Electroencephalography and clinical Neurophysiology, 80 ( 1991) 454-457 © 1991 Elsevier Scientific Publishers Ireland, Ltd. 0168-5597/91/$03.50 ADONIS 016855979100108Z
454
EVOPOT 90639
Short communication
Auditory event-related potentials in obstructive sleep apnea: effects of treatment with nasal continuous positive airway pressure L. Rumbach *, J. Krieger and D. Kurtz Service d'Explorations Fonctionnelles du Syst~me Nerveux, CliniqueNeurologique, CHU, Strasbourg (France) (Accepted for publication: 3 June 1991)
Summary
Event-related potentials (ERPs) were recorded in 47 patients with obstructive sleep apnea (OSA) syndrome prior to and after 6 weeks of treatment with continuous positive airway pressure (CPAP). Compared with a control group, the OSA patients showed ERP abnormalities: lengthened P3 latencies and decreased N2-P3 amplitudes. After 6 weeks of CPAP treatment, there was a highly significant improvement in the abnormal ERPs: the P3 and N2 latencies were shortened, but remained longer than in controls, and the N2-P3 and N1-P2 amplitudes were increased. No correlations could be established with various sleep variables. ERPs may be used as an electrophysiological marker of brain dysfunction; treatment of OSA with CPAP is probably responsible for functional brain modifications. On the other hand, possible relationships between the ERP abnormalities and the neuropsychological disorders observed in OSA remain to be established.
Key words: Sleep apnea; Event-related potentials; Continuous positive airway pressure The obstructive sleep apnea (OSA) syndrome is due to the occurrence of upper airway occlusions resulting in apneas lasting 10-60 sec or more, repeated up to several hundred times during a night's sleep (Guilleminault et al. 1978); the major consequences are repeated asphyxia and sleep fragmentation. Clinical manifestations include daytime somnolence, snoring, and sexual deficiency. Obesity, hypertension and polycythemia are not uncommon. Several recent studies have underscored the incidence of cognitive deficits (Kales et al. 1985; Findley et al. 1986). Therapy essentially relies upon the use of continuous positive airway pressure (CPAP) which has been shown to eliminate apneas, restore a normal sleep pattern and improve the patients' symptoms (Sullivan et al. 1981). Starting with the early studies by Sutton et al. (1965), numerous authors have demonstrated the role played by various neuropsychological tasks in the genesis of auditory event-related potentials (ERPs). Thus, the P3 wave appears to be linked to attention, vigilance and memory; ERPs are of great value as a tool for mental chronometry: both N2 an P3 latencies depend upon vigilance and covary with reaction time fluctuations (McCarthy and Donchin 1981; Timsit-Berthier 1984; Hillyard and Pieton 1987; Polich et al. 1989). In view of these data, it seemed of interest to analyze ERPs in OSA patients before and after treatment with CPAP in order to examine whether they may be considered electrophysiological markers of a brain dysfunction.
(range from 24.0 to 47.2) suffering from a sleep-related breathing disorder (either sleep hypopnea (4 patients) or sleep apnea syndrome). Before treatment was initiated, they all underwent 2 consecutive nocturnal polysomnograms including EEG, EOG, EMG and measurement of respiratory variables: ventilation by means of a Fleisch no. 2 pneumotachograph, with a Godart Statham pressure transducer and electronic integrator; esophageal pressure by use of an esophageal balloon linked to a pressure transducer (Validyne MP 45); and arterial SaP z with an ear oximeter (Ohmeda Biox liD. The first polysomnogram served to establish the diagnosis. During the second, CPAP was applied via a nasal mask. The pressure was increased until apneas and snoring were eliminated. The pressure reached was the pressure used for home treatment. In addition, respiratory function tests, daytime arterial blood gas analysis (at rest in the supine position, while breathing room air) and ERP recordings were performed in each patient. Patients were reevaluated after 6 weeks of home treatment; this included a patient interview, a physical examination, and repeated ERP recording. Data from the ERP recordings were compared with those obtained in a group of 40 healthy, age-matched controls (mean age = 47, range = 23-72 years).
Event-related potentials
Methods Patient evaluation The study comprised 47 patients (44 men) aged 39-73 years (mean age = 53) and with a mean body mass index of 32.1 kg/m 2
* Present address: Service de Neurologie, H6pital J. Minjoz, F-25030 Besan~on Codex, France. Correspondence to: Dr. J. Krieger, Ciinique Neurologique, H6pital Central BP 426, 67091 Strasbourg Codex (France). Tel.: (88) 16.13.12; Fax: (88) 16.30.30.
Subjects sat in a reclining chair. Three electrodes were placed at Fz, Cz and Pz (according to the international 10-20 system), referred to linked ears. Eye movements were also monitored. Auditory stimuli were presented binaurally. Two ERPs were recorded in a standard auditory 'oddball' paradigm in the same session; patients were asked to press a button with the right hand for the first recording and in the second recording to count whenever rare (20% of the stimuli) 'high'-pitched target tones (1.5 kHz, 50 msec duration) occurred among a series of frequent (80% of the stimuli) 'low'-pitched nontarget tones (1 kHz, 50 msec duration); reaction time was not measured. Target and non-target tones were presented in a random order with 1000 msec intervals. A total of 40-50 target tones were presented. The signal was filtered with high pass at 0.05 Hz and low
ERPs IN OBSTRUCTIVE SLEEP APNEA
455
TABLE I Mean latencies (msec) and mean amplitudes (/xV) of evoked potentials at Cz in the P300 paradigm in obstructive sleep apnea (OSA) patients, before and after 6 weeks of continuous positive airway pressure (CPAP) treatment compared to controls. Means + standard errors. Latencies
Controls (a) (n = 40) OSA patients (n = 47) Before CPAP (b) After CPAP (c) P (a) vs. (b) (a) vs. (c) (b) vs. (c)
Amplitudes
N1
P2
N2
P3
P1-N1
N1-P2
N2-P3
102.17 + 2.45
155.28 + 3.43
220.32 + 4.52
308.72 ± 3.96
8.25 + 0.63
8.02 + 0.61
12.51 _ 0.62
100.17 + 2.13 99.49 + 2.15
169.30 + 3.36 165.00 + 2.81
242.02 + 4.33 232.41 + 3.86
347.78 + 4.20 330.95 + 4.64
8.32 + 0.76 8.30 + 0.72
11.20 + 0.69 9.43 + 0.63
10.13 J: 0.62 12.12 + 0.76
ns ns ns
0.005 0.05 ns
0.001 0.05 0.01
0.001 0.001 0.0003
ns ns ns
0.005 ns 0.01
0.01 ns 0.02
pass at 100 Hz. Analysis time was 700 msec, without pre-stimulus delay. Only patients who were able to clearly discriminate between the tones and whose counts were within the actual number of target tones _+2 were included (n = 47).
Data analysis Sleep data.
The polysomnographic recordings were scored for sleep stages using standard criteria (Rechtschaffen and Kales 1968). The number of apneic episodes was counted, according to apnea type (obstructive, central or mixed), and an apnea index was calculated by dividing the number of episodes by the total sleep time (in hours). Hypoxemia during sleep was evaluated using 2 parameters: the absolute minimal SaO 2 for the entire recording (minimal SaO 2) and the mean of the minimal SaO2s reached after each apneic episode (mean lowest SaO2). Event-related potentials. The P3 latency was determined as the most positive point (after 250 msec and before 700 msec) of the average wave form to the target tones; when the component was double-peaked or flat-bottomed, P3 latency was determined by extrapolating lines from the ascending and descending slopes of the component. The latencies of the 3 following components were also measured: N1 (defined as the most negative peak on the montage between 50 and 150 msec); P2 (the most positive peak between 150 and 300 msec); N2 (maximum negative peak between 150 and 350 msec). Peak-to-peak amplitudes were determined for P1-N1, N1-P2 and N2-P3; baseline-to-peak amplitudes could not be measured.
Statistical analysis Values observed in untreated OSA patients were compared with those in controls by means of a Student test for unpaired values. The effect of CPAP treatment in OSA patients was evaluated using the t test for paired values. Correlations were analyzed using Pearson's correlation coefficient.
Results
Patient evaluation The patients spanned a wide range of sleep apnea densities (apnea indices ranging from 0.97 to 121.9; mean = 58.9 a p n e a s / h ) and a wide range of hypoxemia during sleep (mean lowest SaO 2 ranging from 57.7% to 95.3%, mean = 86.9%; minimal SaO 2 ranging from 39% to 92%, mean = 72.4%). Their sleep was disorganized and fragmented, as indicated by a high percentage of light non-rapid eye movement (NREM) sleep (stages 1 + 2, 89% _+2.2%) and low percentages of slow wave sleep (stages 3 +4, 9.3%_+ 1.7%) and REM
sleep (8.1% + 0.9%). CPAP treatment eliminated apneas and snoring and improved the patients' sleep pattern. Similarly, their daytime blood gases covered a wide range (PaO 2 from 52.3 mm Hg to 111 mm Hg, mean = 75.1 mm Hg; PaCO 2 from 22 mm Hg to 50 mm Hg, mean = 38.0 mm Hg).
Event-related potentials Prior to CPAP treatment, there were no differences between the control group and the OSA patients in N1 latency and P1-N1 amplitudes; however, the OSA group had greater P2, N2 and P3 latencies (P < 0.003 and P < 0.001, respectively) and N1-P2 amplitudes (P < 0.005) and shorter N2-P3 amplitudes (P < 0.01; Table I). After 6 weeks of CPAP treatment, in the OSA group, the N1 and P2 latencies were unchanged; however, there was a highly significant shortening of the P3 latency, whether it was measured at Cz, Fz or Pz (P < 0.0003). The same applied to the N2 latency, although to a lesser extent ( P < 0.01). Concomitantly, the N2-P3 amplitudes increased at Cz and Pz (P < 0.02) but not at Fz; finally, the N1-P2 amplitudes decreased at Cz ( P < 0.01; Fig. 1). Post-treatment N2 and P3 latencies remained longer than in controls ( P < 0.05 and P < 0.001 respectively), but N1-P2 and N2-P3 amplitudes were not different from controls.
Correlations between ERPs and sleep parameters Two types of possible correlations were examined. First, the correlations between the pretreatment P3 latency and amplitude and the different pretreatment variables were examined in order to identify factors determining P-wave abnormalities. No significant correlations were found, either with variables indicative of sleep disruption or with variables characterizing diurnal or nocturnal hypoxemia. Second, the correlations between the degree of P3 change with treatment and the afore-mentioned variables were analyzed in search of factors responsible for the modifications observed with treatment. The shortening of the P3 latency was correlated with the latency duration prior to treatment (r = 0.44, P < 0.003), and the increase in P3 amplitude was correlated with the P3 amplitude prior to treatment (r = 0.50, P < 0.01); this indicates that the more prolonged pretreatment latencies were shortened the most, and that the smaller pretreatment amplitudes were increased the most.
Discussion
This study demonstrates the presence of abnormal ERPs in OSA and that some of these components may be modified with therapy, showing a rather rapid improvement.
456
L. RUMBACH ET AL.
.3/
........
",,'
'......-' i
i
140
i
i
280
i
5.pv I i
420
i
|
560
i
1 700 ms
Fig. 1. ERP to target tones recorded at Cz in 3 obstructive sleep apnea patients before (solid lines) and after 6 weeks of continuous positive airway pressure treatment (dashed lines). Positive is downwards. Prior to treatment, there were ERP abnormalities in comparison to a control group, with prolonged N2 and P3 latencies, increased N1-P2 amplitudes and decreased N2-P3 amplitudes. ERPs could represent an electrophysiological marker for brain dysfunction. Aguirre and Broughton (1987) and Ollo et al. (1987) have reported a decreased P3 amplitude in another sleep disorder, narcolepsy-cataplexy. The data show that in both types of disease affecting sleep, there are abnormal daytime ERPs. A priori, the only factor the two diseases have in common is somnolence. It is possible that daytime somnolence is responsible for abnormal ERPs, especially an increased P3 latency. However, all the patients included in this study responded correctly to the target tones in 2 ways (button press and count); nevertheless, the P300 paradigm may be simple enough to allow fluctuations between waking and stage 1, since behavioral responses to stimuli similar to those used in the P300 paradigm are possible during sleep stages 1 and 2 (Ogilvie et al. 1989); however, in normal subjects, there is an increasing rate of failure to respond, averaging 38.8% (9.4-84.4%) in stage 1 and 92.1% (71.4-100%) in stage 2 (Ogilvie et al. 1989). Since patients who did not identify the target stimuli during the test were excluded from the study, we can accept that the data were collected during wakefulness. In addition, there was no difference in P1-N1 amplitudes between control and OSA groups which further suggests that the patients and controls had a similar level of vigilance and that patients did not sleep, since it has been shown that there is a decrease in N1 and P3 amplitudes as well as an increase in P3 latency during sleep ( N i i t i n e n and Picton 1987; Wesensten and Badia 1988). Therefore, we are confident that the changes in ERPs observed in this study do not merely reflect the patients' falling asleep during the performance of the test, but that they are related to some degree of brain dysfunction during the waking state. However, further studies including simultaneous vigilance and attention tests, together with recording of the EEG during the test, would be useful to ascertain the absence of a change in vigilance during the course of the investigation.
Six weeks after the start of CPAP treatment, N2 and P3 latencies decreased and N2-P3 amplitude increased. The shortening of the P3 latency could have been linked to that of N2, but the latency reduction of the latter was less, thus showing that treatment is mainly responsible for a decrease in the N2-P3 duration. Walsleben et al. (1989) have reported a reduction in the latency of P3, but not of N2, following 2 days of CPAP treatment in OSA patients. They have shown that these changes were not related to nocturnal hypoxemia. Our data point in the same direction: there was no correlation between nocturnal or diurnal oxygenation and the various ERP components. In addition, none of the sleep parameters was correlated with the ERP descriptors. One possible explanation for the observed changes in ERPs could be changes in neurotransmitter metabolism associated with sleep apnea; it has been shown that 5-hydroxyindoleacetic acid levels were increased in the cerebrospinal fluid of OSA patients (Cramer et aL 1981), and that P3 latencies may be modified by drugs acting upon various neurotransmitters (Meador et al. 1987). The observation made in our study that the longer pretreatment P3 latencies were the most shortened with CPAP supports the hypothesis of a functional disturbance, On an anatomical level, the observed changes in ERPs could reflect disorders of the temporal region, particularly the limbic system (Halgren et al. 1980; McCarthy et al. 1982; Smith et al. 1986; Stapleton et al. 1987; Daruna et al., 1989; Polich 1989). It is presently accepted that the P3 wave is related to cognitive processes, attention, and memorization (Donchin 1981; Duncan-Johnson and Donchin 1982; Timsit-Berthier, 1984; Johnson 1986; Donchin and Coles 1988), and that the N1 wave may be the expression of several components: attention, tone recognition and recall of its characteristics, and sensory memory build-up (Hansen and Hillyard 1980; N i i t i n e n and Picton 1987). Researchers who have administered various neuropsychological tests to OSA patients have underscored disorders in short-term memory, attention, concentration, and verbal fluency (Kales et al. 1985; Findley et al. 1986; Greenberg et al. 1987; Klonoff
ERPs IN OBSTRUCTIVE SLEEP APNEA et al. 1987; Derderian et al. 1988); these disorders may improve with CPAP (Watson et al. 1985; Bearpark et al. 1987; Derderian et al. 1988). Given these data, it is tempting to establish a link between the improvement in certain ERP components and that of the neuropsychological disorders. However, Walsleben et al. (1989) did not observe significant changes in neuropsychological tests after 2 days of CPAP treatment. This raises the question of the validity of neuropsychological tests coupled with ERP recordings and of the functional significance of the ERP components. One still needs to determine the relationship between changes in an objective, electrophysiological marker such as ERPs and both the anatomical-functional and cognitive disorders in OSA.
References
Aguirre, M. and Broughton, R.J. Complex event-related potentials (P300 and CNV) and MSLT in the assessment of excessive daytime sleepiness in narcolepsy-cataplexy. Electroenceph. clin. Neurophysiol., 1987, 67: 298-316. Bearpark, H., Grunstein, R., Touyz, S., Channon, L. and Sullivan, C. Cognitive and psychological dysfunction in sleep apnea before and after treatment with CPAP. Sleep Res., 1987, 16: 303. Cramer, H., Warter, J.M., Renaud, B., Krieger, J., Marescaux, C. and Hammers, R. Cerebrospinal fluid adenosine 3',5'-monophosphate, 5-hydroxyindolacetic acid and homovanillic acid in patients with sleep apnoea syndrome. J. Neurol. Neurosurg. Psychiat., 1981, 44: 1165-1167. Daruna, J.H., Nelson, A.V. and Green, J.B. Unilateral temporal lobe lesions alter P300 scalp topography. Int. J. Neurosci., 1989, 46: 243-247. Derderian, S.S., Bridenbaugh, H. and Rajagopal, K.R. Neuropsychologic symptoms in obstructive sleep apnea improve after treatment with nasal continuous positive airway pressure. Chest, 1988, 94: 1023-1027. Donchin, E. Surprise!...surprise? Psychophysiology, 1981, 18: 493513. Donchin, E. and Coles, M.G.H. Is the P300 component a manifestation of context updating? Behav. Brain Sci., 1988, 11: 355-425. Duncan-Johnson, C.C. and Donchin, E. The P300 component of the event-related brain potential as an index of information processing. Biol. Psychol., 1982, 14: 1-52. Findley, L.J., Barth, J.T., Powers, D.C., Wilhoit, S.C., Boyd, D.C. and Suratt, P.M. Cognitive impairment in patients with obstructive sleep apnea and associated hypoxemia. Chest, 1986, 90: 686-690. Goodin, D.S., Squires, K.C. and Starr, A. Long latency event-related components of the auditory evoked potential in dementia. Brain, 1978, 101: 635-648. Greenberg, G.D., Squires, R.K. and Deptula, D. Neuropsychological dysfunction in sleep apnea. Sleep, 1987, 10: 254-262. Guilleminault, C., Van den Hoed, J. and Mitler, M.M. Clinical overview of the sleep apnea syndromes. In: C. Guilleminault and W.C. Dement (Eds.), Sleep Apnea Syndromes. Alan R. Liss, New York, 1978: 1-12. Halgren, E., Squires, N., Wilson, C., Rohrbaugh, J., Babb, T. and Crandall, P. Endogenous potentials in the human hippocampal formation and amygdala by infrequent events. Science, 1980, 210: 803-805. Hansen, J.C. and Hillyard, S.A. Endogenous brain potentials associated with selective auditory attention. Electroenceph. clin. Neurophysiol., 1980, 49: 277-290. Hillyard, S.A. and Picton, T.W. Electrophysiology of cognition. In: F. Plum (Volume Ed.), V.B. Mountcastle (Section Ed.), Handbook of Physiology, Vol. V, Section I. American Physiological Society, Bethesda, MD, 1987: 519-584. Issa, F., Grunstein, R., Bruderer, J., Costas, L., McCauley, V. and
457 Berthon-Jones, M. Five years' experience with home nasal continuous positive airway pressure therapy for the obstructive sleep apnea syndrome. In: J.H. Peter, T. Podzsus and P. von Wichert (Eds.), Sleep Related Disorders and Internal Diseases. Springer, Berlin, 1987: 360-365. Johnson, R. A triarchic model of P300 amplitude. Psychophysiology, 1986, 23: 367-385. Kales, A., Caldwell, A.B., Cadieux, R.J., Vella-Bueno, A., Ruch, L.G. and Mayes, S.D. Severe obstructive sleep apnea. II. Associated psychopathology and psychosocial consequences. J. Chron. Dis., 1985, 38: 427-434. Klonoff, H., Fleetham, J., Taylor, D.R. and Clark, C. Treatment outcome of obstructive sleep apnea physiological and neuropsychological concomitants. J. Nerv. Ment. Dis., 1987, 1975: 208-212. Kutcher, S.P., Blackwood, D., St. Clair, D., Gaskell, D.F. and Muir, W.J. Auditory P300 in borderline personality disorder and schizophrenia. Arch. Gen. Psychiat., 1987, 44: 645-650. Legall, D., Hubert, P., Truelle, J.L., Meslier, N., De Bray, J.M. and Racineux, J.L. Le syndrome d'apn6e du sommeil. Une cause curable de d&6rioration mentale. Nouv. Presse M6d., 1986, 15: 260-261. McCarthy, G. and Donchin, E. A metric for thought: a comparison of P300 latency and reaction time. Science, 1981, 211: 77-80. McCarthy, G., Wood, C.C., Allison, T., Goff, W.R., Williamson, P.D. and Spencer, D.D. Intracranial recording of event-related potentials in humans engaged in cognitive tasks. Neurosci. Abst., 1982, 8: 976. N~i~it~inen, R. and Picton, T. The N1 wave of the human electric and magnetic response to sound: a review and an analysis of the component structure. Psychophysiology, 1987, 24: 375-425. Ogilvie, R.D., Wilkinson, R.T. and Allison, S.A. The detection of sleep onset: behavioral, physiological, and subjective convergence. Sleep, 1989, 12: 458-474. Ollo, C., Squires, N., Pass, H., Walsleben, J., Baker, T. and Gujavarty, K. Electrophysiological and neuropsychological assessment of cognitive function in narcolepsy. Sleep Res, 1987, 16: 402. Polich, J. Habituation of P300 from auditory stimuli. Psychobiology, 1989, 17: 19-28. Polich, J., Ehlers, C.E., Otis, S., Mandell, A.J. and Bloom, F.E. P300 reflects the degree of cognitive decline in dementing illness. Electroenceph. clin. Neurophysiol., 1986, 63: 138-144. Rechtschaffen, A. and Kales, A. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. Brain Information Service, Brain Research Institute, Los Angeles, CA, 1968. Smith, M.E., Stapleton, J.M. and Halgren, E. Human medial temporal lobe potentials evoked in memory and language tasks. Electroenceph, clin. Neurophysiol., 1986, 63: 145-159. Stapleton, J.M., Halgren, E. and Moreno, K.A. Endogenous potentials after anterior temporal Iobectomy. Neuropsychologia, 1987, 25: 549-557. Sullivan, C.E., Issa, F.G., Berthon-Jones, M. and Eves, L. Reversal of obstructive sleep apnoea by continuous positive airway pressure applied through the nares. Lancet, 1981, i: 862-865. Sutton, S., Braren, M., Zubin, J. and John, E.R. Evoked potential correlates of stimulus uncertainty. Science, 1965, 150:1187-1188. Timsit-Berthier, M. Variation contingente n6gative et composantes endog~nes du potentiel 6voqu6. Rev. EEG Neurophysiol., 1984, 14: 77-96. Walsleben, J.A., Squires, N.K. and Rothenberger, V.L. Auditory event-related potentials and brain dysfunction in sleep apnea. Electroenceph. clin. Neurophysiol., 1989, 74:297-311. Watson, R., Greenberg, G. and Deptula, D. Neuropsychological performance following treatment of sleep apnea. Sleep Res., 1985, 14: 136. Wesensten, N.J. and Badia, P. The P300 component in sleep. Physiol. Behav., 1988, 44: 215-220.
454
EVOPOT 90639
Short communication
Auditory event-related potentials in obstructive sleep apnea: effects of treatment with nasal continuous positive airway pressure L. Rumbach *, J. Krieger and D. Kurtz Service d'Explorations Fonctionnelles du Syst~me Nerveux, CliniqueNeurologique, CHU, Strasbourg (France) (Accepted for publication: 3 June 1991)
Summary
Event-related potentials (ERPs) were recorded in 47 patients with obstructive sleep apnea (OSA) syndrome prior to and after 6 weeks of treatment with continuous positive airway pressure (CPAP). Compared with a control group, the OSA patients showed ERP abnormalities: lengthened P3 latencies and decreased N2-P3 amplitudes. After 6 weeks of CPAP treatment, there was a highly significant improvement in the abnormal ERPs: the P3 and N2 latencies were shortened, but remained longer than in controls, and the N2-P3 and N1-P2 amplitudes were increased. No correlations could be established with various sleep variables. ERPs may be used as an electrophysiological marker of brain dysfunction; treatment of OSA with CPAP is probably responsible for functional brain modifications. On the other hand, possible relationships between the ERP abnormalities and the neuropsychological disorders observed in OSA remain to be established.
Key words: Sleep apnea; Event-related potentials; Continuous positive airway pressure The obstructive sleep apnea (OSA) syndrome is due to the occurrence of upper airway occlusions resulting in apneas lasting 10-60 sec or more, repeated up to several hundred times during a night's sleep (Guilleminault et al. 1978); the major consequences are repeated asphyxia and sleep fragmentation. Clinical manifestations include daytime somnolence, snoring, and sexual deficiency. Obesity, hypertension and polycythemia are not uncommon. Several recent studies have underscored the incidence of cognitive deficits (Kales et al. 1985; Findley et al. 1986). Therapy essentially relies upon the use of continuous positive airway pressure (CPAP) which has been shown to eliminate apneas, restore a normal sleep pattern and improve the patients' symptoms (Sullivan et al. 1981). Starting with the early studies by Sutton et al. (1965), numerous authors have demonstrated the role played by various neuropsychological tasks in the genesis of auditory event-related potentials (ERPs). Thus, the P3 wave appears to be linked to attention, vigilance and memory; ERPs are of great value as a tool for mental chronometry: both N2 an P3 latencies depend upon vigilance and covary with reaction time fluctuations (McCarthy and Donchin 1981; Timsit-Berthier 1984; Hillyard and Pieton 1987; Polich et al. 1989). In view of these data, it seemed of interest to analyze ERPs in OSA patients before and after treatment with CPAP in order to examine whether they may be considered electrophysiological markers of a brain dysfunction.
(range from 24.0 to 47.2) suffering from a sleep-related breathing disorder (either sleep hypopnea (4 patients) or sleep apnea syndrome). Before treatment was initiated, they all underwent 2 consecutive nocturnal polysomnograms including EEG, EOG, EMG and measurement of respiratory variables: ventilation by means of a Fleisch no. 2 pneumotachograph, with a Godart Statham pressure transducer and electronic integrator; esophageal pressure by use of an esophageal balloon linked to a pressure transducer (Validyne MP 45); and arterial SaP z with an ear oximeter (Ohmeda Biox liD. The first polysomnogram served to establish the diagnosis. During the second, CPAP was applied via a nasal mask. The pressure was increased until apneas and snoring were eliminated. The pressure reached was the pressure used for home treatment. In addition, respiratory function tests, daytime arterial blood gas analysis (at rest in the supine position, while breathing room air) and ERP recordings were performed in each patient. Patients were reevaluated after 6 weeks of home treatment; this included a patient interview, a physical examination, and repeated ERP recording. Data from the ERP recordings were compared with those obtained in a group of 40 healthy, age-matched controls (mean age = 47, range = 23-72 years).
Event-related potentials
Methods Patient evaluation The study comprised 47 patients (44 men) aged 39-73 years (mean age = 53) and with a mean body mass index of 32.1 kg/m 2
* Present address: Service de Neurologie, H6pital J. Minjoz, F-25030 Besan~on Codex, France. Correspondence to: Dr. J. Krieger, Ciinique Neurologique, H6pital Central BP 426, 67091 Strasbourg Codex (France). Tel.: (88) 16.13.12; Fax: (88) 16.30.30.
Subjects sat in a reclining chair. Three electrodes were placed at Fz, Cz and Pz (according to the international 10-20 system), referred to linked ears. Eye movements were also monitored. Auditory stimuli were presented binaurally. Two ERPs were recorded in a standard auditory 'oddball' paradigm in the same session; patients were asked to press a button with the right hand for the first recording and in the second recording to count whenever rare (20% of the stimuli) 'high'-pitched target tones (1.5 kHz, 50 msec duration) occurred among a series of frequent (80% of the stimuli) 'low'-pitched nontarget tones (1 kHz, 50 msec duration); reaction time was not measured. Target and non-target tones were presented in a random order with 1000 msec intervals. A total of 40-50 target tones were presented. The signal was filtered with high pass at 0.05 Hz and low
ERPs IN OBSTRUCTIVE SLEEP APNEA
455
TABLE I Mean latencies (msec) and mean amplitudes (/xV) of evoked potentials at Cz in the P300 paradigm in obstructive sleep apnea (OSA) patients, before and after 6 weeks of continuous positive airway pressure (CPAP) treatment compared to controls. Means + standard errors. Latencies
Controls (a) (n = 40) OSA patients (n = 47) Before CPAP (b) After CPAP (c) P (a) vs. (b) (a) vs. (c) (b) vs. (c)
Amplitudes
N1
P2
N2
P3
P1-N1
N1-P2
N2-P3
102.17 + 2.45
155.28 + 3.43
220.32 + 4.52
308.72 ± 3.96
8.25 + 0.63
8.02 + 0.61
12.51 _ 0.62
100.17 + 2.13 99.49 + 2.15
169.30 + 3.36 165.00 + 2.81
242.02 + 4.33 232.41 + 3.86
347.78 + 4.20 330.95 + 4.64
8.32 + 0.76 8.30 + 0.72
11.20 + 0.69 9.43 + 0.63
10.13 J: 0.62 12.12 + 0.76
ns ns ns
0.005 0.05 ns
0.001 0.05 0.01
0.001 0.001 0.0003
ns ns ns
0.005 ns 0.01
0.01 ns 0.02
pass at 100 Hz. Analysis time was 700 msec, without pre-stimulus delay. Only patients who were able to clearly discriminate between the tones and whose counts were within the actual number of target tones _+2 were included (n = 47).
Data analysis Sleep data.
The polysomnographic recordings were scored for sleep stages using standard criteria (Rechtschaffen and Kales 1968). The number of apneic episodes was counted, according to apnea type (obstructive, central or mixed), and an apnea index was calculated by dividing the number of episodes by the total sleep time (in hours). Hypoxemia during sleep was evaluated using 2 parameters: the absolute minimal SaO 2 for the entire recording (minimal SaO 2) and the mean of the minimal SaO2s reached after each apneic episode (mean lowest SaO2). Event-related potentials. The P3 latency was determined as the most positive point (after 250 msec and before 700 msec) of the average wave form to the target tones; when the component was double-peaked or flat-bottomed, P3 latency was determined by extrapolating lines from the ascending and descending slopes of the component. The latencies of the 3 following components were also measured: N1 (defined as the most negative peak on the montage between 50 and 150 msec); P2 (the most positive peak between 150 and 300 msec); N2 (maximum negative peak between 150 and 350 msec). Peak-to-peak amplitudes were determined for P1-N1, N1-P2 and N2-P3; baseline-to-peak amplitudes could not be measured.
Statistical analysis Values observed in untreated OSA patients were compared with those in controls by means of a Student test for unpaired values. The effect of CPAP treatment in OSA patients was evaluated using the t test for paired values. Correlations were analyzed using Pearson's correlation coefficient.
Results
Patient evaluation The patients spanned a wide range of sleep apnea densities (apnea indices ranging from 0.97 to 121.9; mean = 58.9 a p n e a s / h ) and a wide range of hypoxemia during sleep (mean lowest SaO 2 ranging from 57.7% to 95.3%, mean = 86.9%; minimal SaO 2 ranging from 39% to 92%, mean = 72.4%). Their sleep was disorganized and fragmented, as indicated by a high percentage of light non-rapid eye movement (NREM) sleep (stages 1 + 2, 89% _+2.2%) and low percentages of slow wave sleep (stages 3 +4, 9.3%_+ 1.7%) and REM
sleep (8.1% + 0.9%). CPAP treatment eliminated apneas and snoring and improved the patients' sleep pattern. Similarly, their daytime blood gases covered a wide range (PaO 2 from 52.3 mm Hg to 111 mm Hg, mean = 75.1 mm Hg; PaCO 2 from 22 mm Hg to 50 mm Hg, mean = 38.0 mm Hg).
Event-related potentials Prior to CPAP treatment, there were no differences between the control group and the OSA patients in N1 latency and P1-N1 amplitudes; however, the OSA group had greater P2, N2 and P3 latencies (P < 0.003 and P < 0.001, respectively) and N1-P2 amplitudes (P < 0.005) and shorter N2-P3 amplitudes (P < 0.01; Table I). After 6 weeks of CPAP treatment, in the OSA group, the N1 and P2 latencies were unchanged; however, there was a highly significant shortening of the P3 latency, whether it was measured at Cz, Fz or Pz (P < 0.0003). The same applied to the N2 latency, although to a lesser extent ( P < 0.01). Concomitantly, the N2-P3 amplitudes increased at Cz and Pz (P < 0.02) but not at Fz; finally, the N1-P2 amplitudes decreased at Cz ( P < 0.01; Fig. 1). Post-treatment N2 and P3 latencies remained longer than in controls ( P < 0.05 and P < 0.001 respectively), but N1-P2 and N2-P3 amplitudes were not different from controls.
Correlations between ERPs and sleep parameters Two types of possible correlations were examined. First, the correlations between the pretreatment P3 latency and amplitude and the different pretreatment variables were examined in order to identify factors determining P-wave abnormalities. No significant correlations were found, either with variables indicative of sleep disruption or with variables characterizing diurnal or nocturnal hypoxemia. Second, the correlations between the degree of P3 change with treatment and the afore-mentioned variables were analyzed in search of factors responsible for the modifications observed with treatment. The shortening of the P3 latency was correlated with the latency duration prior to treatment (r = 0.44, P < 0.003), and the increase in P3 amplitude was correlated with the P3 amplitude prior to treatment (r = 0.50, P < 0.01); this indicates that the more prolonged pretreatment latencies were shortened the most, and that the smaller pretreatment amplitudes were increased the most.
Discussion
This study demonstrates the presence of abnormal ERPs in OSA and that some of these components may be modified with therapy, showing a rather rapid improvement.
456
L. RUMBACH ET AL.
.3/
........
",,'
'......-' i
i
140
i
i
280
i
5.pv I i
420
i
|
560
i
1 700 ms
Fig. 1. ERP to target tones recorded at Cz in 3 obstructive sleep apnea patients before (solid lines) and after 6 weeks of continuous positive airway pressure treatment (dashed lines). Positive is downwards. Prior to treatment, there were ERP abnormalities in comparison to a control group, with prolonged N2 and P3 latencies, increased N1-P2 amplitudes and decreased N2-P3 amplitudes. ERPs could represent an electrophysiological marker for brain dysfunction. Aguirre and Broughton (1987) and Ollo et al. (1987) have reported a decreased P3 amplitude in another sleep disorder, narcolepsy-cataplexy. The data show that in both types of disease affecting sleep, there are abnormal daytime ERPs. A priori, the only factor the two diseases have in common is somnolence. It is possible that daytime somnolence is responsible for abnormal ERPs, especially an increased P3 latency. However, all the patients included in this study responded correctly to the target tones in 2 ways (button press and count); nevertheless, the P300 paradigm may be simple enough to allow fluctuations between waking and stage 1, since behavioral responses to stimuli similar to those used in the P300 paradigm are possible during sleep stages 1 and 2 (Ogilvie et al. 1989); however, in normal subjects, there is an increasing rate of failure to respond, averaging 38.8% (9.4-84.4%) in stage 1 and 92.1% (71.4-100%) in stage 2 (Ogilvie et al. 1989). Since patients who did not identify the target stimuli during the test were excluded from the study, we can accept that the data were collected during wakefulness. In addition, there was no difference in P1-N1 amplitudes between control and OSA groups which further suggests that the patients and controls had a similar level of vigilance and that patients did not sleep, since it has been shown that there is a decrease in N1 and P3 amplitudes as well as an increase in P3 latency during sleep ( N i i t i n e n and Picton 1987; Wesensten and Badia 1988). Therefore, we are confident that the changes in ERPs observed in this study do not merely reflect the patients' falling asleep during the performance of the test, but that they are related to some degree of brain dysfunction during the waking state. However, further studies including simultaneous vigilance and attention tests, together with recording of the EEG during the test, would be useful to ascertain the absence of a change in vigilance during the course of the investigation.
Six weeks after the start of CPAP treatment, N2 and P3 latencies decreased and N2-P3 amplitude increased. The shortening of the P3 latency could have been linked to that of N2, but the latency reduction of the latter was less, thus showing that treatment is mainly responsible for a decrease in the N2-P3 duration. Walsleben et al. (1989) have reported a reduction in the latency of P3, but not of N2, following 2 days of CPAP treatment in OSA patients. They have shown that these changes were not related to nocturnal hypoxemia. Our data point in the same direction: there was no correlation between nocturnal or diurnal oxygenation and the various ERP components. In addition, none of the sleep parameters was correlated with the ERP descriptors. One possible explanation for the observed changes in ERPs could be changes in neurotransmitter metabolism associated with sleep apnea; it has been shown that 5-hydroxyindoleacetic acid levels were increased in the cerebrospinal fluid of OSA patients (Cramer et aL 1981), and that P3 latencies may be modified by drugs acting upon various neurotransmitters (Meador et al. 1987). The observation made in our study that the longer pretreatment P3 latencies were the most shortened with CPAP supports the hypothesis of a functional disturbance, On an anatomical level, the observed changes in ERPs could reflect disorders of the temporal region, particularly the limbic system (Halgren et al. 1980; McCarthy et al. 1982; Smith et al. 1986; Stapleton et al. 1987; Daruna et al., 1989; Polich 1989). It is presently accepted that the P3 wave is related to cognitive processes, attention, and memorization (Donchin 1981; Duncan-Johnson and Donchin 1982; Timsit-Berthier, 1984; Johnson 1986; Donchin and Coles 1988), and that the N1 wave may be the expression of several components: attention, tone recognition and recall of its characteristics, and sensory memory build-up (Hansen and Hillyard 1980; N i i t i n e n and Picton 1987). Researchers who have administered various neuropsychological tests to OSA patients have underscored disorders in short-term memory, attention, concentration, and verbal fluency (Kales et al. 1985; Findley et al. 1986; Greenberg et al. 1987; Klonoff
ERPs IN OBSTRUCTIVE SLEEP APNEA et al. 1987; Derderian et al. 1988); these disorders may improve with CPAP (Watson et al. 1985; Bearpark et al. 1987; Derderian et al. 1988). Given these data, it is tempting to establish a link between the improvement in certain ERP components and that of the neuropsychological disorders. However, Walsleben et al. (1989) did not observe significant changes in neuropsychological tests after 2 days of CPAP treatment. This raises the question of the validity of neuropsychological tests coupled with ERP recordings and of the functional significance of the ERP components. One still needs to determine the relationship between changes in an objective, electrophysiological marker such as ERPs and both the anatomical-functional and cognitive disorders in OSA.
References
Aguirre, M. and Broughton, R.J. Complex event-related potentials (P300 and CNV) and MSLT in the assessment of excessive daytime sleepiness in narcolepsy-cataplexy. Electroenceph. clin. Neurophysiol., 1987, 67: 298-316. Bearpark, H., Grunstein, R., Touyz, S., Channon, L. and Sullivan, C. Cognitive and psychological dysfunction in sleep apnea before and after treatment with CPAP. Sleep Res., 1987, 16: 303. Cramer, H., Warter, J.M., Renaud, B., Krieger, J., Marescaux, C. and Hammers, R. Cerebrospinal fluid adenosine 3',5'-monophosphate, 5-hydroxyindolacetic acid and homovanillic acid in patients with sleep apnoea syndrome. J. Neurol. Neurosurg. Psychiat., 1981, 44: 1165-1167. Daruna, J.H., Nelson, A.V. and Green, J.B. Unilateral temporal lobe lesions alter P300 scalp topography. Int. J. Neurosci., 1989, 46: 243-247. Derderian, S.S., Bridenbaugh, H. and Rajagopal, K.R. Neuropsychologic symptoms in obstructive sleep apnea improve after treatment with nasal continuous positive airway pressure. Chest, 1988, 94: 1023-1027. Donchin, E. Surprise!...surprise? Psychophysiology, 1981, 18: 493513. Donchin, E. and Coles, M.G.H. Is the P300 component a manifestation of context updating? Behav. Brain Sci., 1988, 11: 355-425. Duncan-Johnson, C.C. and Donchin, E. The P300 component of the event-related brain potential as an index of information processing. Biol. Psychol., 1982, 14: 1-52. Findley, L.J., Barth, J.T., Powers, D.C., Wilhoit, S.C., Boyd, D.C. and Suratt, P.M. Cognitive impairment in patients with obstructive sleep apnea and associated hypoxemia. Chest, 1986, 90: 686-690. Goodin, D.S., Squires, K.C. and Starr, A. Long latency event-related components of the auditory evoked potential in dementia. Brain, 1978, 101: 635-648. Greenberg, G.D., Squires, R.K. and Deptula, D. Neuropsychological dysfunction in sleep apnea. Sleep, 1987, 10: 254-262. Guilleminault, C., Van den Hoed, J. and Mitler, M.M. Clinical overview of the sleep apnea syndromes. In: C. Guilleminault and W.C. Dement (Eds.), Sleep Apnea Syndromes. Alan R. Liss, New York, 1978: 1-12. Halgren, E., Squires, N., Wilson, C., Rohrbaugh, J., Babb, T. and Crandall, P. Endogenous potentials in the human hippocampal formation and amygdala by infrequent events. Science, 1980, 210: 803-805. Hansen, J.C. and Hillyard, S.A. Endogenous brain potentials associated with selective auditory attention. Electroenceph. clin. Neurophysiol., 1980, 49: 277-290. Hillyard, S.A. and Picton, T.W. Electrophysiology of cognition. In: F. Plum (Volume Ed.), V.B. Mountcastle (Section Ed.), Handbook of Physiology, Vol. V, Section I. American Physiological Society, Bethesda, MD, 1987: 519-584. Issa, F., Grunstein, R., Bruderer, J., Costas, L., McCauley, V. and
457 Berthon-Jones, M. Five years' experience with home nasal continuous positive airway pressure therapy for the obstructive sleep apnea syndrome. In: J.H. Peter, T. Podzsus and P. von Wichert (Eds.), Sleep Related Disorders and Internal Diseases. Springer, Berlin, 1987: 360-365. Johnson, R. A triarchic model of P300 amplitude. Psychophysiology, 1986, 23: 367-385. Kales, A., Caldwell, A.B., Cadieux, R.J., Vella-Bueno, A., Ruch, L.G. and Mayes, S.D. Severe obstructive sleep apnea. II. Associated psychopathology and psychosocial consequences. J. Chron. Dis., 1985, 38: 427-434. Klonoff, H., Fleetham, J., Taylor, D.R. and Clark, C. Treatment outcome of obstructive sleep apnea physiological and neuropsychological concomitants. J. Nerv. Ment. Dis., 1987, 1975: 208-212. Kutcher, S.P., Blackwood, D., St. Clair, D., Gaskell, D.F. and Muir, W.J. Auditory P300 in borderline personality disorder and schizophrenia. Arch. Gen. Psychiat., 1987, 44: 645-650. Legall, D., Hubert, P., Truelle, J.L., Meslier, N., De Bray, J.M. and Racineux, J.L. Le syndrome d'apn6e du sommeil. Une cause curable de d&6rioration mentale. Nouv. Presse M6d., 1986, 15: 260-261. McCarthy, G. and Donchin, E. A metric for thought: a comparison of P300 latency and reaction time. Science, 1981, 211: 77-80. McCarthy, G., Wood, C.C., Allison, T., Goff, W.R., Williamson, P.D. and Spencer, D.D. Intracranial recording of event-related potentials in humans engaged in cognitive tasks. Neurosci. Abst., 1982, 8: 976. N~i~it~inen, R. and Picton, T. The N1 wave of the human electric and magnetic response to sound: a review and an analysis of the component structure. Psychophysiology, 1987, 24: 375-425. Ogilvie, R.D., Wilkinson, R.T. and Allison, S.A. The detection of sleep onset: behavioral, physiological, and subjective convergence. Sleep, 1989, 12: 458-474. Ollo, C., Squires, N., Pass, H., Walsleben, J., Baker, T. and Gujavarty, K. Electrophysiological and neuropsychological assessment of cognitive function in narcolepsy. Sleep Res, 1987, 16: 402. Polich, J. Habituation of P300 from auditory stimuli. Psychobiology, 1989, 17: 19-28. Polich, J., Ehlers, C.E., Otis, S., Mandell, A.J. and Bloom, F.E. P300 reflects the degree of cognitive decline in dementing illness. Electroenceph. clin. Neurophysiol., 1986, 63: 138-144. Rechtschaffen, A. and Kales, A. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. Brain Information Service, Brain Research Institute, Los Angeles, CA, 1968. Smith, M.E., Stapleton, J.M. and Halgren, E. Human medial temporal lobe potentials evoked in memory and language tasks. Electroenceph, clin. Neurophysiol., 1986, 63: 145-159. Stapleton, J.M., Halgren, E. and Moreno, K.A. Endogenous potentials after anterior temporal Iobectomy. Neuropsychologia, 1987, 25: 549-557. Sullivan, C.E., Issa, F.G., Berthon-Jones, M. and Eves, L. Reversal of obstructive sleep apnoea by continuous positive airway pressure applied through the nares. Lancet, 1981, i: 862-865. Sutton, S., Braren, M., Zubin, J. and John, E.R. Evoked potential correlates of stimulus uncertainty. Science, 1965, 150:1187-1188. Timsit-Berthier, M. Variation contingente n6gative et composantes endog~nes du potentiel 6voqu6. Rev. EEG Neurophysiol., 1984, 14: 77-96. Walsleben, J.A., Squires, N.K. and Rothenberger, V.L. Auditory event-related potentials and brain dysfunction in sleep apnea. Electroenceph. clin. Neurophysiol., 1989, 74:297-311. Watson, R., Greenberg, G. and Deptula, D. Neuropsychological performance following treatment of sleep apnea. Sleep Res., 1985, 14: 136. Wesensten, N.J. and Badia, P. The P300 component in sleep. Physiol. Behav., 1988, 44: 215-220.
