Alveolar Dead Space Is Augmented During Exercise in Patients With Heart Failure With Preserved Ejection Fraction

Abstract

Background

Patients with heart failure with preserved ejection fraction (HFpEF) exhibit many cardiopulmonary abnormalities that could result in $\dot{\text{V}}$/$\dot{\text{Q}}$ mismatch, manifesting as an increase in alveolar dead space (VD_alveolar) during exercise. Therefore, we tested the hypothesis that VD_alveolar would increase during exercise to a greater extent in patients with HFpEF compared with control participants.

Research Question

Do patients with HFpEF develop VD_alveolar during exercise?

Study Design and Methods

Twenty-three patients with HFpEF and 12 control participants were studied. Gas exchange (ventilation [$\dot{\text{V}}$_E], oxygen uptake [$\dot{\text{V}}$o₂], and CO₂ elimination [$\dot{\text{V}}$co₂]) and arterial blood gases were analyzed at rest, twenty watts (20W), and peak exercise. Ventilatory efficiency (evaluated as the $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope) also was measured from rest to 20W in patients with HFpEF. The physiologic dead space (VD_physiologic) to tidal volume (VT) ratio (VD/VT) was calculated using the Enghoff modification of the Bohr equation. VD_alveolar was calculated as: (VD / VT × VT) – anatomic dead space. Data were analyzed between groups (patients with HFpEF vs control participants) across conditions (rest, 20W, and peak exercise) using a two-way repeated measures analysis of variance and relationships were analyzed using Pearson correlation coefficient.

Results

VD_alveolar increased from rest (0.12 ± 0.07 L/breath) to 20W (0.22 ± 0.08 L/breath) in patients with HFpEF (P < .01), whereas VD_alveolar did not change from rest (0.01 ± 0.06 L/breath) to 20W (0.06 ± 0.13 L/breath) in control participants (P = .19). Thereafter, VD_alveolar increased from 20W to peak exercise in patients with HFpEF (0.37 ± 0.16 L/breath; P < .01 vs 20W) and control participants (0.19 ± 0.17 L/breath; P = .03 vs 20W). VD_alveolar was greater in patients with HFpEF compared with control participants at rest, 20W, and peak exercise (main effect for group, P < .01). Moreover, the increase in VD_alveolar correlated with the $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope (r = 0.69; P < .01), which was correlated with peak $\dot{\text{V}}$o_2peak (r = 0.46; P < .01) in patients with HFpEF.

Interpretation

These data suggest that the increase in $\dot{\text{V}}$/$\dot{\text{Q}}$ mismatch may be explained by increases in VD_alveolar and that increases in VD_alveolar worsens ventilatory efficiency, which seems to be a key contributor to exercise intolerance in patients with HFpEF.

Take-home Points.

Study Question: Do patients with heart failure with preserved ejection fraction (HFpEF) develop alveolar dead space (VD_alveolar) during exercise?

Results: VD_alveolar was elevated across all conditions (rest, 20W, and peak exercise) and increased during exercise to a greater extent in patients with HFpEF compared with control participants.

Interpretation: Lung regions where the amount of ventilation is high relative to the amount of pulmonary blood flow may develop in patients with HFpEF, resulting in a greater relative distribution of high $\dot{\text{V}}/\dot{\text{Q}}$ lung regions (or VD_alveolar, or both) during exercise.

FOR EDITORIAL COMMENT, SEE PAGE 1233

Heart failure with preserved ejection fraction (HFpEF) is a major public health concern.^1,2 The pathophysiologic features of HFpEF are complex, and this disorder has been characterized as a heterogenous syndrome that exhibits numerous cardiopulmonary abnormalities, including elevated cardiac filling pressures,^{3, 4, 5} pulmonary edema,^6,7 impaired pulmonary vascular recruitment and distension,^{8, 9, 10} and right ventricle-pulmonary artery uncoupling.^{11, 12, 13, 14, 15} Although investigations into the cardiac-related consequences of these cardiopulmonary abnormalities are ongoing, these cardiopulmonary abnormalities also could alter pulmonary blood flow distribution considerably and could contribute to $\dot{\text{V}}/\dot{\text{Q}}$ mismatch, manifesting as an increase in alveolar dead space (VD_alveolar),¹⁶ which is characterized by the development of lung regions with a high $\dot{\text{V}}/\dot{\text{Q}}$ ratio.¹⁶

$\dot{\text{V}}/\dot{\text{Q}}$ mismatch typically is estimated by calculating the physiologic dead space (VD_physiologic; the sum of anatomic dead space [VD_anatomic] and VD_alveolar) to tidal volume (VT) ratio (VD/VT).¹⁷ Studies have demonstrated that VD/VT is increased in patients with HFpEF compared with control participants.^{18, 19, 20} We also demonstrated that VD/VT increased during exercise in patients with HFpEF²¹ and to a greater extent compared with control participants,²² which we hypothesized could be the result of an increase in VD_alveolar secondary to the exacerbation of a variety of cardiopulmonary abnormalities.¹⁶ However, VT also can have a major effect on the VD/VT ratio.²³ Given that VT was smaller in patients with HFpEF compared with control participants in previous studies,^{18, 19, 20} one may argue that the smaller VT at least in part could have contributed to the increased VD/VT ratio observed in patients with HFpEF. As such, the subsequent conclusions that can be drawn from the reported data^{18, 19, 20} regarding the development of VD_alveolar as a mechanism of $\dot{\text{V}}/\dot{\text{Q}}$ mismatch during exercise in patients with HFpEF are limited.

To circumvent the confounding VT problem, it is possible to estimate VD_alveolar from VD/VT, particularly if VD/VT is calculated using the Enghoff modification of the Bohr equation¹⁷ because this equation accounts for the mechanical dead space volume of the breathing apparatus, which can influence the VD/VT ratio measurement independently.²² Next, it is recommended that VD/VT be multiplied by VT to estimate VD_physiologic.²³ This value should approximate 0.15 L regardless of the VT,²³ which theoretically constitutes the VD_anatomic in healthy individuals.²⁴ Therefore, any increase in VD_physiologic above the estimated VD_anatomic represents an estimate of the VD_alveolar.^23,25

Although the increase in VD/VT during exercise reported previously indicates worsening $\dot{\text{V}}/\dot{\text{Q}}$ mismatch,^21,22 the extent to which the increase in VD/VT is the result of an increase in VD_alveolar in patients with HFpEF is unclear. Because $\dot{\text{V}}/\dot{\text{Q}}$ mismatch is associated with exercise intolerance,^22,26,27 such information on the development of VD_alveolar as a mechanism of $\dot{\text{V}}/\dot{\text{Q}}$ mismatch could have a significant public health outcome by allowing for a better understanding of the mechanism(s) that may contribute to exercise intolerance in patients with HFpEF. This becomes especially important given that exercise intolerance is a dominant symptom of HFpEF and is associated with increased mortality in this population.¹⁶ Therefore, the purpose of the present study was to calculate VD_alveolar and to compare VD_alveolar at rest and during exercise in patients with HFpEF and healthy control participants. We hypothesized that VD_alveolar would increase during exercise in patients with HFpEF compared with control participants.

Study Design and Methods

This was a retrospective analysis of previously collected data. Although some of these data have been published elsewhere,^21,22,28 we repeat only the methods and data essential to the new findings presented herein.

Participants

We evaluated 23 patients with HFpEF who were enrolled in our larger ongoing study (ClinicalTrials.gov Identifier: NCT04068844) and 12 control participants who were part of a different study.²⁸ Patients with HFpEF were included if they had signs and symptoms of heart failure based on Framingham criteria,²⁹ an ejection fraction of ≥ 50%, and evidence of volume overload (eg, pulmonary edema) confirmed by a pulmonary capillary wedge pressure of ≥ 25 mm Hg at peak exercise or an increase in pulmonary capillary wedge pressure of ≥ 15 mm Hg from rest to peak exercise. Participants were excluded if they had a diagnosis of severe valvular disease, congenital heart disease, known restrictive or infiltrative cardiomyopathy, acute myocarditis, New York Heart Association class IV chronic heart failure or chronic heart failure that cannot be stabilized with medical therapy, a prior ejection fraction of < 50%, manifest or provocable ischemic heart disease, or significant obstructive lung disease (ie, FEV₁ of < 40% predicted). Control participants did not have a history of smoking of > 5 pack-years, asthma, cardiovascular disease, sleep disorders, or musculoskeletal abnormalities that would preclude exercise or a BMI of > 30 kg/m². Before all testing, written and informed consent was obtained. The experimental procedures were reviewed and approved by the UT Southwestern Institutional Review Board (Identifier: STU2019-0617).

Study Design and Methods

Patients with HFpEF visited the laboratory on two separate occasions. During the first visit, patients with HFpEF underwent preparticipation health screening and performed pulmonary function testing according to American Thoracic Society/European Respiratory Society guidelines³⁰ and a maximum cardiopulmonary exercise test (CPET) to determine participant eligibility. During the second visit, patients with HFpEF underwent pulmonary artery and radial artery catheterizations and performed a 6-min constant-load cycling test and a maximum incremental cycling test on an upright cycle ergometer (Lode BV), as described previously.²² Control participants visited the laboratory on two separate occasions. During the first visit, control participants underwent preparticipation health screening and performed pulmonary function testing according to American Thoracic Society/European Respiratory Society guidelines³⁰ and a maximum CPET to determine participant eligibility. During the second visit, control participants underwent radial artery catheterization and performed a maximum incremental cycling test on an upright cycle ergometer (MedGraphics, model CPE 2000), as described previously.²²

Catheterization Protocol

Catheterizations for patients with HFpEF and control participants were performed as described previously.²² Briefly, patients with HFpEF underwent Swan-Ganz catheter placement in the pulmonary artery via brachial or antecubital vein access under fluoroscopic guidance, which was used to measure pulmonary artery pressure, pulmonary capillary wedge pressure, and central blood temperature at rest, during the final minute of constant-load exercise, and at peak exercise. Control participants did not undergo Swan-Ganz catheter placement in the pulmonary artery. Both patients with HFpEF and control participants underwent radial artery catheterization using a modified Seldinger technique. For patients with HFpEF, arterial blood gases were measured at rest, during the final minute of constant-load exercise, and at peak exercise. For control participants, arterial blood gases were measured at rest, during each work rate increment, and at peak exercise. All blood gas samples were withdrawn over a 10-s period to reduce the fluctuations of blood gas tensions over a given respiratory cycle during the same time frame in which pulmonary gas exchange determinations were made. All blood gas samples were temperature corrected and analyzed for Paco₂, Pao₂, and hemoglobin oxygen saturation (ABL90 Flex; Radiometer).

Measurements Obtained From Patients With HFpEF

Heart rate and rhythm were monitored continuously using 12-lead electrocardiography. BP was monitored via arterial waveform tracings. Invasive central hemodynamic measurements and radial arterial blood gases were measured as described above. Gas exchange, including ventilation ($\dot{\text{V}}$_E), oxygen uptake ($\dot{\text{V}}$o₂), and CO₂ elimination ($\dot{\text{V}}$co₂), was measured using a customized breath-by-breath measurement system (Beck Integrative Physiological System; KCBeck, Physiological Consulting) integrated with a mass spectrometer (Perkin-Elmer, model 1100). These measurements were obtained at rest, during the final minute of constant-load cycling (performed at twenty watts, 20W), and at peak exercise. During the maximum incremental cycling test, a maximum effort was determined by observing: (1) a plateau in $\dot{\text{V}}$o₂ of < 150 mL/min, (2) the inability to maintain a cycling cadence of > 55 rpm for a duration of 5 s, (3) a respiratory exchange ratio of > 1.05, or a combination thereof.

Measurements Obtained From Control Participants

Heart rate and rhythm were monitored continuously using 12-lead electrocardiography. BP was monitored via an automated system. Radial arterial blood gases were measured as described above, and gas exchange, including $\dot{\text{V}}$_E, $\dot{\text{V}}$o₂, and $\dot{\text{V}}$co₂, was measured using a customized breath-by-breath measurement system that was computerized (NEC 486DX) and integrated with a mass spectrometer (Perkin-Elmer, model 1100). These measurements were obtained at rest, the 20W stage (to facilitate between-group comparisons at a given absolute work rate), and at peak exercise. During the maximum incremental cycling test, a maximum effort was determined by observing: (1) a plateau in $\dot{\text{V}}$o₂ of < 150 mL/min, (2) the inability to maintain a cycling cadence of > 55 rpm for a duration of 5 s, (3) a respiratory exchange ratio of > 1.05, or a combination thereof.

Data Analysis

VD/VT was calculated using the Enghoff modification of the Bohr equation¹⁷: VD/VT = [(Paco₂ – Peco₂) / Paco₂] – (mechanical dead space / VT). Partial pressure of mixed expired CO₂ (Peco₂) was calculated as: 863 × $\dot{\text{V}}$co₂ (standard temperature pressure dry, STPD) / $\dot{\text{V}}$_E(body temperature pressure saturated, BTPS) in both groups. Mechanical dead space was estimated to equal 0.18 L and 0.14 L for patients with HFpEF and control participants, respectively.²² VD_alveolar was calculated as: VD_alveolar = (VD/VT × VT) – VD_anatomic.²³ VD_anatomic was estimated as 2 mL/kg body mass in those with normal BMI and VD_anatomic was estimated as 2 mL/kg lean body mass in those with BMI of ≥ 30 kg/m² as recommended,²³ because VD_anatomic varies with body size.²³ In patients with HFpEF, we also calculated the slope of the relationship between the rest-to-20W change in $\dot{\text{V}}$_E and the rest-to-20W change in $\dot{\text{V}}$co₂ (ie, the $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope),^31,32 which reflects ventilatory efficiency. Cardiac output was calculated via the direct Fick formula: cardiac output = $\dot{\text{V}}$o₂ / arteriovenous oxygen difference.

Statistical Analysis

Data were analyzed using SPSS version 22.0 software. Differences in participant characteristics, cardiorespiratory parameters, and arterial blood gas data (rest, 20W, and peak exercise) were compared using independent t tests, and differences in VD_alveolar were compared between groups (patients with HFpEF vs control participants) across conditions (rest, 20W, and peak exercise) using a two-way repeated measures analysis of variance. Where significant interactions were detected, post hoc comparisons using Bonferroni adjustments were performed. Assumptions of normality and homogeneity of variances were assessed using the Shapiro-Wilk test (in addition to an examination of Q-Q plots) and the Levene’s test, respectively. The Shapiro-Wilk test was nonsignificant for independent t tests and the two-way repeated measures analysis of variance when examining all three conditions separately, indicating that the data did not deviate significantly from a normal distribution. The Levene’s test was nonsignificant for the two-way repeated measures analysis of variance, indicating that the variance was not significantly heterogenous. The Levene’s test also allowed us to report the appropriate t test results (ie, the P value for when homogeneity of variances is or is not assumed). Furthermore, relationships between variables were assessed using Pearson correlation coefficient analyses and described as small (0.10 < r < 0.29), medium (0.30 < r < 0.49), and large (r > 0.50).³³ Statistical significance was defined as P < .05. All data are presented as mean ± SD.

Results

Participant Characteristics

Participant characteristics and pulmonary function data are displayed in Table 1. All patients with HFpEF were not smoking at the time of the study. Four patients with HFpEF had a history of smoking (approximately a 16 pack-year history on average). Of the patients with HFpEF studied, 15 had a prior diagnosis of obstructive sleep apnea. Participants were similar in age and height, but patients with HFpEF overall were heavier and demonstrated greater BMI compared with control participants. Also, FVC, FEV₁, and TLC were lower in patients with HFpEF compared with control participants. Despite these differences, pulmonary function was within normal limits for most patients based on lower limits of normal from 95% CIs. One patient with HFpEF demonstrated moderate obstructive lung disease and one patient had undergone a prior lobectomy. Although diffusion capacity of the lung for carbon monoxide (Dlco) was lower in patients with HFpEF compared with control participants, Dlco normalized to alveolar volume was similar between groups. Cardiorespiratory responses are shown in Table 2, and hemodynamic and arterial blood gas responses are shown in Table 3.

Table 1.

Participant Characteristics

Characteristic	Control Participants	Patients With HFpEF
Demographics
Age, y	70 ± 3	71 ± 6
Sex, male/female, no.	7/5	9/14
Height, m	1.72 ± 0.1	1.70 ± 0.1
Weight, kg	72.3 ± 11	107 ± 15a
BMI, kg/m2	27.4 ± 1.9	37.9 ± 6.8a
Pulmonary function
FVC, % predicted	115 ± 17	86 ± 16a
FEV1, % predicted	104 ± 13	85 ± 17a
FEV1 to FVC, %	71 ± 5	75 ± 9
TLC, % predicted	113 ± 13	93 ± 15a
Dlco, % predicted	108 ± 14	69 ± 19a
Dlco normalized to VA, % predicted	109 ± 16	108 ± 25

Data are presented as mean ± SD, unless otherwise indicated. Dlco = diffusing capacity of the lung for carbon monoxide; HFpEF = heart failure with preserved ejection fraction; TLC = total lung capacity; V_A = alveolar volume.

^aP < .01.

Table 2.

Cardiorespiratory Parameters at Rest, During 20W, and at Peak Exercise

Variable	Control Participants	Patients With HFpEF
Rest
$\dot{\text{V}}$E, L/min	11.6 ± 4.0	11.9 ± 2.5
$\dot{\text{V}}$o2, L/min	0.34 ± 0.08	0.31 ± 0.07
$\dot{\text{V}}$co2, L/min	0.28 ± 0.06	0.24 ± 0.06
$\dot{\text{V}}$E/$\dot{\text{V}}$co2 ratio	40.8 ± 9.6	50.2 ± 7.4a
RER	0.83 ± 0.09	0.79 ± 0.05
20W
WR, % peak	19 ± 7	31 ± 11a
$\dot{\text{V}}$E, L/min	19.7 ± 3.0	30.1 ± 4.4a
$\dot{\text{V}}$o2, L/min	0.65 ± 0.08	0.87 ± 0.16a
$\dot{\text{V}}$co2, L/min	0.54 ± 0.07	0.75 ± 0.14a
$\dot{\text{V}}$E/$\dot{\text{V}}$co2 ratio	31.0 ± 5.5	40.6 ± 5.3a
RER	0.84 ± 0.08	0.87 ± 0.05
Peak exercise
WR, W	120 ± 40	76 ± 34a
WR, % predicted	120 ± 27	82 ± 25a
$\dot{\text{V}}$E, L/min	82.9 ± 24.4	58.6 ± 18.9a
$\dot{\text{V}}$E, % predicted MVV	74.7 ± 11.4	72.1 ± 8.7
$\dot{\text{V}}$o2, L/min	1.70 ± 0.51	1.27 ± 0.39b
$\dot{\text{V}}$o2, % predicted	106 ± 21	73 ± 17a
$\dot{\text{V}}$o2, mL/min/kg	23.35 ± 5.45	11.80 ± 3.23a
$\dot{\text{V}}$co2, L/min	2.11 ± 0.58	1.39 ± 0.53a
$\dot{\text{V}}$E/$\dot{\text{V}}$co2 ratio	39.6 ± 5.0	43.8 ± 7.6
RER	1.26 ± 0.09	1.09 ± 0.12b

Data are mean ± SD. 20W = 20 watts; HFpEF = heart failure with preserved ejection fraction; MVV = maximal voluntary ventilation; RER = respiratory exchange ratio; $\dot{\text{V}}$_E = minute ventilation; $\dot{\text{V}}$o₂ = oxygen consumption; $\dot{\text{V}}$co₂ = CO₂ elimination; WR = work rate.

^aP < .01.

^bP < .05.

Table 3.

Hemodynamic and Arterial Blood Gas Data at Rest, During 20W, and at Peak Exercise

Variable	Control Participants	Patients With HFpEF
Rest
HR, beats/min	75 ± 11	74 ± 16
Qc, L/min	—	5.2 ± 1.3
Pao2, mm Hg	99.2 ± 8.6	83.6 ± 5.7a
Paco2, mm Hg	36.2 ± 4.2	40.5 ± 4.0a
HbO2, %	96 ± 1	94 ± 2a
20W
HR, beats/min	109 ± 18	93 ± 16
Qc, L/min	—	8.9 ± 2.0
Pao2, mm Hg	99.4 ± 9.1	85.1 ± 12.9b
Paco2, mm Hg	36.5 ± 4.1	40.6 ± 4.2b
HbO2, %	97 ± 1	95 ± 3b
Peak exercise
HR, beats/min	157 ± 9	122 ± 19a
HR, % predicted maximum	99 ± 6	79 ± 12a
Qc, L/min	—	10.5 ± 2.7
Pao2, mm Hg	108.5 ± 8.1	88.3 ± 15.1a
Paco2, mm Hg	31.6 ± 2.5	38.9 ± 4.7a
HbO2, %	97 ± 1	94 ± 3b

Data are presented as mean ± SD. 20W = 20 watts; HbO₂ = hemoglobin oxygen saturation; HFpEF = heart failure with preserved ejection fraction; HR = heart rate; Qc = cardiac output; — = no data available.

^aP < .01.

^bP < .05.

VD_alveolar

VD_alveolar increased from rest to 20W in patients with HFpEF (P < .01) (Fig 1), whereas VD_alveolar did not change significantly from rest to 20W in control participants (P = .19) (Fig 1). Thus, the increase in VD_alveolar was greater in patients with HFpEF compared with control participants from rest to 20W (patients with HFpEF, 0.11 ± 0.05 L/breath; control participants, 0.04 ± 0.08 L/breath; P = .03). Thereafter, VD_alveolar increased from 20W to peak exercise in both patients with HFpEF (P < .01) (Fig 1) and control participants (P = .03) (Fig 1), the increase of which was similar between groups (patients with HFpEF, 0.13 ± 0.12 L/breath; control participants, 0.13 ± 0.16 L/breath; P = .94). Furthermore, VD_alveolar was greater in patients with HFpEF compared with control participants at rest, 20W, and peak exercise (P < .01, main effect for group) (Fig 1). Furthermore, the change in VD_alveolar from rest to 20W significantly correlated with the $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope (r = 0.69; P < .01) (Fig 2A). Finally, the $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope significantly correlated inversely with peak $\dot{\text{V}}$o_2peak (r = 0.46; P < .01) (Fig 2B).

Figure 1.

Graph showing mean ± SD values for VD_alveolar measured at rest and during exercise (20W and peak exercise) in control participants (blue circles) and patients with HFpEF (red circles). Differences in VD_alveolar measurements were compared between groups (control participants vs patients with HFpEF) across condition (rest, 20W, and peak exercise) using a two-way repeated measures analysis of variance. ∗Significantly different between groups at P < .05. ∗∗Significantly different between groups at P < .01. 20W = 20 watts; HFpEF = heart failure with preserved ejection fraction; VD_alveolar = alveolar dead space.

Figure 2.

A, B, Scatterplots showing the relationships between the change in VD_alveolar from rest to 20W and the $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope (A) and the $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope and $\dot{\text{V}}$o_2peak (B) in patients with heart failure with preserved ejection fraction. Relationships were analyzed using Pearson correlation coefficients. The dashed lines indicate linear regression. 20W = 20 watts; $\dot{\text{V}}$_E = ventilation; $\dot{\text{V}}$co₂ = CO₂ elimination; VD_alveolar = alveolar dead space; $\dot{\text{V}}$o_2peak = peak oxygen uptake; twenty watts = 20W.

Discussion

To our knowledge, this is the first study to examine VD_alveolar in patients with HFpEF. The novel findings from this study are: (1) VD_alveolar was elevated across all conditions (rest, 20W, and peak exercise) and increased during exercise to a greater extent in patients with HFpEF compared with control participants, (2) the increase in VD_alveolar correlated significantly with the $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope in patients with HFpEF, and (3) the $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope correlated significantly with $\dot{\text{V}}$O_2peak in patients with HFpEF.

Patients with HFpEF exhibit many cardiopulmonary abnormalities that could result in $\dot{\text{V}}/\dot{\text{Q}}$ mismatch, particularly during exercise when cardiopulmonary abnormalities are exacerbated and the degree of $\dot{\text{V}}/\dot{\text{Q}}$ mismatch is likely to worsen. Indeed, we previously demonstrated that $\dot{\text{V}}/\dot{\text{Q}}$ mismatch increased during exercise in patients with HFpEF (as evidenced by an increase in VD/VT),^21,22 which we hypothesized could be the result of an increase in VD_alveolar secondary to the exacerbation of a variety of cardiopulmonary abnormalities. Consistent with our hypothesis are the findings of the present study: an increase in VD_alveolar from rest to exercise in patients with HFpEF, the magnitude of which was greater than that observed in control participants. Therefore, the results of the present study extend our previous work^21,22 by providing insight into the mechanism(s) that may be responsible for the increased VD/VT ratio during exercise in patients with HFpEF. Our findings also support those of others^17,34,35 who suggest that a VD/VT ratio that increases, or fails to decrease from rest to exercise, indicates worsening $\dot{\text{V}}/\dot{\text{Q}}$ mismatch possibly because of increasing VD_alveolar. Sullivan et al³⁶ also reported similar results in patients with heart failure with reduced ejection fraction, whereby an elevated VD/VT ratio primarily was the result of an increase in VD_physiologic. Collectively, these findings,^{17,34, 35, 36} combined with that of the present study, provide strong evidence that patients with HFpEF develop profound gas exchange abnormalities during exercise, secondary to an increase in VD_alevolar.

It is well documented that an increase in VD_alevolar contributes to ventilatory inefficiency, which typically is defined as an increase in the $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope.³⁷ A growing body of evidence shows that patients with HFpEF exhibit an increased $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope when compared with control participants.^9,18 Van Iterson et al¹⁹ also demonstrated that an increased $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope can be ascribed primarily to an increased VD/VT ratio in patients with HFpEF.¹⁹ The findings of the present study extend those of Van Iterson et al¹⁹ by showing that the increase in VD_alevolar (from rest to 20W) also was related to the $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope in patients with HFpEF. After a retrospective analysis of alveolar $\dot{\text{V}}$_E, which was calculated as $\dot{\text{V}}$_Ex(1 – VD/VT),³⁶ no significant difference was found in alveolar $\dot{\text{V}}$_E between groups at 20W (patients with HFpEF, 15.98 ± 3.38 L/min; control participants, 14.68 ± 2.14 L/min; P = .25). Therefore, the increased ventilatory demand and thus the $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope in patients with HFpEF must be in response to an increase in VD_alveolar to maintain appropriate alveolar $\dot{\text{V}}$_E and blood gas homeostasis. These findings are similar to what has been observed in patients with heart failure with reduced ejection fraction in whom exercise $\dot{\text{V}}$_E is elevated, yet arterial blood gases are maintained.³⁶ Moreover, multiple studies have observed a link between ventilatory inefficiency and exercise intolerance in patients with a range of cardiopulmonary conditions.^{38, 39, 40, 41} Herein, we also showed, for the first time m that those patients who exhibited a greater magnitude of ventilatory inefficiency demonstrated a lower exercise tolerance. This was evidenced by the fact that the $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope was related inversely to $\dot{\text{V}}$o_2peak in patients with HFpEF. Although the association between the $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope and $\dot{\text{V}}$O_2peak was relatively weak, the strength of the relationship is similar to what has been observed previously in patients with heart failure with reduced ejection fraction.^42,43 Overall, the results of the present study extend that of previous studies and suggest that elevations in VD_alveolar contribute to ventilatory inefficiency in patients with HFpEF, which seems to contribute to exercise intolerance in these patients.

The increase in VD_alveolar has not been shown before in patients with HFpEF. Indeed, it is well known that VD_alveolar is characterized by the development of lung regions with a high $\dot{\text{V}}/\dot{\text{Q}}$ ratio.⁴⁴ Thus, it is conceivable that during exercise, lung regions develop where the amount of ventilation is high relative to the amount of pulmonary blood flow in patients with HFpEF, resulting in a greater relative distribution of high $\dot{\text{V}}/\dot{\text{Q}}$ lung regions (or VD_alveolar, or both).^45,46 Although the mechanism(s) governing a greater relative distribution of high $\dot{\text{V}}/\dot{\text{Q}}$ lung regions during exercise in patients with HFpEF is unclear, it is possible that the aforementioned cardiopulmonary abnormalities associated with HFpEF could alter the distribution of or limit pulmonary perfusion, or both, which ultimately could lead to a greater proportion of alveoli that are underperfused yet are otherwise well ventilated. Moreover, it must be acknowledged that we cannot completely exclude airway-related causes of $\dot{\text{V}}/\dot{\text{Q}}$ mismatch; yet, the lack of spirometric abnormities in the patients with HFpEF studied seems to point to abnormalities relating to pulmonary circulation, rather than the airways, as the primary origin of $\dot{\text{V}}/\dot{\text{Q}}$ mismatch. Finally, we also observed a mild reduction in Dlco at rest in patients with HFpEF relative to control participants. These findings are consistent with those of Olson et al⁹ and likely are the manifestation of pulmonary vascular remodeling, the development of pulmonary edema, a decrease in alveolar volume resulting from obesity, or a combination thereof. Given that a mild reduction in Dlco also can impact alveolar-arterial O₂ and CO₂ exchange, the extent to which these factors contribute to measurements of VD_alveolar during exercise is unclear. However, it should be noted that the $\dot{\text{V}}/\dot{\text{Q}}$ relationships and lung diffusion capacity of patients with HFpEF have not yet been analyzed by means of the multiple inert gas elimination technique,⁴⁷ which remains the only way to quantify the distribution of ventilation and perfusion alongside simultaneous measurements of lung diffusion capacity at rest and during exercise. Separating the roles of $\dot{\text{V}}/\dot{\text{Q}}$ mismatch and diffusion limitation in determining gas exchange abnormalities is beyond the scope of this study, but represents an interesting avenue for further research to determine the mechanism(s) potentially responsible for the development of VD_alveolar in patients with HFpEF.

Aside from an increased VD_alveolar and ventilatory inefficiency, a secondary observation is that most patients with HFpEF seemed to show an inadequate ventilatory response, which is typically defined as Paco₂ of ≥ 38 mm Hg at peak exercise.⁴⁸ Indeed, this was evidenced by patients with HFpEF exhibiting Paco₂ on average of 39 mm Hg at peak exercise. Although $\dot{\text{V}}$_E for a given $\dot{\text{V}}$co₂ (ie, $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ ratio) at peak exercise was slightly higher in patients with HFpEF compared with control participants, it is likely that the $\dot{\text{V}}$_E achieved in patients with HFpEF was not sufficiently high to compensate for the increased VD_alveolar and decrease Paco₂ adequately, that is, to 35 mm Hg, which reflects the upper limit of an adequate ventilatory response.⁴⁸ To confirm this hypothesis, we estimated the $\dot{\text{V}}$_E required to decrease Paco₂ to 35 mm Hg using the following equation: (863 × $\dot{\text{V}}$co₂) / (Paco₂ × [1 – VD/VT]).⁴⁹ By using this equation, we determined that patients with HFpEF would have had to increase $\dot{\text{V}}$_E at peak exercise on average by approximately 9% (ie, from 58.63 ± 18.86 L/min to 63.55 ± 16.54 L/min) to decrease Paco₂ to 35 mm Hg, which would constitute an adequate ventilatory response.⁴⁸ Currently, the reason why patients with HFpEF were unable to generate an adequate ventilatory response is unclear. However, it could be related to a blunted drive to breathe resulting from either the presence of sleep-disordered breathing, obesity hypoventilation syndrome, mechanical ventilatory constraints resulting from increases in fat mass surrounding the chest wall, particularly because the patients with HFpEF in the present study had obesity with a mean BMI of approximately 38 kg/m², or a combination thereof. In support of the latter, we recently showed that the ventilatory response to exercise is reduced in adults with obesity⁵⁰ and that a reduced ventilatory response could be the result of the presence of obesity-related mechanical ventilatory constraints that are imposed on breathing. Therefore, not only do patients with HFpEF exhibit ventilatory inefficiency, but they also exhibit an inadequate ventilatory response, which also could be an important contributor to exercise intolerance in these patients.

The results of the present study have important clinical implications, particularly for patients who are referred for CPET for unexplained dyspnea. CPET is a valuable clinical tool, and assessing $\dot{\text{V}}/\dot{\text{Q}}$ mismatch often is an important reason for completing CPET. Currently, $\dot{\text{V}}/\dot{\text{Q}}$ mismatch typically is estimated by calculating the VD/VT ratio.¹⁷ Although a VD/VT ratio that fails to decrease, or increases from rest to exercise, indicates worsening $\dot{\text{V}}/\dot{\text{Q}}$ mismatch, the mechanism(s) responsible for an increase in VD/VT often are not detectable or assessed during CPET. That is, rarely is it determined whether an elevated VD/VT ratio is ascribed to an increase in VD_alveolar resulting from the potential exacerbation of cardiopulmonary abnormalities or a small VT resulting from mechanical ventilatory constraints that limit VT expansion and blunt the ventilatory response to exercise. Based on the findings of the present study, which suggest that the increase in VD/VT^21,22 largely may be explained by VD_alveolar, we propose that clinicians and physiologists consider deriving VD_alveolar from the VD/VT ratio to obtain a measurement that may provide insight into the potential (patho)physiologic mechanisms governing the development of $\dot{\text{V}}/\dot{\text{Q}}$ mismatch during exercise in patients with HFpEF. Given that $\dot{\text{V}}/\dot{\text{Q}}$ mismatch has been linked with ventilatory inefficiency and exercise intolerance, both of which are used widely as prognostic indicators for heart failure, it is important for clinicians and physiologists to understand the potential (patho)physiologic mechanism(s) responsible for the increase in $\dot{\text{V}}/\dot{\text{Q}}$ mismatch so that novel therapeutics can be developed to improve clinical outcomes in patients with HFpEF.

Moreover, and from a symptomatology perspective, whether VD_avleolar-related increases in ventilatory demand (evidenced by an increase in the $\dot{\text{V}}$_E/$\dot{\text{V}}$co₂ slope) contribute to exertional dyspnea in this population is unknown. However, it is possible that an increased ventilatory demand resulting from increased VD_alveolar may provoke greater mechanical ventilatory constraints (as has been shown in patients with heart failure with reduced ejection fraction⁵¹), which could predispose these patients to a greater potential for exertional dyspnea at a given exercise work rate and consequently result in the premature termination of exercise. Furthermore, we demonstrated that VD_alveolar was elevated (relative to control participants) even at low levels of exercise, which may explain in part why these patients often experience dyspnea and exercise intolerance at low levels of physical activity (eg, activities of daily living). Of note, it must be emphasized that the underlying mechanism(s) responsible for the development of VD_alveolar still need to be uncovered in patients with HFpEF, particularly because these patients exhibit many cardiopulmonary abnormalities that could alter pulmonary blood flow distribution considerably and could contribute to the development of VD_alveolar (eg, elevated cardiac filling pressures,^{3, 4, 5} pulmonary edema,^6,7 impaired pulmonary vascular recruitment or distension,^{8, 9, 10} and right ventricle-pulmonary artery uncoupling^{11, 12, 13, 14, 15}). We propose that this is a critical area that deserves further investigation so that clinicians are able to understand how best to improve gas exchange efficiency and, thus, potentially to manage patients’ symptoms effectively.

We examined patients with HFpEF older than 60 years, which limits the generalization of our findings to only older individuals with HFpEF. Moreover, the submaximal exercise conditions in the present study were different between groups. Also noteworthy is that mechanical dead space was estimated as the effective dead space, rather than volumetric dead space. This is because previous studies suggest that VD/VT calculations that take into account the entire volumetric dead space underestimate VD_physiologic (and thus, VD_alveolar).⁵² As such, we estimated the effective dead space of the breathing apparatus using methods based on Bradley and Younes.⁵² Finally, it should be acknowledged that two patients in the present study harbored concomitant pulmonary abnormalities; that is, one patient had moderate obstructive lung disease and one patient had undergone a prior lobectomy. It is unclear whether these pulmonary abnormalities contribute to VD_alveolar or provoke greater VD_alveolar beyond what is observed in patients with HFpEF with normal lung function (when assessed based on spirometric values). Notably, the VD_alveolar measurements made in these two patients were similar to that measured in the other patients studied and the outcomes of the present study did not change even after removing these two patients from the analysis. Although some pulmonary abnormalities could contribute to the development of VD_alveolar, we do not have the capacity to inform outcomes based on the additive effects of pulmonary abnormalities and HFpEF on the development of VD_alveolar during exercise. This is an area that deserves further investigation.

Interpretation

The results of the present study confirm our original hypothesis by demonstrating that VD_alveolar increased during exercise in patients with HFpEF, the magnitude of which was greater when compared with control participants. We also demonstrated that VD_alveolar correlated with the $\dot{\text{V}}$_E$\dot{\text{V}}$co₂ slope, the latter of which was associated inversely with $\dot{\text{V}}$o_2peak in patients with HFpEF. These data have important clinical implications and suggest that the increase in $\dot{\text{V}}/\dot{\text{Q}}$ mismatch may be explained by increases in VD_alveolar and that increases in VD_alveolar worsens ventilatory inefficiency, which seems to be a key contributor to exercise intolerance in patients with HFpEF. Further investigation is warranted to determine the underlying mechanism(s) governing the development of VD_alveolar in patients with HFpEF and subsequently how VD_alveolar (and $\dot{\text{V}}/\dot{\text{Q}}$ mismatch) might be targeted therapeutically potentially to improve exercise tolerance in these patients.