- Email: [email protected]
Alcohol increases sleep apnea and oxygen desaturation in asymptomatic men
Alcohol Increases Sleep Apnea and Oxygen Desaturation in Asymptomatic Men
VICENTE C. TAASAN,
M.D.*
A. JAY BLOCK, M.D. PHILIP G. BOYSEN, M.D. JAMES W. WYNNE. M.D. with the technical
assistance
of
CAROLE WHITE SHERRY LINDSEY Gainesville, Florida
From the Departments of Medicine and Anesthesiology, College of Medicine, University of Florida, and the Veterans Administration Medical Center, Gainesville, Florida. This study was supported by the Medical Research Service of the Veterans Administration, by Grant HL-22622-03 from the U.S. Public Health Service, by the Parker B. Francis Foundation, and by a Pulmonary Academic Award (K07 Hl 00122) from the National Heart, Lung, and Blood Institute. Requests for reprints should be addressed to Dr. A. Jay Block at the Pulmonary Division, Veterans Administration Medical Center, Gainesville, FLA 32602. Manuscript accepted March 4. 1981. Fellow, Parker B. Francis Foundation. l
240
August 1981
Using standard sleep techniques, we performed a placebo-controlled and randomized study to assess the effect of alcohol ingestion (2 ml/kg of body weight) on breathing and oxygen saturation during sleep. Twenty asymptomatic men volunteered for the two-night study: 11 were given a placebo on night 1, and alcohol on night 2 (group A); nine were given alcohol on night 1 and a placebo on night 2 (group B). We compared the incidence of sleep events (apnea, hypopnea and arterial oxygen desaturation) during the nights the subjects received alcohol and during the nights they received the placebo. Alcohol was associated with significant increases in the occurrence of the following: the number of sleep events (207 to 383, p
The American Journal of Medicine
Volume 71
ALCOHOL INCREASES SLEEP APNEA-TAASAN
MATERIALS
AND METHODS
Subjects. We recruited 20 healthy male volunteers for the study, the initial criterion being the subject’s willingness to undergo a two-night sleep study. They were then screened for abnormal sleep patterns and cardiorespiratory complaints. Subjects with a major illness or a history of alcohol or drug dependence were excluded. The study was approved by the institutional committees for the protection of human subjects, and informed consent was obtained from all participants. General Procedure. Participants were advised to eat light meals on both study nights. Subjects came to the laboratory 1 l/2 hours before their usual bedtime to give them sufficient time to become familiar with the laboratory environment and experimental procedures. The forced vital capacity and forced expiratory volume in 1 second (FEV,) were measured with a Vitalor@ spirometer. We gave the subjects 100 proof vodka in orange juice, and the amount used (2 ml/kg of body weight) was similar to that utilized in previously reported studies [ 2-51. Orange juice alone was the placebo. Eleven subjects were given the placebo on the first night and alcohol on the second (group A). The reverse was true for the other nine subjects (group B). Alcohol and control drinks were consumed within 1 hour. We used an Intoxilyzefl (CMI, Inc., Minturn, CO) to estimate blood alcohol concentrations before sleep on both study nights and also to measure residual concentrations when the subjects awakened the morning after drinking alcohol. Alcohol concentrations before sleep were determined 30 to 45 minutes after the subjects’ last intake of alcohol. Sleep Techniques. Methods used in this study have been reported [g-12]. All variables were continuously measured: oxygen saturation was determined with an ear oximeter (Hewlett-Packard 47201A); oral and nasal air flows were sensed with thermistor clips on one lip and one nostril (Grass Instruments); and chest movement was monitored by impedance pneumography with surface electrodes and pneumograph couplers (Narco Biosystems). All of the foregoing were also continuously recorded (Narco Biosystems Physiograph DMP-4P 4-channel recorder). Electroencephalograms, electro-oculograms and electrocardiograms were simultaneously recorded on a separate recording system (Grass Model 79D polygraphic recorder). Methods of electroencephalograms and sleep stage scoring were based on those of Agnew and Webb [ 131. A technician, who was present throughout the study to attend to the subject’s needs and to assure adequate recordings, noted spurious events (e.g., detached leads, clogged thermistors) and correlated them with the recordings. Definitions. Apnea was defined as a cessation of oral and nasal flows for 10 seconds or longer. Hypopnea was noted when desaturation occurred simultaneously with decreased airflows and chest wall movement. Arterial oxygen desaturation was defined as a L4 percent decrease in oxygen saturation. Sleep-disordered breathing was defined as any occurrence of apnea or hypopnea or both. An event was any occurrence of apnea, hypopnea or desaturation. Sleep-period time was defined as the interval in minutes, from the onset of stage 1 sleep to full awakening the following morning. Experimental Design and Analysis. According to a preestablished code, each subject was randomly assigned to group
ET AL.
A or B. We attempted a single blind study, but the subjects immediately identified the drink, and the breathalyzer maneuver informed the sleep technician of the identity of the drink. Prior to breaking the codes, we analyzed our records in the following manner: Recordings were inspected visually. The number and the duration of desaturation and disordered breathing were recorded. For each night, the lowest saturation, the maximum decrease in saturation and the longest duration of each of the events were recorded for each subject. To control for varying lengths of sleep, the total time (in minutes) of desaturation or disordered breathing was counted and expressed as a percentage of sleep-period time. The frequency of events per hour of sleep-period time was also calculated. The effect of alcohol on sleep structure was analyzed by measuring the duration of the sleep-period time and calculating the percentage of sleep-period time for each stage of sleep. Statistical differences between the nights alcohol was injested (alcohol nights) and the placebo (control night) in terms of desaturation and breathing variables were sought by obtaining the difference between night 2 and night 1 values for each subject. The groups were then compared by the Wilcoxon rank sum test [ 141 using these differences. This nonparametric method was used because the presence of “outlier” values would invalidate usual parametric methods. To determine the residual effects of alcohol, we used a similar analysis to compare the findings on placebo nights that preceded the ingestion of alcohol (group A) with those that followed alcohol ingestion (group B). The effect of alcohol on sleep stage distribution was analyzed using the Hotelling t test [IS]. Correlations between age, weight to height ratios, blood alcohol concentrations and sleep events were determined using Kendall’s Tau [ 161. A correlation was declared significant when it achieved p <0.05 for both nights. The Statistical Analysis Systems (SAS) computer package was used for the data analysis. RESULTS
Eleven of 20 subjects were given alcohol on night 2 (group A, prealcohol control) and nine subjects were given alcohol on night 1 (group B, postalcohol control). Table I shows the anthropometric, spirometric and TABLE I
Anthropometric, Splrometric and Blood Alcohol Concentrations in Two Groups of Normal Men* GroupA
Subjects (no.) Age (yr) Weight (kg) Weight:height (kg/cm) Forced vital capacity (liters) Forced expiratory volume in 1 second (liters/set) Blood alcohol concentration
48.0 78.0 0.43 3.98 3.27
GroupB
11 f 4.5 f 2.8 f 0.01 f 0.27 f 0.21
80.0 f 6.8
48.7 78.6 0.45 4.50 3.52
9 f f f f f
5.4 3.9 0.01 0.32 0.23
87.7 f 8.1
(mgld)) NOTE: Values given are the mean f standard error. * Alcohol was given on night 2 to patients in group A and on night 1 to patients in group B.
August 1981
The American Journal of Medicine
Volume 71
241
ALCOHOL INCREASES SLEEP APNEA-TAASAN
TABLE II
ET AL
TABLE Ill
Arterial Oxygen Desaturation in Two Groups of Normal Men (Group A, 11 Subjects; Group B, 9 Subjects)*
Breathing Variables During Alcohol and Control Nights Group A
GroupA Night 2 Night 1 (alcohol) (placebo) Desaturation (no. events) Desaturation (total duration, min) Desaturation (no./hr) Lowest saturation (%) Longest desaturation (set)
2.6 f 2.5
9.6 f 4.5
0.4 f 0.4
2.5 f
0.4f0.3
1.5f0.7
93 f
4.0 f
1.0
1.5
88 f 2.0
2.8 26.1 f
Group B Night 2 Night 1 (placebo) (alcohol) 13.4 f 6.2 10.0 f 5.6 3.3 f 2.5
2.6 f
2.5f1.2
1.6f0.7
87 f 3.0
1.2
88 f 2.0
12.2 19.1 f 9.6 20.8 f 8.4
NOTE: All values are mean f standard error. ’ p values obtained by ranks sum test are shown in the text.
blood alcohol concentrations of the 20 subjects. Four subjects were 20 years of age or younger, six were 60 years or older. The oldest subject was 65 years of age. Only one subject weighed more than 100 kg (102 kg). Concentration of alcohol in blood before sleep was 1100 mg/dl in seven subjects; the lowest concentration was 40 mg/dl, the highest, 120 mg/dl. In four subjects, 10 to 20 mg/dl of alcohol were detectable after sleep on the nights alcohol was given. Spirometric values in all subjects were within normal limits for our laboratory. Only three subjects were smokers. Alcohol Versus Control Nights. Oxygen saturation: The total number of desaturation events was significantly greater during sleep following alcohol ingestion (p
242
August 1981
Night 1 (placebo) SDB (no. events) SDB (total duration, min) SDB frequency (no./hr SPT) Apnea (no. events)
Night 2 (alcohol)
Group B Night 2 Night 1 (placebo) (alcohol)
2.4 f 1.9 5.4 f 3.0 10.7 f 7.3 6.9 f 5.6 0.8 f 0.6 3.4 f 1.7 3.8 f 2.9 3.3 f 2.8 0.3 f 0.2 0.9 f 0.4
2.0 f
1.2 f 0.8 2.6 f
9.0 f 7.2 0.7 f 0.5
1.3
1.4 1.2 f 0.9
NOTE: SDB = sleep disordered breathing, SPT = sleep-period time. All values are mean f standard error. p values obtained by ranks sum test are shown in the text.
Sleep-disordered breathing: The number (p = 0.03), total duration (p = 0.05) and frequency (p = 0.02) of sleep-disordered breathing were significantly greater after alcohol ingestion. All those who had disordered breathing on control nights (nights they received the placebo) showed increases in the number and duration of disordered-breathing events after the ingestion of alcohol. Apnea was more common than hypopnea with alcohol, and 11 subjects had at least an apneic event with alcohol. Only four subjects were noted to have apnea during control nights. A total of 110 apneic events were counted during alcohol nights (nights they received alcohol); on the control nights, 20. This difference was statistically significant (p KO.01). There were 157 events of either apnea or hypopnea during alcohol nights, 89 during control nights. For descriptive purposes, Table Ill shows the means and standard errors of different breathing variables. Total Number of Sleep Events. Table IV shows the total number of sleep events and the number of subjects who had them. In 18 of the 20 subjects, desaturation or disordered-breathing with alcohol occurred at least once. In nine, it did so with the placebo. With alcohol, the number of sleep events was significantly greater, and almost twice as many sleep events were recorded after the ingestion of alcohol (383) than after the intake of the placebo (207) (p
TABLE IV
Number of Events of Desaturation and Disordered Breathing and Number of Subjects with Sleep Events During Control and Alcohol Nights All Events
Alcohol nights No. subjects Control nights No. subjects p value
383 18 207 9 p
l
Desaturation 226 17 118 9 p
Apnea 110 11 20 4 p
Hypopnea 47 11 69 4 NS+
p values from rank sum test apply to number of events only. + Not significant. l
The American Journalof Medicine Volume 71
ALCOHOL
INCREASES
SLEEP
APNEA-TAASAN
ET AL.
60
CHEST MOVMNT b---
50 40
NOSt-_. AIRFLOW
: _ cs
,--_~,-_ ___~_
30 20 10 0i
STAGE P VALUE
Figure 1. Polygraphic tracing showing a 46-second event of profound desaturation with cessation of airflow and diminution of chest movement. The subject was a 59 year old asymptomatic man who had 126 events of desaturation or abnormal breathing in a single night’s sleep.
more common event. In addition, the number of apneic events and the number of sleep events per hour of sleep (frequency) were significantly greater with alcohol (p
TABLE V
Arterial Oxygen Saturation and Breathing Variables During Prealcohol and Postalcohol Control Nights for Two Groups of Subjects
Variable Desaturation (no. of events) Desaturation (total duration) Desaturation (no./hr SPT) Disordered breathing (no. events) Disordered breathing (no./hr SPT)
GroupA (Prealcohol)
GroupB (Postalcohol)
2.6 0.4 0.4 2.4 0.3
10.0 2.6 1.6 6.9 1.2
f f f f f
2.5 0.4 0.3 1.9 0.2
f f f f f
5.6 1.2 0.7 5.6 0.9
NOTE: SPT = sleep-period time-the interval between the onset of stage 1 sleep and full awakening. All values are mean f standard error.
0
I
NS
< 01
2
3
4
R
NS
NS
NS
NS
rfgure z. c‘omparrson of sleep-stage percentages for placebo and alcohol nights for both groups. No significant differences were noted except for stage 1, which in group A was significantly lower during alcohol nights. P values for differences between sleep-stage percentages for period A and group B control nights are given in the text. SPT = sleep period time, the interval, in minutes, between the onset of stage 7 sleep and full awakening. Group A subjects were given alcohol on night 2, Group B subjects were given alcohol on night 1.
whereas it occurred in seven of the nine subjects in group B. The maximum change in saturation was greater (p = 0.02) and the lowest saturation (p = 0.02) was also significantly lower in group B. However, no significant differences in disordered breathing were noted for the two groups. Table V shows descriptive data for groups A and B for both desaturation and breathing events, indicating the persistence of desaturation on the control nights that followed alcohol nights. Distribution of Sleep. Figure 2 shows the distribution of sleep during control and alcohol nights. All non-REM and REM stages of sleep were identified in all subjects. For both nights, sleep-period time was shorter than previously established normal values [ 171. Between the two nights, the.sleep-period time for all subjects showed no significant change. The average sleep period times of both groups for both nights ranged from 3 16 to 344 minutes. The shortest sleep time was 3 hours, the longest was 7 hours and 20 minutes. In addition, analysis of the latencies of stage 2, slow wave sleep and REM sleep showed no significant changes with alcohol. Sleep-stage percentages during postalcohol control nights showed significant differences from prealcohol control nights. Sleep stages 0 (p = 0.03) and 1 (p = 0.01) were significantly greater during postalcohol control nights, but sleep stage 4 was significantly reduced (p = 0.04). Stages 2 and REM sleep showed no significant change with alcohol. Effect of Age, Weight to Helght Ratios and Concentrations of Alcohol in Blood. Correlation between
August 1991
The American Journal of Medlclne
Volume 71
243
ALCOHOL INCREASES SLEEP APNEA-TAASAN
ET AL.
sleep events during alcohol nights and age, weight to height ratios and concentrations of alcohol in blood showed that increased age had a positive relation to sleep events. Increased age correlated positively with the number of events of disordered breathing and desaturation (p = 0.05 and 0.03), respectively, and with the duration of disordered breathing (p = 0.03). For control nights, a correlation was noted between increasing age and low saturation for group A. The number of desaturation events also correlated with age in group B. Weight to height ratios and blood alcohol concentrations did not statistically correlate with sleep events on either night. COMMENTS This study confirms our previous findings that asymptomatic men tend to have oxygen desaturation and disordered breathing during sleep. In addition, we have found that alcohol ingestion is associated with an increased frequency and severity of these sleep events, and this increase persists through an additional night after alcohol intake. In this study nine of the 20 men had 207 sleep events during alcohol-free nights. In a previously reported study [l] 20 of 30 male subjects had 264 events. Other similarities were noted between the types of sleep events in this and the previous study: the most common event was arterial oxygen desaturation, and the least common was apnea. In this study, control night saturations were significantly lower in a group of subjects as age advanced, and elderly subjects had more events during sleep after alcohol ingestion. Because no subject in this study was obese, we cannot comment on the previously reported effect of weight on the frequency of sleep events [ 11. Notably, the simultaneous occurrence of desaturation and an altered breathing pattern was frequent with alcohol. Inspection of our records shows that of 226 episodes of desaturation during alcohol nights, 114 were accompanied by a simultaneous occurrence of apnea or hypopnea. The three lowest levels of saturation were all accompanied by disordered breathing. Of the 110 apneic events with alcohol, 71 were associated with an episode of desaturation. Several episodes of desaturation were further noted to be associated with at least some decreases in air flow or chest movement, but they failed to meet our predetermined criteria and were not counted as disordered breathing. Not all episodes of desaturation, however, were accompanied by an episode of disordered breathing. Thus, the sleep events with alcohol in this study were related to alterations of ventilatory patterns, which on most occasions were accompanied by episodes of desaturation. The increased incidence and severity of oxygen desaturation and sleep-disordered breathing after alcohol ingestion could not be explained by differences in the
244
August 1981
The American Journal of Medicine
length of sleep or by an increase in stages 2 and REM sleep. This does not mean that the state of sleep itself is not involved in the occurrence of more sleep events with alcohol. On the contrary, we believe that sleep has a permissive role in these events. The persistence of desaturation on the placebo control night after alcohol ingestion is a surprising finding. Although its significance is unknown to us, a similar persistence of alcohol effects has been reported for catecholamine metabolism [ 181. Brohult et al. [ 181 showed that catecholamine excretion is increased for an entire week after a single large dose of ethanol. In a study of sleep in young adults, Rundell et al. [5], used a dose similar to that in our study, showed that on the night after alcohol ingestion sleep latency increased and REM latency decreased [ 51. Thus, as has been noted, alcohol ingestion has sustained effects that persist beyond the measurable presence of alcohol in the blood. Data have often shown that moderate doses of alcohol have a potential for respiratory depression. These data were gathered in the laboratory setting, and the practical implications of these findings are unclear. More sensitive tests, e.g., ventilatory response to oxygen and carbon dioxide, reveal that a moderate dose of alcohol also depresses chemosensitivity [ 19,201. Furthermore, direct application of alcohol to brain stem structures in the cat has been shown to cause apnea [21]. “Respiratory arrhythmia” [ 191 has been reported in awake subjects given moderate doses of alcohol, and episodic apnea has been reported in a child with alcohol intoxication [22], findings of great interest in relation to the fivefold increase in apnea in this study. A recent study has shown that small to moderate amounts of alcohol potentiate GABA-mediated inhibition of single cortical neurons in cats [ 231, analogous to the cortical depression caused by barbiturates and diazepines [23]. Thus, there is evidence that even with moderate doses of alcohol, respiratory chemosensitivity is depressed and ventilatory patterns are altered. On the other hand, some studies reveal that alcohol has inconsistent or in fact stimulatory effects on respiration [24,25]. Sahn et al. [26] have recently shown that ingestion of moderate doses of alcohol is not an important cause of hypoventilation in patients with chronic obstructive lung disease. It is worthwhile mentioning that their study was carried out with awake subjects. Thus, the practical implications of the laboratory data pointing to a potential for respiratory depression by small to moderate doses of alcohol remain unclear. In all of these studies, however, only the effects of alcohol ingestion on the awake subject were considered. The possibility exists that wakefulness overrides the depressant effects of alcohol. We know of no previous study in which respiratory behavior was observed
Volume 71
ALCOHOL INCREASES SLEEP APNEA-TAASAN
during sleep after a moderate dose of alcohol. In this study it has clearly been shown that the ingestion of 2 ml/kg of 100 proof vodka markedly increases the frequency and severity of nocturnal oxygen desaturation and sleep-disordered breathing in asymptomatic men. At this point, it is difficult to understand the practical implications of our findings. Asymptomatic subjects have not as yet been shown to suffer cardiovascular consequences of the brief episodes of apnea or desaturation. However, patients with chronic obstructive lung disease who ingest moderate amounts of alcohol
ET AL.
may suffer cardiopulmonary sequelae. Previous studies have shown that nocturnal oxygen desaturation is associated with cardiovascular deterioration in such patients [ 11,271. Because breathing abnormalities occur during sleep in patients with chronic obstructive lung disease [S-12], moderate doses of alcohol could be expected to accentuate concomitant arrhythmias and pulmonary hypertension. These findings in normal subjects thus open the way to research into how moderate amounts of alcohol affect breathing during sleep of patients with serious pulmonary and cardiovascular disease.
REFERENCES
2.
3.
4. 5.
6.
7. 8.
9.
10.
11.
12.
13.
Block AJ. Boysen PG. Wynne JW, Hunt LA: Sleep apnea, hypopnea and oxygen saturation in normal subjects, a strong male predominance. N Engl J Med 1979; 300: 513-517. Gresham SC, Webb WB, Williams RL: Alcohol and caffeine: effect on inferred visual dreaming. Science 1963; 140: 1266. Yules RB, Freedman DX, Chandler KA: The effect of ethyl alcohol on man’s electroencephalographic sleep cycle. Electroencephalogr Clin Neurophysiol 1966; 20: 109111. Knowles JB, Laverty SG, Keuchler HA: Effects of alcohol on REM sleep. Q J Stud Alcohol 1968; 29: 342-349. Rundell OH, Lester BK, Griffiis WJ, Williams HL: Alcohol and sleep in young adults. Psychopharmacologia (Berlin) 1972; 26: 201-218. Wynne JW, Block AJ, Hemenway J, Hunt LA, Flick MR: Disordered breathing and arterial oxygen desaturation during sleep in patients with chronic obstructive lung disease. Am J Med 1979; 66: 573-579. Lugaresi E: Snoring. Electroencephalqr Clin Neurophysiol 1975; 39: 59-64. Lugaresi E, Coccagna G, Cirignotta F: Snoring and its clinical implications. In: Guilleminault C, Dement WC, eds. Sleep apnea syndromes. New York: Alan R Liss, 1978; 13-2 1. Flick MR, Block AJ: Continuous in vivo monitoring of arterial oxygenation in chronic obstructive lung disease. Ann Intern Med 77; 86: 725-730. Boysen PG. Block AJ. Wynne JW, Hunt LA, Flick MR: Nocturnal pulmonary hypertension in patients wlth chronic obstructive pulmonary disease. Chest 1979; 76: 536-542. Wynne JW, Block AJ, Hunt LA, Flick M(: Disordered breathing and oxygen desaturation during daytime naps. Johns Hopkins Med J 1978; 143: 3-7. Flick M1, Block AJ: Continuous in vivo measurement of arterial oxygen saturation by oximetry. Heart Lung 1977; 6: 990-993. Agnew HW Jr, Webb WB: Sleep stage scoring (American
14. 15. 16. 17.
18. 19. 20.
21.
22. 23.
24. 25. 26.
27.
Augud lggl
Psychological Association Abstract Series, Manuscript No. 293). Washington, DC, American Psychological Association, 1972. Hollander MD. Wolfe DA: Nonparametric statistical methods. New York: John Wiley and Son, 1973; 68-75. Morrison DF: Multivariate statistical analysis. New York: McGraw-Hill, 1967; 145-146. Siegal S: Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill, 1956; 213-223. Williams RL, Karacan I, Hursch Cl: Electroencephalography (EEG) of human sleep: Clinical applications. New York: Wiley, 1974. Brohult J, Levi L, Richard H: Urinary excretion of adrenal hormones in man. Acta Med Stand 1970; 188: 5-13. Johnstone RE, Reier CE: Acute respiratory effects of ethanol in man. Clin Pharmacol Ther 1973; 14: 501-508. Sahn SA, Lakshminarayan S, Pierson DJ, Weil JV: Effect of ethanol on the ventilatory responses to oxygen and carbon dioxide in man. Clin Sci Mol Med 1974; 49: 33-38. Rosenstein R. McCarthy LE. Borison HL: Respiratory effects of ethanol and procaine injected into the cerebrospinal fluid of the brain-stem in cats. J Pharmacol Exp Ther 1968; 164: 174-181. Dickerman JD, Bishop W. Marks JF: Acute ethanol intoxication in a child. Pediatrics 1968; 42: 837-840. Nestoros J: Ethanol selectively potentiates GABA-mediated inhibition of single feline cortical neurons. Life Sci 1979; 26: 519-523. Hitchcock FA: Alterations in respiration caused by alcohol. Cl J Stud Alcohol 1942; 2: 641-660. Johnstone RE, Witt RL: Respiratory effects of ethyl alcohol intoxication. JAMA 1972; 222: 468. Sahn SA, Scoggin CH, Chernow B: Moderate alcohol dose and chronic obstructive pulmonary disease, not a cause of hypoventilation. Arch Intern Med 1979; 139: 429-431. Flick MR. Block AJ: Nocturnal versus diurnal arrhythmias in patients with chronic obstructive lung disease. Chest 1979; 76: 536-543.
The American Journal of Medlclne
Volume 71
245
VICENTE C. TAASAN,
M.D.*
A. JAY BLOCK, M.D. PHILIP G. BOYSEN, M.D. JAMES W. WYNNE. M.D. with the technical
assistance
of
CAROLE WHITE SHERRY LINDSEY Gainesville, Florida
From the Departments of Medicine and Anesthesiology, College of Medicine, University of Florida, and the Veterans Administration Medical Center, Gainesville, Florida. This study was supported by the Medical Research Service of the Veterans Administration, by Grant HL-22622-03 from the U.S. Public Health Service, by the Parker B. Francis Foundation, and by a Pulmonary Academic Award (K07 Hl 00122) from the National Heart, Lung, and Blood Institute. Requests for reprints should be addressed to Dr. A. Jay Block at the Pulmonary Division, Veterans Administration Medical Center, Gainesville, FLA 32602. Manuscript accepted March 4. 1981. Fellow, Parker B. Francis Foundation. l
240
August 1981
Using standard sleep techniques, we performed a placebo-controlled and randomized study to assess the effect of alcohol ingestion (2 ml/kg of body weight) on breathing and oxygen saturation during sleep. Twenty asymptomatic men volunteered for the two-night study: 11 were given a placebo on night 1, and alcohol on night 2 (group A); nine were given alcohol on night 1 and a placebo on night 2 (group B). We compared the incidence of sleep events (apnea, hypopnea and arterial oxygen desaturation) during the nights the subjects received alcohol and during the nights they received the placebo. Alcohol was associated with significant increases in the occurrence of the following: the number of sleep events (207 to 383, p
The American Journal of Medicine
Volume 71
ALCOHOL INCREASES SLEEP APNEA-TAASAN
MATERIALS
AND METHODS
Subjects. We recruited 20 healthy male volunteers for the study, the initial criterion being the subject’s willingness to undergo a two-night sleep study. They were then screened for abnormal sleep patterns and cardiorespiratory complaints. Subjects with a major illness or a history of alcohol or drug dependence were excluded. The study was approved by the institutional committees for the protection of human subjects, and informed consent was obtained from all participants. General Procedure. Participants were advised to eat light meals on both study nights. Subjects came to the laboratory 1 l/2 hours before their usual bedtime to give them sufficient time to become familiar with the laboratory environment and experimental procedures. The forced vital capacity and forced expiratory volume in 1 second (FEV,) were measured with a Vitalor@ spirometer. We gave the subjects 100 proof vodka in orange juice, and the amount used (2 ml/kg of body weight) was similar to that utilized in previously reported studies [ 2-51. Orange juice alone was the placebo. Eleven subjects were given the placebo on the first night and alcohol on the second (group A). The reverse was true for the other nine subjects (group B). Alcohol and control drinks were consumed within 1 hour. We used an Intoxilyzefl (CMI, Inc., Minturn, CO) to estimate blood alcohol concentrations before sleep on both study nights and also to measure residual concentrations when the subjects awakened the morning after drinking alcohol. Alcohol concentrations before sleep were determined 30 to 45 minutes after the subjects’ last intake of alcohol. Sleep Techniques. Methods used in this study have been reported [g-12]. All variables were continuously measured: oxygen saturation was determined with an ear oximeter (Hewlett-Packard 47201A); oral and nasal air flows were sensed with thermistor clips on one lip and one nostril (Grass Instruments); and chest movement was monitored by impedance pneumography with surface electrodes and pneumograph couplers (Narco Biosystems). All of the foregoing were also continuously recorded (Narco Biosystems Physiograph DMP-4P 4-channel recorder). Electroencephalograms, electro-oculograms and electrocardiograms were simultaneously recorded on a separate recording system (Grass Model 79D polygraphic recorder). Methods of electroencephalograms and sleep stage scoring were based on those of Agnew and Webb [ 131. A technician, who was present throughout the study to attend to the subject’s needs and to assure adequate recordings, noted spurious events (e.g., detached leads, clogged thermistors) and correlated them with the recordings. Definitions. Apnea was defined as a cessation of oral and nasal flows for 10 seconds or longer. Hypopnea was noted when desaturation occurred simultaneously with decreased airflows and chest wall movement. Arterial oxygen desaturation was defined as a L4 percent decrease in oxygen saturation. Sleep-disordered breathing was defined as any occurrence of apnea or hypopnea or both. An event was any occurrence of apnea, hypopnea or desaturation. Sleep-period time was defined as the interval in minutes, from the onset of stage 1 sleep to full awakening the following morning. Experimental Design and Analysis. According to a preestablished code, each subject was randomly assigned to group
ET AL.
A or B. We attempted a single blind study, but the subjects immediately identified the drink, and the breathalyzer maneuver informed the sleep technician of the identity of the drink. Prior to breaking the codes, we analyzed our records in the following manner: Recordings were inspected visually. The number and the duration of desaturation and disordered breathing were recorded. For each night, the lowest saturation, the maximum decrease in saturation and the longest duration of each of the events were recorded for each subject. To control for varying lengths of sleep, the total time (in minutes) of desaturation or disordered breathing was counted and expressed as a percentage of sleep-period time. The frequency of events per hour of sleep-period time was also calculated. The effect of alcohol on sleep structure was analyzed by measuring the duration of the sleep-period time and calculating the percentage of sleep-period time for each stage of sleep. Statistical differences between the nights alcohol was injested (alcohol nights) and the placebo (control night) in terms of desaturation and breathing variables were sought by obtaining the difference between night 2 and night 1 values for each subject. The groups were then compared by the Wilcoxon rank sum test [ 141 using these differences. This nonparametric method was used because the presence of “outlier” values would invalidate usual parametric methods. To determine the residual effects of alcohol, we used a similar analysis to compare the findings on placebo nights that preceded the ingestion of alcohol (group A) with those that followed alcohol ingestion (group B). The effect of alcohol on sleep stage distribution was analyzed using the Hotelling t test [IS]. Correlations between age, weight to height ratios, blood alcohol concentrations and sleep events were determined using Kendall’s Tau [ 161. A correlation was declared significant when it achieved p <0.05 for both nights. The Statistical Analysis Systems (SAS) computer package was used for the data analysis. RESULTS
Eleven of 20 subjects were given alcohol on night 2 (group A, prealcohol control) and nine subjects were given alcohol on night 1 (group B, postalcohol control). Table I shows the anthropometric, spirometric and TABLE I
Anthropometric, Splrometric and Blood Alcohol Concentrations in Two Groups of Normal Men* GroupA
Subjects (no.) Age (yr) Weight (kg) Weight:height (kg/cm) Forced vital capacity (liters) Forced expiratory volume in 1 second (liters/set) Blood alcohol concentration
48.0 78.0 0.43 3.98 3.27
GroupB
11 f 4.5 f 2.8 f 0.01 f 0.27 f 0.21
80.0 f 6.8
48.7 78.6 0.45 4.50 3.52
9 f f f f f
5.4 3.9 0.01 0.32 0.23
87.7 f 8.1
(mgld)) NOTE: Values given are the mean f standard error. * Alcohol was given on night 2 to patients in group A and on night 1 to patients in group B.
August 1981
The American Journal of Medicine
Volume 71
241
ALCOHOL INCREASES SLEEP APNEA-TAASAN
TABLE II
ET AL
TABLE Ill
Arterial Oxygen Desaturation in Two Groups of Normal Men (Group A, 11 Subjects; Group B, 9 Subjects)*
Breathing Variables During Alcohol and Control Nights Group A
GroupA Night 2 Night 1 (alcohol) (placebo) Desaturation (no. events) Desaturation (total duration, min) Desaturation (no./hr) Lowest saturation (%) Longest desaturation (set)
2.6 f 2.5
9.6 f 4.5
0.4 f 0.4
2.5 f
0.4f0.3
1.5f0.7
93 f
4.0 f
1.0
1.5
88 f 2.0
2.8 26.1 f
Group B Night 2 Night 1 (placebo) (alcohol) 13.4 f 6.2 10.0 f 5.6 3.3 f 2.5
2.6 f
2.5f1.2
1.6f0.7
87 f 3.0
1.2
88 f 2.0
12.2 19.1 f 9.6 20.8 f 8.4
NOTE: All values are mean f standard error. ’ p values obtained by ranks sum test are shown in the text.
blood alcohol concentrations of the 20 subjects. Four subjects were 20 years of age or younger, six were 60 years or older. The oldest subject was 65 years of age. Only one subject weighed more than 100 kg (102 kg). Concentration of alcohol in blood before sleep was 1100 mg/dl in seven subjects; the lowest concentration was 40 mg/dl, the highest, 120 mg/dl. In four subjects, 10 to 20 mg/dl of alcohol were detectable after sleep on the nights alcohol was given. Spirometric values in all subjects were within normal limits for our laboratory. Only three subjects were smokers. Alcohol Versus Control Nights. Oxygen saturation: The total number of desaturation events was significantly greater during sleep following alcohol ingestion (p
242
August 1981
Night 1 (placebo) SDB (no. events) SDB (total duration, min) SDB frequency (no./hr SPT) Apnea (no. events)
Night 2 (alcohol)
Group B Night 2 Night 1 (placebo) (alcohol)
2.4 f 1.9 5.4 f 3.0 10.7 f 7.3 6.9 f 5.6 0.8 f 0.6 3.4 f 1.7 3.8 f 2.9 3.3 f 2.8 0.3 f 0.2 0.9 f 0.4
2.0 f
1.2 f 0.8 2.6 f
9.0 f 7.2 0.7 f 0.5
1.3
1.4 1.2 f 0.9
NOTE: SDB = sleep disordered breathing, SPT = sleep-period time. All values are mean f standard error. p values obtained by ranks sum test are shown in the text.
Sleep-disordered breathing: The number (p = 0.03), total duration (p = 0.05) and frequency (p = 0.02) of sleep-disordered breathing were significantly greater after alcohol ingestion. All those who had disordered breathing on control nights (nights they received the placebo) showed increases in the number and duration of disordered-breathing events after the ingestion of alcohol. Apnea was more common than hypopnea with alcohol, and 11 subjects had at least an apneic event with alcohol. Only four subjects were noted to have apnea during control nights. A total of 110 apneic events were counted during alcohol nights (nights they received alcohol); on the control nights, 20. This difference was statistically significant (p KO.01). There were 157 events of either apnea or hypopnea during alcohol nights, 89 during control nights. For descriptive purposes, Table Ill shows the means and standard errors of different breathing variables. Total Number of Sleep Events. Table IV shows the total number of sleep events and the number of subjects who had them. In 18 of the 20 subjects, desaturation or disordered-breathing with alcohol occurred at least once. In nine, it did so with the placebo. With alcohol, the number of sleep events was significantly greater, and almost twice as many sleep events were recorded after the ingestion of alcohol (383) than after the intake of the placebo (207) (p
TABLE IV
Number of Events of Desaturation and Disordered Breathing and Number of Subjects with Sleep Events During Control and Alcohol Nights All Events
Alcohol nights No. subjects Control nights No. subjects p value
383 18 207 9 p
l
Desaturation 226 17 118 9 p
Apnea 110 11 20 4 p
Hypopnea 47 11 69 4 NS+
p values from rank sum test apply to number of events only. + Not significant. l
The American Journalof Medicine Volume 71
ALCOHOL
INCREASES
SLEEP
APNEA-TAASAN
ET AL.
60
CHEST MOVMNT b---
50 40
NOSt-_. AIRFLOW
: _ cs
,--_~,-_ ___~_
30 20 10 0i
STAGE P VALUE
Figure 1. Polygraphic tracing showing a 46-second event of profound desaturation with cessation of airflow and diminution of chest movement. The subject was a 59 year old asymptomatic man who had 126 events of desaturation or abnormal breathing in a single night’s sleep.
more common event. In addition, the number of apneic events and the number of sleep events per hour of sleep (frequency) were significantly greater with alcohol (p
TABLE V
Arterial Oxygen Saturation and Breathing Variables During Prealcohol and Postalcohol Control Nights for Two Groups of Subjects
Variable Desaturation (no. of events) Desaturation (total duration) Desaturation (no./hr SPT) Disordered breathing (no. events) Disordered breathing (no./hr SPT)
GroupA (Prealcohol)
GroupB (Postalcohol)
2.6 0.4 0.4 2.4 0.3
10.0 2.6 1.6 6.9 1.2
f f f f f
2.5 0.4 0.3 1.9 0.2
f f f f f
5.6 1.2 0.7 5.6 0.9
NOTE: SPT = sleep-period time-the interval between the onset of stage 1 sleep and full awakening. All values are mean f standard error.
0
I
NS
< 01
2
3
4
R
NS
NS
NS
NS
rfgure z. c‘omparrson of sleep-stage percentages for placebo and alcohol nights for both groups. No significant differences were noted except for stage 1, which in group A was significantly lower during alcohol nights. P values for differences between sleep-stage percentages for period A and group B control nights are given in the text. SPT = sleep period time, the interval, in minutes, between the onset of stage 7 sleep and full awakening. Group A subjects were given alcohol on night 2, Group B subjects were given alcohol on night 1.
whereas it occurred in seven of the nine subjects in group B. The maximum change in saturation was greater (p = 0.02) and the lowest saturation (p = 0.02) was also significantly lower in group B. However, no significant differences in disordered breathing were noted for the two groups. Table V shows descriptive data for groups A and B for both desaturation and breathing events, indicating the persistence of desaturation on the control nights that followed alcohol nights. Distribution of Sleep. Figure 2 shows the distribution of sleep during control and alcohol nights. All non-REM and REM stages of sleep were identified in all subjects. For both nights, sleep-period time was shorter than previously established normal values [ 171. Between the two nights, the.sleep-period time for all subjects showed no significant change. The average sleep period times of both groups for both nights ranged from 3 16 to 344 minutes. The shortest sleep time was 3 hours, the longest was 7 hours and 20 minutes. In addition, analysis of the latencies of stage 2, slow wave sleep and REM sleep showed no significant changes with alcohol. Sleep-stage percentages during postalcohol control nights showed significant differences from prealcohol control nights. Sleep stages 0 (p = 0.03) and 1 (p = 0.01) were significantly greater during postalcohol control nights, but sleep stage 4 was significantly reduced (p = 0.04). Stages 2 and REM sleep showed no significant change with alcohol. Effect of Age, Weight to Helght Ratios and Concentrations of Alcohol in Blood. Correlation between
August 1991
The American Journal of Medlclne
Volume 71
243
ALCOHOL INCREASES SLEEP APNEA-TAASAN
ET AL.
sleep events during alcohol nights and age, weight to height ratios and concentrations of alcohol in blood showed that increased age had a positive relation to sleep events. Increased age correlated positively with the number of events of disordered breathing and desaturation (p = 0.05 and 0.03), respectively, and with the duration of disordered breathing (p = 0.03). For control nights, a correlation was noted between increasing age and low saturation for group A. The number of desaturation events also correlated with age in group B. Weight to height ratios and blood alcohol concentrations did not statistically correlate with sleep events on either night. COMMENTS This study confirms our previous findings that asymptomatic men tend to have oxygen desaturation and disordered breathing during sleep. In addition, we have found that alcohol ingestion is associated with an increased frequency and severity of these sleep events, and this increase persists through an additional night after alcohol intake. In this study nine of the 20 men had 207 sleep events during alcohol-free nights. In a previously reported study [l] 20 of 30 male subjects had 264 events. Other similarities were noted between the types of sleep events in this and the previous study: the most common event was arterial oxygen desaturation, and the least common was apnea. In this study, control night saturations were significantly lower in a group of subjects as age advanced, and elderly subjects had more events during sleep after alcohol ingestion. Because no subject in this study was obese, we cannot comment on the previously reported effect of weight on the frequency of sleep events [ 11. Notably, the simultaneous occurrence of desaturation and an altered breathing pattern was frequent with alcohol. Inspection of our records shows that of 226 episodes of desaturation during alcohol nights, 114 were accompanied by a simultaneous occurrence of apnea or hypopnea. The three lowest levels of saturation were all accompanied by disordered breathing. Of the 110 apneic events with alcohol, 71 were associated with an episode of desaturation. Several episodes of desaturation were further noted to be associated with at least some decreases in air flow or chest movement, but they failed to meet our predetermined criteria and were not counted as disordered breathing. Not all episodes of desaturation, however, were accompanied by an episode of disordered breathing. Thus, the sleep events with alcohol in this study were related to alterations of ventilatory patterns, which on most occasions were accompanied by episodes of desaturation. The increased incidence and severity of oxygen desaturation and sleep-disordered breathing after alcohol ingestion could not be explained by differences in the
244
August 1981
The American Journal of Medicine
length of sleep or by an increase in stages 2 and REM sleep. This does not mean that the state of sleep itself is not involved in the occurrence of more sleep events with alcohol. On the contrary, we believe that sleep has a permissive role in these events. The persistence of desaturation on the placebo control night after alcohol ingestion is a surprising finding. Although its significance is unknown to us, a similar persistence of alcohol effects has been reported for catecholamine metabolism [ 181. Brohult et al. [ 181 showed that catecholamine excretion is increased for an entire week after a single large dose of ethanol. In a study of sleep in young adults, Rundell et al. [5], used a dose similar to that in our study, showed that on the night after alcohol ingestion sleep latency increased and REM latency decreased [ 51. Thus, as has been noted, alcohol ingestion has sustained effects that persist beyond the measurable presence of alcohol in the blood. Data have often shown that moderate doses of alcohol have a potential for respiratory depression. These data were gathered in the laboratory setting, and the practical implications of these findings are unclear. More sensitive tests, e.g., ventilatory response to oxygen and carbon dioxide, reveal that a moderate dose of alcohol also depresses chemosensitivity [ 19,201. Furthermore, direct application of alcohol to brain stem structures in the cat has been shown to cause apnea [21]. “Respiratory arrhythmia” [ 191 has been reported in awake subjects given moderate doses of alcohol, and episodic apnea has been reported in a child with alcohol intoxication [22], findings of great interest in relation to the fivefold increase in apnea in this study. A recent study has shown that small to moderate amounts of alcohol potentiate GABA-mediated inhibition of single cortical neurons in cats [ 231, analogous to the cortical depression caused by barbiturates and diazepines [23]. Thus, there is evidence that even with moderate doses of alcohol, respiratory chemosensitivity is depressed and ventilatory patterns are altered. On the other hand, some studies reveal that alcohol has inconsistent or in fact stimulatory effects on respiration [24,25]. Sahn et al. [26] have recently shown that ingestion of moderate doses of alcohol is not an important cause of hypoventilation in patients with chronic obstructive lung disease. It is worthwhile mentioning that their study was carried out with awake subjects. Thus, the practical implications of the laboratory data pointing to a potential for respiratory depression by small to moderate doses of alcohol remain unclear. In all of these studies, however, only the effects of alcohol ingestion on the awake subject were considered. The possibility exists that wakefulness overrides the depressant effects of alcohol. We know of no previous study in which respiratory behavior was observed
Volume 71
ALCOHOL INCREASES SLEEP APNEA-TAASAN
during sleep after a moderate dose of alcohol. In this study it has clearly been shown that the ingestion of 2 ml/kg of 100 proof vodka markedly increases the frequency and severity of nocturnal oxygen desaturation and sleep-disordered breathing in asymptomatic men. At this point, it is difficult to understand the practical implications of our findings. Asymptomatic subjects have not as yet been shown to suffer cardiovascular consequences of the brief episodes of apnea or desaturation. However, patients with chronic obstructive lung disease who ingest moderate amounts of alcohol
ET AL.
may suffer cardiopulmonary sequelae. Previous studies have shown that nocturnal oxygen desaturation is associated with cardiovascular deterioration in such patients [ 11,271. Because breathing abnormalities occur during sleep in patients with chronic obstructive lung disease [S-12], moderate doses of alcohol could be expected to accentuate concomitant arrhythmias and pulmonary hypertension. These findings in normal subjects thus open the way to research into how moderate amounts of alcohol affect breathing during sleep of patients with serious pulmonary and cardiovascular disease.
REFERENCES
2.
3.
4. 5.
6.
7. 8.
9.
10.
11.
12.
13.
Block AJ. Boysen PG. Wynne JW, Hunt LA: Sleep apnea, hypopnea and oxygen saturation in normal subjects, a strong male predominance. N Engl J Med 1979; 300: 513-517. Gresham SC, Webb WB, Williams RL: Alcohol and caffeine: effect on inferred visual dreaming. Science 1963; 140: 1266. Yules RB, Freedman DX, Chandler KA: The effect of ethyl alcohol on man’s electroencephalographic sleep cycle. Electroencephalogr Clin Neurophysiol 1966; 20: 109111. Knowles JB, Laverty SG, Keuchler HA: Effects of alcohol on REM sleep. Q J Stud Alcohol 1968; 29: 342-349. Rundell OH, Lester BK, Griffiis WJ, Williams HL: Alcohol and sleep in young adults. Psychopharmacologia (Berlin) 1972; 26: 201-218. Wynne JW, Block AJ, Hemenway J, Hunt LA, Flick MR: Disordered breathing and arterial oxygen desaturation during sleep in patients with chronic obstructive lung disease. Am J Med 1979; 66: 573-579. Lugaresi E: Snoring. Electroencephalqr Clin Neurophysiol 1975; 39: 59-64. Lugaresi E, Coccagna G, Cirignotta F: Snoring and its clinical implications. In: Guilleminault C, Dement WC, eds. Sleep apnea syndromes. New York: Alan R Liss, 1978; 13-2 1. Flick MR, Block AJ: Continuous in vivo monitoring of arterial oxygenation in chronic obstructive lung disease. Ann Intern Med 77; 86: 725-730. Boysen PG. Block AJ. Wynne JW, Hunt LA, Flick MR: Nocturnal pulmonary hypertension in patients wlth chronic obstructive pulmonary disease. Chest 1979; 76: 536-542. Wynne JW, Block AJ, Hunt LA, Flick M(: Disordered breathing and oxygen desaturation during daytime naps. Johns Hopkins Med J 1978; 143: 3-7. Flick M1, Block AJ: Continuous in vivo measurement of arterial oxygen saturation by oximetry. Heart Lung 1977; 6: 990-993. Agnew HW Jr, Webb WB: Sleep stage scoring (American
14. 15. 16. 17.
18. 19. 20.
21.
22. 23.
24. 25. 26.
27.
Augud lggl
Psychological Association Abstract Series, Manuscript No. 293). Washington, DC, American Psychological Association, 1972. Hollander MD. Wolfe DA: Nonparametric statistical methods. New York: John Wiley and Son, 1973; 68-75. Morrison DF: Multivariate statistical analysis. New York: McGraw-Hill, 1967; 145-146. Siegal S: Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill, 1956; 213-223. Williams RL, Karacan I, Hursch Cl: Electroencephalography (EEG) of human sleep: Clinical applications. New York: Wiley, 1974. Brohult J, Levi L, Richard H: Urinary excretion of adrenal hormones in man. Acta Med Stand 1970; 188: 5-13. Johnstone RE, Reier CE: Acute respiratory effects of ethanol in man. Clin Pharmacol Ther 1973; 14: 501-508. Sahn SA, Lakshminarayan S, Pierson DJ, Weil JV: Effect of ethanol on the ventilatory responses to oxygen and carbon dioxide in man. Clin Sci Mol Med 1974; 49: 33-38. Rosenstein R. McCarthy LE. Borison HL: Respiratory effects of ethanol and procaine injected into the cerebrospinal fluid of the brain-stem in cats. J Pharmacol Exp Ther 1968; 164: 174-181. Dickerman JD, Bishop W. Marks JF: Acute ethanol intoxication in a child. Pediatrics 1968; 42: 837-840. Nestoros J: Ethanol selectively potentiates GABA-mediated inhibition of single feline cortical neurons. Life Sci 1979; 26: 519-523. Hitchcock FA: Alterations in respiration caused by alcohol. Cl J Stud Alcohol 1942; 2: 641-660. Johnstone RE, Witt RL: Respiratory effects of ethyl alcohol intoxication. JAMA 1972; 222: 468. Sahn SA, Scoggin CH, Chernow B: Moderate alcohol dose and chronic obstructive pulmonary disease, not a cause of hypoventilation. Arch Intern Med 1979; 139: 429-431. Flick MR. Block AJ: Nocturnal versus diurnal arrhythmias in patients with chronic obstructive lung disease. Chest 1979; 76: 536-543.
The American Journal of Medlclne
Volume 71
245