
caused by smoking.
ications for heart failure,39 which causes dyspnea on
exertion, limiting exercise tolerance.40 Approxi-
Other Correlates
mately 42% of men and women who reported heart
failure were receiving an ACE inhibitor, and one half Previous studies using the 6-min walk have not
were receiving digitalis. Use of these medications is included measures of symptoms of depression, men-
probably an indicator of more clinically severe CHF; tal status (MMSE), limitations of ADL, self-reported
however, men receiving lipid-lowering medications general health, or grip strength, all of which we
walked approximately 7% farther than other men. found were independently associated with the
Approximately 25% of men with a history of MI 6MWD. A 1-U higher MMSE score was associated
and 29% of those with claudication were receiving with walking 6 m farther in women, and a similar
lipid-lowering medications. Although these medi- relationship was seen in men. The 7% lower mean
cations act as markers of vascular disease, perhaps 6MWD values found at two of the four clinics could
their use improves walking distance in such pa- be due to technicians at those clinics less frequently
tients; a future study of these medications should excluding frail participants, or giving less encourage-
investigate this possibility by including 6MWD as ment to walk farther (despite the attempt to stan-
an outcome measure.41 dardize the messages); or it could be due to un-
In the bivariate analyses, there were only nonsig- measured differences in the populations in those
nificant trends for participants with echocardio- communities.
graphic abnormalities to walk shorter distances, and Grip strength is a direct measure of skeletal
only wall motion abnormalities in women were inde- muscle strength of the hands, but it is also an index
pendently associated with 6MWD in the models. of overall muscle strength, endurance, and dis-
This lack of a relationship of echocardiographic ability.46 It remained a strong, independent (linear)
variables with exercise ability was previously re- predictor of 6MWD in our models for both women
ported in patients with overt heart failure,42 perhaps and men. Mean grip strength was 23 kg for women
because echocardiographic measurements are done and 40 kg for men in the 65-to 69-year age groups.
while resting. For a 10-kg increment in grip strength, men and
Both clinical and subclinical peripheral vascular dis- women walked an average of 14 m farther. A study
ease (claudication and low AAI) were bivariately asso- from Finland demonstrated that muscle strength and

www.chestjournal.org CHEST / 123 / 2 / FEBRUARY, 2003 395

walking speed can be increased substantially by and stretching exercises before beginning the test,
physical exercise (either endurance or strength train- walked in groups of three to six outside on a
ing) in elderly women.47 rectangular track, and were instructed to “walk as
A limitation of this study is that one third of the fast as they comfortably could, trying to cover the
participants were excluded from the walk, and they maximum distance possible,” while our participants
had generally poorer health than those who per- had no warm-up, walked alone, and were instructed
formed the test; however, their exclusion only de- to walk from end to end of the hallway at their own
creased the power of the analyses to detect associa- pace, in order to cover as much ground as possible.
tions with these diseases. Those entered into the Differences in participant recruitment and test
study, and those who came to the clinics for the instructions probably account for the lower distances
examination between 1996 and 1997 (the only time walked by our participants when compared to the
the 6MWT was done) were survivors, and were two other studies. The ATS has recently published
healthier than older persons in institutions or hos- detailed guidelines for 6MWT procedures49 that
pitals.
should be followed by investigators studying care-
fully selected healthy persons. This new document
Reference Equations for the 6MWD states that “a practice test is not needed in most
The mean 6MWD in the healthy subset of partic- clinical settings, but should be considered.” The
ipants was 367 m (CI, 249 to 479 m) for women, and CHS 6MWT done in 1996 was performed exactly
400 m (CI, 280 to 532 m) for men. Age, gender, race, according to the 2002 ATS guidelines, except that
height, and weight were all statistically significant the instructions given to the CHS participants were
predictors of 6MWD in the healthy subgroup, sug- to “walk from end to end of the hallway at your own
gesting that these factors should be considered when pace, in order to cover as much ground as possible.”
comparing the 6MWD of an individual patient to According to the new ATS guidelines, patients
healthy elderly persons; however, the reference should be told to “Remember that the object is to
equations obtained from this model explained only walk as far as possible for six minutes, but don’t run
20% of the variation in 6MWD. The fifth percentile or jog.” This seemingly small difference in the
of the 6MWD for the healthy participants, which instructions may have caused the elderly CHS par-
may be considered the lower limit of the normal ticipants not to walk as quickly as they would have if
range, was approximately 75% of the predicted value the new ATS recommended instructions had been
(mean minus lower limit of the normal range: given to them.
400 m ⫺ 100 m ⫽ 300 m for men). The 6MWT should be useful for measuring
Our reference equation gives predicted (mean) changes in functional status (preintervention and
6MWDs that are substantially lower than those postintervention) in the clinical setting, but consid-
published by previous investigators.37,48 Our refer- erable caution is needed when using currently avail-
ence equations predicts distances of 430 m and able reference equations to determine if a given
464 m for a 67-year-old white woman and man of
patient’s 6MWD is normal or low. According to the
average height and weight, respectively, while a
ATS review of previously published 6MWT studies,
study of 290 healthy adults in Tucson, AZ,37 predicts
the increases due to the learning effect ranged from
distances of 466 m and 544 m; the study of Rikli and
Jones,48 which enrolled 7,183 older adults from 21 a mean of zero to 17%. Performance usually reaches
states, predicts distances approximately 50% greater a plateau after two tests done within a week. The
than ours (624 m and 689 m, respectively, for the reproducibility results from one study of 112 patients
same woman and man). with stable, severe COPD suggest that an improve-
Our choice of criteria for excluding participants ment of ⬎ 70 m in the 6MWD after an intervention
from the healthy subgroup is probably not the reason is necessary to be 95% confident that the improve-
for our lower distances, since we excluded more than ment was significant.50
two thirds of the cohort, yet the healthy subgroup In summary, most community-dwelling elderly
only walked an average of 10.5% farther than the persons can perform the 6MWT. Factors associated
entire cohort who completed the test. The study of with a shorter distance walked are similar to those
Rikli and Jones48 recruited participants using ad- associated with a reduced oxygen uptake at maximal
vertisements in newspapers, magazines, and jour- exercise reported by previous investigators. Unique
nals (resulting in a recruitment bias toward highly factors measured by this study included cognitive
fit individuals), while our participants were a function, symptoms of depression, limitations of
community-based sample. The participants in the ADL, indexes of inflammation, and the use of car-
study by Rikli and Jones did 8 to 10 min of warm-up diovascular medications.

396 Clinical Investigations

 Appendix: Participating Institutions and coagulation factors VII and VIII and fibrinogen in adults over
Principal Investigators 65 years old: results for the Cardiovascular Health Study. Ann
Epidemiol 1992; 2:509 –519
Wake Forest University School of Medicine, Gregory L. Burke, 15 Tracy RP, Lemaitre R, Psaty BM, et al. Relationship of
MD; ECG Reading Center, Wake Forest University, Pentti C-reactive protein to risk of cardiovascular disease in the
Rautaharju, MD, PhD; University of California, Davis, John Cardiovascular Health Study. Arterioscler Thromb Vasc Biol
Robbins, MD, MHS; The Johns Hopkins University, Linda P. 1997; 17:1121–1127
Fried, MD, MPH; MRI Reading Center, The Johns Hopkins 16 Gardin JM, Wong ND, Bommer W, et al. Echocardiographic
University, Nick Bryan, MD, PhD, and Norman J. Beauchamp, design of a multicenter investigation of free-living elderly
MD; University of Pittsburgh, Lewis H. Kuller, MD; Echocardi- subjects: The Cardiovascular Health Study. J Am Soc Echo-
ography Reading Center (baseline), University of California, cardiogr 1992; 5:63–72
Irvine, Julius M. Gardin, MD; Echocardiography Reading Center 17 Furberg CD, Manolio TA, Psaty BM, et al. Major electrocar-
(follow-up), Georgetown Medical Center, John Gottdiener, MD; diographic abnormalities in persons aged 65 years and older:
Ultrasound Reading Center, New England Medical Center, The Cardiovascular Health Study. Am J Cardiol 1992; 69:
Boston, Daniel H. O’Leary, MD; Central Blood Analysis Labo- 1329 –1335
ratory, University of Vermont, Russell P. Tracy, PhD; Pulmonary 18 O’Leary DH, Polak JF, Wolfson SK, et al. Use of sonography
Reading Center, University of Arizona, Tucson, Paul Enright, to evaluate carotid atherosclerosis in the elderly: Cardiovas-
MD; Retinal Reading Center, University of Wisconsin, Ron cular Health Study. Stroke 1991; 22:1155–1163
Klein, MD; Coordinating Center, University of Washington, 19 Ives DG, Fitzpatrick AL, Bild DE, et al. Surveillance and
Richard A. Kronmal, PhD; and National Heart, Lung, and Blood ascertainment of cardiovascular events: The Cardiovascular
Institute Project Office, Diane Bild, MD, MPH. Health Study. Ann Epidemiol 1995; 5:278 –285
20 American Thoracic Society. Standardization of spirometry,
1987 update. Am Rev Respir Dis 1987; 136:1285–1298
21 Enright PL, Kronmal RA, Higgins M, et al. Spirometry
References reference values for women and men 65– 85 years of age:
1 Guralnik J, Branch L, Cummings S, et al. Physical perfor- Cardiovascular Health Study. Am Rev Respir Dis 1993;
mance measures in aging research. J Gerontol Med Sci 1989; 147:125–133
44:M141–M146 22 American Thoracic Society. Recommended respiratory dis-
2 Peeters P, Mets T. The 6 minute walk as an appropriate ease questionnaires for use with adults and children in
exercise test in elderly patients with chronic heart failure. J epidemiological research. Am Rev Respir Dis 1978; 6(part
Gerontol 1996; 51A:M147–M151 2):7–23
3 Jette AM, Jette DU, Ng J, et al. Are performance-based 23 Butland RJA, Pang J, Gross ER, et al. Two, six, and 12 minute
measures sufficiently reliable for use in multicenter trials? J walking tests in respiratory disease. BMJ 1982; 284:1607–1608
Gerontol 1999; 54A:M3–M6 24 Wilson RC, Jones PW. A comparison of the visual analogue
4 McGavin CR, Artvinli M, Naoe H, et al. Dyspnea, disability, scale and modified Borg scale for the measurement of
and distance walked: comparison of estimates of exercise dyspnea during exercise. Clin Sci 1989; 76:277–282
performance in respiratory disease. BMJ 1978; 2:241–243 25 Tibshirani R. Estimating transformations for regression via
5 Bittner V, Weiner DH, Yusuf S, et al. Prediction of mortality additivity and variance stabilization. J Am Stat Assoc 1988;
and morbidity with a six minute walk test in patients with left 83:394 – 405
ventricular dysfunction. JAMA 1993; 270:1702–1707 26 Norusis MJ. SPSS for Windows, base system user’s guide,
6 Hirsch CH, Fried LP, Harris T, et al. Correlates of perfor- release 9.0. Chicago, IL: SPSS, 1998
mance-based measures of muscle function in the elderly: The 27 S-Plus guide to statistical and mathematical analysis, version
Cardiovascular Health Study. J Gerontol 1997; 52A:M192– 3.2. Seattle, WA: StatSci, 1993
M200 28 Cooper KH. A means of assessing maximal oxygen intake.
7 Fried LP, Borhani NO, Enright PL, et al. The Cardiovascular JAMA 1968; 203:201–204
Health Study: design and rationale. Ann Epidemiol 1991; 29 McGavin CR, Gupta SP, McHardy GJR. Twelve minute
1:263–276 walking test for assessing disability in chronic bronchitis. BMJ
8 Tell GS, Fried LP, Hermanson B, et al. Recruitment of adults 1976; 1:822– 823
65 years and older as participants in the Cardiovascular 30 Langenfeld H, Schneider B, Grimm W, et al. The six minute
Health Study. Ann Epidemiol 1993; 3:358 –366 walk test: an adequate exercise test for pacemaker patients?
9 Psaty BM, Lee M, Savage PJ, et al. Assessing the use of Pacing Clin Electrophysiol 1990; 13:1761–1765
medications in the elderly: methods and initial experience in 31 Spence DPS, Hay JG, Carter J, et al. Oxygen desaturation and
the Cardiovascular Health Study. J Clin Epidemiol 1992; breathlessness during corridor walking in COPD: effect of
45:683– 692 oxitropium bromide. Thorax 1993; 48:1145–1150
10 Orme J, Reis J, Herz E. Factorial and discriminate validity of 32 Bernstein ML, Despars JA, Singh NP, et al. Re-analysis of
the CES-D scale. J Clin Psychol 1986; 42:28 –33 the 12 minute walk in patients with COPD. Chest 1994;
11 Cohen S, Mermelstein R, Kamarck T, et al. Measuring the 105:163–167
functional components of social support. In: Saranson IG, 33 Montgomery PS, Gardner AW. The clinical utility of a six
Saranson BR, eds. Social support: theory, research and minute walk test in peripheral arterial occlusive disease
application. Dordrecht, The Netherlands: Martinus Nijhoff, patients. J Am Geriatr Soc 1998; 46:706 –711
1985 34 Fleg JL, Lakatta EG. Role of muscle loss in the age-associated
12 Teng EL, Chui HC. The modified Mini-Mental State (3MS) reduction in Vo2max. J Appl Physiol 1988; 65:1147–1151
Examination. J Clin Psychol 1987; 48:314 –318 35 Tolep K, Kelsen SG. Effect of aging on respiratory skeletal
13 Newman AB, Siscovick DS, Manolio TA, et al. Ankle-arm muscles. Clin Chest Med 1993; 3:363–378
index as a marker of atherosclerosis in the Cardiovascular 36 Teramoto S, Ohga E, Ishii T, et al. Reference value of
Health Study. Circulation 1993; 88:837– 845 six-minute walking distance in healthy middle-aged and older
14 Tracy RP, Bovill EG, Fried LP, et al. The distribution of subjects [letter]. Eur Respir J 2000; 15:1132–1133

www.chestjournal.org CHEST / 123 / 2 / FEBRUARY, 2003 397

37 Enright PL, Sherrill DL. Reference equations for the six effort, and perceived breathlessness in patients with airflow
minute walk in healthy adults. Am J Respir Crit Care Med limitation. Thorax 1993; 48:33–38
1998; 158:1384 –1387 45 Wijkstra PJ, TenVergert EM, vanderMark ThW, et al. Rela-
38 Troosters T, Gosselink R, Decramer M. Six minute walking tion of lung function, maximal inspiratory pressure, dyspnoea,
distance in healthy elderly subjects. Eur Respir J 1999; and quality of life with exercise capacity in patients with
14:270 –274 COPD. Thorax 1994; 49:468 – 472
39 DeBrock V, Mets T, Romagnoli M, et al. Captopril treatment 46 Rantanen T, Guralnik JM, Rantala RS, et al. Disability,
of chronic heart failure in the very old. Gerontology 1994; physical activity, and muscle strength in older women: The
49:148 –152 Women’s Health and Aging Study. Arch Phys Med Rehab
40 Lipkin DP, Scrivin AJ, Crake T, et al. Six minute walking test 1999; 80:130 –135
for assessing exercise capacity in chronic heart failure. BMJ 47 Sipila S, Multanen J, Kallinen M, et al. Effects of strength and
1986; 292:653– 655 endurance training on isometric muscle strength and walking
41 Bittner V. Six-minute walk test in patients with cardiac speed in elderly women. Acta Physiol Scand 1996; 156:457– 464
dysfunction. Cardiologia 1997; 42:897–902 48 Rikli RE, Jones CJ. Functional fitness normative scores for
42 Franciosa JA, Levine TB. Lack of correlation between exer- community-residing older adults, ages 60 –94. J Aging Phys
cise capacity and indices of resting LV performance in heart Activity 1999; 7:162–181
failure. Am J Cardiol 1981; 47:33–39 49 American Thoracic Society. Guidelines for the six minute
43 Montgomery PS, Gardner AW. The clinical utility of a walk test. Am J Respir Crit Care Med 2002; 166:111–117
six-minute walk test in peripheral arterial occlusive disease 50 Redelmeier DA, Bayoumi AM, Goldstein RS, et al. Interpret-
patients. J Am Geriatr Soc 1998; 46:706 –711 ing small differences in functional status: the six minute walk
44 Mak VHF, Bugler JR, Roberts CM, et al. Effect of arterial test in chronic lung disease patients. Am J Respir Crit Care
oxygen desaturation on six minute walk distance, perceived Med 1997; 155:1278 –1282

398 Clinical Investigations