Prognostic Value of Self-Reported Subjective Exercise Capacity in Patients With Acute Dyspnea

Background Self-reported exercise capacity is a well-established prognostic measure in stable ambulatory patients with cardiac and pulmonary disease. Objectives The authors aimed to directly compare the prognostic accuracy of quantified self-reported exercise capacity using the Duke Activity Status Index (DASI) with the established objective disease-severity marker B-type natriuretic peptide (BNP) in patients presenting with acute dyspnea to the emergency department. Methods The DASI was obtained in a prospective multicenter diagnostic study recruiting unselected patients presenting with acute dyspnea to the emergency department. The prognostic accuracy of DASI and BNP for 90-day and 720-day all-cause mortality was evaluated using C-index. Results Among 1,019 patients eligible for this analysis, 75 (7%) and 297 (29%) patients died within 90 and 720 days after presentation, respectively. Unadjusted hazard ratios (HRs) and multivariable adjusted hazard ratios (aHRs) for 90- and 720-day mortality increased continuously from the fourth (best self-reported exercise capacity) to the first DASI quartile (worst self-reported exercise capacity). For 720-day mortality the HR of the first quartile vs the fourth was 9.1 (95% CI, 5.5-14.9) vs (aHR: 6.1, 95% CI: 3.7-10.1), of the second quartile 6.4 (95% CI: 3.9-10.6) vs (aHR: 4.4, 95% CI: 2.6-7.3), while of the third quartile the HR was 3.2 (95% CI: 1.9-5.5) vs (aHR: 2.4, 95% CI: 1.4-4.0). The prognostic accuracy of the DASI score was high, and higher than that of BNP concentrations (720-day mortality C-index: 0.67 vs 0.62; P = 0.024). Conclusions Quantification of self-reported subjective exercise capacity using the DASI provides high prognostic accuracy and may aid physicians in risk stratification. (Basics in Acute Shortness of Breath EvaLuation [BASEL V] Study [BASEL V]; NCT01831115)

7][8] Likely the best validated one is the Duke Activity Status Index (DASI), a simple self-assessment tool with 12 questions for estimating exercise capacity. 9DASI-scores, with higher values indicating higher exercise capacity, correlate very well with peak oxygen uptake and are validated measures of functional status, particularly in outpatients with heart failure and in the preoperative setting. 9,10 is unknown, whether beyond providing important insights regarding functional health-related quality of life, quantitative assessment of selfreported subjective exercise capacity using the DASI may also help physicians in the risk stratification of patients presenting to the emergency department (ED) with acute dyspnea due to either acute heart failure (AHF) or pulmonary disease.Given the lack of established disease-independent tools for accurately risk stratifying patients with acute dyspnea, the DASI may well have clinical utility.Therefore, we aimed to test this hypothesis by directly comparing the prognostic accuracy of quantified self-reported exercise capacity using the DASI with the established objective disease-severity marker B-type natriuretic peptide (BNP) in a large prospective multicenter study.Self-Reported Exercise Capacity in Acute Dyspnea NT-proBNP concentration at presentation, betablocker intake on entry, and besides, sex was added. 18Higher event rate allowed to additionally adjust the Cox regression model of 720-day mortality, based on relevant clinical findings and medical knowledge, to: systolic blood pressure and peripheral oxygen saturation at presentation, leg edema on entry, history of hypertension, diabetes mellitus, coronary artery disease, atrial fibrillation, previous heart failure and obstructive pulmonary disease, serum creatinine and sodium level, intake of angiotensin converting enzyme inhibitors/angiotensin receptor blockers, and diuretics. 18,19Missing predictor values were imputed using the Markov Chain Monte Carlo method.The number of imputed datasets was 20.The Cox regression analysis was performed on all 20 data sets and the results were pooled using Rubin's rules. 20Findings were confirmed in a sensitivity analysis using the original data.Schoenfeld's global test was used to test the proportional hazards assumption of the Cox regression models.Cox regression models were internally bootstrap validated.Kaplan-Meier curves of 90-and 720-day allcause mortality were plotted, and comparisons were performed by log-rank tests.Censored patients were displayed in tables including number at risk with the use of "survminer" package R statistical software.

METHODS
PROGNOSTIC ACCURACY.2][23][24][25] The prognostic accuracy of the DASI score was also compared to a previously described risk score derived from the compact mortality model, further referred to as the "Voors score", ranging from 0 to 5. 18 Calculations were performed in the original, not imputed data.Moreover, the internal bootstrap validation of the fully adjusted model ( 19variables plus DASI score) was performed in a single imputed dataset.All prognostic accuracies were reported as optimism corrected C-indices.Correlated C-indices were compared using the "compareC" package R statistical software. 26Calibration curves showed observed vs predicted probability for 720-day all-cause mortality of BNP level on admission, the Voors score, the DASI quartiles, and the DASI score are part of the supplemental material ("rms"). 27inical usefulness was assessed by a decision curve analysis ("rmda") calculating the clinical net benefit (true-positive counts minus false-positive counts weighted by the respective threshold probability).For a specific threshold probability, a larger net benefit indicates a greater number of true-positive predictions without increase of false positives.Possible time dependencies were studied using timedependent area under the receiver operating characteristic curve which accommodates censored data ("timeROC"). 28BGROUP ANALYSES.Finally, a subgroup analysis was performed in patients with an adjudicated final diagnosis of AHF.Interaction tests were conducted between gender and the DASI score to evaluate a potential sex-specific effect on mortality in general and in inpatients with adjudicated AHF in particular.

RESULTS
Among 2,153 patients enrolled in BASEL V, 1,019 patients with a median age of 74 years comprising 43% women were eligible for this analysis, as they completed the DASI questionnaire shortly after presentation (median, 2 days; IQR: 1-5 days) (Supplemental Figure 1).Overall, baseline patient characteristics were comparable among patients completing the DASI questionnaire vs those who did not (Supplemental Table 1).The median-modified DASI score was 24.95 (IQR: 15.45-39.45).The most common adjudicated final diagnosis as the cause of acute dyspnea was AHF in 529 (52%) patients.Table 1 displays the baseline characteristics of patients included in this analysis and Supplemental Table 2 shows an overview of the covariates with missing values and summary statistics.  2 and 3).

Self-Reported Exercise Capacity in Acute Dyspnea
M A Y 2 0 2 3 : 1 0 0 3 4 2 and the aHR 2.4 (95% CI: 1.4-4.0)(Figure 3).When analyzing DASI as a continuous variable again the HR and aHR indicated a higher risk for mortality for lower DASI scores, and therefore, lower exercise capacity (Supplemental Tables 3 and 4).For 90-day mortality the HR and aHR were 1.05 (95% CI: 1.03-1.07)and 1.04 (95% CI: 1.02-1.07),respectively.These findings were confirmed in sensitivity analyses only including not imputed data with 851 and 823 patients for the 90-day and 720-day analyses, respectively (data not shown).
PROGNOSTIC ACCURACY.The prognostic accuracy of the DASI score was high, and higher compared to BNP and NT-proBNP levels on admission: for 720-day mortality prediction C-index 0.67 vs 0.62 (n ¼ 743; P ¼ 0.024) and 0.69 vs 0.65 (n ¼ 998; P ¼ 0.075).The prognostic performance of DASI was significantly higher than that of the previously described Voors score (C-index 0.68 vs 0.64, n ¼ 854; P < 0.001).These findings were also reported using timedependent area under the receiver operating characteristic curve (Figure 4) and Kaplan-Meier curves (Supplemental Figure 2).The fully adjusted model showed a high accuracy (C-index of 0.75).Calibration plots display observed vs predicted mortality probability (Supplemental Figures 3 and 4).Corrected C-indices and estimates for optimism are part of the supplement (Supplemental Table 5).The clinical usefulness was assessed by decision curve analysis.Supplemental Figure 5 compares the clinical net benefit for prediction of 720-day all-cause mortality of BNP concentration on admission, the Voors score, DASI quartiles, and DASI score to the extreme cases of intervention for none or all patients.This decision curve analysis showed that for a threshold probability of 0.3 the DASI score (net benefit 0.093) predicted 49 more true positives per 1,000 patients compared to BNP (net benefit 0.044).Values are median (IQR) or n (%).Patients stratified according to DASI into quartiles.P values for trend calculated using Cochran-Armitage test for categorical and linear regression for continuous variables.P < 0.05 were considered statistically significant.All hypothesis testing was 2-tailed.
Self-Reported Exercise Capacity in Acute Dyspnea SUBGROUP ANALYSIS.Similar findings to those of the overall cohort were observed in the 529 (52%) patients with an adjudicated final diagnosis of AHF, the aHR for 90-and 720-day mortality increased continuously from the fourth (reference) to the first DASI quartile (Supplemental Tables 6 and 7).Again, the prognostic accuracy of the DASI score for 90-and 720-day mortality was moderate (C-index 0.65 and 0.64) (Supplemental Figure 6).There was no significant interaction between DASI and gender (Supplemental Table 8).

DISCUSSION
This study directly compared the prognostic accuracy of quantified, self-reported exercise capacity using the DASI with the established objective diseaseseverity BNP in a large prospective multicenter study of patients presenting with acute dyspnea to the ED.We found that the patients in the lowest DASI quartile (0-15.25),and therefore with the lowest exercise capacity, were at significantly higher risk of death within 90 as well as 720 days, with a cumulative mortality of almost 50% after 720 days.Furthermore, the prognostic accuracy of the DASI score for all-cause mortality was high, and higher than BNP as a well-validated possible reference standard.It was even numerically higher vs NT-proBNP concentrations, a biomarker that also includes aspects of renal dysfunction as a second prognostic variable. 13The prognostic value of the DASI persisted after multivariable adjustment for other prognostic variables including age, sex, history of heart failure, chronic obstructive pulmonary disease, pulse oximetry, serum creatinine, hemoglobin, and NT-proBNP concentrations.Patients with low exercise capacity in the days preceding hospitalization (first DASI quartile) had a 7-fold higher risk of death within 720 days vs patients with high exercise capacity (fourth DASI quartile).Comparing the prognostic performance between DASI and an established mortality risk score for patients with heart failure (Voors score) showed a significantly higher performance, emphasizing the substantial prognostic value of self-reported exercise capacity as quantified by DASI.0][31][32][33] In a large study including 1,401 patients undergoing major non-cardiac surgery, DASI scores even outperformed cardiopulmonary exercise testing and NT-proBNP concentrations in the prediction of death or myocardial infarction within 30 days of surgery. 5,10Given the facts that its prognostic value seemed to be largely independent of the acute disease causing the acute dyspnea (cardiac vs pulmonary) in the ED setting, sex had no modifying effect and it was also independent STUDY DESIGN AND POPULATION.BASEL V (Basics in Acute Shortness of Breath EvaLuation) was a prospective, multicenter, diagnostic, and prognostic study enrolling adult patients presenting with acute dyspnea to the ED in 2 University Hospitals in Switzerland (Basel & Zurich). 11-15Patients were included irrespective of renal function, whereas patients with terminal kidney failure on chronic dialysis were excluded.For this analysis, patients were eligible if they had completed the DASI questionnaire within the first days after presentation.The investigation conforms with the principles outlined in the Declaration of Helsinki and the study was approved by the local ethics committee.The authors designed the study, gathered, and analyzed the data according to the STARD (Standards for Reporting Diagnostic accuracy studies) guidelines for studies of diagnostic accuracy.QUANTIFIED SELF-REPORTED EXERCISE CAPACITY: DASI.After patients had provided written informed consent, they received paper forms with 12 questions regarding exercise capacity during daily life representing the DASI and were asked to complete the forms within the first days of hospitalization.Initially, the original 12-item DASI version was translated into German and used (Figure 1). 9 After 2 patients were severely offended by the question regarding sexual activities, in consultation with the ethics committee we decided to omit this question from all further questionnaires.The maximum score of this modified version of the DASI, which was later applied for all patients in this analysis, was 52.95 vs 58.2 in the original version.ADJUDICATION OF FINAL DIAGNOSIS.The final diagnosis of the main disorder responsible for acute dyspnea was centrally adjudicated by 2 independent cardiologists/internists who had access to all patients' medical records including clinical history, physical examination, 12-lead electrocardiogram, laboratory findings, chest x-ray, echocardiography, lung function testing, computed tomography, the response to therapy, and also autopsy data for patients who had died in hospital.All laboratory findings obtained through the clinician's routine diagnostic workup were available for this study.These findings included one of the natriuretic peptides (BNP or N-Terminal pro-BNP [NT-proBNP]) that current guidelines recommended for diagnosing AHF with a Class I recommendation. 1,2In situations of disagreement about the final diagnosis, cases were reviewed and adjudicated by a third cardiologist.

FOLLOW
-UP.Patients were contacted 3, 12, and 24 months after discharge by telephone or in written A B B R E V I A T I O N S A N D A C R O N Y M S aHR = adjusted hazard ratio AUROC = area under the receiver operating characteristic curve BNP = B-type natriuretic peptide BUN = blood urea nitrogen COPD = chronic obstructive pulmonary disease DASI = Duke Activity Status Index ED = emergency department NT-proBNP = N-terminal pro-B-type natriuretic peptide The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors' institutions and Food and Drug Administration guidelines, including patient consent where appropriate.For more information, visit the Author Center.Manuscript received September 26, 2022; revised manuscript received January 19, 2023, accepted February 9, 2023.Belkin et al JACC: ADVANCES, VOL. 2, NO. 3, 2023 Self-Reported Exercise Capacity in Acute Dyspnea M A Y 2 0 2 3 : 1 0 0 3 4 2 form.Specific questions regarding dyspnea, possible rehospitalizations, and relevant diagnostics (eg, electrocardiogram, echocardiography .)performed after discharge were assessed.Furthermore, patients were asked to fill out another DASI questionnaire which was sent in paper form.Information regarding death during follow-up was obtained from the hospital medical records, the general practitioner, and the national mortality registry.GENERAL STATISTICAL METHODS.Continuous variables are presented as median (IQR) and categorical variables are expressed by numbers (percentages).Linear regression and Cochran-Armitage tests were used to calculate P values for trend across the quartiles for continuous and categorical characteristics, respectively. 16,17All hypothesis testing was 2-tailed and P values of <0.05 were considered to indicate statistical significance.Adjustment for multiple testing was not performed.Statistical analysis was operated using SPSS for Windows 26.0 (SPSS Inc) and R statistical Software Version 3.4.3(MathSoft).DASI-SCORE FOR RISK STRATIFICATION.Unadjusted and multivariable adjusted Cox regression analyses were performed to quantify the independent effect of the DASI on 90-and 720-day all-cause mortality.The DASI was considered as a continuous score as well as stratified into quartiles with the fourth DASI-quartile including patients with the highest exercise capacity and serving as reference group.When treated as a continuous variable DASI score was reversed, thus a higher score indicated lower exercise capacity.The Cox regression model of 90-day mortality was adjusted for previously published covariates, further referred to as the "compact model", including age, the natural logarithm blood urea nitrogen, hemoglobin level, the natural logarithm of

FIGURE 2
FIGURE 2 DASI in Predicting Future Risk of 90-Day Death From Any Cause of the dominant comorbidity in the preoperative setting, the generalizability of implementing the DASI into clinical routine seems high.As the questions within the DASI substantially overlap with the initial CENTRAL ILLUSTRATION The DASI Predicting Future Risk of 720-Day Death From Any Cause Belkin M, et al.JACC Adv.2023;2(3):100342.Patients stratified into DASI quartiles and referenced to fourth quartile (highest self-reported exercise capacity).Kaplan-Meier curves with P value (log-rank test), unadjusted HRs with corresponding 95% CI and numbers at risk.DASI ¼ Duke Activity Status Index; ED ¼ emergency department.J A C C : A D V A N C E S , V O L . 2 , N O . 3 Self-Reported Exercise Capacity in Acute Dyspnea nurse-lead functional and social assessment routinely performed in hospitalized patients in many institutions, the extra workload associated with the clinical implementation of the DASI may be rather small, further increasing the attractiveness of this tool.Prospective implementation studies seem warranted to evaluate the best strategies on how to implement the DASI into clinical routine.In this study of mostly older patients presenting with acute dyspnea to the ED, the question related to sexual activity severely offended 2 patients leading to the withdrawal of this question for all remaining patients.As assessing sexual activities may be delicate and also controversial in many other settings and not all patients are sexual active with partners, the findings of this study, based on the modified DASI version without the question on sexual relations, may have major clinical implications as they highlight the opportunity of using this modified, so far not validated DASI version also

FIGURE 3
FIGURE 3 DASI in Predicting Future Risk of 720-Day Death From Any Cause

J
A C C : A D V A N C E S , V O L . 2 , N O . 3 , 2 0 2 3 elsewhere.Although, a recent study assessing whether all 12 DASI items were equally important, omitted the same sensitive question and showed that the discriminative ability of the DASI was not substantially weakened, further studies with the modified DASI version externally validating our findings are warranted. 34STUDY LIMITATIONS.As this study required informed consent, we cannot comment on the

FIGURE 4
FIGURE 4 Prognostic Accuracy of DASI for All-Cause Mortality Within 720 days of Follow-Up Compared to Natriuretic Peptide Peptide Concentrations on Admission and to the Voors Score

TABLE 1
Baseline Characteristics in the Overall Cohort Grouped According to DASI Quartiles