Genetic variation in alpha₂-adrenoreceptors and heart rate recovery after exercise

Abstract

Heart rate recovery (HRR) after exercise is an independent predictor of adverse cardiovascular outcomes. HRR is mediated by both parasympathetic reactivation and sympathetic withdrawal and is highly heritable. We examined whether common genetic variants in adrenergic and cholinergic receptors and transporters affect HRR. In our study 126 healthy subjects (66 Caucasians, 56 African Americans) performed an 8 min step-wise bicycle exercise test with continuous computerized ECG recordings. We fitted an exponential curve to the postexercise R-R intervals for each subject to calculate the recovery constant (k_r) as primary outcome. Secondary outcome was the root mean square residuals averaged over 1 min (RMS_1min), a marker of parasympathetic tone. We used multiple linear regressions to determine the effect of functional candidate genetic variants in autonomic pathways (6 ADRA2A, 1 ADRA2B, 4 ADRA2C, 2 ADRB1, 3 ADRB2, 2 NET, 2 CHT, and 1 GRK5) on the outcomes before and after adjustment for potential confounders. Recovery constant was lower (indicating slower HRR) in ADRA2B 301–303 deletion carriers (n = 54, P = 0.01), explaining 3.6% of the interindividual variability in HRR. ADRA2A Asn251Lys, ADRA2C rs13118771, and ADRB1 Ser49Gly genotypes were associated with RMS_1min. Genetic variability in adrenergic receptors may be associated with HRR after exercise. However, most of the interindividual variability in HRR remained unexplained by the variants examined. Noncandidate gene-driven approaches to study genetic contributions to HRR in larger cohorts will be of interest.

heart rate recovery (HRR) following exercise, the gradual decrease in heart rate from peak exercise-induced tachycardia toward resting heart rate, predicts mortality independently of other cardiovascular risk factors both in patients with cardiovascular disease and in asymptomatic healthy subjects (6, 16, 33). HRR after exercise is mediated by the interplay between the sympathetic and the parasympathetic limbs of the autonomic nervous system. After exercise, parasympathetic reactivation and sympathetic withdrawal both act to lower heart rate toward pre-exercise values (24).

The extent and rate of HRR after exercise differ among individuals, and the mechanisms underlying these interindividual differences are not known. The heritability for HRR is considerable (estimated to be 0.34) (15), suggesting that genetic factors contribute significantly to interindividual variability. However, there is little information about the specific genetic variants involved. Since sympathetic and parasympathetic mechanisms mediate HRR, functional genetic variation in these pathways may contribute to interindividual variability in HRR.

There are common functional variants in the genes encoding the β₁- (ADRB1) and β₂-adrenergic receptors (ARs) (ADRB2), which mediate the increase in heart rate in response to catecholamines (3, 4). However, ADRB1 and ADRB2 variants were not associated with heart rate at 3 min after completion of a standardized exercise test (15). The effects of variants in other candidate genes that affect sympathetic response on HRR have not been studied. In particular, functional genetic variants in α₂-adrenergic receptors (α₂ARs) that regulate the sympathetic and the parasympathetic tone in the central nervous system (α_2AAR and α_2CAR) (23, 32, 44, 45, 48) and variants in the norepinephrine transporter (NET), responsible for the rapid reuptake of norepinephrine from the synapse (13, 18, 19), could affect HRR. Moreover, a functional variant in the G protein-coupled receptor kinase 5 (GRK5) affects β-AR-mediated signaling (25).

Less is known about genetic variation in proteins regulating the parasympathetic nervous system and their functional effects. The choline transporter (CHT) mediates the reuptake of choline from the synaptic cleft into presynaptic cholinergic neurons of autonomic ganglia; this constitutes the rate-limiting step in the availability of acetylcholine in synaptic vesicles and may serve as a key regulator of parasympathetic tone (9, 34). Mice that are genetically modified to have reduced expression of CHT have slower HRR after treadmill exercise (9). Several genetic variants in human CHT have been characterized and are functional in vitro (35). In addition, variant allele carriers of a CHT 3′-untranslated region polymorphism were reported to have a higher parasympathetic tone in vivo as determined by heart rate variability measures (31). However, the effect of these genetic variants on parasympathetic tone during HRR is not known.

Thus, we addressed the hypothesis that common functional polymorphisms in candidate genes involved in regulation of adrenergic and cholinergic signaling (ADRA2A, ADRA2B, ADRA2C, ADRB1, ADRB2, NET, GRK5 and CHT) contribute to the interindividual variability in HRR and parasympathetic tone after exercise.

METHODS

Subjects

This study was approved by the Institutional Review Board of Vanderbilt University Medical Centre, Nashville, TN, and all subjects gave written informed consent. We studied 126 healthy subjects, taking no medications, all of whom also contributed data to previous studies (21). Details of the subject recruitment are described elsewhere (19, 21). In short, unrelated American white and black subjects residing in the Nashville area were eligible to participate if they fulfilled the following inclusion criteria: Aged between 18 and 50 yr, nonsmokers, and no clinically significant abnormality based on medical history, physical examination, electrocardiogram, and routine laboratory testing. Patients were free of medications for at least 1 wk and received a controlled alcohol-free and caffeine-free diet (providing 150 mmol of sodium, 70 mmol of potassium, and 600 mmol of calcium daily) for 6 days before the study. Each subject reported the intensity, duration, and frequency of his/her physical activity over the last 4 wk, and these data were converted into a measure of physical activity [metabolic equivalents (METS) * hour] using standard charts (1). The exercise score for each subject was defined as the mean number of METS*h/wk.

Protocol

On the morning (8:00–10:00 AM) of the study day, subjects were admitted to a temperature-controlled room (22–23°C) in the Vanderbilt University Clinical Research Center after an overnight fast. A 20 G intravenous cannula was inserted into an antecubital arm vein for blood sampling. A continuous electrocardiogram (ECG) recording (Gould ECG/Biotach amplifier; Gould, Cleveland, OH) was digitized (DI720USB; DATAQ Instruments, Akron, OH) on a personal computer at 500 Hz using Windaq software (v. 2.20, DATAQ Instruments). After 30 min of supine rest, subjects moved to a semirecumbent position on an adjacent electronically braked supine bicycle ergometer (ER 800L; Ergoline, Bitz, Germany). The exercise protocol consisted of biking at a constant rate (60 rpm) at sequentially increasing workloads of 25, 50, 75, and 100 W for 2 min each. A blood sample for the determination of plasma norepinephrine concentration was drawn immediately after peak exercise.

Genotyping

Subjects were genotyped for 21 variants in eight adrenergic and cholinergic pathway genes encoding receptors (6 variants in ADRA2A, 1 in ADRA2B, 4 in ADRA2C, 2 in ADRB1, and 3 in ADRB2), presynaptic transporters (2 variants each in CHT and NET), and one kinase (1 variant in GRK5, Table 1). These particular variants were chosen because they were either common [minor allele frequency (MAF) >5%], located in the coding or promoter region, or were shown to be functional in previous studies. To capture best genetic variability, markers within the same gene were chosen only if they were not in high linkage disequilibrium (r² ≥ 0.80) in at least one of the races. Genotyping methods and quality control for variants in ADRA2A, ADRA2B, ADRA2C, ADRB1, ADRB2, NET, and GRK5 have been described previously (8, 19–22, 30, 46). ADRB2 rs1800888 (Thr164Ile) and two CHT variants [rs333229 (G/T) and rs1013940 (Ile89Val)] were genotyped in the Vanderbilt Center for Human Genetics Research DNA Resources Core by allelic discrimination with TaqMan 5′-nuclease assays (26) on an ABI 7900 HT real-time polymerase chain reaction system (Applied Biosystems, Foster City, CA) using validated TaqMan probes and a 95% quality value threshold (22). Call rates ranged from 89.7 to 100.0%. For quality assurance, we repeated genotyping for 20% of the samples, yielding concordant results.

Table 1.

Genetic variants in candidate genes in adrenergic and parasympathetic pathways

			Caucasians		African Americans		Total Cohort
Gene	Variant, rs number	Nucleotide/Amino Acid Change, position	MAF, %	Genotype Distribution, 0/1/2 variant alleles	MAF, %	Genotype Distribution, 0/1/2 variant alleles	MAF, %	Genotype Distribution, 0/1/2 variant alleles
NET	rs28386840	A-3081T1	25.4	33/26/3	53.2	13/18/16	36.7	49/45/19
	rs2242446	T-182C1	27.2	28/27/2	19.4	31/17/1	23.6	61/46/3
CHT	rs333229	G/T2	18.3	41/16/3	34.0	20/22/5	24.5	63/40/8
	rs1013940	Ile89Val3	7.0	52/8/0	2.0	47/2/0	5.0	103/10/0
ADRA2A	rs1800035	Asn251Lys3	0.0	65/0/0	5.5	48/6/0	2.4	117/6/0
	rs1800544	C-1297G1	29.2	35/22/8	64.3	6/28/22	46.4	42/50/33
	rs2484516	C-727G	5.0	53/6/0	4.8	47//3/1	5.7	102/11/1
	rs1800545	G-262A4	12.0	50/14/1	28.5	28/24/4	19.6	81/39/5
	rs3750625	C1802A2	7.7	49/9/0	17.5	36/12/3	12.4	88/22/3
	rs553668	G1780A2	16.9	46/16/3	26.8	29/24/3	22.8	76/41/8
ADRA2B		del301–3033	37.1	28/22/12	16.0	36/17/0	27.7	65/42/12
ADRA2C	rs9790683	C-2865T1	8.3	50/10/0	8.0	48/7/1	8.75	100/19/1
	rs13118771	C-2579T1	5.0	50/12/0	8.3	48/8/0	9.0	100/22/0
		del322_3253	3.2	59/2/1	48.1	15/25/13	24.3	75/30/14
	rs7434444	G-1736C1	20.3	38/18/3	66.1	8/22/26	42.8	47/42/30
ADRB1	rs1801252	Ser49Gly3	13.0	48/17/	26.4	29/23/3	18.9	80/41/3
	rs1801253	Arg389Gly3	25.6	36/26/3	41.1	18/30/8	31.5	58/56/11
GRK5	rs17098707	Gln49Leu3	0.0	65/0/0	29.5	23/27/6	16.5	91/28/7
ADRB2	rs1042713	Arg16Gly3	72.3	10/29/26	56.2	9/31/16	58.8	21/61/43
	rs1042714	Gln27Glu3	44.7	22/28/15	18.8	37/17/2	31.6	63/45/17
	rs1800888	Thr164Ile3	1.5	63/2/0	0.0	56/0/0	0.8	123/2/0

The total cohort includes Caucasians, African Americans, and 4 subjects of other racial background. MAF, minor allele frequency. ¹Promoter variant; ²variant in the 3′-untranslated region (UTR); ³variant in coding region; ⁴variant in the 5′-UTR.

Plasma Catecholamine Determination

Blood was collected in cooled heparin-coated tubes, which were immediately placed on ice until centrifuged at 4°C for 10 min at 3,000 rpm. Plasma was harvested and stored in tubes containing 40 μl of reduced glutathione (6%) at −20°C until assayed. Norepinephrine and epinephrine concentrations were measured by high-performance liquid chromatography with electrochemical detection. Dihydroxybenzylamine was used as an internal standard (14).

Measurements of HRR

After exercise, heart rate decreases from peak heart rate in an exponential manner. Several measures have been used to quantify HRR. The simplest measure is the absolute decrease in heart rate during the first minute after cessation of exercise (ΔHR_1min) (6, 16, 33). This measure captures sympathetic withdrawal and parasympathetic reactivation during early recovery. However, since heart rate declines exponentially, ΔHR_1min depends on the peak exercise heart rate achieved and is not an optimal marker to assess HRR during submaximal exercise.

An integrative measure of HRR during the whole recovery period is the recovery constant k_r, representing the decay constant in the equation describing exponential decay (39). This recovery constant is independent of the peak heart rate and equivalent to the elimination rate constant k_e that describes the exponential decrease in plasma concentration over time of a drug with first-order pharmacokinetics. The recovery constant k_r has been previously validated for the description of the exponential HRR after submaximal exercise (39), and we prospectively defined it as our primary outcome.

As a secondary outcome, we also assessed root mean square residuals (RMS), a validated marker of parasympathetic tone during recovery from exercise (11). RMS measures the deviation of the R-R intervals from a straight line and thus is a measure of the typical oscillations of successive RR intervals, which reflect parasympathetic activity. At peak exercise, when vagal tone is very low, RMS is close to 0, but it rises quickly during the first minutes after cessation of exercise as vagal tone recovers.

Data Analysis and Statistics

Heart rate analyses from continuous ECG recordings were performed using customized software (Physiowave) written in PV-WAVE (Visual Numerics, Houston, TX) and MATLAB environments (Mathworks, Natick, MA). We used an algorithm described by Pan and Tompkins for QRS detection (36). Additionally, we visually inspected all QRS detections to identify atrial and ventricular premature beats; R-R intervals preceding and following premature beats were removed from analysis.

We estimated the recovery constant k_r for each subject after fitting an exponential curve to a graph of RR intervals over time, using y = a*(1 − e^−kr*t) + c, where y represented R-R intervals (ms), a the magnitude of R-R interval changes, k_r the recovery constant, t the time (s), and c the minimal R-R interval at peak exercise (Fig. 1). Average heart rate and RMS values were calculated from R-R intervals over successive 30 s segments for the resting period (10 to 5 min before exercise), and during early recovery at 1 min after exercise (30–90 s postexercise) (11).

Fig. 1.

Representative heart rate recovery curve. For each subject, we fitted an exponential curve to the heart rate recovery curve, using y = a*(1 − e^−kr*t) + c, where y represented R-R intervals (ms), a the magnitude of R-R interval changes, k_r the recovery constant, t the time (s), and c the minimal R-R interval at peak exercise.

Normally distributed data were expressed as mean ± standard deviation (SD) or 95% confidence interval. We used multiple linear regression models to analyze the effect of polymorphisms in adrenergic and cholinergic candidate genes on different prospectively defined measures of HRR (primary outcome, k_r; secondary outcome, RMS_1min). All models included as potential confounders age, sex, race, and exercise score.

We assumed an additive mode of inheritance, and genotypes were coded 0–2 according to the number of variant alleles. We tested for Hardy-Weinberg equilibrium using χ² test with 1 degree of freedom. Statistical analyses were performed using the statistical software STATA v. 10.0 (StataCorp, College Station, TX). All tests were two tailed, and P < 0.05 was considered significant. In this preliminary study, we did not adjust for multiple comparisons.

RESULTS

Subject Characteristics and Genotypes

The demographic characteristics, resting cardiovascular measures, and genotyping results for the 126 subjects (66 Caucasians and 56 African Americans; 70 women) are summarized in Tables 1 and 2. All genotypes conformed to Hardy-Weinberg equilibrium in each ethnic group (all P > 0.05) except for the ADRA2A C-727G (rs2484516) variant in African Americans (χ² = 6.95, P = 0.01) and ADRA2C del322-325 variant in Caucasians (χ² = 14.5, P < 0.001), a finding possibly attributable to genetic admixture in African Americans and paucity of deletion allele carriers in Caucasians. Minor allele frequencies were in the expected range.

Table 2.

Demographic characteristics, resting cardiovascular measures, and HRR parameters

Parameter	Mean ± SD (n)
Age, yr	26.0 ± 6.1
Female sex	70 (56%)
Weight, kg	74.3 ± 17.1
Heigh, m	1.71 ± 0.10
BMI, kg/m2	25.7 ± 5.1
Race
Caucasian	66 (52.4%)
African American	56 (44.4%)
Others	4 (3.2%)
Exercise activity (METS*hr/week), median (IQR)	10.8 (0.0–21.6)
Resting heart rate, beats/min	67.7 ± 7.8
Resting systolic BP, mmHg	111.4 ± 12
Resting diastolic BP, mmHg	65.2 ± 7.4
RMS1 min, ms	23.9 ± 12.0
(range)	(4.4–68.5)
Recovery constant kr, s/min	0.013 ± 0.008
(range)	(0.0003–0.043)

HRR, heart rate recovery; BMI, body mass index; METS, metabolic equivalent; IQR, interquartile range; BP, blood pressure; RMS_{1 min}, Root mean square averaged over the first minute of recovery.

Exercise Test and HRR

Among 126 participants, 119 completed all four exercise stages up to 100 W, while seven subjects reached only 75 W and finished prematurely due to physical exhaustion. Mean peak exercise heart rate was 131.1 ± 20.4 beats/min, and mean plasma NE concentration at peak exercise was 490 ± 322 pg/ml. Results for the markers of HRR (k_r and RMS_1min) were within the previously reported range (Table 2) (7, 11).

Measures of HRR and Genotypes

Recovery constant k_r.

The ADRA2B 301–303 deletion was significantly associated with lower k_r and, thus, slower HRR (Table 3, Fig. 2). k_r was 28.6% lower in ADRA2B 301–303 deletion homozygotes (n = 12) than insertion homozygotes (n = 65), with heterozygous subjects (n = 42) having intermediate values (P = 0.01, Fig. 2). This association remained significant after adjustment for age, sex, race, and exercise score in a multiple linear regression model (P = 0.01, Table 3). Adding the 301–303 deletion variant to this linear regression model, with k_r as dependent variable and age, sex, race, and exercise score as covariates, increased the coefficient of determination R² from 6.5 to 10.1% (P = 0.01). Other covariates that were significantly associated with k_r included race (P = 0.02) and exercise score (P = 0.04). A race-stratified sensitivity analysis showed that the association between k_r and the del301-303 deletion was significant in Caucasians (P = 0.02) but not in African Americans, likely because none of the deletion homozygotes were African American.

Table 3.

Associations of HRR parameters with genetic variants

	kr		RMS1 min
Polymorphism	β Coeff.	P Value*	β Coeff.	P Value*
NET
rs28386840	−0.002	0.09	−2.1	0.15
rs2242446	−0.002	0.102	−2.3	0.22
CHT
rs333229 (G/T)	−0.001	0.34	−2.5	0.16
rs1013940 (Ile89Val)	−0.001	0.56	0.31	0.94
ADRA2A
rs1800035 (Asn251Lys)	0.006	0.053	9.7	0.01
rs1800544 (C/G)	−0.0005	0.61	−1.4	0.31
rs2484516 (C-727G)	0.002	0.44	−2.1	0.51
rs1800545 (G-262A)	−0.002	0.10	−1.63	0.36
rs3750625 (C1802A)	−0.002	0.10	−3.7	0.10
rs553668 (C/T)	0.001	0.23	−0.38	0.82
ADRA2B
del301–303	−0.003	0.01	−2.8	0.07
ADRA2C
rs9790683 (C-2865T)	0.002	0.14	4.2	0.05
rs13118771 (C-2579T)	0.002	0.17	5.8	0.025
del322_325	0.0006	0.59	1.2	0.49
rs7434444 (G-1736C)	0.001	0.29	0.15	0.92
ADRB1
rs1801252 (Ser49Gly)	0.002	0.15	5.3	0.005
rs1042713 (Arg389Gly)	−0.001	0.26	−1.43	0.36
GRK5
rs17098707 (Gln49Leu)	0.001	0.45	2.9	0.16
ADRB2
rs1042713 (Arg16Gly)	0.001	0.18	1.3	0.38
rs1042714 (Gln27Glu)	0.001	0.20	2.20	0.14
rs1800888 (Thr164Ile)	−0.003	0.55	−7.1	0.38

P values are adjusted for age, sex, race, and the fitness score.

Fig. 2.

Recovery constant (k_r) and ADRA2B 301–303 deletion genotype. The deletion genotype was associated with smaller k_r (indicating slower heart rate recovery; P = 0.01). Squares and error bars represent the means and SEs, respectively.

None of the other genetic variants was associated with k_r. Similarly, k_r was not associated with the peak norepinephrine concentrations at the end of exercise before or after adjustment for potential confounders like age, sex, race, and fitness (unadjusted P = 0.35, adjusted P = 0.84).

Measures of Parasympathetic Effects (RMS_1min)

None of the single nucleotide polymorphisms in the CHT gene was associated with RMS_1min. In contrast, variants in ADRA2A rs1800035 (Asn251Lys), ADRA2C rs13118771, and ADRB1 rs1801252 (Ser49Gly) were associated with higher RMS_1min (Table 3, Fig. 3): compared with noncarriers, RMS_1min was 46% higher in carriers of one ADRA2A rs1800035 (Asn251Lys) allele (P = 0.01 after adjustment for age, sex, race, and exercise score) and 26% higher in carriers of one ADRA2C rs13118771 allele (adjusted P = 0.025). Similarly, subjects with two variants of ADRB1 rs1801252 had a 24% larger RMS_1min than subjects without any minor allele, with heterozygous subjects having intermediate values (adjusted P = 0.005).

Fig. 3.

Root mean square residuals at 1 min after exercise (RMS_1min) and ADRA2A Asn251Lys, ADRA2C rs13118771, and ADRB1 Ser49Gly genotypes. The diamonds and the error bars represent means and SE, respectively.

None of the other polymorphisms or other confounders was associated with RMS_1min before and after adjustment (all P > 0.34).

DISCUSSION

To our knowledge, this is the first study performed under highly controlled conditions to focus on the genetic contribution to HRR after exercise in healthy subjects. Our findings provide evidence that some genetic variants in the sympathetic pathway may be associated with HRR. A variant in ADRA2B was associated with slower HRR, while variants in ADRB1, ADRA2A, and ADRA2C were associated with higher RMS_1min, a marker of increased parasympathetic tone. However, associations were generally weak, and the variants tested explained only a small proportion of the interindividual variability in HRR measures.

Several studies have explored the genetic contribution to cardiovascular parameters during exercise (15, 41); however, little is known about the genetic contribution to HRR after exercise. A cross-sectional population study in the Framingham Offspring cohort examined the association of 235 variants in 14 candidate genes with cardiovascular parameters (blood pressure and heart rate) during exercise and recovery (15). HRR was weakly associated with a variant in the α_1B-AR. However, HRR was assessed by only one parameter (heart rate at 3 min after exercise). Moreover, candidate genes were selected focusing on regulation of blood pressure rather than of heart rate, and only two of the candidate genes in our study (encoding the β₁- and β₂-ARs) were included.

In our study, a number of variants in α₂-ARs, important regulators and mediators of sympathetic tone, were associated with HRR. A common deletion variant in the α_2BAR was associated with a lower recovery constant and thus slower HRR after exercise. α_2BAR is the major α₂-subtype mediating vasoconstriction in mice and most likely in humans (5, 17). The ADRA2B del301-303 variant encodes a receptor that is markedly resistant to agonist-induced desensitization in vitro (43) and in vivo (29), and this variant has been associated with higher resting sympathetic tone in healthy humans (47). Thus, deletion allele carriers would be expected to have a higher sympathetic tone, compatible with our observation of slower HRR after exercise.

A number of genetic variants in ARs (e.g., in ADRA1A, ADRA2C del322-325, and ADRB2) (27, 28) have been associated with HRV at rest, but little is known about the effect of genetic variants on parasympathetic tone during recovery from exercise. We found three variants in adrenergic receptors [ADRB1 rs1801252 (Ser49Gly), ADRA2A rs1800035 (Asn251Lys), and ADRA2C rs13118771] to be associated with RMS_1min, a measure of parasympathetic tone, during the first minute of recovery. Parasympathetic and sympathetic pathways are directly linked (e.g., through inhibition of norepinephrine release from sympathetic neurons through presynaptic muscarinic receptors) and interact closely in the regulation of heart rate (2), but the functional level at which AR variants could affect vagal tone during HRR is unclear. Interestingly, in a previous study the Ser49 allele of ADRB1 (rs1801252) was associated with less heart rate reduction in response to the beta-blocker atenolol in healthy subjects (21). α_2A-ARs located in the central nervous system and presynaptically on sympathetic nerve terminals are regulators of both the sympathetic and the parasympathetic tone. Their activation results in inhibition of norepinephrine release into the synaptic cleft and thus a reduction in sympathetic tone (17, 38). In addition, stimulation with clonidine, an agonist at the α₂-ARs, elicits a profound increase in cardiac parasympathetic activity (48).

None of the CHT variants were associated with any of the HRR markers. However, the prevalence of the nonsynonymous rs1013940 variant, previously associated with a reduction of 50% in choline transport and the diagnosis of attention-deficit hyperactivity disorder (10) and major depression (12), was low in our cohort (5%), and none of our subjects was homozygous for the variant allele.

The clinical significance of interindividual differences in HRR and parasympathetic tone has been explored by several recent clinical trials showing that these measures (assessed by ΔHR and RMS) are independent predictors of cardiovascular mortality (6, 7, 16, 33, 40, 42, 49). We studied healthy, comparatively young subjects; therefore, our findings cannot be automatically extrapolated to patients with cardiac disease. Future studies exploring the association of the genetic variants identified with HRR and cardiovascular morbidity and mortality in patients with heart disease will be of interest.

Our study had some limitations. Peak hearts were relatively low because we did not study HRR after maximum work capacity was reached and because bicycle ergometry produces lower heart rates when performed supine than when performed upright with the same exercise intensity (37). However, using the recovery constant k_r allows the assessment of HRR even without reaching maximal heart rate, and k_r is more reliably derived after submaximal compared with maximal exercise (39). The amount of self-reported physical exercise was only marginally associated with measures of HRR in our study. However, the majority of our subjects performed little or no physical exercise, and thus the range of physical fitness represented was small, reducing our chances to detect an association. Our study was exploratory, and in view of our sample size we did not perform statistical corrections for multiple comparisons. With this in mind, a post hoc power analysis revealed that our sample size (n = 126) provided 91% power to detect a mean difference of 0.006 in the recovery constant k_r (corresponding to 75% of its SD) between carriers and noncarriers of a polymorphism with an MAF of at least 0.1 (α = 0.05). Similarly, our study design had the advantage of studying a relatively homogenous cohort under closely controlled conditions, but the relatively small sample size limited power to detect small genotype effects, especially for less common variants. Thus, we cannot exclude the possibility that some of the variants had smaller effects on the outcomes that we did not detect. The sample size also provided only limited power for separate race-specific analyses. However, sensitivity analyses in racial subgroups generally showed effects of similar magnitude and similar trends when the variants were prevalent enough in both races. Moreover, combined analyses of genetically diverse ethnic subgroups may be helpful in identifying truly causal variants rather than markers merely associated with causal variants by linkage disequilibrium (50). In this preliminary study, our objective was to identify markers that substantially contribute to the population variability in HRR; we therefore studied only common (MAF > 5%) or functional polymorphisms in candidate genes associated with sympathetic and cholinergic pathways. Substantially larger cohorts would be necessary to fully cover genetic diversity in these candidate genes (by haplotype analyses) or for a noncandidate driven, systematic approach (e.g., genome-wide association studies).

In conclusion, this exploratory, candidate-gene approach to the study of genetic contribution to HRR provides evidence for an association of variants in ARs with measures of HRR and parasympathetic tone after exercise. However, most of the interindividual variability in HRR remained unexplained. Future studies in larger samples are necessary to validate our results, define additional genetic markers, and extrapolate our findings to patients with heart disease in order to examine the predictive value of the identified markers for cardiovascular morbidity and mortality.

GRANTS

This study was supported by Vanderbilt Clinical and Translational Science Award Grant 1 UL 1 RR-024975 from the National Center for Research Resources and National Heart, Lung, and Blood Institute Grant P01 HL-56693. C. M. Stein is the recipient of the Dan May Chair in Medicine.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: U.K., A.D., P.J.K., M.K.H., B.A.E., R.D.B., C.M.S., and D.K. conception and design of research; U.K., M.M., G.G.S., and D.K. performed experiments; U.K., A.D., and D.K. analyzed data; U.K., A.D., P.J.K., C.M.S., and D.K. interpreted results of experiments; U.K. prepared figures; U.K. drafted manuscript; U.K., A.D., P.J.K., M.M., G.G.S., M.K.H., B.A.E., R.D.B., C.M.S., and D.K. edited and revised manuscript; U.K., A.D., P.J.K., M.M., G.G.S., M.K.H., B.A.E., R.D.B., C.M.S., and D.K. approved final version of manuscript.