Walking Impairment Questionnaire Improves Mortality Risk Prediction Models in a High-Risk Cohort Independent of Peripheral Arterial Disease Status

Abstract

Background

The Walking Impairment Questionnaire (WIQ) is a subjective measure of patient-reported walking performance developed for peripheral arterial disease. The purpose of this study is to examine whether this simple tool can improve the predictive capacity of established risk models and whether the WIQ can be used in patients without peripheral arterial disease.

Methods and Results

At baseline we assessed the walking distance, stair-climbing, and walking speed WIQ category scores among individuals who were undergoing coronary angiography. During a median follow-up of 5.0 years, there were 172 mortalities among 1417 study participants. Adjusted Cox proportional hazards models showed that all 3 WIQ categories independently predicted future all-cause and cardiovascular mortality, including among individuals without peripheral arterial disease (P<0.001). Compared with the cardiovascular risk factors model, we observed significantly increased risk discrimination with a C-index of 0.741 (change in C-index, 0.040; 95% confidence interval, 0.011–0.068) and 0.832 (change in C-index, 0.080; 95% confidence interval, 0.034–0.126) for all-cause and cardiovascular mortality, respectively. Examination of risk reclassification using the net reclassification improvement index showed a 48.4% (P<0.001) improvement for all-cause mortality and a 77.4% (P<0.001) improvement for cardiovascular mortality compared with the cardiovascular risk factors model.

Conclusions

All 3 WIQ categories independently predicted future all-cause and cardiovascular mortality. Importantly, we found that this subjective measure of walking ability could be extended to patients without peripheral arterial disease. The addition of the WIQ scores to established cardiovascular risk models significantly improved risk discrimination and reclassification, suggesting broad clinical use for this simple, inexpensive test.

Cardiovascular disease affects >80 million Americans,¹ with >8 million of these individuals having peripheral arterial disease (PAD).² Given that atherosclerotic diseases, such as PAD, have a prolonged incubation period and that therapies exist to delay disease progression, clinical tools to identify those at risk could improve cardiovascular disease outcomes.³

The Walking Impairment Questionnaire (WIQ) is a subjective measure of patient-perceived walking performance developed for individuals with PAD.⁴ This test has been shown to be a valid and reliable correlate of objective walking ability.^5–9 Among those with PAD, objective measures of walking performance (eg, the 6-minute and 4-minute walking tests) have been shown to predict clinical outcomes, such as future mortality.^10,11 Previous studies have examined the association of the subjective WIQ with clinical outcomes and found associations with cardiovascular events and mortality in patients with PAD, but not in those without PAD.^12–14

Although the WIQ was developed for patients with PAD, objective walking ability as a predictor of clinical outcomes is not limited to patients with PAD. The 6-minute walking test has been shown to predict future mortality in patients with both congestive heart failure¹⁵ and pulmonary hypertension.¹⁶ In addition, the long-distance corridor walk,¹⁷ which is comparable with the 6-minute walking test,¹⁸ and gait speed¹⁹ have been associated with future cardiovascular disease and mortality in older community-dwelling adults. Therefore, it is possible that subjective measures of walking ability, such as the WIQ, may be useful predictors of risk in patients even in the absence of PAD.

Here, we test whether lower WIQ category scores are associated with future all-cause and cardiovascular mortality among those with PAD, as previously shown, and whether this simple test can be extended to those without PAD. Additionally, we examine for the first time whether the WIQ can improve existing clinical tools and independently improve the predictive capacity, discriminatory ability, and net reclassification parameters of established cardiovascular risk prediction models in an at-risk cohort.

Methods

Study Population

The Genetic Determinants of Peripheral Arterial Disease (GenePAD) study is composed of individuals (n=1755) who underwent coronary angiography at Stanford University and Mount Sinai Medical Centers between January 1, 2004, and March 1, 2008.^20–22 All individuals provided written informed consent. The GenePAD study was funded by the National Heart, Lung, and Blood Institute and approved by the Stanford University and Mount Sinai School of Medicine Committees for the Protection of Human Subjects.

Inclusion Criteria

Individuals were included in the study sample if complete data were available on all relevant covariates: age, sex, race, smoking history, body mass index, systolic blood pressure (SBP), use of lipid-lowering and antihypertensive medications, use of insulin or oral hypoglycemic agents, total cholesterol, high-density lipoprotein–cholesterol, and ankle-brachial index (ABI). Remaining eligible patients with an ABI>1.4²³ were excluded (n=12). Using these criteria, 1428 patients were identified. Complete WIQ data were available in >99% of these individuals for all 3 WIQ category scores (n=1417).

Walking Impairment Questionnaire

Patients completed the WIQ at enrollment with a trained nurse or research assistant. The WIQ consists of 3 primary categories assessing walking distance, stair-climbing, and walking speed, as previously described.⁷ Individuals are asked to rate the degree of difficulty of various activities with responses ranging from 0 (unable) to 4 (none). Walking distance questions range from walking indoors to walking 1500 feet. Stair-climbing questions range from climbing 1 flight of stairs to climbing 3 flights of stairs. Walking speed questions range from walking 1 block slowly to jogging 1 block. Questions within each category are then weighted based on the degree of difficulty, according to the approximate number of feet, stairs, or miles per hour for the distance, stair-climbing, and speed scores, respectively. Scores are then divided by the maximum number of points and presented on a scale of 0% to 100%, where 0% represents the lowest possible score (ie, answering “unable” for all questions in that category) and 100% represents the highest possible score (ie, indicating “none” with regard to difficulty for all questions in that category).

Outcomes

The outcomes of interest in this analysis were death from any cause and death from cardiovascular causes. Cardiovascular deaths were those attributed to myocardial infarction, cardiac arrest, stroke, heart failure, or aneurysm rupture. Ascertainment of mortality was achieved through phone or postal communication with the participant or the participants’ listed contacts and medical record review. Additional mortalities and confirmation of the date of death for reported mortalities were obtained via patient linkage to the Social Security Death Index. New mortalities were identified through March 31, 2012.

Covariates

Detailed information on all included covariates was obtained by a trained nurse or clinical research assistant at enrollment. Age, sex, race, and smoking history were acquired by self-report, and body mass index and SBP were measured. The use of lipid-lowering and antihypertensive medications was evaluated by direct medication inventory. Diabetes mellitus status was classified as self-reported use of insulin or oral hypoglycemic agents. Total and high-density lipoprotein–cholesterol levels were measured at the time of coronary angiography.

Before the coronary angiogram, posterior tibial, dorsalis pedis, and brachial artery systolic pressures were measured using a 5-MHz Doppler ultrasound. The ABI for each patient was calculated by dividing the higher ankle pressure of each leg over the higher of the left or right brachial pressures. Each patient was then classified as having PAD by an ABI of <0.9 in either leg or not having PAD with an ABI ≥0.9 in both legs.

Statistical Methods

For all survival analyses, the follow-up time was defined as the period between the enrollment interview and the last confirmed follow-up or date of death. Adjusted models included the following covariates: age, sex, race, smoking history (ever or never), body mass index, SBP, use or nonuse of lipid-lowering and antihypertensive medications at enrollment, diabetes mellitus status, total cholesterol, high-density lipoprotein–cholesterol, and ABI. All covariates were continuous except race (categorical), smoking, use or nonuse of lipid-lowering and anti-hypertensive medications, and diabetes mellitus status (dichotomous).

Kaplan–Meier curves were constructed to examine the risk of all-cause mortality in those reporting any deficit (score<100%) compared with those reporting no deficit (score=100%) in each WIQ category score. Cox proportional hazards regression models were used to calculate unadjusted and adjusted hazard ratios for this binary score exposure to determine the use of the WIQ as a binary measure. Separate models were calculated for each WIQ category score.

The associations per standard deviation (SD) decrease in WIQ category scores with death from all causes and death from cardiovascular causes were investigated using unadjusted and adjusted Cox proportional hazards regression models calculated individually for each WIQ category score. Associations with mortality per SD decrease were evaluated in subgroup analysis among individuals without PAD. Additionally, the interaction of each WIQ category score with PAD status was evaluated in the adjusted model to determine whether the association of the WIQ category score with all-cause and cardiovascular mortality significantly differed (P<0.05) according to PAD status. Proportional hazards assumptions were evaluated by Schoenfeld residuals tests.

The C-index, the integrated discrimination improvement (IDI), and the net reclassification improvement (NRI) were evaluated to determine whether the WIQ category scores significantly improved risk discrimination and reclassification for all-cause and cardiovascular mortality over a baseline model. These analyses were conducted per SD change in WIQ score and used time-to-event data as previously described.²⁴ In this diverse population at high risk for cardiovascular events, we used a comprehensive baseline model consisting of risk factors for cardiovascular disease and death, including age, sex, race, smoking history, body mass index, SBP, use or nonuse of lipid-lowering and antihypertensive medications at enrollment, diabetes mellitus status, total cholesterol, high-density lipoprotein–cholesterol, and ABI.^25–27 Additionally, secondary analyses were conducted using risk variables from the European Systematic Coronary Risk Evaluation (SCORE) risk model to evaluate model improvement against an established risk score.²⁶ This model was established for cardiovascular mortality and includes age, sex, smoking history, SBP, and total cholesterol.

The C-index was used to quantify improvements in model discrimination with the addition of the individual WIQ category scores to the baseline model.²⁸ In survival analysis, the C-index is an extension of the area under the regional operating characteristic curve or C statistic, while allowing for censored data.^29,30 A 1% increase in the C-index would indicate that the correct order of failure (eg, mortality) is accurately predicted in an additional 1 in every 100 pairs of randomly selected individuals compared with the baseline model.

Model performance with the addition of the WIQ category scores to the baseline model was further evaluated using the IDI estimate.²⁴ The IDI compares 2 models according to the average difference in predicted risk between those who have the event and those who do not. If the new model assigns a higher risk to those who will have a mortality and a lower risk to those who will not, as compared with the baseline model, the IDI will be >0. Therefore, the IDI can be interpreted as the average net improvement in the predicted risk of the outcome in the new model compared with the baseline model.

The NRI was used to evaluate the proportion of correct risk re-classification when adding the WIQ category scores to the baseline model.²⁴ The category-free NRI was used in this study because it has been suggested to be the most objective and reproducible measure of improvement in risk prediction, particularly when established a priori risk categories do not exist.³¹ This measure quantifies the degree of correct upward or downward absolute risk reclassification with the addition of a new variable or variables to a baseline model. Furthermore, the NRI was calculated separately among individuals with and without an event (eg, all-cause or cardiovascular mortality) during follow-up.

Calibration was assessed on all models using the Grønnesby–Borgan test to evaluate goodness-of-fit (P>0.05) by comparing predicted mortalities with observed mortalities as described for survival analysis.³²

Tests were considered significant if the 2-sided P value was <0.05. All analyses were performed using Stata version 12.0 (StataCorp, College Station, TX). Study data were collected and managed using REDCap electronic data capture tools hosted at Stanford University.³³

Results

Study Population Characteristics

Enrollment characteristics of the 1417 participants in this study are displayed in Table 1. There were 172 mortalities (12%), of which 47 were known to be from cardiovascular causes, during a median follow-up period of 5.0 years (inter-quartile range, 4.0–6.3).

Table 1.

Baseline Study Population Characteristics (n=1417)

Characteristic	Value
Age, mean y (SD)	67 (10)
Women, n (%)	495 (35)
Ethnicity
White	790 (56)
Black	175 (12)
Hispanic	158 (11)
Asian	114 (8)
Other*	180 (13)
Systolic blood pressure, mean mm Hg (SD)	139 (21)
Body mass index, mean kg/m2 (SD)	29 (6)
Lipids, mean mg/dL (SD)
Total cholesterol	139 (37)
High-density lipoprotein–cholesterol	41 (12)
Ever smoker, n (%)	813 (57)
Use of cholesterol-lowering medication, n (%)	965 (68)
Use of antihypertensive therapy, n (%)	1133 (80)
Use of insulin or oral hypoglycemics, n (%)	429 (30)
Peripheral arterial disease, n (%)	234 (17)
Walking Impairment Questionnaire, mean score % (SD)
Distance score	70 (37)
Stair-climbing score	53 (39)
Speed score	51 (32)

^*Includes Asian-Indian, Pakistani, Middle Eastern, and Pacific Islander.

SD indicates standard deviation.

Mortality Prediction Models

There were 634 individuals reporting no deficit in any WIQ category, 341 in 1 category, 217 in 2 categories, and 225 in all 3 categories. Kaplan–Meier curves demonstrated increased cumulative all-cause mortality (Figure) among individuals reporting any deficit (score<100%) in the distance, stair-climbing, or speed scores compared with those reporting no deficit (score=100%). Unadjusted and adjusted hazard ratios for this binary score exposure are presented in Table 2. In adjusted models, a significantly increased risk of mortality was observed among individuals reporting any deficit in the distance (hazard ratio, 2.24; 95% confidence interval [CI], 1.56–3.23), stair-climbing (hazard ratio, 1.90; 95% CI, 1.19–3.04), and speed scores (hazard ratio, 2.25; 95% CI, 1.22–4.14). Additionally, a 56% (hazard ratio, 1.56; 95% CI, 1.27–1.91; P<0.001) increased risk of mortality was observed for each additional WIQ category with any reported deficit.

Figure.

Kaplan–Meier cumulative all-cause mortality curves for the Walking Impairment Questionnaire distance (A), stair-climbing (B), and speed (C) scores among individuals reporting no deficit in Walking Impairment Questionnaire score (blue) compared with those reporting any deficit in Walking Impairment Questionnaire score (red).

Table 2.

Hazard Ratios for All-Cause Mortality by Walking Impairment Questionnaire Category for Individuals With Any Reported Deficit (Score<100%) Compared With No Reported Deficit (Score=100%)

	Unadjusted		Adjusted*
	HR (95% CI)	P Value	HR (95% CI)	P Value
Distance score	2.95 (2.11–4.10)	<0.001	2.24 (1.56–3.23)	<0.001
Stair-climbing score	2.74 (1.77–4.26)	<0.001	1.90 (1.19–3.04)	0.007
Speed score	3.38 (1.88–6.08)	<0.001	2.25 (1.22–4.14)	0.009

CI indicates confidence interval; and HR, hazard ratio.

^*Adjusted for age, sex, race, smoking history, body mass index, systolic blood pressure, use of lipid-lowering medication, use of antihypertensive medication, diabetes mellitus status, total cholesterol, high-density lipoprotein-cholesterol, and ankle-brachial index.

The association per SD decrease in WIQ category scores with all-cause and cardiovascular mortality is presented in Table 3. The distance, stair-climbing, and speed scores were all significantly associated (P<0.001) with increased risk of all-cause and cardiovascular mortality in the fully adjusted models. The adjusted estimated increase in risk observed per SD decrease in WIQ category scores ranged from 54% to 73% for all-cause mortality and 83% to 191% for cardiovascular mortality. Schoenfeld residuals tests demonstrated that the proportional hazards assumption was met for all models.

Table 3.

Hazard Ratios Per SD Decrease in Walking Impairment Questionnaire Category Score

	Unadjusted		Adjusted*
	HR (95% CI)	P Value	HR (95% CI)	P Value
All-cause mortality
Distance score	1.73 (1.50–1.99)	<0.001	1.54 (1.32–1.81)	<0.001
Stair-climbing score	1.91 (1.60–2.27)	<0.001	1.73 (1.42–2.10)	<0.001
Speed score	1.95 (1.63–2.33)	<0.001	1.68 (1.37–2.06)	<0.001
Cardiovascular mortality
Distance score	2.10 (1.59–2.77)	<0.001	1.83 (1.34–2.51)	<0.001
Stair-climbing score	3.29 (2.13–5.07)	<0.001	2.91 (1.82–4.66)	<0.001
Speed score	3.04 (2.02–4.58)	<0.001	2.67 (1.69–4.22)	<0.001

CI indicates confidence interval; and HR, hazard ratio.

^*Adjusted for age, sex, race, smoking history, body mass index, systolic blood pressure, use of lipid-lowering medication, use of antihypertensive medication, diabetes mellitus status, total cholesterol, high-density lipoprotein-cholesterol, and ankle-brachial index.

The all-cause and cardiovascular mortality associations per SD decrease in WIQ category scores remained significant (P<0.001) when examined among individuals without PAD (Table I in the online-only Data Supplement). Additionally, the test for interaction between the presence or absence of PAD and the WIQ category scores demonstrated that the observed associations did not significantly differ according to PAD status for either all-cause (P>0.2 for all categories) or cardiovascular (P>0.4 for all categories) mortality.

C-Index

Results for the C-index analysis are presented in Table 4. Significant increases in the C-index were noted with the independent addition of the distance, stair-climbing, and speed scores to the cardiovascular risk factors model for all-cause and cardiovascular mortality. The addition of all 3 WIQ scores yielded the largest increase with a C-index of 0.741 (change in C-index [ΔC], 0.040; 95% CI, 0.011–0.068) and 0.832 (ΔC, 0.080; 95% CI, 0.034–0.126) compared with the baseline models for all-cause (C-index, 0.701; 95% CI, 0.660–0.743) and cardiovascular (C-index, 0.752; 95% CI, 0.685–0.819) mortality, respectively. CIs around each reported C-index are presented in Table II in the online-only Data Supplement.

Table 4.

C-Index and Integrated Discrimination Improvement Over the Baseline Cardiovascular Risk Factors Model

Model		C-Index		IDI
	C*	ΔC (95% CI)	P Value	IDI, % (95% CI)	P Value
All-cause mortality
BRF†	0.701	ref	ref	ref	ref
BRF + distance score	0.735	0.034 (0.010 to 0.058)	0.006	1.2 (0.4 to 2.0)	0.002
BRF + stair-climbing score	0.731	0.030 (0.004 to 0.056)	0.023	1.6 (0.9 to 2.4)	<0.001
BRF + speed score	0.731	0.030 (0.007 to 0.053)	0.012	1.1 (0.5 to 1.8)	0.001
BRF + all 3 scores	0.741	0.040 (0.011 to 0.068)	0.006	1.8 (1.0 to 2.7)	<0.001
Cardiovascular mortality
BRF	0.752	ref	ref	ref	ref
BRF + distance score	0.797	0.045 (−0.004 to 0.094)	0.073	0.9 (−0.1 to 1.8)	0.083
BRF + stair-climbing score	0.821	0.070 (0.026 to 0.113)	0.002	1.8 (0.8 to 2.8)	<0.001
BRF + speed score	0.819	0.067 (0.024 to 0.110)	0.002	0.7 (−0.2 to 1.6)	0.108
BRF + all 3 scores	0.832	0.080 (0.034 to 0.126)	0.001	1.8 (0.7 to 2.8)	0.001

BRF indicates baseline risk factors; ΔC, change in C-index from the baseline risk factors model; C, C-index; CI, confidence interval; IDI, integrated discrimination improvement; and ref, reference.

^*C-index for all models was significantly >0.5 at P<0.001.

^†Age, sex, race, smoking history, body mass index, systolic blood pressure, use of lipid-lowering medication, use of antihypertensive medication, diabetes mellitus status, total cholesterol, high-density lipoprotein-cholesterol, and ankle-brachial index.

Integrated Discrimination Improvement Index

The IDI demonstrated a significant average net improvement in the predicted risk for all-cause and cardiovascular mortality with the addition of the WIQ category scores (Table 4). The stair-climbing score exhibited the largest IDI among models, with the addition of individual WIQ category scores for both all-cause (IDI, 1.6%; 95% CI, 0.9–2.4) and cardiovascular (IDI, 1.8%; 95% CI, 0.8–2.8) mortality.

Category-Free Net Reclassification Improvement

The category-free NRI indicated that for all models tested, there was a statistically significant (P<0.001) improvement in the net proportion of risk reclassification as compared with the cardiovascular risk factors model (Table 5). In models with individual WIQ category scores, the stair-climbing score had the largest NRI for all-cause mortality (NRI, 47.5%), where the net proportion of individuals correctly reclassified was 33.7% and 13.7% for those with and without an event, respectively. The individual addition of the stair-climbing score to the baseline model for cardiovascular mortality demonstrated an NRI of 81.9% with an event NRI of 58.7% and a nonevent NRI of 23.2%.

Table 5.

Category-Free Net Reclassification Improvement Over the Baseline Cardiovascular Risk Factors Model

Model	Overall NRI		Event NRI, %	Nonevent NRI, %
	NRI, %	P Value
All-cause mortality
BRF*	ref	ref	ref	ref
BRF + distance score	44.4	<0.001	9.3	35.1
BRF + stair-climbing score	47.5	<0.001	33.7	13.7
BRF + speed score	32.2	<0.001	19.8	12.4
BRF + all 3 scores	48.4	<0.001	24.4	24.0
Cardiovascular mortality
BRF	ref	ref	ref	ref
BRF + distance score	60.3	<0.001	23.4	36.9
BRF + stair-climbing score	84.0	<0.001	61.7	22.3
BRF + speed score	51.5	0.001	27.7	23.9
BRF + all 3 scores	77.4	<0.001	48.9	28.5

BRF indicates baseline risk factors; NRI, net reclassification improvement; and ref, reference.

^*Age, sex, race, smoking history, body mass index, systolic blood pressure, use of lipid-lowering medication, use of antihypertensive medication, diabetes mellitus status, total cholesterol, high-density lipoprotein-cholesterol, and ankle-brachial index.

Calibration

Assessment of calibration using the Grønnesby–Borgan statistic demonstrated good fit for all models with and without the WIQ category scores (P≥0.12).

Secondary Analyses

The results of the addition of WIQ category scores to the model consisting of SCORE risk variables are presented in Tables III and IV in the online-only Data Supplement. These analyses demonstrated statistically significant improvement for all measures of risk discrimination and reclassification for all-cause and cardiovascular mortality using the C-index, IDI, and NRI. Additionally, the rates of all-cause and cardiovascular mortality did not significantly differ between individuals included and excluded from the current study (P≥0.26).

Discussion

We found that lower WIQ distance, stair-climbing, and speed scores all independently predicted future all-cause and cardiovascular mortality among high-risk individuals even when accounting for several potential confounders, including ABI. Importantly, even though the WIQ was developed for use in patients with PAD, the associations were significant among patients without PAD and did not significantly differ according to PAD status. For all 3 WIQ categories, we also found that any reported deficit (score<100%) was associated with a significantly increased risk of death from any cause. The evaluation of the C-index, IDI, and NRI demonstrated that the WIQ category scores significantly improved risk discrimination and reclassification for all-cause and cardiovascular mortality over the baseline model of cardiovascular risk factors and the model of SCORE risk variables. In total, these data suggest that the WIQ is associated with survival in a high-risk population, correlates with survival even among those without PAD, and can significantly improve the prognostic capacity of established risk prediction models.

This study is the first to evaluate the WIQ in a high-risk cardiovascular cohort referred for coronary angiography and to show that the WIQ can predict future all-cause and cardiovascular mortality among individuals without PAD. Additionally, this is the first study to show that the WIQ improves risk prediction beyond a model of cardiovascular risk factors. Therefore, the WIQ may be a clinically useful tool to evaluate the risk of future mortality among individuals at high risk for cardiovascular events regardless of PAD status at the time of WIQ administration. These data suggest that the WIQ may reflect overall cardiopulmonary fitness and reserve and therefore may have broad applicability to the cardiovascular field. The finding that any reported deficit in the 3 WIQ categories is associated with an increased risk of death underscores the potential clinical use of the WIQ for risk evaluation as a simple binary measure.

The WIQ has been validated against objective measures of walking ability in a number of populations,^5,6,34 but few studies have examined its clinical use in predicting outcomes.^12–14 Schiano et al¹² followed 760 patients, of whom 60 were diagnosed with PAD, over a 24-month period and found that the stair-climbing and walking speed scores were significantly associated with cardiovascular events in the 60 PAD patients only. No association with cardiovascular events was observed for the walking distance score or for any WIQ scores among individuals without PAD. Similarly, Gardner et al¹³ examined the ability of the WIQ to predict future mortality among 434 patients with PAD and found an association for the stair-climbing score only (P=0.023).

In a larger study, Jain et al¹⁴ evaluated the association of the WIQ with all-cause and cardiovascular mortality in 1048 individuals with and without PAD during a median follow-up period of 4.5 years. They found that the stair-climbing score predicted all-cause and cardiovascular mortality among the 638 individuals diagnosed with PAD only. However, no significant associations were observed for the distance or speed scores among patients with PAD or for any of the 3 WIQ categories among individuals without PAD. The results of these previous studies, taken with the observation that in this study the stair-climbing score was often the strongest predictor of risk, suggests that the stair-climbing WIQ category may be the most sensitive predictor of future outcomes.

Strengths

This is the largest known study used to investigate the association of WIQ scores with outcomes. Detailed enrollment characteristics allowed for the adjustment of a large number of potential confounders. Finally, WIQ scores were available for >99% of eligible individuals.

Limitations

This study was conducted among individuals referred for coronary angiography, and therefore the findings presented may not be generalizable to lower risk populations. Additionally, although a large number of covariates were included in the adjusted models, unknown and residual confounding cannot be completely excluded as explanations for the observed associations. In this analysis, we were unable to directly compare the WIQ to objective measures of walking ability. Finally, this study was not sufficiently powered to determine whether a risk model including all 3 WIQ category scores is superior to a simpler model with a single component.

Conclusions

Walking ability has been shown to predict clinical outcomes across multiple diseases and in community settings. This study extends previous work by showing that all 3 WIQ categories can predict mortality risk, independent of established risk factors. More importantly, we provide evidence that in addition to being an independent predictor of risk, the WIQ actually improves the accuracy of established risk prediction models. The WIQ is a patient-reported measure of walking ability that has been validated in multiple populations and could potentially be administered as a simple, safe, and economical tool to improve risk prognostication, even in patients without PAD. Future investigations will be needed to confirm these findings, and randomized prospective trials examining improvements in outcomes will be necessary to determine whether the addition of the WIQ to clinical models should be used to guide clinical intervention.

WHAT IS KNOWN

• Subjective measures of walking capacity, such as the Walking Impairment Questionnaire, have previously been shown to be associated with cardiovascular events in patients with peripheral arterial disease.

• It is unclear whether this clinical tool can be applied to a broader patient population and whether it can provide information beyond established risk models.

WHAT THE STUDY ADDS

• This study shows that the Walking Impairment Questionnaire is an independent predictor of all-cause and cardiovascular mortality among individuals undergoing coronary angiography, even in those without peripheral arterial disease.

• When added to established risk models, this questionnaire significantly improves mortality risk discrimination and reclassification.

• Taken together, these data suggest that simple and economical tools that correlate with exercise capacity may be able to enhance our ability to prognosticate risk.