The Prognostic Value of the Diastolic Stress Test in Patients Undergoing Treadmill Stress Echocardiography

The Prognostic Value of the Diastolic Stress Test in Patients Undergoing Treadmill Stress Echocardiography Benjamin T. Fitzgerald, MBBS, Jeffrey J. Presneill, MBBS, PhD, MBiostat, Isabel G. Scalia, BSc, MD, Casey L. Hawkins, BSc, MD, Yael Celermajer, MD, William M. Scalia, BSc (Hons), MD, and Gregory M. Scalia, MBBS, MMedSc, Auchenflower, Chermside, Brisbane, and Gold Coast, Queensland; and Melbourne, Victoria, Australia

Background: Exercise stress echocardiography (SE) is well validated for the evaluation of myocardial ischemia. Diastolic stress testing (DST) is recommended in the 2016 American Society of Echocardiography and European Association of Cardiovascular Imaging Guidelines for unexplained dyspnea. This study’s aim was to prognostically evaluate the DST prospectively in a large stress testing population. Methods: Patients underwent SE with mitral E/e’ measured before and after maximal treadmill exertion to estimate diastolic function. Patients were divided into four groups: group 1 (n = 201)—ischemic; group 2 (n = 1,563)— negative DST (E/e’pre < 12, E/e’post < 12); group 3 (n = 68)—positive DST (E/e’pre < 12, E/e’post $ 12); group 4 (n = 314)—high baseline E/e’ (E/e’pre $ 12). Results: Consecutive patients (n = 2,201, 770 [35%] female; 58 6 12 years) were followed after SE for 27,964 patient-months. Time to first heart failure event (composite of heart failure admission, worsening New York Heart Association class, worsening ejection fraction, or cardiovascular death) was analyzed and adjusted using Cox proportional hazards regression. Ischemic patients hazard ratio (HR) was 28, 95% CI, 17-44, P < .0005, for subsequent heart failure compared with negative DST patients. Nonischemic, positive DSTs were highly predictive (HR = 4.2; 95% CI, 1.6-11.0; P = .001); while high E/e’pre was not predictive (HR = 1.3; 95% CI, 0.7-2.4; P = .49) of future heart failure events. Conclusions: DST differentiates heart failure prognosis in patients with induced diastolic dysfunction. Ischemia predictably portends the worst heart failure outcomes, and nonischemic, positive diastolic stress tests predicted more events compared with negative tests. These prognostic data support and add to the recommendations of the 2016 guidelines. (J Am Soc Echocardiogr 2019;-:---.) Keywords: Diastolic stress test, Stress echocardiography, Guidelines, Mitral E/e’ ratio, Prognosis

Left ventricular diastolic function can be evaluated noninvasively using Doppler echocardiography. The latest American Society of Echocardiography and European Association of Cardiovascular Imaging Guidelines (the Guidelines) for the evaluation of diastolic function endorse the use of the ratio of the resting mitral E-wave velocity and the resting mitral annular e’-wave velocity measured by Doppler tissue imaging as one measure of left ventricular filling presFrom HeartCare Partners, GenesisCare (B.T.F., G.M.S.), and Wesley Hospital (B.T.F., G.M.S.), Auchenflower; Prince Charles Hospital (B.T.F., G.M.S.), Chermside; Bond University, Gold Coast (Y.C.); University of Queensland (I.G.S., C.L.H., G.M.S.), Brisbane (I.G.S., C.L.H., G.M.S.), Queensland; and Royal Melbourne Hospital (J.J.P.), University of Melbourne (J.J.P., W.M.S.), and Monash University (J.J.P.), Melbourne, Victoria, Australia. Conflicts of Interest: None. Reprint requests: Dr. Benjamin T. Fitzgerald, MBBS, HeartCare Partners, GenesisCare, Level 5, Sandford Jackson Building, 30 Chasely Street, Auchenflower, Queensland, Australia 4051 (E-mail: [email protected]). 0894-7317/$36.00 Copyright 2019 by the American Society of Echocardiography. https://doi.org/10.1016/j.echo.2019.05.021

sures.1 The diastolic stress test (DST) is well described in the Guidelines, measuring these velocities at rest and at peak exercise, and is recommended for use when patients have unexplained dyspnea, especially with exertion.1 Myocardial ischemia is one form of exercise-induced diastolic dysfunction, which can be identified using stress echocardiography (SE).2 Indeed, changes in the mitral E/e’ ratio during stress testing have been shown to assist with the detection of myocardial ischemia, probably due to diastolic changes occurring earlier than systolic abnormalities in the ischemic cascade.3,4 Inducible diastolic dysfunction in the absence of inducible ischemia is hypothesized to impart adverse outcomes. Postexercise measurements of diastolic parameters are easiest to perform using supine bicycle exercise but can be performed before and after treadmill stress.1,5 Normal values for the DST have previously been determined4-6 and validated.7,8 The Guidelines reference four studies where the DST has been comprehensively studied, each in small numbers of patients.5-7,9 This study was designed to evaluate the DST in a large stress testing population, with subsequent assessment of adverse heart failure events during prolonged patient follow-up. 1

2 Fitzgerald et al

Abbreviations

BP = Blood pressure

Journal of the American Society of Echocardiography - 2019

METHODS

Consecutive patients referred to GenesisCare clinical testing DST = Diastolic stress testing, facilities for SE between June test 2011 and June 2014 were studied DE/e’ = Change in E/e’ with prospectively. Indications for exercise or delta E/e’ testing were typically dyspnea or chest pain. Standard Bruce protoE/e’ ratio = Mitral inflow Ewave velocity to mitral annular col treadmill testing with digital e’-wave velocity ratio gated echocardiography before and after exercise was performed. ePLAR = Echocardiographic Patients requiring dobutamine SE pulmonary to left atrial ratio and those with resting left bundle E/e’pre = Mitral inflow velocity branch block, a paced rhythm or E to mitral annular velocity e’ atrial fibrillation, reduced ejection ratio at rest fractions (<50%), or mitral valve E/e’post = Mitral inflow replacements or repairs were velocity E to mitral annular excluded. Patients with moderate velocity e’ ratio after peak or worse valvular heart disease exertion (stenosis or regurgitation) were also excluded, as were patients HR = Hazard ratio with severe mitral annular calcifiMETs = Metabolic cation. equivalents General Electric medical grade MPHR = Maximum predicted treadmills using Case systems heart rate (220 – patient age in (Milwaukee, WI) were used to years) replicate and quantitate exercise. Standard Bruce protocols were NYHA = New York Heart used to produce exercise stress in Association a controlled environment and in SE = Stress a reproducible manner. Imaging echocardiography, was performed using various echocardiogram echocardiography machines TR = Tricuspid regurgitation including the General Electric velocity Vivid e9 (Horton, Norway) and Vivid 7 (Horton, Norway), Siemens SC2000 and SC2000 Prime (Mountain View, CA), and the Phillips ie33 (Best, The Netherlands) scanners. All tests were supervised and read by cardiologists with subspecialty training in SE and an exercise physiologist. The echocardiogram was performed by cardiac sonographers with subspecialty training in SE. Results were then overread, standardized, and recorded by a SE specialized cardiologist. The echocardiographic images were acquired in the parasternal long-axis, short-axis, and apical four, two, and three chambers. Ejection fraction was measured by Simpson’s method. Mitral inflow pulsed-wave Doppler including E-wave and A-wave velocities were measured and compared with the pulsed-wave tissue Doppler imaging e’-wave velocity at the medial and lateral mitral annulus, as per the Guidelines in the apical four-chamber view. The mitral E/e’ ratio was then averaged. These measurements were recorded at rest, and in recovery postexertion, immediately after the ventricular wall assessment (see Figure 1).1 Changes in mitral E-wave velocity and the annular e’-wave velocity are said to persist for at least several minutes after the end of exertion and were measured after the regional wall assessment.7,10-12 The diastolic parameters in this study were acquired between 90 and 180 seconds after the completion of exercise.

Other baseline markers of diastolic function as recommended in the Guidelines1 were recorded, including the indexed left atrial volume and tricuspid regurgitant velocity (TR)13 and echocardiographic pulmonary to left atrial ratio or echocardiographic pulmonary to left atrial ratio (ePLAR; maximum TR/E/e’ in m/sec). Baseline characteristics also recorded were age, Framingham risk, body surface area, body mass index, and ejection fraction. Blood pressure (BP) was recorded at rest and at peak exertion, and the percentage of maximum predicted heart rate (MPHR) achieved and exercise workload completed in metabolic equivalents (METs) were documented. Left atrial volume was measured in the apical four- and two-chamber views and averaged using the method of discs. TRs were estimated before and after exercise (where possible), in the view with the highest measurement (typically, but not restricted to the apical four-chamber window).1,9 Blood tests were not performed as part of this study. Ischemic stress tests (group 1; see Figure 2) were defined as those with new regional wall motion abnormalities in two contiguous segments or cavity dilatation with a lack of cardiac augmentation (direct comparison of the pre- and postexercise echocardiographic images). A subset of patients underwent anatomic testing for coronary obstruction (invasive or computed tomography coronary angiography), as determined by the treating physician. Nonischemic SEs (groups 24) were defined as having no evidence of myocardial ischemia on the SE (no new regional wall motion abnormalities, no cavity dilatation, and appropriate augmentation of cardiac contractility postexercise). Group 2, classified as the negative DSTs, had a nonischemic SE. Their resting E/e’ (E/e’pre) was <12, and the peak exercise value (E/ e’post) was also <12. Group 3 had a positive DST. They also had nonischemic tests but had an increase in the peak exercise diastolic parameters (E/e’pre < 12, E/e’post $ 12). Group 4 was defined as having high baseline E/e’ with E/e’pre $ 12. Patients’ medical records were subsequently reviewed for up to 5 years after the SE data collection was commenced. Predetermined end points were admission to hospital with a diagnosis of heart failure, worsening New York Heart Association (NYHA) class, a reduction in ejection fraction > 10%, and cardiovascular death. Worsening NYHA class was determined by the treating cardiologist at the time of follow-up. This assessment was made independent of the study, and the treating specialist was blinded to the results. The time interval to a patient’s first heart failure event (a composite of the above) was analyzed using Cox proportional hazards regression, adjusted for patient age, sex, exercise capacity, and pretest Framingham risk, at the time of the SE. The study design and methodology were reviewed and approved by the HeartCare Partners (GenesisCare) echocardiographic working group ethics subcommittee. The research protocol was carried out in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki). All patients provided written consent. Statistical Analyses A multivariable model incorporating the resting and peak stress mitral E/e’ ratio for the prediction of subsequent heart failure events in adult patients referred for SE was developed.14 Data were exported in Microsoft Excel format for subsequent statistical analysis using Stata Statistical Software Release14 (StataCorp LP, College Station, TX). Demographic and baseline data were tabulated and summarized using descriptive statistics. No imputation for missing data was

Fitzgerald et al 3

Journal of the American Society of Echocardiography Volume - Number -

RESULTS

HIGHLIGHTS  Guidelines currently recommend diastolic stress testing based on limited data.  Diastolic assessment (E/e’ pre- and post-exertion) can be done in most stress echoes.  DST identifies ischemic, nonischemic positive DST, and nonischemic negative DST.  Ischemic stress tests portend extremely poor heart failure outcomes.  Nonischemic positive DST predict worse outcomes compared to negative DST.

performed, and the number of analyzed observations was reported for each summary proportion. Unpaired t-tests were used for cohort analysis of continuous data for separate populations. Paired t-tests were used to compare continuous data for dependent variables. Data expression was presented as the mean 6 SD. The outcome for time-to-event analyses within each patient was the first observed element of the study composite outcome of predetermined end points. Results were presented graphically using Kaplan-Meier curves assessed by log-rank tests, as well as tabulated unadjusted and adjusted hazard ratios (HRs) with 95% CI returned by Cox proportional hazard regression models. Variables included in adjusted Cox models were the patient’s echocardiographic category (four categories), gender, age in years at the time of baseline assessment, exercise capacity in METs, and pretest Framingham risk. The proportional hazards assumption was evaluated using scaled Schoenfeld residuals and visual assessment of log-log plots.15

There were 2,201 consecutive SEs available with complete data for analysis from 770 female (35%), and 1,431 male patients. Mean 6 SD age was 58 6 12 years. The baseline characteristics are detailed in Table 1. Total follow-up was for 27,964 patients-months (up to 5 years per patient, for a mean of 13 6 19 months). Ischemic versus Nonischemic In the ischemic cohort (group 1, n = 201 with 51 females, 26% of the total cohort), the mitral E/e’ ratio increased significantly from the resting value (E/e’pre) to the postexercise measurements (E/e’post, see Figure 1B and Tables 2 and 3). The largest difference was seen for the medial E/e’ ratio (E/e’pre 9.8 6 3.3 to E/e’post 12.1 6 4.3, P < .0001), but all three measurements increased significantly (lateral E/e’pre 7.7 6 2.6 to E/e’post 9.5 6 3.4, P < .0001; averaged E/e’pre 8.6 6 2.7 to E/e’post 10.5 6 3.6; P < .0001). In the nonischemic cohort (groups 2-4, n = 2,000, including 790 or 40% females, mean age 57 6 12 years), all peak SEs showed a normal hemodynamic response to exercise,16 with an increase in ejection fraction and stroke volume, with no new regional wall motion abnormalities or cavity dilatation (see Figure 1A and Table 2). The DST—Rationale for Selection of E/e’ Threshold Values The 2016 Guidelines suggest that a normal DST response results in an average E/e’post < 10 and that an abnormal response is an average E/ e’post > 14, or a medial E/e’post >15.1 This leaves a ‘‘gray zone’’ of measurements that are yet to be defined. These Guidelines are based on previous data that have used various values for E/e’post as abnormal, from >10 through to >15.3,6-9,17 Some studies used medial E/e’, while others used average E/e’post. For this study, assuming a normal distribution, the nonischemic patients (groups 2-4,

Figure 1 (A) Normal (negative DST) and (B) abnormal (positive DST) E/e’pre and E/e’post mitral inflow and Doppler tissue imaging waveforms of the DST.

4 Fitzgerald et al

Journal of the American Society of Echocardiography - 2019

Figure 2 Diagram of the SE cohort and the breakdown into the different groups.

Table 1 Baseline characteristics for the 2,201 Patients (770 females [35%], 2,000 Nonischemic SE, 201 Ischemic SE) Characteristics

Mean age, yrs BSA, m2

Overall cohort

Nonischemic patients (n = 2,000)

Ischemic patients (n = 201)

58 6 12

57 6 12

65 6 9

1.99 6 0.24

1.99 6 0.23

2.00 6 0.30

P value

<.0001 NS

BMI, kg/m2

28.8 6 5.0

28.8 6 5.0

28.3 6 4.1

NS

Framingham risk score, %

16.7 6 12.0

16.2 6 11.9

22.2 6 11.6

.018

Ejection fraction, %

64 6 5

64 6 5

62 6 6

NS

Indexed left atrial volume

31 6 8

30 6 8

34 6 9

<.0001

TR, m/sec

2.42 6 0.29

2.40 6 0.29

2.40 6 0.29

NS

ePLAR, m/sec

0.28 6 0.09

0.28 6 0.09

0.27 6 0.10

NS

BP pre

125/78

125/78

129/77

NS

BP post

177/78

NS

MPHR% METs

95 6 9% 10.7 6 3.4

177/78

176/80

94 6 9

91 6 12

NS

10.8 6 3.4

8.9 6 3.5

<.0001

The P values compare the ischemic patients to the nonischemic patients. BMI, Body mass index; BSA, body surface area.

n = 2,000) demonstrated a 90th percentile value for the upper limit of normal18-20 for E/e’post of 12. An average E/e’post >14 or a medial E/e’post >15 resulted in very few cases for a statistical analysis. The DE/e’ values were not discriminatory for evaluating outcomes. A receiver operating characteristic analysis of medial E/e’ cutoff values resulted in an area under the curve of 0.71. The optimal cutoff value for a medial E/e’post of 12 yielded a likelihood of correct classification of 89% (see Figure 3). Based on these data, medial E/ e’post $ 12 was selected as the definition of an abnormal result (the positive DST) in this study. The medial E/e’ ratio was used as it was the most reproducible and easiest to perform. In post hoc analysis, it also provided the largest variation between rest and postexertion E/e’. Based on these criteria, the nonischemic cohort (groups 2-4, n = 2,000) was divided into three groups. Group 2 (n = 1,633) represented patients with a negative DST (not suggestive of an inducible increase in filling pressures, E/e’post < 12). Group 3 (n = 68) represented the positive DST (E/e’pre < 12, E/e’post $ 12). Group 4 (n = 314) had high pretest E/e’ (E/e’pre $12). The baseline patient characteristics for these groups are presented in Table 3. An overall comparison of the E/e’ values pre- and postexertion for the four groups is displayed in Figure 4. It is also possible to describe the change in E/e’ ratio (DE/e’). In group 1 (ischemic SEs), the E/e’ increased by an average of 2 from

rest to peak stress (DE/e’ of +2 6 3, P < .0001). For nonischemic tests (groups 2-4), the E/e’ minimally decreased (DE/e’ of –0.3 6 0.4, P < .0001). The ischemic and nonischemic DE/e’ values were statistically different (P < .0001). Heart Failure Event Analysis Raw event rates for the four groups are displayed in Table 4. Ischemic patients (group 1) not surprisingly had the worst heart failure outcomes. It should be noted that at baseline, these patients were older and had larger atrial volumes, reduced exercise tolerance (as measured by METs), and higher Framingham risk scores compared with negative DST patients (see Table 1). The adjusted HR in this group for the composite heart failure end point was 28 (95% CI, 17-44; P < .0005) compared with patients with a negative DST (see Figure 5 and Tables 4 and 5). The positive DST group (group 3) were also at increased risk of subsequent heart failure events and had significantly poorer outcomes compared with patients with a negative DST. These patients were older, had decreased exercise tolerance, and had higher Framingham risk scores compared with group 2 patients (see Table 3). When the analysis was adjusted for age, gender, exercise capacity, and Framingham risk score, the positive DST

Fitzgerald et al 5

Journal of the American Society of Echocardiography Volume - Number -

Table 2 E/e’ Ratio Pre- and Postexercise for 2,000 Consecutive Nonischemic SEs (Groups 2-4), with Breakdown by Decade Medial

Average

Lateral

n

Pre

Post

P

Pre

Post

P

Pre

Post

P

2,000

9.1 6 3.0

8.6 6 2.6

<.0001

8.3 6 2.7

7.7 6 2.4

<.0001

7.4 6 2.6

7.2 6 2.7

.015

<30 years

56

6.6

6.9

.18

5.7

5.9

.48

4.9

5.6

.002

30-39

97

7.1

7.1

.96

6.5

6.5

.85

5.9

6.2

.09

40-49

307

7.9

7.7

.04

7.1

6.9

.03

6.2

6.2

.97

50-59

559

8.7

8.4

.01

7.9

7.4

<.001

7.0

6.8

.03

60-69

657

9.6

8.8

<.001

8.7

7.8

<.001

7.8

7.5

.005

70-79

271

10.7

10.0

<.001

9.7

9.0

<.001

8.8

8.8

.54

>79

149

12.8

11.7

.09

11.8

11.1

.25

10.8

Group

All Age

11

.71

Table 3 Baseline characteristics of Group 2 (negative DST) Compared with Group 3 (positive DST) and Group 4 (high baseline filling) with E/e’ values before and after stress testing Characteristics

n (% female) Mean age

Group 2 (negative DST)

1,563 (38) 56 6 12 years

Group 3 (positive DST)

P value (Group 3)

68 (45)

Group 4 (high E/e’pre)

P value (Group 4)

314 (40)

64 6 11

<.0001

66 6 9

<.0001

BSA

1.99 6 0.23

1.98 6 0.23

.73

1.97 6 0.24

.34

BMI

27.8 6 4.9

28.2 6 5.3

.52

29.3 6 5.1

<.0001

Framingham risk score

15.7 6 11.4

24.3 6 16.4

<0.0001

20.8 6 12.7

<.0001

Ejection fraction

64 6 5

65 6 4

.20

65 6 6

.32

Indexed left atrial volume

30 6 8

32 6 7

.052

32 6 10

.02

TR

2.3 6 0.26

2.5 6 0.25

.0006

2.5 6 0.32

<.0001

ePLAR

0.29 6 0.07

0.25 6 0.05

.0045

0.18 6 0.03

<.0001

BP pre

124 6 16

128 6 18

.041

131 6 17

<.0001

BP post

176 6 23

179 6 27

.33

179 6 25

.08

MPHR% METs

92 6 11% 11.0 6 3.4

E/e’ pre

8.2 6 3.2

E/e’ post

7.8 6 1.6 P < .001

90 6 10

.18

93 6 11

.47

9.1 6 3.0

<.0001

8.7 6 2.8

<.0001

9.5 6 1.9

.11

14.3 6 3.3

<.0001

12.7 6 1.2

<.0001

11.7 6 3.9

<.0001

The P values in the baseline data are compared with the group 2 values. For the E/e’ data, the P value compares the group 3 or the group 4 value to the group 2 data. BMI, Body mass index; BSA, body surface area.

patients were still shown to have significantly worse heart failure outcomes (HR = 4.2; 95% CI, 1.6-11.0; P = .001; see Figure 5 and Table 4). Patients with high baseline E/e’ (group 4) also appeared statistically different from patients with negative DSTs (group 2) at baseline (see Table 3). They were older and larger in size and had higher baseline BP, larger atrial volumes (although still within the normal range), lower exercise capacity, and higher Framingham scores. Despite their apparent increased risk, they did not have more heart failure events compared with group 2 (HR = 1.3; 95% CI, 0.7-2.4, P = .49, see Figure 5 and Tables 4 and 5). Despite the presumed high filling pressure at rest, the E/e’pre decreased on average from 14.3 6 3.3 to E/e’post 11.7 6 3.9, P < .0001 (see Figure 4 and Table 3). Patients with an E/e’pre >12 who had an increase in the E/e’post were too few in number for a meaningful statistical analysis.

DISCUSSION SE is a well validated investigation for the assessment of cardiac status and pathologies, especially for the detection of obstructive coronary artery disease.5,21,22 One of the manifestations of myocardial ischemia is abnormal relaxation and diastolic dysfunction.1,23,24 Diastolic abnormalities occur earlier in the ischemic cascade than systolic changes.1,3 Assessment of diastolic parameters during exercise has been shown to be feasible and valuable.1,5,7,8 In normal subjects, exertion results in the mitral inflow E-wave velocity and the mitral annular e’-wave velocity increasing proportionally, with minimal change to their ratio.1,5,21 For patients with diastolic dysfunction, abnormalities of myocardial relaxation mean that the annular velocities are restricted more than mitral inflow, and the mitral E/e’ ratio will increase. Filling pressures need to increase to augment the needed increase in cardiac output.1,5,21 Assessing

6 Fitzgerald et al

Journal of the American Society of Echocardiography - 2019

Cutpoint >= 10>= 11>= 12>= 13>= 14>= 15>= 20-

Sensivity 60.87% 32.61% 30.43% 23.91% 19.57% 15.22% 6.52%

Specificity 72.16% 85.05% 90.14% 94.80% 96.87% 97.99% 99.58%

Correctly Classified 71.89% 83.80% 88.72% 93.12% 95.03% 96.01% 97.36%

LR+ 2.1867 2.1809 3.0860 4.6020 6.2542 7.5526 15.3750

LR0.5422 0.7924 0.7718 0.8026 0.8303 0.8653 0.9388

Figure 3 Smooth fitting receiver operating characteristic curve with 95% CIs developed to assess first heart failure event for various medial E/e’ results. The LR+ is the ratio of the probability of a positive test among the truly positive subjects to the probability of a positive test among the truly negative subjects. The LR is the ratio of the probability of a negative test among the truly positive subjects to the probability of a negative test among the truly negative subjects. LR+, positive likelihood ratio; LR-, negative likelihood ratio.

Figure 4 Box and whiskers plot of the E/e’pre and E/e’post for the four groups.

diastolic parameters therefore has a role in stress testing and is an early marker of myocardial ischemia.1,5,7,8 The 2016 Guidelines endorse the use of the DST in selected circumstances1; however, the recommendations are based on limited studies and small numbers of patients.5-8 Stress E/e’ can be assessed using several techniques. The 2016 Guidelines suggest using the

average of the medial and lateral measurements.1 The research presented here settled on using the medial measurement as it proved to be the most practical and robust in normal daily clinical practice. The data presented here support this simplified approach, with significant differences between rest and postexercise producing the strongest statistical results.

Fitzgerald et al 7

Journal of the American Society of Echocardiography Volume - Number -

Figure 5 Kaplan-Meier curves estimating heart failure events after DST.

This study was designed prospectively to evaluate the DST in a very large stress testing population. In subjects with a nonischemic SE and negative DST (group 2), the mitral E/e’ ratio did not change or decreased by a small amount (see Figure 4). Predictably, patients with an ischemic SE (group 1, those suggestive of myocardial ischemia and therefore inducible diastolic dysfunction) demonstrated a significant rise in these markers of diastolic dysfunction. Stress diastology provided additional information for the evaluation of myocardial ischemia. A significant increase in the ratio suggests diastolic dysfunction and may indicate an ischemic SE. The results of this study may assist in that assessment.17,25 The data also suggest that these patients have significantly more heart failure events, reflecting a poorer prognosis (see Figure 5 and Table 5). Among the patients with a nonischemic SE, there was a group whose E/e’post increased (group 3, the positive DST patients). These patients were observed to have more heart failure events than patients with a normal E/e’post (group 2 or the negative DSTs; see Figure 5 and Table 5). These data suggest that for patients with a history of dyspnea, investigation with DST may be warranted. This study suggested a poorer prognostic outcome for these patients and potentially identifies a group for whom medical intervention may be warranted. This is consistent with the current Guidelines. In the data presented here, the cutoff for a positive DST was lower than the Guidelines suggest, with an E/e’post of 12 or more being associated with significantly worse heart failure outcomes. One of the surprising results was that the high E/e’ at rest group (group 4) did not have increased adverse heart failure events, compared with those in the negative DST group (group 2). A closer evaluation showed that while the cohort by our definition started with E/e’pre $ 12, the majority would not have met the 2016 Guidelines criteria high filling pressures (when incorporating the measurement of LA volume and/or TR). Additionally, they had a reduction in the E/e’ with exercise (E/e’pre 14.3 6 3.3 vs E/e’post 11.7 6 3.9, P < .0001; see Figure 4). Their DE/e’ results were comparable to group 2 (patients with a negative DST), even though they started from a higher E/e’pre. The observed event rates reflected this (see Figure 5 and Table 5). Those patients who had an increase in the E/e’post were very small in number, and no meaningful statistical analysis could be performed. Contrary to the Guidelines, this may suggest that performing the DST in patients with high E/e’ at rest may

Table 4 Heart failure (HF) events by group Events

Event rate, %

HF admission

[ NYHA class

Y Ejection fraction

CV death

1

79

39.5

11

63

4

1

2

26

1.7

1

21

4

3

6

8.8

1

5

4

14

4.4

Group

14

CV, Cardiovascular.

Table 5 Adjusted cox proportional hazard results for combined heart failure outcomes by group compared with group 2 (nonischemic and negative DST) Group Number

Adjusted* Name

HR (95% CI)

P value

2

Negative DST

1.0

1

Ischaemic

28 (17-44)

3

Positive DST

4.2 (1.6-11.0)

.001

4

High baseline E/e’

1.3 (0.7-2.4)

.49

Age†

1.3 (1.1-1.4)

<.0005

Female

1.1 (0.8-1.7)

.53

<.0005

*Effect estimated adjusted for all other variables in the table. † Age centered at 60 years; effect is for a 5-year increment.

still be worthwhile. If the E/e’post decreased from baseline, they may be at lower risk of events. While not definitively shown in this cohort, an increase in the ratio with exertion may be the more important finding, almost regardless of the increased E/e’ at rest. This may further suggest that an increase in the E/e’post is the more significant factor in predicting future heart failure events. There are multiple limitations in this observational cohort. Ideally this research would be conducted with patients undergoing an anatomical test to confirm the diagnosis and measure pressure invasively, blindly after the stress test. From a practical and clinical

8 Fitzgerald et al

viewpoint, this is not feasible, as it would negate the noninvasive attribute of SE. There would also be ethical implications for low- to intermediate-risk patients undergoing an invasive test.2,26-28 Previous studies have already documented the accuracy of SE for the detection of myocardial ischemia.5-8 In most clinical scenarios, SE is used to attempt to avoid an invasive test. A limited number of patients in this study had a clinical reason to evaluate their anatomy, and in those patients, the effective and accurate role of stress imaging in this setting was confirmed. In follow-up, the ischemic patients (group 1) had significantly worse outcomes, further suggesting the ‘‘true’’ positive nature of this group. Patients were not exclusively referred for assessment of dyspnea or heart failure. Ideally, for assessment of heart failure outcomes, patient selection would have been limited to these indications. The likely lower risk of heart failure in this study group was reflected in the low number of adverse events. However, the broader referral population in this study setting does extend the utility of DST to a general stress test population. While baseline clinical characteristics were documented, blood tests were not. Ideally, documenting natriuretic peptides and troponins before testing and during follow-up would have added further to the interpretation of the results. The E/e’post cannot be performed in all patients. Exercise-induced tachycardia results in fusion of the E- and A-waves (and the annular e’and a’-waves), making assessment difficult. By performing the measurements after the ventricular assessment, this problem was minimized. It is also reasonable to wait until the signals separate to perform the assessments.1 Only a small percentage of patients in this study could not have the E/e’ measured. The E/e’post was able to be performed in >97% of these SEs. In this study, patients were not excluded or stratified based on their resting diastolic assessment. The nonischemic group was a heterogeneous stress testing population containing varying levels of E/e’ ratios at rest. Future research could look at differentiating the E/e’ ratios at rest, to see how different levels of diastolic function at rest change with exertion. In this nonrandomized, single-center cohort study, there were some apparent differences between the groups at baseline (see Table 1). Despite presentation of the adjusted estimates, these baseline differences may have influenced the outcomes presented here. Patients with an E/e’pre $ 12 showed the most differences at baseline compared with group 2 (negative DST), but did not statistically have more events. Their E/e’post decreased with exertion, in a manner similar to the negative DST patients. This could suggest that the baseline differences were less important that the change in E/e’ with stress, further emphasizing the value of DST. The final analysis did adjust for the baseline differences. Evaluation of the medical records was a potential source of ascertainment bias. The possibility exists for events occurring elsewhere and potentially not being included in the patient review. The ischemic patients (group 1) may have been assessed more closely due to the results of the stress test, resulting in a higher reporting of events. The event rate in the negative DST patients (group 2) was, however, very low, so that an underdetection of events in this group would not be expected to influence these results. The positive DST patients (group 3) were unlikely to have been preferentially scrutinized, as the DST results were not communicated to the clinicians because the implications were not apparent at the time of data collection. While failure to appropriately account for missing data in analyses may lead to bias and loss of precision, imputation of missing results also requires assumptions. The present analysis of cases with complete data rather than imputation of missing values may not be

Journal of the American Society of Echocardiography - 2019

associated in an epidemiological context with substantial bias in reported regression estimates.29 The multivariable prediction model described in the present study was derived from echocardiographic observations at a single center. There were significantly fewer women than men (a common problem in cardiac research). This does reflect real world experiences. The reviewers were not blinded to the results of the stress test, making it possible for ascertainment or other biases to occur. Overall event rates were low, especially in the nonischemic patients. The present model would benefit from validation within an independently collected data set.

CONCLUSION The 2016 American Society of Echocardiography and the European Association of Cardiovascular Imaging Recommendations for the Evaluation of Left Ventricular Diastolic Function by Echocardiography Guidelines define and discuss the role of the DST. The results presented in this study generally validate the DST in a very large stress testing population. These new data suggest an expansion in the role of the DST, with minor changes to the definitions suggested in the 2016 guidelines. Measuring the E/e’ ratio at rest and peak exercise is a useful additional element of the SE and may provide incremental prognostic value. The DST identifies patients with significantly worse heart failure outcomes. It suggests that while an ischemic SE predicts markedly increased heart failure events, nonischemic, positive DSTs also confer adverse heart failure consequences, compared with patients without induced diastolic dysfunction. Patients with markers of elevated filling at rest (in this study definition), but who have a normal diastolic stress response, have similar heart failure risk to patients without an elevated E/e’ at rest and after exertion. These data confirm, corroborate, and expand upon the latest Guidelines and establish the prognostic value of the DST.

REFERENCES 1. Nagueh NF, Otto A, Appleton CP, Byrd BF, Dokainish H, Edvardsen T, et al. Recommendations for the evaluation of left ventricular diastolic function by echocardiography: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr 2016;29:277-314. 2. Nesto RW, Kowalchuk GJ. The ischemic cascade: temporal sequence of hemodynamic, electrocardiographic and symptomatic expressions of ischemia. Am J Cardiol 1987;59:23C-30C. 3. Podolec P, Rubis P, Tomkiewicz-Pajak L, Kopec G, Tracz W. Usefulness of the evaluation of left ventricular diastolic function changes during stress echocardiography in predicting exercise capacity in patients with ischemic heart failure. J Am Soc Echocardiogr 2008;21:834-40. 4. Ha JW, Lulic F, Bailey KR, Pellikka PA, Seward JB, Tajik AJ, et al. Effects of treadmill exercise on mitral inflow and annular velocities in healthy adults. Am J Cardiol 2003;91:114-5. 5. Marwick TH. Stress echocardiography—Its role in the diagnosis and evaluation of coronary artery disease. 2nd ed. Boston: Kluwer Academic Publishers; 2003. 6. Schiano-Lomoriello V, Santoro C, de Simone G, Trimarco B, Galderisi M. Diastolic bicycle stress echocardiography: normal reference values in a middle age population. Int J Cardiol 2015;191:181-3. 7. Burgess MI, Jenkins C, Sharman JE, Marwick TH. Diastolic stress echocardiography: hemodynamic validation and clinical significance of estimation of ventricular filling pressure with exercise. J Am Coll Cardiol 2006;47: 1891-900.

Journal of the American Society of Echocardiography Volume - Number -

8. Holland DJ, Prasad SB, Marwick TH. Prognostic implications of left ventricular filling pressure with exercise. Circ Cardiovasc Imaging 2010;3: 149-56. 9. Talreja DR, Nishimura RA, Oh JK. Estimation of left ventricular filling pressure with exercise by Doppler echocardiography in patients with normal systolic function: a simultaneous echocardiographic-cardiac catheterization study. J Am Soc Echocardiogr 2007;20:477-9. 10. Feigenbaum’s echocardiography. 7th ed. Philadelphia: Wolters Kluwer Health/Lippincott Williams & Wilkins; 2010. 11. Grewal J, McCully RB, Kane GC, Lam C, Pellikka PA. Left ventricular function and exercise capacity. JAMA 2009;301:286-94. 12. Ishii K, Imai M, Suyama T, Maenaka M, Nagai T, Kawanami M, et al. Exercise-induced post-ischemic left ventricular delayed relaxation or diastolic stunning: is it a reliable marker in detecting coronary artery disease? J Am Coll Cardiol 2009;53:698-705. 13. Scalia GM, Scalia IG, Kierle R, Beaumont R, Cross DB, Feenstra J, et al. ePLAR—the echocardiographic pulmonary to left atrial ratio—a novel non-invasive parameter to differentiate pre-capillary and post-capillary pulmonary hypertension. Int J Cardiol 2016;212:379-86. 14. Moons KG, Altman DG, Reitsma JB, Ioannidis JP, Macaskill P, Steyerberg EW, et al. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): explanation and elaboration. Ann Intern Med 2015;162:W1-73. 15. Grambsch PM, Therneau TM. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika 1994;81:515-26. 16. Fitzgerald BT, Ballard EL, Scalia GM. Estimation of the blood pressure response with exercise stress testing. Heart Lung Circ 2019;28:742-51. 17. Ha JW, Oh J, Pellikka PA, Ommen SR, Stussy VL, Bailey KR, et al. Diastolic stress echocardiography: a novel noninvasive diagnostic test for diastolic dysfunction using supine bicycle exercise Doppler echocardiography. J Am Soc Echocardiogr 2005;18:63-8. 18. Herrera L. The precision of percentiles in establishing normal limits in medicine. Transl Res 1958;52:34-42. 19. Falkner B, Daniels SR, Flynn JT, Gidding S, Green LA, Ingelfinger JR, et al. The fourth report on the diagnosis, evaluation, and treatment of high blood pressure in children and adolescents. Pediatrics 2004;114(Suppl 2):555-76.

Fitzgerald et al 9

20. Dubowitz T, Ghosh-Dastidar M, Eibner C, Slaughter ME, Fernandes M, Whitsel EA, et al. The women’s health initiative: the food environment, neighborhood socioeconomic status, BMI, and blood pressure. Obesity (Silver Spring) 2012;20:862-71. 21. Pellikka PA, Nagueh S, Elhendy AA, Kuehl CA, Sawada SG. American Society of Echocardiography recommendations for performance, interpretation, and application of stress echocardiography. J Am Soc Echocardiogr 2007;20:1021-41. 22. Douglas PS, Garcia MJ, Haines DE, Lai WW, Manning WJ, Patel AR, et al. Appropriate use criteria for echocardiography. A report of the American College of Cardiology Foundation Appropriate Use Criteria Task Force, American Society of Echocardiography, American Heart Association, American Society of Nuclear Cardiology, Heart Failure Society of America, Heart Rhythm Society, Society for Cardiovascular Angiography and Interventions, Society of Critical Care Medicine, Society of Cardiovascular Computed Tomography, Society for Cardiovascular Magnetic Resonance American College of Chest Physicians. J Am Soc Echocardiogr 2011;24: 229-67. 23. Nagueh SF, Sun H, Kopelen HA, Middleton KJ, Khoury DS. Hemodynamic determinants of the mitral annulus diastolic velocities by tissue Doppler. J Am Coll Cardiol 2001;37:278-85. 24. Ommen SR, Nishimura RA. A clinical approach to the assessment of left ventricular diastolic function by Doppler echocardiography: Update 2003. Heart 2003;89:iii18-23. 25. Fitzgerald B, Scalia G. Delta DTI improves stress echocardiography in the real world. Heart Lung Circ 2012;21:S196. 26. Braunwald’s heart disease: A textbook of cardiovascular medicine. 10th ed. Philadelphia: Saunders; 2015. 27. Textbook of cardiovascular medicine. 2nd ed. Philadelphia: Lippincott Williams & Wilkins; 2002. 28. Do D, West JA, Morise A, Atwood JE, Froelicher V. An agreement approach to predict severe angiographic coronary artery disease with clinical and exercise test data. Am Heart J 1997;134:672-9. 29. Hughes RA, Heron J, Sterne JAC, Tilling K. Accounting for missing data in statistical analyses: multiple imputation is not always the answer. Int J Epidemiol 2019; https://doi.org/10.1093/ije/dyz032.