Acute effects of β-hydroxybutyrate on left ventricular function in young, healthy adults

Abstract

Interest in ketones as a cardiac “super fuel” has grown significantly following reports of a marked increase in cardiac output after exogenous ketone administration in heart failure. However, the extent to which this increase in cardiac output is related to changes in cardiac contractility, and dependent on the presence of heart failure, remains incompletely understood. Therefore, we performed a randomized, double-blind, placebo-controlled study of oral ketone ester in young healthy volunteers. Baseline cardiac magnetic resonance imaging was performed and repeated every 15 min for 60 min after ketone and placebo ingestion to assess changes in left ventricular function. As expected, circulating β-hydroxybutyrate increased rapidly after ketone ingestion, but did not change with placebo (interaction: P < 0.001). Consistent with prior investigations, ketone ingestion resulted in an average 1 L/min increase in cardiac output after 60 min that did not occur with placebo (interaction: P = 0.026). This increase in cardiac output was primarily driven by an increase in heart rate after ketone ingestion (interaction: P = 0.018), with only a modest increase in stroke volume (interaction: P = 0.037). Changes in left ventricular strain and twist mechanics were limited. Taken together, the increase in cardiac output following an acute elevation in circulating β-hydroxybutyrate is primarily driven by changes in cardiac chronotropy, with minimal inotropic contribution.

NEW & NOTEWORTHY In this randomized, double-blind, placebo-controlled study of oral ketone ester in young healthy volunteers, we show a marked increase in cardiac output (∼1 L/min), driven primarily by changes in chronotropy. The cardiac magnetic resonance imaging data support the limited role for inotropy.

INTRODUCTION

The heart is the most metabolically active organ in the body, meeting its high energy demand by metabolizing diverse substrates, such as fatty acids, glucose, lactate, ketones, and amino acids (1, 2). Although most studies have focused on fatty acid and carbohydrate metabolism (3, 4), there is growing interest in ketones being a cardiac “super fuel,” leading to the hypothesis that preferential oxidation of β-hydroxybutyrate by the heart could improve cardiovascular outcomes in disease (5). Indeed, acute administration of β-hydroxybutyrate increases cardiac output by >1 L/min in patients with heart failure with reduced ejection fraction (6). However, the failing heart may have increased reliance on ketones that could contribute to this response (7–11). Whether acute ketone administration causes similar increases in cardiac output in healthy adults (i.e., a physiological vs. pathophysiological response) remains incompletely understood.

Recently, Selvaraj et al. (12) reported that a single, high-dose administration of oral ketone ester enhanced left ventricular deformation after just 30 min in healthy adults. However, nearly half of the participants in this single-arm study experienced gastric discomfort from the high-dose ketone ester, which may have independently contributed to the results via a stress-induced response. Furthermore, the reported changes in stroke volume were small and measured at a single time-point. Therefore, to further examine the time-course by which ketone administration augments cardiac systolic function in healthy adults, we performed a randomized, double-blind, placebo-controlled study of moderate-dose oral ketone ester, using serial cardiac magnetic resonance imaging (cMRI).

METHODS

Participants

A total of 15 participants (8 men and 7 women) were enrolled in this study. Participants were eligible for enrollment if they were between 18 and 30 yr of age, asymptomatic with no history or diagnosis of cardiovascular or metabolic disease, had a body mass index (BMI) between 18.5 and 30 kg/m², blood pressure <140/90 mmHg, were sedentary or recreationally active (<3 days of vigorous aerobic exercise each week), not taking medications that might influence cardiovascular function, and were nonsmokers. Premenopausal women were excluded if they were using oral contraceptives, pregnant, or planning to become pregnant. All subjects provided written informed consent before being enrolled to participate in the present study, as approved by the Institutional Review Board at the University of Texas Southwestern Medical Center (UTSW; ethical approval No. STU-112017-013), and in accordance with the standards set by the Declaration of Helsinki.

Study Protocol

This randomized, double-blind, placebo-controlled study consisted of two visits on nonconsecutive days. Each subject’s pair of visits were scheduled at the same time of day to avoid diurnal variations in metabolism. Before each study visit, subjects arrived at UTSW’s Advanced Imaging Research Center (AIRC) fasted for at least 8 h, and withdrawn from exercise, alcohol, and caffeine for at least 24 h. Upon arrival at the AIRC, an intravenous (iv) catheter was inserted into an antecubital vein for repeated blood sampling. Cardiac magnetic resonance imaging (cMRI) was then performed to measure LV morphology and function. Once baseline images were collected, subjects were removed from the bore of the magnet but instructed to remain still on the bed of the MRI. While lying still, baseline blood was drawn immediately before the ingestion of either a placebo or ketone supplement in randomized order. The ketone supplement consisted of D-β-hydroxybutyrate-(R)-1,3 butanediol (KetoneAid Inc, Falls Church, Virginia; 0.45 mL/kg body mass or 483 mg/kg body mass) with added water and flavoring (Sweet Drops, SweetLeaf, Wisdom Natural Brands, Gilbert, Arizona) in a total volume of 100 mL. The placebo consisted of solely water and equal amount of flavoring as the ketone supplement in 100 mL volume. Immediately after ingesting the ketone supplement or placebo, subjects were given 20 mL of calorie-free sports drink (G Zero, Gatorade, PepsiCo, Inc., Purchase, New York) to eliminate any remaining flavor. Then, subjects were returned to the bore of the magnet for repeat imaging and blood draws. Imaging was completed within 5–7 min and timed so that each period ended immediately before each 15 min blood draw. Blood was drawn while the subjects remained in the MRI, after which blood collection tubes were centrifuged and samples transferred to aliquots for immediate storage at −80°C until analysis.

Measurements

Left ventricular morphology and function.

cMRI was performed on a Philips Achieva 3 T MRI scanner (Philips Medical Systems, Best, Netherlands). Arterial blood pressure was measured inside the bore of the MRI using an MRI-compatible automated blood pressure monitor. LV morphology was assessed using a stack of short axis ECG‐gated, steady-state free precision images, acquired at end-expiration using the following parameters: 3.0 ms repetition time, 1.49 ms echo time, 40° flip angle, 10 mm slice thickness with a 4-mm gap between slices, 232 × 227 matrix, and 350 × 350 field of view. Data were analyzed offline by an experienced observer using commercially available analysis software (CVI42, version 5.13.5, Circle Cardiovascular Imaging, Calgary, AB, Canada). Epicardial and endocardial borders of each slice were traced manually at end-diastole and end-systole to calculate LV mass and volume. Effective arterial elastance (Ea) was calculated by dividing end-systolic pressure (0.9 × systolic blood pressure) by stroke volume. End-systolic elastance (Ees) was calculated by dividing end-systolic pressure (0.9 × systolic blood pressure) by end-systolic volume. The ratio of the two (Ea/Ees) was then used to assess ventricular-arterial coupling.

LV deformation was assessed using three evenly spaced short-axis cardiac MR tissue tagging images spanning the LV using a standard gradient echo sequence. Typical imaging parameters included an 8 mm slice thickness, 9 mm grid tags, 5.4 ms repetition time, 3.2 ms echo time, 140 × 126 matrix, 10° flip angle, and 300 × 300 field of view. Tagged cMR images were analyzed using commercial software (HARP, Diagnosoft 3.0; Palo Alto, CA) to determine circumferential strain, strain rate, twist, twisting rate, and torsion. Tag analysis was semiautomated, with user input limited to tracing the endo- and epicardium at a single reference cardiac phase for each slice.

Bioassays.

All metabolic analyses were performed by UTSW’s Metabolic Phenotyping Core facility. Plasma glucose, triglyceride, and cholesterol were determined using a fully automated OCD Vitros 350 dry chemistry analyzer following the protocols provided by the reagent kit manufacturer (Ortho Clinical Diagnostics, Raritan, NJ). Plasma insulin was quantified using a commercially available sandwich ELISA (ALPCO, Salem, NH, Cat. No. 80-INSHU-E01.1). Plasma nonesterified fatty acids (NEFA) and β-hydroxybutyrate were quantified by enzymatic colorimetric assays following the protocols described by the manufacturer (Fujifilm Healthcare Americas Corporation, Lexington, MA; Cat. Nos. 999-34691, 991-34891, 997-76491, 417-73501, 413-73601, and 412-73791).

Statistical Analysis

Data are presented as means ± standard error mean. Some data points were not collected due to subjects terminating the study early to void their bladder or poor data quality, but no variables contained more than 3.6% missing data. Although a significant Little’s Missing Completely At Random test was found (P = 0.034), only one pattern of missingness was detected so our data were assumed missing at random. Expectation maximization was used to impute missing data. A two-way repeated measures ANOVA was performed to compare the effect of condition (ketone and placebo) over time on our study variables. When significant interactions were observed, simple main effects were assessed using Fisher’s least significant difference tests. Study variables that did not meet required assumptions were normalized with log transformation. All statistical hypothesis testing was two-sided with alpha set to 0.05. All statistical analyses were performed using SPSS Version 29 for Windows.

RESULTS

Of the 15 participants enrolled in the study, two were lost to attrition before completing their second visit, one was excluded for noncompliance with prestudy instructions, and one was excluded based on poor image quality. The remaining 11 participants (five men/six women, 22 ± 2 yr of age, BMI: 25.4 ± 3.1 kg/m²) completed the study without incident.

Blood Biomarkers

As described in Table 1, β-hydroxybutyrate increased rapidly within 15 min of ketone ingestion and remained elevated for the duration of the experiment. Circulating glucose and NEFA gradually decreased from baseline within 15 min after ketone ingestion, whereas insulin initially increased after 15 min before returning to baseline at 60 min post-ketone ingestion. Although no major changes in circulating blood markers occurred following the ingestion of the placebo, NEFA did increase and remained elevated after 45 min, consistent with the prolonged fast.

Table 1.

Changes in metabolic variables

	Ketone					Placebo					P Value
	0 min	15 min	30 min	45 min	60 min	0 min	15 min	30 min	45 min	60 min	Condition	Time	Interaction
Glucose, mg/dL	93 ± 3	91 ± 3*	89 ± 3*	86 ± 3*†	85 ± 3*†	92 ± 3	91 ± 3	92 ± 3	92 ± 3	90 ± 2			0.029
Cholesterol, mg/dL	151 ± 6	152 ± 6	154 ± 6	156 ± 6	156 ± 7	156 ± 7	156 ± 7	159 ± 6	161 ± 7	156 ± 6	0.064	0.025	0.503
Triglycerides, mg/dL	68 ± 6	67 ± 6	65 ± 6	64 ± 7	68 ± 6	76 ± 8	76 ± 7	77 ± 7	78 ± 7	78 ± 7	0.055	0.617	0.222
Nonesterified fatty acids, mEq/dL	0.52 ± 0.06	0.44 ± 0.05*	0.33 ± 0.03*†	0.25 ± 0.02*†	0.27 ± 0.04*†	0.47 ± 0.05	0.50 ± 0.07	0.52 ± 0.06	0.54 ± 0.06*	0.54 ± 0.06*			<0.001
β-Hydroxybutyrate, µmol/L	243 ± 68	1,282 ± 263*†	1,837 ± 283*†	2,119 ± 302*†	1,675 ± 257*†	263 ± 76	232 ± 64	202 ± 58	268 ± 67	310 ± 76			<0.001
Insulin, µlU/mL	7.1 ± 1.8	12.8 ± 3.2*†	12.3 ± 2.5*†	10.5 ± 1.8*	9.3 ± 1.4†	7.2 ± 1.4	7.2 ± 1.4	7.7 ± 1.3	8.2 ± 1.4	6.3 ± 1.0			0.012

Data reported as means ± SE for n = 11 young, healthy adults. *P <0.05 vs. 0 min, †P <0.05 vs. placebo (significant values in bold).

Cardiac Hemodynamics

Mean arterial pressure increased modestly over time during both ketone and placebo conditions (Table 2). Cardiac output increased significantly above baseline within 30 min of ketone ingestion, and remained elevated for the duration of the experiment, whereas not changing following the ingestion of the placebo (Fig. 1A). The increase in cardiac output following ketone ingestion was primarily driven by heart rate, which increased significantly above baseline within 30 min and to a greater extent than the placebo condition after 45 min and 60 min (Fig. 1B). Indeed, stroke volume increased only modestly 15 min following ketone ingestion but remained elevated for the duration of the experiment (Fig. 1C). Neither end-diastolic nor end-systolic volumes were affected by either condition, but there was a small, albeit significant increase in ejection fraction within 15-min of ketone ingestion that persisted until the end of the study (Table 2). Neither stroke volume nor ejection fraction were affected by the placebo.

Table 2.

Changes in mean arterial pressure and cMRI outcome variables

	Ketone					Placebo					P Value
	0 min	15 min	30 min	45 min	60 min	0 min	15 min	30 min	45 min	60 min	Condition	Time	Interaction
Mean arterial pressure, mmHg	69 ± 2	71 ± 2	71 ± 3	71 ± 3	72 ± 3	68 ± 3	72 ± 3	71 ± 3	73 ± 3	71 ± 3	0.810	0.037	0.616
LV morphology
EDV, mL	152 ± 7	155 ± 7	154 ± 7	153 ± 8	152 ± 7	154 ± 8	155 ± 7	152 ± 7	153 ± 8	150 ± 7	0.519	0.253	0.394
ESV, mL	62 ± 4	61 ± 3	59 ± 3	58 ± 3	59 ± 4	63 ± 4	65 ± 4	64 ± 4	64 ± 4	62 ± 4	0.023	0.171	0.263
LV function
Ejection fraction, %	59 ± 1	61 ± 1†	61 ± 1*†	62 ± 1*†	61 ± 1*	59 ± 1	58 ± 1	58 ± 1	58 ± 1	59 ± 1			0.033
End-systolic elastance, mmHg/mL	1.50 ± 0.10	1.55 ± 0.11	1.60 ± 0.11	1.64 ± 0.11	1.64 ± 0.12	1.43 ± 0.09	1.46 ± 0.10	1.47 ± 0.09	1.49 ± 0.09	1.51 ± 0.10	0.008	0.002	0.479
Arterial elastance, mmHg/mL	1.01 ± 0.05	0.98 ± 0.05	0.99 ± 0.05	1.00 ± 0.06	1.03 ± 0.06	0.99 ± 0.05	1.03 ± 0.05	1.04 ± 0.05	1.05 ± 0.06	1.05 ± 0.05	0.129	0.018	0.052
LV-arterial coupling, Ea/Ees	0.69 ± 0.03	0.65 ± 0.02†	0.63 ± 0.03*†	0.62 ± 0.02*†	0.64 ± 0.03*	0.70 ± 0.03	0.72 ± 0.03	0.72 ± 0.03	0.72 ± 0.03	0.71 ± 0.03			0.030
Peak circumferential strain, %	−19.8 ± 0.6	−20.1 ± 0.6	−20.2 ± 0.7	−20.1 ± 0.6	−19.7 ± 0.6	−19.4 ± 0.6	−19.6 ± 0.5	−19.3 ± 0.5	−19.3 ± 0.4	−19.3 ± 0.6	0.040	0.524	0.508
Systolic circumferential strain rate, %/s	−99 ± 2	−100 ± 3	−105 ± 3†	−105 ± 3†	−105 ± 4†	−97 ± 2	−97 ± 2	−96 ± 2	−95 ± 2	−96 ± 2			0.034
Early diastolic circumferential strain rate, %/s	152 ± 6	152 ± 8	150 ± 8	150 ± 8	147 ± 8	148 ± 7	141 ± 6	136 ± 5	137 ± 5	138 ± 6	0.036	0.058	0.617
Torsion, °/cm	3.0 ± 0.1	3.1 ± 0.1	3.1 ± 0.1	3.4 ± 0.2	3.3 ± 0.1	3.0 ± 0.2	3.1 ± 0.1	3.0 ± 0.1	3.1 ± 0.2	3.3 ± 0.2	0.337	0.030	0.268
Peak twisting rate, °/s	39 ± 2	40 ± 2	41 ± 2	45 ± 2	47 ± 2	41 ± 2	40 ± 2	41 ± 2	41 ± 2	41 ± 2	0.326	0.042	0.057
Peak untwisting rate, °/s	−60 ± 3	−57 ± 3	−60 ± 4	−64 ± 4	−67 ± 4†	−59 ± 2	−60 ± 3	−55 ± 3	−59 ± 3	−56 ± 4			0.026

Data reported as means ± SE for n = 11 young, healthy adults. Ea, arterial elastance; EDV, end-diastolic volume; Ees, end-systolic elastance; ESV, end-systolic volume. *P <0.05 vs. 0 min, †P <0.05 vs. placebo (significant values in bold).

Figure 1.

Hemodynamic changes with ketone and placebo ingestion. A: cardiac output increased 30 min after ketone ingestion but did not change after placebo ingestion. B: heart rate increased after 30 min in both conditions but increased to a greater extent 45 and 60 min after ketone ingestion. C: stroke volume increased 15 min after ketone ingestion but did not change after placebo ingestion. Data are shown as means ± SE; n = 11. *P < 0.05 vs. 0 min, †P < 0.05 vs. placebo.

Cardiac Mechanics

Although peak circumferential strain was unaffected by either condition, systolic circumferential strain rate in the ketone condition was greater than the placebo condition 30 min, 45 min, and 60 min after the experiment began (Table 2). Systolic circumferential strain rate did not increase from baseline following ketone ingestion. Instead, end-systolic elastance, peak torsion, and peak twisting rate increased over time but did not differ by condition.

Although both end-systolic and arterial elastance increased over time, regardless of condition, there was a small decrease in ventricular-arterial coupling within 30 min of ketone ingestion that remained lower for the duration of the experiment (Table 2). Ventricular-arterial coupling was unaffected by the placebo.

DISCUSSION

The results from this randomized, double-blind, placebo-controlled study extend prior observations demonstrating a marked increase in cardiac output following acute ketone administration in healthy adults (6, 12). That the increase in cardiac output was primarily driven by an increase in heart rate argues against ketones being a “super fuel” for the heart. Rather, we interpret the heightened chronotropy to be a neurally mediated phenomenon. Indeed, ketones may evoke this chronotropic response through either direct effects on autonomic neurons (13, 14) or reflex-mediated (i.e., baroreflex) responses secondary to ketone-induced vasodilation (15–18).

These results are somewhat inconsistent with that of Selvaraj et al. (12), the only other study to date that has investigated the impact of acute ketone ingestion on cardiac hemodynamics and function in healthy adults. For instance, we observed a similar, small increase in stroke volume, but only limited changes in cardiac mechanics. The reason for this difference is likely multifactorial. First, we chose to use a lower dose to avoid gastric distress. Second, we used MR tissue tagging to evaluate cardiac mechanics, which, by convention, evaluates LV mechanics in the short-axis orientation. However, that stroke volume increased, without major changes in circumferential strain or torsion, supports the increase in longitudinal shortening reported by Selvaraj et al. (12). Third, we included a placebo control arm, which may have helped to account for changes beyond ketone ingestion.

Although caution is warranted when making direct comparisons, the increase in cardiac output observed in this study is remarkably consistent with prior studies in patients with heart failure (6, 19). We interpret this to suggest that the response is “physiological,” rather than “pathophysiological.” Nevertheless, randomized control trials have produced evidence that SGLT2-inhibition in patients with heart failure reduces the risk of cardiovascular death and hospitalizations (20), for which ketones may play a part. Indeed, SGLT2-inhibition evokes endogenous ketone production, but this effect is one of many potential mechanisms by which SGLT2-inhibitors may impart their beneficial effects (21). Whether or not ketone supplementation, in the absence of these concomitant mechanisms, can also reduce cardiovascular risk remains to be elucidated.

This study is not without limitations. First, the sample size is relatively small. However, n = 11 subjects provided adequate power to detect meaningful differences in our primary outcomes. Second, the single-dose design of the present study limits our understanding of the dose-response relationship. Third, the acute nature of the experiment limits our understanding beyond 60 min and/or with chronic, repeat dosing. Lastly, although we report ketone-induced changes in cardiac hemodynamics, the exact mechanism driving the chronotropic response is beyond the scope of the present investigation. More work is therefore needed.

Taken together, we conclude that the increase in cardiac output following an acute elevation in circulating β-hydroxybutyrate is primarily driven by changes in cardiac chronotropy, with minimal inotropic contribution.

DATA AVAILABILITY

Data will be made available upon reasonable request.

GRANTS

This work was supported by research grants from the National Institutes of Health (R01HL136601) and American Heart Association (PRE835833).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.P.O., B.E.Y., and M.D.N. conceived and designed research; A.P.O. performed experiments; A.P.O. and D.J.C. analyzed data; A.P.O., B.E.Y., V.Z., and M.D.N. interpreted results of experiments; A.P.O. and M.D.N. prepared figures; A.P.O. and M.D.N. drafted manuscript; A.P.O., B.E.Y., D.J.C., V.Z., and M.D.N. edited and revised manuscript; A.P.O., B.E.Y., D.J.C., V.Z., and M.D.N. approved final version of manuscript.