Effects of Alcohol on Sympathetic Activity, Hemodynamics, and Chemoreflex Sensitivity

We studied 16 normal male subjects aged 26±4 (SD) years. None was taking any medications. Informed written consent was obtained from all subjects. The study was approved by the Institutional Human Subjects Review Committee.

Systolic, diastolic, and mean BPs were measured every minute with a sphygmomanometer (Lifestat 200, Physio Control Corp). Electrocardiogram, respiration (pneumograph), oxygen saturation (Nellcor N-100 C Pulse oxymeter), and end-tidal CO₂ (Hewlett-Packard 47210A Capnometer) were recorded on a Gould 2800 S recorder. Ventilatory rate and minute ventilation were determined with a Bourns LS-75 monitor. Subjects breathed through a mouthpiece with a nose clip to ensure exclusive mouth breathing. Forearm blood flow was measured by venous-occlusion plethysmography. An air-filled latex cuff of known volume and compliance was placed around the midforearm and connected to a low-pressure transducer (Validyne). Venous outflow from the forearm was occluded intermittently while hand blood flow was excluded with a wrist cuff inflated to suprasystolic pressures. The rate of increase in forearm volume enclosed by the bladder during intermittent occlusion of venous outflow was taken as the rate of blood flow into the enclosed tissues. SNA to muscle was recorded continuously by obtaining multiunit recordings of postganglionic sympathetic activity to muscle, measured from a nerve fascicle in the peroneal nerve posterior to the fibular head as described previously.^{12 13} Electrical activity in the nerve fascicle was measured with tungsten microelectrodes (shaft diameter, 200 μm, tapering to an uninsulated tip of 1 to 5 μm). A subcutaneous reference electrode was inserted 2 to 3 cm away from the recording electrode, which was inserted into the nerve fascicle. The neural signals were amplified, filtered, rectified, and integrated to obtain a mean voltage display of SNA.

The randomized, double-blind, and vehicle-controlled study design had two experimental sessions: a vehicle session and an alcohol session. Both were performed at the same times on 2 separate days. All studies were performed with subjects in the fasting state to avoid changes in muscle SNA and in alcohol absorption induced by food intake.¹⁴ All subjects voided immediately before the studies. A condom urinary catheter was then placed on all subjects for voiding as needed. Subjects were fed and hospitalized in the Clinical Research Center for approximately 4 hours after alcohol intake, after which they were driven home.

The protocol used for determination of chemoreflex responses to hypoxia and hypercapnia was identical to that used in previous studies.^{15 16} Measurements were taken during a 5-minute baseline period of stable ventilation while subjects breathed room air. Peripheral chemoreflex activation was then achieved by having subjects breathe an isocapnic hypoxic gas mixture (10% O₂ in N₂ with CO₂ titrated to maintain isocapnia; n=10) for 5 minutes or a hypercapnic hyperoxic gas mixture (7% CO₂/93% O₂; n=6) for 5 minutes.

After 5 minutes of baseline measurements, followed by measurement of chemoreflex sensitivity (hypoxia or hypercapnia), alcohol (1.0 g/kg body wt, diluted in 400 mL water) or vehicle (400 mL water) was administered orally over 30 minutes. A flavoring (Crystal Light) was added to these solutions to prevent the subjects from distinguishing the alcohol from the vehicle session. Five-minute baseline recordings were performed 45 minutes after the beginning of the ingestion of either alcohol or vehicle. The identical protocol (ie, 5 minutes of baseline followed by 5 minutes of hypoxia or hypercapnia) was then repeated 60 minutes after the beginning of alcohol or vehicle intake. On average, the recordings continued until 85 minutes after the beginning of vehicle or alcohol consumption. A loss of sympathetic nerve recording occurred during one of either the vehicle or alcohol sessions in six subjects. Consequently, we completed technically excellent studies examining the effects of alcohol and vehicle on muscle SNA in 10 subjects. Eight of these complete recordings of muscle SNA were obtained during studies of the chemoreflex responses to hypoxia, and two were obtained during studies of chemoreflex responses to hypercapnia. Because of a consistently lower oxygen saturation during hypoxia after alcohol compared with vehicle, we carried out an additional study of measurements of oxygen saturation and arterial Po₂ during hypoxia before and after alcohol in six subjects. The purpose of this study was to determine whether alcohol decreased the partial pressure of oxygen or whether alcohol altered oxygen saturation at a given level of Po₂. In these six subjects, an arterial line was placed in the radial artery. Arterial blood was obtained for Po₂ measurements at baseline (room air) and at minutes 3 and 5 of hypoxia. Arterial blood gas measurements were repeated during room air breathing and at minutes 3 and 5 of hypoxia after alcohol intake at 1.0 g/kg.

Sympathetic bursts were identified by a careful inspection of the mean voltage neurogram, and SNA was calculated as bursts per minute. The amplitude of each burst was determined, and SNA was calculated as bursts per minute multiplied by mean burst amplitude and expressed as a percentage of change from baseline. Measurements were made by a single observer (P. van de B.) in a blinded fashion. The intraobserver and interobserver variabilities in our laboratory are 4.3±0.3%¹³ and 5.4±0.5%.¹⁷ FVR was calculated by dividing mean arterial pressure by blood flow and is expressed in arbitrary units.

The effect of alcohol on HR variability was determined 45 minutes after vehicle and alcohol ingestion in 16 subjects. During this period, the subjects were at rest and breathing freely. Measurements obtained 45 minutes after alcohol ingestion were compared with those obtained on a separate day 45 minutes after vehicle ingestion. Analog-to-digital conversion was performed over 10 minutes at 600 samples per second for both the electrocardiogram and respiratory signals. Data were then analyzed off-line with a personal computer (IBM 433DX/T). Analysis of HR variability was also carried out in a blinded fashion. The principles of the software for data acquisition and autoregressive spectral analysis have been described elsewhere.^{18 19 20} Stationary segments devoid of arrhythmias (150 to 300 RR intervals) were analyzed with autoregressive algorithms. These algorithms provide the number, center frequency, and power of the oscillatory components. Akaike’s test and Anderson’s test^{18 19 20} allowed determination of the optimal model order fitting the data and verified that all information contained in the time series had been extracted in the computation. Previous studies¹⁸ have shown that two major oscillatory components are usually detectable in short-term RR interval variability. One of the oscillatory components is synchronous with respiration and is called high-frequency oscillation. The other component is described as low-frequency oscillation and has a center frequency of about 0.10 Hz, which can vary considerably (from 0.04 to 0.15 Hz).^{18 19 20} The ratio of the low-frequency to the high-frequency oscillations has been suggested as an index of sympathetic-vagal modulation to the heart.^{18 19 20} Variability data could not be analyzed in two subjects because of technical problems during one of the sessions. In addition, the respiratory variability had a significant low-frequency component during both alcohol and vehicle sessions in three subjects (0.07 and 0.04 Hz in subject E.B., 0.06 and 0.08 Hz in subject J.H., and 0.05 and 0.08 Hz in subject L.G., respectively). The presence of low-frequency components in respiration prevented interpretation of the ratios of low to high frequency because of overlap of the respiratory oscillation with the low-frequency oscillation. Consequently, we obtained meaningful data on the effects of both alcohol and vehicle on HR variability in 11 subjects.

Statistical analysis was performed by an independent statistician and consisted of a repeated measures ANOVA. A single contrast determined (1) whether alcohol affected the baseline measurements differently than vehicle, and (2) whether alcohol and vehicle affected the responses to the interventions (hypoxia and hypercapnia) differently. Comparisons of the results obtained by spectral analysis were performed with Student’s paired t tests (two-tailed). Correlations were estimated with the Pearson coefficient. Significance was assumed at a value of P<.05.

Plasma alcohol was undetectable during the vehicle sessions. During the alcohol sessions, plasma alcohol increased from 0 to 23.2±1.5 mmol/L (107±7 mg/dL) at 60 minutes after alcohol intake and 20.2±1 mmol/L (93±4 mg/dL) at 85 minutes after alcohol intake (P<.0001).

Resting cardiovascular and respiratory measurements were obtained at the start of the study and at 45, 60, and 80 minutes after alcohol and vehicle intake. Mean BP did not differ significantly between the alcohol and vehicle sessions throughout the study (Table, Fig 1). Alcohol increased baseline HR from 59±2 to 65±2 beats per minute at 45 minutes (P<.0001). This increase in HR by alcohol was accompanied by an increase in the ratio of low-frequency to high-frequency HR variability (6.3±2.5 during alcohol versus 1.4±0.4 during vehicle, P<.05) (Fig 2). The increase in HR induced by alcohol persisted throughout the study (Fig 1).

Alcohol also elicited a marked increase in SNA (Table, Figs 1 and 3). SNA increased from 22±2 to 34±3 (P<.0001), 33±3 (P<.001), and 35±3 (P<.01) bursts per minute at 45, 60, and 80 minutes, respectively, after alcohol intake. Furthermore, alcohol also increased the total amplitude of SNA by 184±20% (P<.0001), 200±20% (P<.01), and 239±22% (P<.01) 45, 60, and 80 minutes after alcohol intake (Table, Fig 1). The magnitude of increase in the total amplitude of SNA was positively correlated with the magnitude of increase in plasma alcohol (r=+.60, P=.007, after correction for the effects of vehicle). No other correlations were found between changes in cardiovascular measures and plasma alcohol levels.

FVR did not differ at 45 minutes after alcohol intake compared with vehicle (Table). FVR decreased from 35±4 U at baseline to 33±4 at 60 minutes and 33±4 at 80 minutes (both P<.05 compared with the effects of vehicle).

Alcohol did not affect minute ventilation (Figs 4 and 5), oxygen saturation, or end-tidal CO₂ while subjects breathed room air during the baseline sessions.

Ventilatory and sympathetic responses to both hypoxia (Fig 4) and hypercapnia (Fig 5) were not affected by either alcohol or vehicle. However, although alcohol did not affect the ventilatory responses to hypoxia (Fig 4), there was a modest but significantly greater reduction in oxygen saturation during hypoxia after alcohol. Oxygen saturation during hypoxia after alcohol was 4±1% lower than it was during hypoxia after vehicle (P<.05) (Fig 4), despite identical levels of inspired oxygen and similar ventilatory responses.

To determine whether a reduced affinity of hemoglobin for oxygen or a reduction in arterial blood gas oxygen tension with alcohol was responsible for this observation, we repeated the study in six subjects in whom three arterial blood gas samples were taken, at baseline and at minutes 3 and 5 of hypoxia, before and after alcohol. This study revealed that during hypoxia, arterial blood gas tension in oxygen averaged 66±5 mm Hg before alcohol and 66±6 after alcohol (P=NS). In contrast, oxygen saturation during hypoxia, averaged 89±2% before alcohol and 86±2% after alcohol(P=.0001). Thus, alcohol decreased the affinity of hemoglobin for oxygen but did not affect the partial pressure of oxygen, the actual stimulus sensed by the chemoreceptors. Blood pH before alcohol averaged 7.411±0.005 and after alcohol, 7.401±0.004 (P=.01). Base excess was 0.761±0.14 before alcohol and 0.144±0.29 after alcohol (P=.007). Alcohol did not affect the ventilatory response to hypoxia (minute ventilation increased by 325±52 mL per decrease in millimeters of mercury of oxygen before alcohol and by 336±92 mL per decrease in millimeters of mercury of oxygen after alcohol; P=NS).

Alcohol did not affect the BP, HR, and muscle SNA responses to hypoxia or hypercapnia.

Hypoxia and hypercapnia, acting primarily via the peripheral and central chemoreceptors, respectively, elicit increases in muscle SNA, BP, and minute ventilation.^{15 16} Alcohol did not affect the ventilatory, sympathetic, or pressor responses to hypoxia or hypercapnia in our healthy volunteers. Our data support those of Krol et al,⁷ who suggested that alcohol does not alter ventilation during room air breathing and hypercapnic rebreathing. However, our study suggests that alcohol decreases the affinity of hemoglobin for oxygen, resulting in lower oxygen saturation. Alcohol does not affect the partial pressure of oxygen, the stimulus sensed by the peripheral chemoreceptors. Thus, similar ventilatory and sympathetic responses to hypoxia during alcohol and vehicle indicate that alcohol does not depress peripheral chemoreflex sensitivity. Hence, chemoreflex effects of alcohol are unlikely to explain the exacerbation of obstructive sleep apnea after alcohol ingestion.^{1 2 3 4 5}

Effects of alcohol and vehicle on muscle SNA responses to hypercapnia were able to be measured on both occasions in only two subjects. Minute ventilation during hypercapnia was measured in all six subjects on both occasions and did not change after alcohol. Since the predominant response to hypercapnia is an increase in minute ventilation,¹⁶ it is unlikely that alcohol would have selectively changed the muscle SNA response to hypercapnia without affecting the ventilator response. This is further supported by the absence of any effect of alcohol on the pressor response to hypercapnia.

An additional limitation of this study is that responses were tested only in normal volunteers. Chemoreflex sensitivity in individuals with sleep apnea may differ from those in normal humans. We therefore cannot completely exclude the possibility that alcohol may affect chemoreflex sensitivity in individuals with sleep apnea.

Our study demonstrates the new finding that alcohol may affect the oxygen-hemoglobin dissociation curve. The reason for this effect is unknown. It is possible that the fall in blood pH, though very small, may be implicated. This observation may explain the association between chronic alcohol abuse and nocturnal hypoxemia defined as a fall in oxygen saturation.²¹ This effect of alcohol, evident during a significant hypoxic stress, may result in greater oxygen desaturation during sleep apnea, independent of any effect on the ventilatory response,²² and could result in an overestimation of the severity of sleep apnea after alcohol intake.

The exact mechanisms responsible for the association between alcohol and hypertension reported in epidemiological studies remain unclear.⁹ Previous studies have reported conflicting results regarding the acute effects of alcohol on BP, with studies reporting an increase,^{10 11} a decrease,^{23 24} or no changes²⁵ in BP after alcohol intake. Conflicting results were also reported when plasma catecholamines were measured after acute alcohol intake.⁹ However, in studies by Grassi et al¹⁰ and Randin et al,¹¹ oral¹⁰ and intravenous¹¹ alcohol administration induced a pressor response that was accompanied by an increase in efferent SNA to skeletal muscle. The latter study did not use a vehicle control. These studies^{10 11} have suggested that alcohol promotes hypertension through sympathetic activation. While a significant increase in BP was seen after alcohol in our study, a similar increase was seen after vehicle (Fig 1, P<.05 after 45, 60, and 80 minutes during both alcohol and vehicle sessions compared with baseline values).

Epidemiological studies have suggested that the J-shaped relation between alcohol intake and BP is due to a vasodilator effect at low doses but to a pressor effect at higher doses.^{9 26} The plasma levels of alcohol achieved in our study were approximately 1.6¹¹ to 2²¹ times higher than those reported in the two studies in which alcohol raised both SNA and BP. Moreover, greater increases in plasma alcohol levels resulted in larger rises in SNA in our study (r=+.60). Thus, one would have expected that BP would increase even more in our study than in the studies by Grassi et al¹⁰ and Randin et al.¹¹ The strengths of our study include the use of a vehicle control and simultaneous measurements of muscle SNA, FVR, and spectral analysis of HR as well as the avoidance of effects of bladder distention. In our study, alcohol increased muscle SNA traffic, HR, and the ratio of low- to high-frequency HR variability. Thus, our data are consistent with a cardiac sympathetic activation and an increased sympathetic drive to muscle circulation after oral alcohol ingestion. However, FVR increased with vehicle but not with alcohol, although alcohol increased sympathetic drive to muscle circulation. This finding suggests that the direct vasodilator action of alcohol on peripheral blood vessels opposed the vasoconstrictor effects of sympathetic activation induced by alcohol, thus limiting any rise in BP. In contrast, it may be that when lower levels of alcohol are achieved,^{10 11} the increased SNA overrides completely the direct vasodilator property of alcohol and induces an increase in BP.

Alcohol induces an acute diuresis,²⁷ and bladder distention increases both muscle SNA and BP.²⁸ The particular attention we paid in our study to avoid bladder distention could also have limited any pressor effect observed after alcohol intake.

These data indicate the following: (1) Alcohol induces sympathetic activation to the heart and to muscle blood vessels; the sympathetic vasoconstrictor activation is opposed by a direct vasodilator effect of alcohol. (2) Alcohol does not affect chemoreflex responses to hypoxia or hypercapnia; thus, alterations in chemoreflex sensitivity are unlikely to explain the worsening of obstructive sleep apnea by alcohol. (3) During significant hypoxic stress, alcohol is accompanied by greater levels of oxygen desaturation. The partial pressure of oxygen is unchanged, suggesting that alcohol may affect the characteristics of the oxygen-hemoglobin dissociation curve.