ROLE OF DIASTOLIC FUNCTION IN PRESERVED EXERCISE CAPACITY IN PATIENTS WITH REDUCED EJECTION FRACTION

Abstract

Background

Some patients with a markedly reduced ejection fraction (EF < 35%) have preserved exercise performance greater than predicted for age and gender. Because diastolic function maybe a determinant of exercise performance, we hypothesized that patients with preserved exercise tolerance despite EF < 35% may have relatively normal diastolic function.

Methods

We retrospectively examined 65 subjects with EF < 35% who underwent exercise Doppler echocardiography and had no inducible ischemia. Forty-five subjects with normal EF (> 60%) and preserved exercise capacity were analyzed as a control group.

Results

Sixteen of 65 patients with EF < 35% had greater than predicted normal exercise capacity for their age and gender, and the remaining 49 patients had reduced exercise capacity. Patients with reduced EF and preserved exercise capacity had an E/e’ (10±4) similar to that of controls (10±3) and less than those with reduced exercise tolerance (16±8, p < 0.01). In addition, they had better diastolic filling patterns and smaller left atrial sizes than patients with EF < 35% and reduced exercise capacity. Multivariate logistic regression analyses indicated that an E/e’ was an independent predictor of a preserved exercise capacity among patients with reduced EF.

Conclusions

Relatively intact diastolic function contributes to preserved exercise capacity in patients who have a reduced EF < 35%.

INTRODUCTION

Reduced exercise tolerance is a key symptom of patients with heart failure (HF) and is common among those with heart failure with a reduced ejection fraction (EF). Normally, during exercise, an increase in left ventricular (LV) suction allows a larger stroke volume to fill the LV without an increase in left atrial pressure to abnormal levels.¹ This normal response is reduced in the presence of heart failure, both with a reduced or preserved EF.^2-4 This suggests that diastolic dysfunction may be one of the factors that contribute to exercise intolerance in patients with HF with a reduced EF (HFrEF).

Some patients with a reduced EF have preserved exercise capacity.^5-7 Furthermore, some patients with reduced EF do not have any clinical manifestations of HF, i.e. Stage B HF.⁸ If left ventricular (LV) diastolic function is an important determinant of exercise tolerance, we hypothesized that relatively normal diastolic function is necessary for patients with reduced EF to have preserved exercise capacity.

METHODS

Patients

This study was approved by the institutional review board at Wake Forest School of Medicine. We retrospectively analyzed patients with an LV EF < 35% who underwent clinically indicated treadmill exercise stress echocardiography between June 2006 and March 2011 at Wake Forest Baptist Medical Center. Patients with inducible ischemia, mitral stenosis, severe mitral regurgitation, prosthetic mitral valve, a history of mitral valve repair, or aortic stenosis were excluded. Sixty-five subjects met our criteria and were included in this study. Forty-five subjects who underwent exercise echocardiography during the same period with normal EF (> 60%) and preserved age- and gender-predicted exercise capacity were studied as the control group. They underwent exercise echocardiography to evaluate chest pain but were found to have no inducible ischemia. These subjects were similar in age (61±9 years) and gender (male 31%) to the patients with reduced EFs.

Echocardiography

Patients underwent symptom-limited treadmill exercise stress echocardiography using the modified Bruce incremental exercise protocol.^9,10 The predicted exercise capacity based on age and gender in METS was calculated according to a previous report.¹¹ Exercise capacity of the patients in METs from the stage reached was compared with the predicted exercise capacity. Preserved exercise capacity was defined as ≥ 100% of the predicted exercise capacity based on age and gender.¹¹

A complete 2D Doppler echocardiogram was performed prior to exercise using an iE33 ultrasound system with a multiple frequency transducer (Philips Medical Systems, Andover, MA).¹² An experienced cardiologist (TO) measured the images without knowledge of clinical or exercise data. The grade of mitral regurgitation was evaluated qualitatively (none, trace, mild, moderate, and severe). LV end-systolic and end-diastolic volumes, stroke volume (SV) and EF were calculated with a modified Simpson method.¹² Transmitral and tissue Doppler parameters were measured from the apical 4-chamber view using standard ASE criteria.¹³ Doppler gain and filter settings are optimized to facilitate the clearest demarcation of velocity profiles.¹⁴ Mitral annular peak velocities at early (e’) and late (a’) diastole and systole (s’) were measured using pulse wave tissue Doppler as the mean of septal and lateral values. Diastolic function was assessed as: grades 1 (impaired relaxation), 2 (pseudo-normal filling), and 3 (restrictive filling) using the e’ and E/e’ values according to ASE/EAE guidelines.¹³ LV chamber stiffness (K_LV) was calculated using mitral inflow deceleration time as previously reported.¹⁵

$$KLV={\left(\frac{tdec-0.02s}{0.07}\right)}^{-2}mmHg∕mL$$

The LV mass was calculated using 2D parameters.¹⁶ To evaluate right ventricular function, tricuspid annular plane systolic excursion (TAPSE) was measured using the 2D frames of the apical 4-chamber view at end-diastole and end-systole.¹⁷

Hemodynamics

Heart rate and non-invasively measured systolic and diastolic blood pressure (BP) were recorded just before exercise and during exercise. Systemic vascular resistance was estimated by the formula:

$$\text{Systemic vascular resistance}=\text{mean BP}\cdot 80∕\text{cardiac output}(\mathrm{dyne}\cdot \sec \cdot {\mathrm{cm}}^{-5})$$

where mean BP was calculated as diastolic pressure plus one third of brachial pulse pressure, and the cardiac output was calculated as SV times the heart rate.

The ratio of SV to brachial pulse pressure was used as an indirect measure of total systemic arterial compliance which is indicative of a pulsatile component of LV afterload.¹⁸

$$\text{Systemic arterial compliance}=\mathrm{SV}∕\text{pulse pressure}(\mathrm{mL}∕\mathrm{mmHg})$$

Effective aortic elastance, a measure of the pulsatile and non-pulsatile LV afterload, was estimated by the formula:¹⁹

$$\text{Aortic effective elastance}=\text{end-systolic BP}∕\mathrm{SV}(\mathrm{mmHg}∕\mathrm{mL})$$

where end-systolic BP was estimated as 0.9 · systolic BP.

Statistics

Mean values of numerical variables and their standard deviations were compared between groups using an unpaired t-test or ANOVA with Tukey's post-hoc test as appropriate, while categorical values between groups are analyzed using a Fisher's exact test or chi-square test, as appropriate. A probability value (p) < 0.05 was accepted as significant. Logistic regression analyses were performed to predict preserved exercise tolerance. For multivariate logistic regression models, we selected the variables which were significant on univariate analyses. Because there were only 16 patients with preserved exercise tolerance, we constructed 2 multivariate models to avoid over-fitting; model 1 included E/e’ and the clinical variables (diuretics usage, heart rate at rest); model 2 included E/e’ and the echo variables (left atrial diameter and EF during peak exercise). Goodness-of-fit of the model was evaluated using the Nagelkerke's R² value. Because there was a strong correlation between E wave velocity and E/e’, only E/e’ was included into the final model. Since only 59% of the patients had the B-type natriuretic peptide (BNP) value, we did not include BNP data in the final models. A receiver operating characteristic analysis was performed to evaluate E/e’ to diagnose patients with preserved exercise capacity. Sensitivity analyses were performed by changing the definition of preserved exercise capacity to ≥ 90% and ≥ 80% of normal predicted capacity.

RESULTS

Among 65 patients with reduced EFs, 16 had preserved exercise capacity (Figure 1A), and the remaining 49 had reduced exercise capacity (Figure 1B). Table 1 summarizes the baseline clinical demographics of these groups and control subjects. The patients with EF < 35% with preserved exercise capacity were more likely to have coronary artery disease (94% vs 62%, p = 0.01), less likely to be current smokers (6% vs 37%, p = 0.03), less likely to be receiving diuretics (13% vs 62%, p < 0.01), and had lower baseline BNP [median 68 and interquartile range (46-256) pg/mL vs 436 (115-939) pg/mL, p < 0.01].

Figure 1.

(A) A patient with a reduced ejection fraction and greater-than-predicted exercise capacity. Forty-year-old male with an ejection fraction of 23% and exercise time of 16 minutes which was 170% of the predicted value. He had a relatively preserved diastolic function with an E/e’ value of 5.2.

(B) A patient with a reduced ejection fraction and reduced exercise capacity. Forty-nine-year-old female with an ejection fraction of 20% and exercise time of 32 seconds which was 26% of the predicted value. She had an elevated E/e’ value of 17.5.

Table 1.

Clinical Characteristics.

	Control (n = 45)	Preserved Ex Capacity (n = 16)	Reduced Ex Capacity (n = 49)	p *
Clinical background
Age, y	61±9	62±10	60±13	0.74
Male gender	31 (69)	11 (69)	38 (78)	0.60
Diabetes mellitus	4 (9)	2 (13)	12 (24)	0.11
Height, cm	177±10	174±9	176±9	0.75
Weight, kg	86±16	79±17	84±19	0.44
BMI, kg/m2	28±5	27±5	27±5	0.87
BSA, m2	2.0±0.2	2.0±0.2	2.0±0.3	0.67
Hypertension	29 (64)	8 (50)	31 (63)	0.57
Smoking	9 (20)	1 (6)	18 (37)	0.03
ICD	0 (0)	8 (50)‡	14 (29)‡	< 0.01
Pulmonary disease	0 (0)	2 (13)	10 (20)‡	0.01
Peripheral artery disease	0 (0)	0 (0)	1 (2)	0.53
Orthopedic disease	0 (0)	2 (13)	0 (0)	< 0.01
Cerebrovascular disease	3 (7)	1 (6)	0 (0)	0.19
Electrocardiogram
Atrial fibrillation	0 (0)	1 (6)	2 (4)	0.31
Left bundle branch block	0 (0)	0 (0)	7 (14)‡	0.01
Paced rhythm	0 (0)	4 (25)‡	5 (10)	0.01
Laboratory data
BNP, pg/mL	24 (5.3-90)	68 (46-256)§	612 (141-1029)‡	< 0.01†
Hemoglobin, g/dL	14.2±1.4	13.4±1.5	13.2±1.8‡	0.04
Creatinine, mg/dL	1.0±0.2	1.0±0.2	1.2±1.0	0.23
Drugs
ACE-I/ARB	15 (33)	15 (94)‡	40 (82)‡	< 0.01
Aldosterone blocker	0 (0)	2 (13)	10 (20)‡	0.01
Beta blocker	15 (33)	13 (81)‡	40 (82)‡	< 0.01
Calcium channel blocker	3 (7)	5 (31)	6 (12)	0.04
Digitalis	1 (2)	3 (19)	5 (10)	0.09
Diuretics	13 (29)	2 (13)§	31 (63)‡	< 0.01
Nitrate	2 (4)	4 (25)	6 (12)	0.07

Values are mean±SD, median (interquartile range) or n (%), as appropriate

^*ANOVA

^†compared using the logarithmically transformed value

^‡p < 0.05 vs. control

^§p < 0.05 vs. reduced Ex capacity.

Ex, exercise; N, number; SD, standard deviation; ICD, implantable cardioverter-defibrillator; BNP, b-type natriuretic peptide; ACE-I, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker.

The 2D and Doppler echocardiographic measures are shown in Tables 2 and 3. Although both groups had reduced EF, patients with preserved exercise tolerance had slightly better EF than patients with reduced exercise capacity (30±3% vs 27±5%, p = 0.01) (Table 3). Measures of LV diastolic function were significantly different. The patients with preserved exercise capacity had lower E-wave velocity (63±22 cm/s vs 90±29 cm/s, p < 0.01) and included fewer patients with grade 3 diastolic dysfunction (19% vs 57%, p = 0.01). Both e’ and s’ were reduced in patients with reduced EF and preserved exercise capacity compared with the control subjects (Figure 2). However, among the patients with reduced EF, E/e’ was significantly lower in the preserved exercise capacity group compared to patients with reduced exercise capacity (10.2±3.5 cm/s vs 16.5±8.2 cm/s, p < 0.01) and was similar to the normal control subjects (Table 2 and Figure 2). Prevalence of moderate mitral regurgitation was higher in the patients with reduced exercise capacity; however, even when we analyzed only the patients with no or trace MR, E/e’ was significantly lower in the preserved exercise capacity group compared to patients with reduced exercise capacity (10.0±2.6 cm/s vs 15.9±9.4 cm/s, p < 0.01). In addition, the patients with preserved exercise tolerance had smaller left atrial diameters (41±6 mm vs 46±7 mm, p = 0.01). Right ventricular function evaluated as TAPSE was similarly reduced in the preserved and reduced exercise capacity group (12.3±4.5 mm, 12.5±4.4 mm, respectively) compared with the control subjects (20.4±4.5 mm) (Table 2).

Table 2.

Echocardiographic Characteristics.

	Control (n = 45)	Preserved Ex Capacity (n = 16)	Reduced Ex Capacity (n = 49)	p *
Interventricular septal wall thickness, mm	10.1±1.9	8.8±2.0	9.7±2.7	0.16
Left ventricular end-diastolic diameter, mm	45±5	58±10‡	59±8‡	< 0.01
Posterior wall thickness, mm	9.4±1.7	9.8±1.2	10.3±2.0‡	0.04
Left ventricular end-systolic diameter, mm	30±5	50±8‡	49±10‡	< 0.01
Left atrial diameter, mm	37±6	41±6§	46±7‡	< 0.01
Left ventricular mass, g	148±34	213±51‡	239±54‡	< 0.01
Left ventricular mass index, g/m2	72±17	116±27‡	122±27‡	< 0.01
Mitral regurgitation				< 0.01†
None-trace	45 (100)	13 (81)	26 (53)
Mild	0 (0)	0 (0)	3 (6)
Moderate	0 (0)	3 (19)	20 (41)
E, cm/s	80±19	63±22‡,§	91±30	< 0.01
A, cm/s	75±23	66±21	67±29	0.23
E/A	1.13±0.32	1.17±0.92	1.71±1.19‡	0.01
Deceleration time, ms	237±49	200±58	192±56‡	< 0.01
KLV, mmHg/mL	0.12±0.06	0.28±0.46‡	0.22±0.12	0.01
e′, cm/s	8.4±2.2	6.3±1.8‡	6.2±2.1‡	< 0.01
a′, cm/s	9.9±2.0	7.4±2.7‡	6.7±2.6‡	< 0.01
s′, cm/s	8.3±2.0	5.6±1.0‡	5.2±1.0‡	< 0.01
E/e′, cm/s	10.0±2.9	10.2±3.5§	16.5±8.4‡	< 0.01
Diastolic dysfunction grade				< 0.01†
Normal	11 (24)	0 (0)§	1 (2)‡
Grade 1, impaired relaxation	5 (11)	6 (38)	3 (6)
Grade 2, pseudo-normal filling	21 (47)	7 (44)	18 (37)
Grade 3, restrictive filling	8 (18)	3 (19)	27 (55)
TAPSE, mm	20.4±4.5	12.3±4.5‡	12.5±4.4‡	< 0.01

Values are mean±SD or n (%) as appropriate

^*ANOVA unless otherwise indicate

^†chi-square test

^‡p < 0.05 vs. control

^§∥, p < 0.05 vs. reduced Ex capacity.

Ex, exercise; K_LV, left ventricular chamber stiffness; N, number; SD, standard deviation; TAPSE, tricuspid annular plane systolic excursion.

Table 3.

Hemodynamic Variables During Exercise Test.

	Control (n = 45)	Preserved Ex Capacity (n = 16)	Reduced Ex Capacity (n = 49)	p *
Exercise duration, s	516±122	496±162§	261±115‡	< 0.01
Maximum exercise capacity, METs	10.9±2.4	10.0±3.4§	5.9±2.0‡	< 0.01
%normal exercise capacity, %	130±23	117±18§	68±19‡	< 0.01
Heart rate, bpm
At rest	80±14	67±11‡,§	84±15	< 0.01
Peak exercise	153±17†	126±20†,‡	132±22†,‡	< 0.01
%predicted heart rate, %	96±10	80±11‡	83±13‡	< 0.01
Δ	73±18	60±19	49±22‡	< 0.01
Systolic blood pressure, mmHg
At rest	142±20	133±18	136±19	0.14
Peak exercise	175±19†	158±22†,‡	161±21†,‡	< 0.01
Δ	32±14	26±17	24±19	0.06
Diastolic blood pressure, mmHg
At rest	79±9	76±10	82±10	0.06
Peak exercise	81±11	81±10†	84±13	0.35
Δ	2±11	5±8	2±11	0.65
Left ventricular end-diastolic volume, mL
At rest	93±21	161±64‡	157±45‡	< 0.01
Peak exercise	95±20	162±57‡	163±54‡	< 0.01
Δ	2±13	1±38	5±43	0.88
Left ventricular end-diastolic volume index, mL/m2
At rest	46±8	83±31‡	79±24‡	< 0.01
Peak exercise	47±8	84±22‡	83±25‡	< 0.01
Δ	2±6	1±16	3±21	0.92
Left ventricular end-systolic volume, mL
At rest	36±12	112±44‡	113±36‡	< 0.01
Peak exercise	26±10†	101±39‡	113±43‡	< 0.01
Δ	−10±8	−10±29	−2±35	0.27
Left ventricular end-systolic volume index, mL/m2
At rest	18±4	58±21‡	57±19‡	< 0.01
Peak exercise	12±4	54±14‡	58±20‡	< 0.01
Δ	−5±3	−3±12	0±17	0.34
Left ventricular stroke volume, mL
At rest	56±13	49±21	44±13‡	< 0.01
Peak exercise	69±15†	60±24†	52±22†,‡	< 0.01
Δ	12±10	12±20	8±19	0.39
Left ventricular stroke volume index, mL/m2
At rest	28±6	25±11	22±7‡	< 0.01
Peak exercise	35±6	30±12	26±10‡	< 0.01
Δ	7±5	5±11	4±10	0.38
Ejection fraction, %
At rest	61±7	30±3‡	28±5‡	< 0.01
Peak exercise	73±7†	37±10†,‡	32±10†,‡	< 0.01
Δ	12±6	7±10	4±9‡	< 0.01
Cardiac output, L/min
At rest	4.5±1.3	3.3±1.6‡	3.7±1.5‡	< 0.01
Peak exercise	10.6±2.5†	7.8±3.6†,‡	6.8±3.1†,‡	< 0.01
Δ	6.0±1.9	4.5±3.1	3.2±2.8‡	< 0.01
Cardiac index, L/min/m2
At rest	2.2±0.6	1.6±0.7‡	1.9±0.8	0.02
Peak exercise	5.2±1.1	3.9±1.9‡	3.3±1.4‡	< 0.01
Δ	3.0±1	2.3±1.6	1.5±1.3‡	< 0.01
Systemic vascular resistance, dyne·sec·cm−5
At rest	1917±541	2775±1179‡	2471±1029‡	< 0.01
Peak exercise	904±244†	1465±1030†,‡	1594±860†,‡	< 0.01
Δ	−1002±452	−1309±1366	−915±830	0.25
Total aortic compliance, mL/mmHg
At rest	0.94±0.31	0.90±0.41	0.88±0.42	0.79
Peak exercise	0.75±0.22†	0.78±0.30	0.70±0.30†	0.51
Δ	−0.19±0.26	−0.12±0.27	−0.19±0.38	0.72
Aortic effective elastance, mmHg/mL
At rest	2.37±0.53	2.84±1.17	3.07±1.13‡	< 0.01
Peak exercise	2.39±0.56	2.90±1.67	3.30±1.52‡	< 0.01
Δ	0.03±0.46	0.06±1.72	0.23±1.26	0.67

Values are mean±SD

^*ANOVA

^†p < 0.05 vs. at rest

^‡p < 0.05 vs. control by Tukey's post-hoc test

^§p < 0.05 vs. reduced Ex capacity by Tukey's post-hoc test.

Ex, exercise; N, number; SD, standard deviation; Δ, value during peak exercise - value at rest.

Figure 2.

Distribution of e’ (A), s’ (B) and E/e’ (C) by groups. Lines show the mean and 95% confidence interval. *, p < 0.05 by Tukey's post-hoc test.

Patients with preserved exercise tolerance had a lower resting heart rate than both the control subjects and patients with reduced exercise capacity (Table 3). Compared with controls, both groups with reduced EF and preserved and reduced exercise capacity had comparably lower peak heart rate, peak systolic blood pressure, resting and peak EF, cardiac output, larger resting and peak end-diastolic and systolic volumes, and systemic vascular resistance. Patients with reduced exercise tolerance had lower resting and peak stroke volume and higher aortic effective elastance values than that of controls.

Echocardiographic and Clinical Predictors of Preserved Exercise Capacity

Univariate analyses identified several variables that were significantly associated with preserved exercise tolerance, including a lower E/e’ ratio, smaller left atrial diameter as well as nonuse of diuretic, and lower resting heart rate (Table 4). Neither EF at rest or moderate mitral regurgitation at rest predicted patients with preserved exercise tolerance in the univariate analyses (Table 4).

Table 4.

Logistic Regression Analysis to Predict Preserved Exercise Tolerance in patients with a reduced ejection fraction.

	Univariate analysis			Multivariate model 1*			Multivariate model 2†
	HR	95% CI	p	HR	95% CI	p	HR	95% CI	p
Diuretics	0.08	(0.02-0.41)	< 0.01	0.09	(0.01-0.60)	0.01
Left atrial diameter, per 1 mm	0.90	(0.82-0.99)	0.02				0.92	(0.83-1.01)	0.08
E/e′, per 1 cm/s	0.80	(0.67-0.95)	0.01	0.69	(0.51-0.93)	0.02	0.80	(0.66-0.96)	0.02
Heart rate at rest, per 1 bpm	0.89	(0.83-0.95)	< 0.01	0.87	(0.79-0.96)	0.01
Ejection fraction at rest, per 1%	1.13	(0.98-1.31)	0.09
Ejection fraction during peak exercise, per 1%	1.06	(1.00-1.12)	0.07				1.03	(0.97-1.10)	0.33
Moderate mitral regurgitation at rest	0.34	(0.08-1.33)	0.12

HR, hazard ratio; CI, confidence interval

^*R² value of the model was 0.69

^†R² value of the model was 0.37.

Lower E/e’ independently predicted preserved exercise capacity in a multivariate model including clinical variables (diuretics usage and resting heart rate) (Table 4). The lower E/e’ was also the only predictor for preserved exercise capacity in a multivariate model including the echocardiographic variables (left atrial diameter and EF during peak exercise).

Receiver operator characteristics curve demonstrated that E/e’ identified preserved exercise capacity with optimal cut-off values of E/e’ < 10.8 (sensitivity = 75%, specificity = 76%) (Figure 3). We performed sensitivity analyses by altering the definition of preserved exercise capacity to greater than or equal to 90% and 80% of normal predicted capacity. The diagnostic capability of E/e’ was unchanged with the area under the curve being 0.77 (p < 0.001) and 0.69 (p = 0.007), respectively. The resting E/e’ ratio was negatively correlated with relative exercise capacity in patients with reduced EFs and preserved and reduced exercise capacity (Figure 4).

Figure 3.

Receiver operating characteristics curve for E/e’ to detect preserved exercise capacity.

Figure 4.

Relationship between E/e’ at rest and relative exercise capacity.

DISCUSSION

Reduced exercise capacity is an important manifestation of HFrEF. Since patients with HFrEF have clear evidence of systolic dysfunction, the clinical manifestations have been attributed to the systolic LV dysfunction. However, some patients with systolic dysfunction apparent as a markedly reduced EF may not have symptomatic HF and have preserved exercise capacity (i.e. Stage B HF).^5-8 Consistent with this concept, we found that nearly a quarter of ambulatory patients with EF < 35% undergoing exercise testing had an exercise capacity that exceeded the predicted normal for their age and gender. These patients with preserved exercise tolerance despite EF < 35% had better diastolic performance than those with a reduction in their exercise capacity. This was apparent as less abnormal filling pattern, smaller left atrium, and lower E/e’ than those with the expected reduction in exercise capacity. In fact, in patients with reduced EF and preserved exercise tolerance, E/e’ was similar to subjects with normal EF and exercise tolerance. The lower E/e’ is associated with lower LV filling pressure as well as preserved LV diastolic function and suction.^20,21 These findings indicate the important contribution of diastolic function to determining exercise capacity in patients with reduced EF. Our results are consistent with the previous studies showing the relationship between diastolic function and exercise capacity in patients with a reduced EF^22-24 and preserved EF.^25,26 However, the previous studies focused on the influence of abnormal diastolic function on the impaired exercise capacities in the patients with reduced EF.^22-24 Since multiple factors contribute to exercise intolerance in HF and all patients with reduced EF have systolic dysfunction, there has continued to be uncertainty about the importance of diastolic function in HFrEF.²⁷ Our study advances our understanding of this issue, by finding that adequate diastolic function is necessary for a preserved exercise capacity despite reduced EF and systolic function. This suggests that improvement in diastolic function may be an important therapeutic target in HFrEF.

It is of note that although patients with preserved exercise capacity had lower E/e’, they had comparably reduced e’ and increased K_LV to those with reduced exercise capacity (Table 2, Figure 2). Exercise capacity depends on the ability to decrease LV early diastolic pressures in response to stress (suction), allowing an increase in LV stroke volume without increase in left atrial pressure to abnormal levels.²⁰ LV suction is determined by LV ejection, relaxation, elastic recoil, and the fluid dynamics of LV inflow.²⁰ The lack of an increase of E/e’ and LA size with preserved exercise tolerance indicates that resting LA pressures were not elevated. A reduced e’ indicates impaired LV relaxation, one of the determinants of suction.²⁸. However, an elevation of E/e’ does not result from increased left atrial pressure alone, but can also indicate a failure in LV suction to increase in response to stress.²⁹ This may be why E/e’, not e’ was different between patients with preserved and reduced exercise tolerance.

Prevalence of moderate mitral regurgitation was higher in the patients with reduced exercise capacity, which might confound the results (Table 2). However, E/e’ was significantly lower in the preserved exercise capacity group compared to patients with reduced exercise capacity, when we analyzed only the patients with no or trace mitral regurgitation.

The patients who had preserved exercise capacity had slightly higher resting EF and greater exercise augmentation of EF (Table 2). However, these parameters were not significant in either univariate or multivariate analysis as predictors of preserved exercise capacity. Systolic function evaluated as s’ was also reduced in the patients with preserved exercise capacity comparable to those with reduced exercise tolerance. These findings are consistent with the notion that systolic function is not a strong determinant of exercise capacity in patients with a reduced EF.

Reduced exercise capacity in patients with HFrEF is multifactorial, resulting from cardiac, vascular, muscular, reflex, and other abnormalities.²⁷ The lack of a strong correlation of exercise capacity in HFrEF to EF has been attributed to the importance of non-cardiac factors, including muscle and vascular function. Our findings suggest that differences in diastolic dysfunction may contribute to the lack of correlation of EF and exercise tolerance. Furthermore, we found no difference in the resting or exercise systemic vascular resistance, aortic compliance, arterial effective elastance, or peak heart rate between patients with reduced EF with preserved or decreased exercise tolerance. Peripheral non-cardiac factors such as skeletal muscle perfusion or oxygen extraction by the active muscles may influence exercise capacity of these patients.^7,30 Although we did not measure skeletal muscle blood flow or metabolism in this study, diastolic dysfunction may interact with peripheral vascular function and/or musculoskeletal function.³⁰ Right ventricular function has prognostic impact on the patients with HF.³¹ The patients with preserved exercise tolerance in this study had reduced right ventricular function evaluated as TAPSE, which was comparable to the patients with reduced exercise tolerance (Table 2).

The patients with preserved exercise capacity had lower BNP levels and were less likely being treated with diuretics, suggesting that these patients, despite their reduced EF, were well compensated and had little or no pulmonary congestion.

Limitations

This is a retrospective, single-center study with all the associated limitations. For example, pulmonary artery pressures, exercise diastolic function, strain rate and apical rotational dynamics were not measured. In addition, the control subjects were patients undergoing testing for chest pain but found to have normal exercise tolerance and LV function and no ischemia. They were not age-and gender-matched to the low EF patients although the mean ages and gender distribution were similar. We studied only patients who could complete symptom limited exercise echocardiography. Thus we compared patients with markedly reduced EF and preserved exercise capacity to compensated patients with HFrEF with mild to moderate exercise intolerance and moderate elevations of BNP. The differences between the two groups would likely have been accentuated if we had compared patients with preserved exercise capacity to patient with more severe HFrEF.

Since we did not obtain Doppler measures at peak exercise, we cannot exclude exercise-induced mitral regurgitation as a mechanism for exercise intolerance.³² We excluded patients with inducible ischemia, however those with greater exercise capacity were more often diagnosed with coronary artery disease; a finding that should favor the low exercise capacity group.

We determined the optimal cut-off value of E/e’ <10.8 using a receiver operator characteristics curve analysis and showed a satisfactory diagnostic ability. However, this should be confirmed in an independent population.

Conclusion

Some patients with reduced EF have well-preserved exercise capacity despite clear systolic LV dysfunction. These patients have less diastolic dysfunction than patients with HFrEF who have mild to moderate exercise intolerance. This finding is consistent with the concept that diastolic function is an important contributor to exercise tolerance in patients with HFrEF.

ABBREVIATIONS

BNP

B-type natriuretic peptide

BP

blood pressure

EF

ejection fraction

HF

heart failure

HFrEF

heart failure with a reduced ejection fraction

KLV

left ventricular chamber stiffness

LV

left ventricle

SV

stroke volume

TAPSE

tricuspid annular plane systolic excursion