Author + information
- Received February 25, 2008
- Revision received August 18, 2008
- Accepted September 10, 2008
- Published online March 24, 2009.
- Mary M. McDermott, MD*,†,* (, )
- Jack M. Guralnik, MD, PhD‡,
- Lu Tian, ScD†,
- Kiang Liu, PhD†,
- Luigi Ferrucci, MD, PhD§,
- Yihua Liao, MS†,
- Leena Sharma, MD* and
- Michael H. Criqui, MD, MPH∥
- ↵*Reprint requests and correspondence:
Dr. Mary M. McDermott, 750 North Lake Shore Drive, 10th Floor, Chicago, Illinois 60611
Objectives We studied associations of borderline and low normal ankle-brachial index (ABI) values with functional decline over a 5-year follow-up.
Background The associations of borderline and low normal ABI with functional decline are unknown.
Methods The 666 participants included 412 with peripheral arterial disease (PAD). Participants were categorized as follows: severe PAD (ABI <0.50), moderate PAD (ABI 0.50 to 0.69), mild PAD (ABI 0.70 to 0.89), borderline ABI (0.90 to 0.99), low normal ABI (1.00 to 1.09), and normal ABI (ABI 1.10 to 1.30). Outcomes were assessed annually for 5 years. Mobility loss was defined as loss of the ability to walk one-quarter mile or walk up and down 1 flight of stairs without assistance among participants without baseline mobility impairment. Becoming unable to walk for 6 min continuously was defined as stopping during the 6-min walk at follow-up among those who walked for 6 min continuously at baseline. Results were adjusted for age, sex, race, comorbidities, and other confounders.
Results Hazard ratios (HRs) for mobility loss according to ABI category were as follows: severe PAD, HR: 4.16 (95% confidence interval [CI]: 1.58 to 10.92); moderate PAD, HR: 3.82 (95% CI: 1.66 to 8.81); mild PAD, HR: 3.22 (95% CI: 1.43 to 7.21); borderline ABI, HR: 3.07 (95% CI: 1.21 to 7.84); and low normal ABI, HR: 2.61 (95% CI: 1.08 to 6.32; p trend = 0.0018). Similar associations were observed for becoming unable to walk for 6 min continuously (p trend < 0.0001).
Conclusions At 5-year follow-up, persons with borderline ABI values have a higher incidence of mobility loss and becoming unable to walk for 6 min continuously compared with persons who have a normal baseline ABI. A low normal ABI is associated with an increased incidence of mobility loss compared with persons who have a normal ABI.
Eight million men and women in the U.S. have lower extremity peripheral arterial disease (PAD). The prevalence of PAD is expected to increase as the population survives longer with chronic disease (1,2). PAD is common among patients age 50 years and older in primary care medical practices and is frequently underdiagnosed (3,4). Among community-dwelling populations, the prevalences of low normal and borderline ankle-brachial index (ABI) values are similar to or higher than PAD (5,6). Persons with PAD, defined as an ABI <0.90, have greater functional impairment and faster rates of functional decline than do persons without PAD (7,8). However, associations of borderline and low normal ABI values with functional decline are unknown.
The ABI is a ratio of Doppler-recorded systolic pressures in the lower and upper extremities. In persons without PAD, arterial pressures increase with greater distance from the heart, because of increasing impedance with increasing arterial taper (9). This phenomenon results in higher systolic pressures at the ankle compared with the brachial arteries in persons without PAD. Thus, persons without lower extremity atherosclerosis have an ABI >1.00. An ABI <0.90 is highly sensitive and specific for angiographically-diagnosed PAD (10). However, recent data show that even persons with borderline ABI values (ABI 0.90 to 0.99) and low normal ABI values (ABI 1.00 to 1.09) have higher prevalences of subclinical atherosclerosis in the coronary and cerebrovascular arterial beds than do persons with an ABI of 1.10 to 1.30 (5). Prevalences of intermittent claudication and atypical exertional leg pain are higher among persons with borderline ABI values than among persons with ABI values of 1.10 to 1.40 (11). Thus, persons with borderline and low normal ABI values may experience higher rates of adverse outcomes than persons with ABI values of 1.10 to 1.30. However, the associations of borderline and low normal ABI values with functional decline are unknown.
In this prospective, observational study, we describe associations between baseline ABI values and annually measured functional outcomes, assessed up to 5 years after baseline measures, in a large cohort of persons with and without PAD. We hypothesized that participants with low normal and borderline baseline ABI values would have rates of functional decline that are less than participants with PAD but greater than persons with a normal ABI at baseline.
The Institutional Review Boards of Northwestern University and Catholic Health Partners Hospital approved the protocol. Participants gave written informed consent. The funding source for this study played no role in the design, conduct, reporting of the study, or decision to submit the manuscript.
Participants were part of the WALCS (Walking and Leg Circulation Study), a prospective, observational study designed to identify predictors of functional decline in persons with and without PAD (7,8,12). Participants underwent baseline assessment and returned annually for follow-up. Participants unable to return for follow-up were interviewed by telephone for the mobility outcome measure. Mean follow-up was 54 months.
Participants with PAD were identified from among consecutive patients ages 55 years and older diagnosed with PAD in 3 Chicago-area noninvasive vascular laboratories (7,8). The PAD participants were identified from among consecutive patients with PAD in the noninvasive vascular laboratory because this is an efficient method to identify large numbers of PAD patients with a wide range of PAD severity. Participants without PAD were identified from among persons with normal lower extremity arterial studies at the 3 noninvasive vascular laboratories and from among consecutive patients with appointments in a large general internal medicine practice at Northwestern. Identifying non-PAD participants from the noninvasive vascular laboratory allowed us to include non-PAD participants with a higher prevalence of leg symptoms and comorbidities who were relatively comparable to PAD participants except for the presence versus absence of PAD. Identifying non-PAD participants from general medicine practice allowed us to include non-PAD participants more typical of the non-PAD patients encountered by practicing clinicians. Participants from general internal medicine who had a low ABI at their study visit were included among PAD participants.
Exclusion criteria for the WALCS study have been reported (7,8,12) and include dementia, recent major surgery, above or below knee amputations, nursing home residence, and wheelchair confinement. Non–English-speaking patients were excluded because investigators were not fluent in non-English languages. Persons with PAD diagnosed in the noninvasive vascular laboratory were excluded if their baseline visit ABI was >0.95. Similarly, persons whose noninvasive vascular laboratory testing showed no PAD were excluded if their ABI was ≤0.95 at their study visit. This ABI threshold was used because this is the ABI threshold used to define PAD in Northwestern's noninvasive vascular laboratory. This exclusion criterion helped minimize misclassification bias, for example, in participants who may have had a low toe-brachial index in the noninvasive vascular laboratory but had a normal ABI at their study visit. Participants with baseline ABI >1.30 were excluded from analyses. Participants with ABI >1.30 have noncompressible lower extremity vessels and therefore could not be accurately classified into our ABI categories of interest (5,13).
A hand-held Doppler probe (Nicolet Vascular Pocket Dop II, Nicolet Biomedical, Golden, Colorado) was used to obtain systolic pressures in the right and left brachial, dorsalis pedis, and posterior tibial arteries (7,8,14). Each pressure was measured twice. The ABI was calculated by dividing the mean of the dorsalis pedis and posterior tibial pressures in each leg by the mean of the 4 brachial pressures (15). Zero values for the dorsalis pedis and posterior tibial pulses were set to missing for the ABI calculation. Average brachial pressures in the arm with highest pressure were used when 1 brachial pressure was higher than the opposite brachial pressure in both measurement sets and the 2 brachial pressures differed by 10 mm Hg or more in at least 1 measurement set, because in such cases subclavian stenosis was possible (15). The lowest leg ABI was used in analyses. The ABI categories were defined a priori as follows (5,8): severe PAD (ABI <0.50), moderate PAD (ABI 0.50 to 0.69), mild PAD (ABI 0.70 to 0.89), borderline PAD (ABI 0.90 to 0.99), low normal ABI (ABI 1.00 to 1.09), and normal ABI (ABI 1.10 to 1.30). These ABI categorizations were defined based on prior study (5,8) and allowed us to assess associations of ABI with study outcomes across a wide range of ABI values.
Our primary outcome measures were mobility loss and becoming unable to walk for 6 min continuously without stopping (8,12,16). These outcomes were assessed annually at each follow-up visit and were selected because they represent clearly defined, discrete end points. These outcomes avoid a possible “floor” effect that may occur with annual changes in distance achieved in the 6-min walk, for example, when a participant becomes unable to walk but cannot further deteriorate in functional performance. Secondary outcomes were a 15% decline in 6-min performance and a 20% decline in 6-min walk performance.
The 6-min walk was administered at baseline and at each annual follow-up visit. The 6-min walk test has excellent test-retest reliability in persons with PAD (17,18). Following a standardized protocol (7,8,12,17–19), participants walk up and down a 100-foot hallway for 6 min after instructions to cover as much distance as possible. The interviewer administering the test records whether the participant stopped during the 6-min walk. Participants who stopped at baseline were excluded from analyses of becoming unable to walk for 6 min continuously.
A critical factor in an older person's ability to function independently in the community is mobility, defined as the ability to walk or climb stairs without assistance (16). Older people who lose mobility are less likely to remain in the community, have higher rates of morbidity, mortality, and hospitalizations, and experience a poorer quality of life (16,19,20). Based on previous study, mobility loss was defined as becoming unable to walk up and down 1 flight of stairs or walk one-quarter mile without assistance among patients without mobility impairment at baseline (20,21). At baseline and at each follow-up visit, participants were asked to indicate whether they were able to walk one-quarter mile and whether they could climb up and down 1 flight of stairs: 1) on their own; 2) with assistance; or 3) not at all (20,21).
Comorbidities assessed were diabetes mellitus, angina, myocardial infarction, heart failure, cancer, chronic lung disease, lower extremity arthritis, spinal stenosis, spinal disk disease, and stroke. Disease-specific algorithms that combine data from patient report, medical record review, medications, laboratory values, and a questionnaire completed by the participant's primary care physician were used to verify and document baseline comorbidities other than knee and hip arthritis, on the basis of criteria previously developed (22). American College of Rheumatology criteria were used to document the presence of knee and hip osteoarthritis (23,24).
Height and weight were measured at the study visit. Body mass index (BMI) was calculated as (weight [kg]/height [m])2. Cigarette smoking history was determined with patient report. At baseline, participants were asked to report the number of blocks they walked during the previous week. At each follow-up visit, we used patient report, a primary care physician questionnaire, and medical record review to identify lower extremity revascularizations each year.
Baseline characteristics of participants in each ABI category were compared using general linear models for continuous variables and chi-square tests for categorical variables. Functional outcomes were mobility loss among participants without mobility impairment at baseline and loss of the ability to walk continuously for 6 min without stopping among participants who completed the 6-min walk test without stopping at baseline. Secondary functional outcome measures were 15% and 20% declines, respectively, in 6-min walk performance.
Cox regression analyses were used to compare rates of each functional outcome across ABI categories, adjusting for age, sex, race, and baseline values for comorbidities, smoking, body mass index, and patient-reported blocks walked during the prior week. The ABI categories were entered as dummy variables into the regression analyses. Participants who met our primary outcome definitions at baseline (mobility impairment or inability to walk for 6 min without stopping) were excluded from these respective analyses. Person-time was calculated as the number of months from the baseline visit to the date of the most recent visit (last seen) or the date of the visit during which each functional outcome of interest was first reported, whichever came first. Participants who died before experiencing an outcome measure or who underwent lower extremity revascularization during follow-up were censored at the date of their last visit before these events. These analyses were repeated, in which death was considered as an outcome if it occurred before mobility loss or stopping during the 6-min walk, respectively. Because lower extremity arthritis and spinal disk disease may influence rates of functional decline, Cox regression analyses were repeated within the subset of participants with no history of knee arthritis, hip arthritis, disk disease, or spinal stenosis. We also tested for interactions between the presence versus absence of knee arthritis, hip arthritis, spinal disk disease, or spinal stenosis and the association of the ABI with functional outcomes.
We tested the proportional hazards assumption for mobility loss and the stop during the 6-min walk analyses using Martingale residuals based methods, and we did not find any significant deviation from the proportional hazards assumption (25). Analyses were performed using SAS statistical software, version 9.1 (SAS Institute, Cary, North Carolina).
Of 731 men and women who completed baseline testing for the WALCS study and had a baseline ABI <1.30, 698 (95%) completed at least 1 follow-up visit. Of these, 18 underwent lower extremity revascularization before their first annual follow-up visit, leaving 680 participants. Of these, 516 (284 with PAD) completed the 6-min walk test at baseline without stopping and participated in at least 1 annual follow-up 6-min walk test. These 516 participants were included in analyses of becoming unable to walk for 6 min without stopping. The corresponding number of participants who were free of mobility impairment at baseline was 647. A total of 666 participants (412 with PAD) met criteria for inclusion in either the 6-min walk or mobility loss analyses.
Table 1shows characteristics associated with each ABI category at baseline among the 666 participants. Lower ABI values were associated with older age, lower BMI, greater pack-years of cigarette smoking, higher prevalences of diabetes and history of cardiac or cerebrovascular diseases, lower prevalences of lower extremity arthritis or spinal disk disease, and fewer blocks walked during the previous week.
Figure 1shows adjusted associations of baseline ABI levels with incident mobility loss among participants without mobility impairment at baseline. Adjusting for age, sex, race, comorbidities, smoking, BMI, and physical activity level, lower ABI values were associated with significantly higher rates of mobility loss (p trend = 0.0018). Compared with participants with a normal baseline ABI, those with a baseline ABI <0.50 had a significantly increased risk of mobility loss (p = 0.0038) during annual follow-up visits up to 5-years after baseline (Fig. 1). Compared with the reference category with normal baseline ABI, increased rates of mobility loss were observed among participants with ABI 0.50 to 0.69 (p = 0.0016), ABI 0.70 to 0.89 (p = 0.0046), ABI 0.90 to 0.99 (p = 0.0187), and ABI 1.00 to 1.09 (p = 0.033), adjusting for confounders (Fig. 1). No significant interactions were observed between presence versus absence of knee arthritis, hip arthritis, spinal stenosis, or spinal disk disease and the association of the ABI with mobility loss (data not shown). Results were not substantially changed when deaths occurring before mobility loss were included as outcomes (data not shown).
Adjusting for age, sex, race, smoking, comorbidities, body mass index, and physical activity level, a significant interaction was observed between presence versus absence of knee arthritis, hip arthritis, spinal stenosis, or spinal disk disease and the association of the ABI with stopping during the 6-min walk test (p value for interaction term = 0.023). In this interaction, associations of ABI with becoming unable to walk continuously for 6 min were not significant among participants with history of knee or hip arthritis, spinal stenosis, or spinal disk disease (data not shown). In contrast, lower ABI values were associated with increased risk of becoming unable to walk continuously for 6 min among participants without history of knee or hip arthritis, spinal stenosis, or spinal disk disease (p trend <0.001) (Fig. 2).Among participants without history of knee or hip arthritis, spinal stenosis, or spinal disk disease, compared with the reference group with ABI of 1.10 to 1.30 at baseline, borderline PAD (ABI 0.90 to 0.99) was associated with increased risk of becoming unable to walk for 6 min continuously without stopping (hazard ratio [HR]: 5.88, 95% confidence interval [CI]: 1.20 to 28.89; p = 0.029) (Fig. 2). Compared with the reference group, a low normal ABI at baseline (ABI 1.00 to 1.10) was associated with an increased risk of becoming unable to walk for 6 min continuously (HR: 3.26, 95% CI: 0.69 to 15.45; p = 0.14), but findings were not statistically significant (Fig. 2). Results were not substantially changed when analyses were repeated in which deaths occurring before a participant stopping during the 6-min walk were included as an outcome. However, in these additional analyses the association of borderline ABI with becoming unable to walk for 6 min continuously did not reach statistical significance (HR: 1.77, 95% CI: 0.93 to 3.36; p = 0.08).
Adjusting for age, sex, race, smoking, comorbidities, body mass index, and physical activity level and compared with the reference group with a normal baseline ABI, borderline and low normal ABI values were not associated significantly with higher risk of experiencing a 15% or greater decline in 6-min walk performance or with higher risk of experiencing a 20% or greater decline in 6-min walk performance (data not shown).
Participants with ABI 1.10 to 1.30 were significantly less likely to develop an ABI <0.90 during follow-up compared with patients whose baseline ABI was 0.90 to 0.99 and those whose baseline ABI was 1.00 to 1.09 (Table 2).We therefore repeated analyses in Figures 1 and 2with additional adjustment for change in ABI during follow-up. Associations of borderline and low normal ABI values with mobility loss and becoming unable to walk for 6 min continuously were attenuated or no longer statistically significant after this additional adjustment (data not shown).
Data presented here show that among 666 men and women with and without PAD, those with borderline ABI values of 0.90 to 0.99 had significantly higher rates of mobility loss and significantly higher rates of becoming unable to walk for 6 min continuously at 5-year follow-up, compared with persons with a normal baseline ABI of 1.10 to 1.30. In addition, participants with low normal ABI values of 1.00 to 1.09 had significantly higher rates of mobility loss at 5-year follow-up, compared with persons who had a normal baseline ABI. These findings are important because the ABI threshold typically considered clinically important is ABI <0.90. However, data presented here demonstrate that even ABI values of 0.90 to 1.09 are associated with higher rates of functional outcomes compared with ABI values of 1.10 to 1.30.
An ABI <0.90 is highly sensitive and specific for diagnosing PAD, compared with lower extremity angiography (10). However, because of increasing impedance with increasing arterial taper, systolic pressures normally increase with greater distance from the heart. Thus, a truly normal ABI is >1.00 (9). Our finding that persons with low normal and borderline ABI values have higher rates of functional decline for some outcomes is consistent with prior studies showing higher prevalences of subclinical atherosclerosis and mortality among persons with low normal and borderline ABI values (5,13).
In contrast to our findings for mobility loss and loss of the ability to walk for 6 min continuously, we found no significant associations of borderline or low normal ABI values with a 15% or 20% decline in 6-min walk performance during 5-year follow-up. Reasons for these discrepancies across our outcome measures are not clear. However, it is possible that becoming unable to walk for 6 min continuously during follow-up is a more specific outcome for walking impairment related to lower extremity ischemia than are the outcomes of 15% or 20% decline in 6-min walk performance. Participants with borderline or low normal ABI values who slow their 6-min walking speed sufficiently to experience 15% or 20% declines in 6-min walk performance but do not stop during the 6-min walk at follow-up may be experiencing walking disability due to comorbidities other than leg ischemia.
The mechanism of higher rates of becoming unable to walk for 6 min continuously and/or mobility loss at 5-year follow-up among persons with low normal and borderline ABI values at baseline cannot be discerned from data presented here. Several mechanisms are possible. First, our results show that participants with low normal and borderline ABI values are more likely to progress to an ABI <0.90 during 5-year follow-up than are participants with normal baseline ABI values. Our findings suggest that a higher rate of progression of lower extremity atherosclerosis among participants with low normal and borderline ABI values is a plausible causal pathway for findings reported here. Alternatively, low normal and borderline ABI values may be markers for other characteristics associated with higher rates of functional decline, such as presence of atherosclerosis in other arterial beds (5).
It is not clear from our data why an interaction with presence of knee arthritis, hip arthritis, disk disease, or spinal stenosis was present for the association of baseline ABI with the inability to walk continuously for 6 min at follow-up, whereas no such interaction was observed for the association of baseline ABI with mobility loss. Differences in these associations relate to the 6-min walk test being an objective measure, and possibly a more accurate measure of actual performance, whereas the mobility loss outcome is subjective. Alternatively, knee or hip arthritis and/or disk disease may affect performance on the 6-min walk to a greater degree than mobility.
First, participants were identified from academic medical centers, and it is unclear whether our findings are generalizable to persons outside of academic medical centers. Second, by necessity, participants with mobility impairment at baseline and those unable to walk for 6 min continuously at baseline were excluded from analyses of mobility loss and incident inability to walk for 6 min continuously at follow-up, respectively. Thus, this study does not assess functional changes in these excluded persons. Third, unmeasured variables may account for associations of borderline and low normal ABI values with functional decline in persons with PAD. Fourth, despite careful identification and confirmation of comorbidities and adjustment for comorbidities in our analyses, we cannot rule out the possibility that comorbidities associated with low normal and borderline ABI values may contribute to findings reported here. Fifth, the relatively small number of events for becoming unable to walk for 6 min continuously limits statistical power. Finally, our results do not include data for disease-specific health-related quality-of-life measures.
At 5-year follow-up, persons who have borderline and low normal ABI values are at higher risk for mobility loss compared with persons who do not have PAD. Participants with borderline ABI values are also at higher risk for becoming unable to walk for 6 min continuously at 5-year follow-up. Further study is needed to confirm the findings reported here and to determine whether interventions such as exercise can prevent functional decline in persons with a borderline or low normal ABI.
This work was supported by grants R01-HL58099, R01-HL64739, R01-HL071223, and R01-HL076298 from the National Heart, Lung, and Blood Institute and by grant RR-00048 from the National Center for Research Resources, and supported in part by the Intramural Research Program, National Institute on Aging, National Institutes of Health. Heather Gornik, MD, served as Guest Editor for this article.
- Abbreviations and Acronyms
- ankle-brachial index
- body mass index
- confidence interval
- hazard ratio
- peripheral arterial disease
- Received February 25, 2008.
- Revision received August 18, 2008.
- Accepted September 10, 2008.
- American College of Cardiology Foundation
- Rosamand W.,
- Flegal K.,
- Friday G.,
- et al.
- McDermott M.M.,
- Liu K.,
- Criqui M.H.,
- et al.
- Newman A.B.,
- Siscovick D.S.,
- Manolio T.A.,
- et al.
- Fung Y.C.
- Wang J.C.,
- Criqui M.H.,
- Denenberg J.O.,
- McDermott M.M.,
- Golomb B.A.,
- Fronek A.
- Resnick H.E.,
- Lindsay R.S.,
- McDermott M.M.,
- et al.
- Shadman R.,
- Criqui M.H.,
- Bundens W.P.,
- et al.
- Guyatt G.H.,
- Sullivan M.J.,
- Thompson P.J.,
- et al.
- Guralnik J.M.,
- Simonsick E.M.,
- Ferrucci L.,
- et al.
- Guralnik J.M.,
- Fried L.P.,
- Simonsick E.M.,
- et al.
- Lin D.Y.,
- Wei L.J.,
- Ying Z.