Oxygen Saturation during Breath-Holding and during Apneas in Sleep*

Transcription

1 Oxygen Saturation during Breath-Holding and during Apneas in Sleep* Kingman P. Strohl, M.D.;t and Murray D. Altose, M.D., F.C.C.P. The rate of fall in oxygen saturation is said to be greater during obstructive apneas than during breath-holding in wakefulness. Using an ear oximeter, a face maslc: and flowmeter, and measurements of thoracoabdominal motion, we determined in six healthy subjects the rate of fall in arterial oxygen saturation (SaO.) during breath-holding which simulated obstructive and nonobstructive apneas. Breath-holding maneuvers were performed during progressive isocapnic hypoxia and were initiated at the same e n ~ expiratory thoracoabdominal configuration. We found that at any given initial Sa 1 the rate of fall in Sa 1 was similar during simulated obstructive (y = x; r =.83) and nonobstructive (y = x; r =.92) apneas. In two healthy subjects and 13 patients with obstructive and nonobstructive apneas during sleep, the rate of fall in Sa 1 at any initial Sa 1 was similar to that found in healthy subjects during breath-holding in wakefulness. We conclude that during wakefulness the presence or absence of respiratory efforts does not affect the rate of fall in Sa 1 during breath-holding and that the rate of fall of Sa 1 during sleep apnea is largely dependent on the initial Sa 1 at the onset of apnea. D uring apneas in sleep, oxygen saturation falls. The decrease in oxygen saturation may be abrupt and severe. It has been suggested that the rate of fall of oxygen saturation during apneas and, in particular, during obstructive apneas is greater than what would be expected with breath-holding during wakefulness and that factors such as ventilation-perfusion mismatching, a decrease in lung volume, or altered cardiac output might accelerate the fall in oxygen saturation during apneas in sleep. t-3 In the present study, we examined the rate of fall in arterial oxygen saturation (SaOJ during breath-holding in wakefulness and during apneas in sleep. Breathholding during wakefulness was performed by healthy subjects who were trained to simulate both obstructive and nonobstructive (central) apneas. During wakefulness the presence or absence of respiratory effurts did not affect the rate offall in Sa 2 during breath-holding. The rate of fall in Sa 2 during sleep apneas was within the range of values observed during wakefulness. We conclude that during breath-holding and in sleep apnea, the rate of fall in Sa 2 is determined largely by initial Sa 2 at the onset of apnea. METHODS Studies were performed in eight healthy subjects (aged 21 to 32 years) during wakefulness, in two healthy subjects (aged 24 and 26 years) during sleep, and in 13 patients (aged 42 to 72 years) with apneas during sleep. All healthy subjects were asymptomatic, and all *From the Department of Medicine, Cleveland Metropolitan General Hospital and' Case Western Reserve University, Cleveland. Supported by grants from the American Heart Association and the National Institutes of Health (HL 2583 and the High Altitude Physiology Study HL 1412). trecipient of a Clinical Investigator Award (HL 167). Manuscript received May 31; revision acce_pted August 1. Reprint requem: Dr: Alto1e, Cleveland Metropolitan General H o l p3395 i ~ Scranton. Road, Cleveland 4419 patients were clinically stable at the time of the study. Patients were referred fur measurement of respiratory events during sleep because of excessive daytime sleepiness or observed apneas during sleep (or both). Six patients (cases 1, 4, 5, 8, 1, and 13) were obese (greater than 1.2 percent of ideal body weight) and were hypercapnic (arterial carbon dioxide tension [ PaC 1 ] greater than 5 mm Hg) and hypoxemic (arterial oxygen pressure [Pa 2 ] less than 6 mm Hg) during wakefulness. Four patients (cases 2, 6, 9, and ll) were obese and normocapnic. Three patients were of normal body weight. Two patients (cases 3 and 12) were in chronic renal failure requiring hemodialysis three times per week. Studies were performed with the subject in the supine or semirecumbent posture. We monitored arterial oxygenation, airflow at the nose and mouth, and rib-cage and abdominal motion. Oxygenation was measured using an ear oximeter (Hewlett-Packard). The ear oximeter used in this study had been previously calibrated against Sa 1 measured spectrophotometrically and was found to have an accuracy of :t 2 percent to an Sa 1 of 6 percent, as previously reported. Airflow was determined either by a pneumotachygraph connected to a face mask or by two thermistors placed in front of the nose and mouth, respectively. Rib-cage and abdominal dimensions were monitored either by two pairs of linearized magnetometers or by a DC-coupled respiratory inductance plethysmograph placed at the level of the mid-chest and umbilicus, respectively. These enable monitoring of thoracoabdominal movements and of the end-expiratory position of the rib cage and abdomen and can distinguish the type of apnea during wakefulness or sleep. Studies during Wakefulness Six healthy subjects were trained to simulate nonobstructive and obstructive apneas. Breath-holds lasting 1 to 25 seconds were initiated from functional residual capacity. Rib-cage and abdominal movements were carefully monitored to ensure that all breath-holds were initiated from the same end-expiratory pulmonary volume and the same initial thoracoabdominal configuration. Breath-holding. maneuvers were performed before and during progressive isocapnic hypoxia produced by the method of Rebuck and Campbell. To slow the rate of fall in Sa 1 during rebreathing, oxygen was added to the rebreathing bag at a rate of 5 to 1 mllmin. To mimic nonobstructive apnea, subjects made no effort to breathe. To simulate obstructive apneas, the subjects made in- CHEST I 85 I 2 I FEBRUARY,

2 z i= cr "' ::;) 1- ~ 1 zz... g 1.4 A 1.2 ~ ~.8 x, Ot- ~ z 6 o ~..JCZ:...11&.1 ~ ~ 4 ~... I- g: "' B oo eo ~ 8 o.2 c:l o o o INITIAL OXYGEN SATURATION (PERCENT) BREATH HOLD TIME (SECONDS) FIGURE 1. A (left), Effects of reduction in initial Sa 2 on rate offal) in Sa 2 for one subject. B (right), Rate of fall in Sa 1 is independent ofbreath-holding time. Symbols represent values for simulated nonobstructive apneas (squares) and for simulated obstructive apneas performed with either rib-cage (solid circles) or abdominal paradox (open circles). Line in Figure 1A represents linear regression through all points (r =.92). creasingly greater inspiratory effurts against a closed airway. Efforts were one to two seconds in duration and were repeated every three to four seconds. Airway pressure during the first effort was 5 to 1 em H 1 in magnitude, and each successive effi>rt was 5 to 1 em H 2 greater than the previous effurt. Mouth pressure was monitored by a pressure transducer (Validyne) ( ± 1 em H 2 ) and was displayed to the subject on an oscilloscope. In three subjects, obstructed effurts were selectively performed by using predominantly rib-cage or diaphragmatic muscles. The presence of either abdominal or ribcage paradoxic motion was confirmed by oscilloscopic x-y display of rib-cage and abdominal movements during the inspiratory efforts. In a separate series of experiments, five healthy subjects performed ten-second to 15-second breath-holds while making continuous expiratory or inspiratory effurts against a closed airway to generate either positive ( + 8 to 12 em H 2 ) or negative pressures (-8 to -12 em H 2 ) at the mouth. Initial Sa 2 during these maneuvers was maintained between 91 and 95 percent. Studies during Sleep Healthy subjects were monitored on two successive nights, and the data reported are from the second night of study. All patients except one were studied on one occasion. Patient 4 was studied on three occasions, each separated by several days. The electroencephalogram, electro-oculogram, and chin-muscle electromyogram were monitored in order to distinguish wakefulness from slow-wave (NREM) sleep and rapid-eye-movement (REM) sleep. 1 In the two healthy subjects the level of inspired oxygen was NONOBSTRUCTIVE lowered by blending nitrogen with room air entrained through a Venturi mask. The oxygen concentration of the inspired air was measured by an analyzer (Beckman) and was kept above 15 percent. All patients were studied while breathing room air. In addition, in four patients (cases 6, 9, 12, and 13), the inspired oxygen concentration was increased using a nasal cannula and an oxygen How rate of 2 to 5 Umin. Apneas of greater than ten seconds were identified by the absence of air8ow at the nose and mouth. Obstructive apneas were distinguished from nonobstructive apneas by the presence of paradoxical rib-cage or abdominal movements. Analysis of Data The length of the apneas, either simulated or spontaneous, was measured to the nearest.5 second. The fall in Sa 2 with apneas was determined from the pre-apneic value and the nadir ofsa 1 after the apnea. Measurements of Sa 2 were made to the nearest whole percentage. The rate of change ofsa 2 is expressed as the change in percentage of Sa 2 per second. Results are expressed as the mean ± the standard deviation ( ± SD). Linear regression analysis was performed using the method of least squares. RESULTS Studies during Wakefulness The effects of reductions in Sa 2 on the rate of 12 OBSTRUCTIVE z ~ w CIC ~ ~ 8 ~!!..J z.6 ~ ~ ~... II.."' ~ ~!«2 2 "'(/) a:: ~ ~ 8 ~ ~ 8 7 ~ ~ 8 ~ ~ 9 1 ~ ~ ~ ~ 98 ~ ~ 35 -~- ~ ~ 89~ 57~ -~- 9 ~ ~ 87 ~ ~ 9 ~ ~ 9 ~ ~ 9 ~ 3 ~ ~ 9 ~ INITIAL OXYGEN SATURATION (%) FIGURE 2. Relationship between initial Sa 1 and rate of fall in Sa 1 during simulated nonobstructive (left) and obstructive (right) apneas. Symbols connected by lines represent linear regression between most extreme values of initial Sa 1 Each symbol represents one subject. Regression line for all values in all subjects: nonobstructive apneas, y=6.8+(-.7) x; r=.92; obstructive apneas, y=5.5+(-.6) x; r= Oxygen Saturation during Sleep Apnea (Strohl, AltDse)

3 Table I-Rate offau in SaO, during Breath-Holding in Healthy Subject. n 97 percent Nonobstructive 6.16:t.11 Obstructive 6.15:!:.12 Rib-cage paradox 3.13:!:.4 Abdominal paradox 3.11 :!:.3 Breath-hold positive pressure 5 Breath-hold negative pressure 5 *No. of subjects. tpercentlsecond (mean :t SD). Initial SaO, t 93 percent 89 percent.42:!:.1.72:!:.16.45:!:.14.76:!:.15.47:!:.1.47:!:.3.44:!:.18.49:!:.19 change in Sa 2 with breath-holding during wakefulness is shown fur one subject in Figure 1. As the initial Sa 2 decreases with rebreathing, the rate of decline in Sa 2 during breath-holding progressively increases (r =. 92). The rate of change in Sa 2 with time is no different during simulated central apneas than during simulated obstructive apneas. Figure 1B illustrates that the rate of change in Sa 2 during breath-holding is not dependent on the breath-holding time. Figure 2 shows the relationship between initial Sa 2 and the rate of fall in Sa 2 for each subject during simulated nonobstructive and obstructive apneas. The lines describing the relationship in each subject were determined by the method of least squares from the values during 12 to 16 breath-holds. The correlation coefficient fur all linear regressions was.91±.6. The slope of the relationship during simulated nonobstructive or obstructive apneas is similar fur each subject, and there are no significant differences taking either individual or group values. Table 1 lists the mean values for the rate of fall in ~ ~ a: :::> ~ z z 1 wo ~ ~ ><"' o;::... z ow u ::l a: 5 : ~. 1&1... C( a: INITIAL OXYGEN SATURATION (PERCENT) FIGURE 3. Relationship between initial Sa 1 and rate offall ofsa 1 is shown for two healthy subjects during sleep. Closed circles represent values from subject 7; open circles are those from subject 8. Une represents linear regression through all points (r =.88). I Sa 2 fur all studies in subjects during wakefulness. During either simulated nonobstructive or obstructive apneas, the lower the initial Sa 2, the greater the rate of fall in Sa 2 (p<.5); however, at any given initial Sa 2, there is no significant difference in the r a of t ~ decline in Sa 2 between the maneuvers. In the three subjects in whom obstructive apneas were simulated by producing either rib-cage or abdominal paradoxic motion, results are similar during the two maneuvers. In the five subjects who maintained continuous positive or negative airway pressure during breathholding, the rate of decline in Sa 2 at an initial Sa 2 of 93 ± 2 percent was no different from the values at 93 percent during simulated obstructive and nonobstructive apneas. Additionally, there were no differences in the rate of decline in Sa 2 during breath-holds with positive or negative airway pressure. Studies during Sleep Periodic breathing during sleep was produced in two healthy subjects by reducing the inspired oxygen concentration. One of the subjects also exhibited brief nonobstructive apneas in stages 1 and 2 of sleep while breathing room air; however, only one apneic episode was greater than ten seconds in duration. In both Table 2-Valua in lbtienta with Apnea during Sleep Initial Rate of Patient, Sa 2 in Duration Fall in Type Sex, Sleep, of Apnea Sa 1, No. of of Age (yr) percent sec percent/sec Apneas Apneast 1, M, 48 89:!:.9 17:t2.78:!:.2 18 M 2, M, 58 96:!:1.2 14:t2.4:!:.15 1 N 95:!:1.4 2l:t6.33:!: , M, 63 94:!:.9 23:t5.58:!:.1 8 N 4, M, 66 86:!:1.5 17:t8 1.68:!: :!:1.2 23:!:5 1.48:!: :!:.7* 28:!:4 1.:!: :!:1. 32:t6.98:!: , F, 72 89:!:2. 15:!:4 1.14:!: , M, 49 91:!:.9 28:t2 1.:!: :!: :!:4.8:!: , M, 51 95:!:.5 16:!:3.22:!: , M, : t l. 18:t3 ~ 1.27:!: :!:2. 25:!:2 1.21:!: , M, 56 9l:t 1. 19:!:3.31:!: :t.6fl 28:t2.1:!: , M, 62 82:tl.O 16:t5.75:!: ll,m,53 93:!:.5* 18:!:2.47:!:.5 56 M 92:!:.5 38:tl2.51:!: , F, 52 92:!:.3 19:!:3.31:!:.2 3 N 98:!: :!:2.9:!:.3 12 N 13, F, 66 78:tl.O 16:!:8 1.12:!:.2 4, M 82:t3.fl 21.8:!:9.68:!:.1 26 M *18ble values are means :t SD unless otherwise indicated. tm, Mixed apneas; N, nonobstructive apneas; and, obstructive apneas. *Slow-wave sleep. Rapid-eye-movement sleep. IIDuring administration of oxygen. CHEST I 85 I 2 I FEBRUARY,

4 INITIAL OXYGEN SATURATION (PERCENT) FIGURE 4. Relationship between initial Sa 1 and rate of fall of Sa 1 during sleep apneas in 23 trials in 13 patients. Circlu represent mean and ban represent ±I SD ror each trial. Points with Dare derived from one patient (case 4) studied on three occasions; all other patients were studied during single night SolUI Une represents mean values and dotted Unu represent 95 percent confidence limits derived from all studies in healthy subjects during wakefulness. For 1Ud Une, y=6.1+(-.6) x; r=o.so. subjects, the rate of decrease in Sa 2 during apneas increased as baseline Sa 2 fell (Fig 3). At any given initial Sa 2 the rate of decline in Sa 2 in these two subjects fell within one standard deviation of the mean value in the normal subjects during simulated nonobstructive apneas in wakefulness. Thble 2lists values of initial Sa 2, duration of apnea, and number and type of apneas observed in patients during sleep. Nonobstructive and obstructive apneas were considered separately, as were apneic episodes during NREM and REM sleep. Also because of the differences in the initial Sa 2 while breathing room air and supplemental oxygen, these conditions were also separately considered. Patient 4 was studied on three occasions, each separated by several days; on one occasion, results from NREM and REM sleep are reported separately. For all other patients, data are derived from one sleep session. The relationships between initial Sa 2 and the rate of fall of Sa 2 fur each subject during each condition are shown in Figure 4. The interrupnng lines represent the 95 percent confidence limits of the values in normal subjects during breath-holding. In general, the values in patients fell within the 95 percent confidence limits of values in healthy subjects. The one patient (case 4) in whom the rate of fall of Sa 2 was greater than normal also had atrial fibrillation. In three patients the rate of fall of Sa 2 during REM sleep was not different from that during NREM sleep. Administration of oxygen (in fuur patients) resulted in an increase in baseline oxygenation and a lower rate of fall in Sa 2, although apneas were generally longer. DISCUSSION The results of our studies indicate that the rate of fall of Sa 2 during breath-holding in wakefulness is a function of the initial Sa 2 at the onset of breathholding and that Sa 2 falls at approximately the same rate during breath-holds simulating obstructive and nonobstructive apneas. Also, the results of the present study show no systematic difference in the rate offall in Sa 2 during breath-holding in wakefulness in healthy subjects when compared to that during apneas in sleep in patients. During breath-holding the fall in and rate of fall of oxygen saturation can be altered by changing stores of oxygen or by increasing the rate of metabolic oxygen consumption. 8 Of these two factors, stores of oxygen, which are influenced by alterations in pulmonary volume, are more important. 3 The increased rate of fall in Sa 2 during breath-holding at lower initial values fur Sa 2 could possibly be explained by an increased metabolic rate, but this seems unlikely, since subjects were tested at rest, and the fall in Sa 2 was similar during breath-holds with and without respiratory efforts. Also, in our healthy subjects during wakefulness, progressive hypoxemia was produced under isocapnic conditions, and pulmonary volume at the onset of all breath-holds was kept constant. We employed relatively brief breath-holds performed at progressively lower Sa 2 so as to minimize the distress of the subjects. As shown in Figure 5, the fall in Sa 2 with time during a prolonged breath-hold is also curvilinear. The lower the Sa 2, the greater the instantaneous rate of fall in Sa 2 FIGURE 5. Changes in sao. during course of single prolonged voluntary breath-hold in one subject. As- 99% OXYGEN terish represent start and termination of prolonged breath-hold. Initial sao. 1s 98 percent. Early in breath- SATURATION hold, Sa 1 drops slowly, but thereafter there is increasingly rapid decline in Sa 1 as breath-holding duration increases Gl 1U * * 1 sec Oxygen Salunlllon during Sleep Apnea {Strohl, A l t o M ~ )

5 Changes in intrathoracic pressure can affect cardiac performance and cardiac output; 9 however, we could not demonstrate that the presence or absence of intermittent or continuous respiratory effi>rts altered the rate of fall in Sa 1 at any given initial Sa 2 Furthermore, three subjects simulated obstructive apneas, creating either rib-cage or abdominal paradox. In this instance, although negative pressures at the mouth are the same, abdominal and regional pleural pressures are differenfl and thus might have differing effects on cardiac output. We could not determine any differences in the rate of decline ofsa 2 between these two maneuvers performed at an initial Sa 2 of 93 ± 2 percent. Weil et al 1 showed that during wakefulness in healthy subjects at altitude with an initial oxygen saturation of approximately 88 percent, oxygen saturation fell to lower levels during inspiratory (- 2 em H 2 ) breath-holds than during expiratory ( + 2 em H 2 ) breath-holds. Breath-holding time or the rate of fall in oxygen saturation was not reported in the study. In the present study, in five subjects at an Sa 2 of approximately 93 percent, there were no differences in the rate of fall of Sa 2 during breath-holds with positive or negative airway pressures. This seeming discrepancy could be explained if in the study by Weil et al 1 breath-holding time was some six to eight seconds longer during the inspiratory effi>rt than the expiratory effi>rt. Alternatively, it is possible that changes in cardiac output consequent to variations in intrathoracic pressure 9 and the resulting changes in the rate of fall of oxygen saturation are more pronounced when hypoxia is more severe. There was a general correspondence between the rate of fall of Sa 2 during breath-holding in wakefulness and that observed in patients during sleep apnea. Previous studies have focused on the role of venous admixture, mismatching of ventilation and perfusion, changing pulmonary volumes, and alterations in cardiac output as factors which would affect oxygen desaturation during sleep apnea. The present study does not exclude these effects but does suggest that the initial Sa 2 is a major determinant of the rate of fall in Sa 2 during sleep apnea. Henderson-Smarfl reported that in the newborn, oxygen saturation falls more abruptly during apneas in sleep than in wakefulness, presumably because of a decrease in end-expiratory pulmonary volume or possibly because of an increase in oxygen consumption during REM sleep. In our study, there was no systematic difference in the rate of fall of Sa 2 during the different stages of sleep, presumably because of the larger stores of oxygen in the adult. It is of interest that one patient (patient 4) demonstrated a rate of fall in Sa 2 during sleep apnea outside the range to be expected from studies in normal subjects during wakefulness. This patient also had atrial fibrillation, and it is possible that in this instance, circulatory adjustments to swings in intrathoracic pressure during airway obstruction might have been impaired. Compensatory changes in cardiac output may play a role in minimizing the fall in oxygen saturation during apneas. Two previous reports have shown that the level to which oxygen saturation falls during obstructive apneas might be greater than that during breath-holding or nonobstructive apneas. u These clinical observations could be explained by a difference in the duration of apneas or in the rate offal} in oxygen saturation (or both). With these factors in mind, we found that the rate of fall in Sa 2 during sleep apnea generally fell within the range of values found in simulated apneas during wakefulness; however, because of the small number of patients with nonobstructive apneas and the variability in the rate of oxygen desaturation among subjects, we are not able to conclude with certainty that the rate of desaturation is not influenced to some degree by the type of sleep apnea. It is likely that factors other than the initial oxygen saturation do contribute to the rate of fall in oxygen saturation during a cessation of breathing. Studies of breath-holding during wakefulness reveal a consistency for each individual in the value for the rate of fall in Sa 2, regardless of the protocol for breathholding; however, for the group, variability for values is wide. Results in sleeping patients show a similar variability (Fig 4). Consequently, a precise quantification of all of the influences affecting changes in oxygen saturation in a given individual would require comparisons in the same subjects during both wakefulness and sleep, as well as a more detailed assessment of metabolic, respiratory, and circulatory factors involved in determining stores of oxygen in the individual subject. In summary, the present study has demonstrated that in adults the rate of fall of Sa 2 during breathholding in wakefulness is not affected by respiratory effi>rts and depends primarily on the initial level of Sa 2 The rate of fall of Sa 2 in patients with sleep apnea is similar to that observed during breath-holding in wakefulness. ACKNOWLEDGMENTS: We thank Ms. Sandy Shaft and Ms. Mary Puryear for their help in the preparation and typing of this manuscript. REFERENCES 1 Wei! JV, Kryger MH, Scoggin CH. Sleep and breathing at high altitude. In: Guilleminault C, Dement WC, eds. Sleep apnea syndromes. New York: Alan R Liss, Inc, 1978: TIIkian AG, Guilleminault C, Schroeder JS, Lehrman KL, Simmons FB, Dement WC. Sleep-induced apnea syndrome: hemodynamic studies during wakefulness and sleep. Ann Intern Med 1976; 85: Henderson-Smart DJ. Vulnerability to hypoxemia in the newborn. Sleep 198; 3: CHEST I 85 I 2 I FEBRUARY,

6 4 Saunders NA, Powles ACP, Rebuck AS. Ear oximetry: accuracy and practicability in the assessment of arterial oxygenation. Am Rev Respir Dis 1976; 113: Sharp JT, Druz WS, Foster JR, Wicks MS, Chokroverty S, et al. Use of the respiratory magnetometer in diagnosis and classification of sleep apnea. Chest 198; 77: Rebuck AS, Campbell EJM. A clinical method for assessing the ventilatory response to hypoxia. Am Rev Respir Dis 1974; 19: Rechtschaffen A, Kales A. A manual of standardized terminology: techniques and scoring system for sleep stages of human subjects. Washington, DC: Public Health Service, Farhi LE. Gas stores of the body. In: Fenn WO, Rahn H, eds. Handbook of physiology (vol I): respiration. Washington, DC: American Physiological Society, 1964: Buda Aj, Pinsky MR, Ingels NB Jr, Daughters GT III, Stinson EB, Alderman EL. Effect of intrathoracic pressure on left ventricular performance. N Eng! J Med 1979; 31: Saunders NA, Kreitzer SM, Ingram RH Jr. Ribcage deformation during static inspiratory efforts. J Appl Physiol1979; 46: Oxygen Salurallon during Sleep Apnea (Strohl, Alfos6)