INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Many of the patients (45 to 57%) with obstructive sleep apnea syndrome (OSAS) have been reported to suffer from systemic hypertension (1). Many patients with OSAS (17 to 42%) also suffer from daytime pulmonary hypertension (PH) (4). A number of studies have claimed that the incidence and severity of complications in the cardiovascular system such as systemic hypertension and coronary vascular disease can be important prognostic factors in OSAS (2, 9). Moreover, it is likely that the involvement of PH and/or right-sided cardiac failure influences the mortality in OSAS, although this argument has so far not been clearly resolved. In fact, in patients with chronic obstructive pulmonary disease, the association with PH during wakefulness was reported to show a poorer prognosis than those without PH (10). Therefore, it seems important to determine in every patient with OSAS whether deteriorating pulmonary circulation is coexistent. From the viewpoint of prognostic and therapeutic aspects, this information may have indispensable value.

In addition to daytime PH, transient elevation of pulmonary artery pressure (Ppa) has also been reported in association with sleep apnea by many investigators (11). This phenomenon is especially prominent during rapid eye movement (REM) sleep, as it was reported that systolic Ppa could attain a level of 80 mm Hg (12), resulting in an increase in right cardiac afterload (14), cardiac arrhythmia (15), or ischemic heart disease (9). So far, it has not been resolved whether nocturnal PH can cause daytime PH, or whether the existence of daytime PH influences nocturnal PH. Ppa elevation under hypoxic condition is mainly caused by hypoxic pulmonary vasoconstriction (16), which is also considered to play an important role in a transient elevation of Ppa during sleep apnea in patients with OSAS. During REM sleep, Ppa elevation occurs in response to hypoxic pulmonary vasoconstriction more prominently than during non-rapid eye movement (NREM) sleep, because prolonged apnea is usually observed in association with marked arterial oxygen desaturation. However, the effect of state-specific change, i.e., REM or NREM sleep, on the pulmonary circulation has not been well explored in OSAS, whereas the effect of sleep stages on the systemic circulation has been reported (19, 20). The aim of this study is to determine whether there are state-specific changes in pulmonary hemodynamics in OSAS, and whether these differ in patients with and without PH.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All patients were referred to our hospital for the purpose of confirming the diagnosis of OSAS, and also for assessing the severity of OSAS, performing right cardiac catheterization, and introducing nasal continuous positive airway pressure (CPAP). Informed consent for right cardiac catheterization was obtained from all patients at the outpatient clinic prior to admission. They all underwent full overnight polysomnographic sleep study for two successive nights, as well as a respiratory function test as a part of their routine clinical investigations. Right cardiac catheterization was performed as soon as polysomnographic sleep studies were completed.

Subjects

Thirty-one patients underwent polysomnographic study with continuous Ppa monitoring by right cardiac catheterization. In 19 patients continuous Ppa measurement was successfully achieved throughout NREM (stage I-II) and REM sleep stages. The other 12 patients were eliminated from the analysis because REM sleep stage could not be well observed or recorded satisfactorily. Ppa changes during slow wave sleep (stage III-IV) were not analyzed in this study because slow wave sleep could only be recorded in a few patients. The 19 patients had mild to severe OSAS with apnea-hypopnea indices ranging from 11.6 to 95.5. Characteristics of the 19 patients are shown in Table 1.

Sleep Studies

All patients underwent first and second night polysomnographic study including electroencephalograms based on the criteria of Rechtschaffen and Kales (21), surface electromyogram of genioglossal muscle, electro-oculogram (EOG), electrocardiogram (ECG), oral and nasal air flow by flow thermistor (Polysomnograph System, NEC San-ei, Tokyo, Japan), arterial oxygen saturation by a pulse oximeter (Sp_O2) (Biox 3700, Ohmeda, Louisville, CO; response time, 3 s), and rib cage and abdominal wall movement by impedance plethysmography (Respitrace Systems; Respitrace, Ardsley, NY). The severity of sleep-disordered breathing was determined from the record of the second-night study.

Respiratory Function Test

Conventional spirography was performed and static lung volumes were measured using the close circuit helium dilution technique (Fudac-60; Fukuda Denshi, Tokyo, Japan) in a sitting position. Arterial blood gases and acid/base status during room air breathing were measured by blood gas analyzer (model 1312; Instrumental Laboratory, Milano, Italy). Both examinations were performed in the morning after all patients were fully awake.

Right Cardiac Catheterization During Wakefulness and Sleep

Catheterization was performed at rest in a supine position while breathing room air. A Swan-Ganz catheter 7.5 Fr in diameter (Guidewire TD Catheter; Baxter, Irvine, CA) was introduced via the right internal jugular vein under the control of distally obtained pressure curves continuously recorded on a pressure monitor (Recti-Horiz-12K pen recorder; NEC San-ei, Tokyo, Japan), and placed in the stem of the pulmonary artery. The pressure was recorded with a transducer (Viggo-Spectramed; Singapore) placed one-third of the anteroposterior diameter below the sternal angle. After the placement of the catheter, patients rested for 15 min with room air breathing, and then systolic and diastolic pulmonary artery pressure was measured. Mean Ppa () was calculated from the following equation: = (systolic Ppa  diastolic Ppa)/3 + diastolic Ppa. of more than 20 mm Hg was defined as PH. Patients with daytime PH were classified as the PH(+) group, and the others as the PH() group. Subsequently, pulmonary arterial occlusive pressure was recorded, and at least three intermittent measurements of cardiac output were performed by thermodilution method (Oxmetric3; Abbott, North Chicago, IL) in order to assess the mean daytime cardiac output. Cardiac index was then calculated to normalize cardiac output by anthropometric parameters. Pulmonary vascular resistance was calculated from the following equation: pulmonary vascular resistance = (Ppa  pulmonary arterial occlusive pressure)/cardiac output.

After right cardiac function was evaluated during wakefulness, patients underwent polysomnography while continuously recording Ppa via the indwelling catheter under natural sleeping conditions. The polysomnographic recordings were analyzed by conventional method and Ppa was evaluated in the respective sleep stages, namely NREM sleep and REM sleep, which were determined by electroencephalogram (EEG), EOG, and surface electromyogram (EMG) of the genioglossal muscle. Ppa was measured for every apneic period in order to obtain basal values (-basal), recorded at the beginning of, or early in the apnea/hypopnea phase after the hyperventilatory phase, and peak values were obtained at a point corresponding to the end of the apneic period (-peak). Each value was assessed when the intrathoracic negative pressure generated by the inspiratory effort against upper airway collapse was released, i.e., when inspiratory effort was absent. The Sp_O2 level was also measured in relation to corresponding , that is, the Sp_O2 basal level corresponding to the hyperventilatory phase (Sp_O2-basal) and the Sp_O2 lowest level corresponding to the termination of apnea (Sp_O2-lowest). values (basal and peak) and Sp_O2 values (basal and lowest) were respectively determined. In four patients of the PH(+) group, while simultaneously monitoring EEG, EOG, and surface EMG of the genioglossal muscle during polysomnography with catheterization, we measured cardiac output and pulmonary arterial occlusive pressure when both NREM and REM sleep stages were confirmed, and thereafter pulmonary vascular resistance was calculated during each sleep stage. Also in those four patients, catecholamine concentrations in blood from the stem of the pulmonary artery during both sleep stages were measured.

Statistical Analysis

Data were expressed as mean ± SD. Group comparisons between PH(+) and PH() were performed by unpaired t-test, and those between REM and NREM were done by paired t-test. When multiple comparisons were needed, one-way analysis of variance (ANOVA) followed by Fisher's PLSD test was used. In every patient, response to hypoxia was analyzed by linear regression curves during NREM and REM sleep, respectively (see RESULTS).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Right Cardiac Function and Respiratory Function During Wakefulness

Results obtained from the catheterization study during wakefulness are shown in Table 2. All patients in the PH(+) and PH() groups had normal pulmonary arterial occlusive pressure values of less than 12 mm Hg. Cardiac output, cardiac index, and pulmonary vascular resistance were significantly higher in the PH(+) than in the PH() group.

Body mass index, awake pH and Pa_CO2 levels were significantly higher in the PH(+) than in the PH() group. One second forced expiratory volume in percentage of forced expiratory volume (FEV1.0%) and functional residual capacity in percentage of the predicted value (%FRC) revealed no significant differences between the two groups. On the other hand, vital capacity in percentage of predicted value (%VC) was significantly lower in the PH(+) than in the PH() group (Table 1).

Ppa Alteration During Wakefulness and Sleep

One representative polysomnographic recording in a patient with daytime PH is shown in Figure 1. Ppa elevation accompanied by Sp_O2 fall was synchronously observed in every apneic period. During each apnea, the decrease in Sp_O2 level induced the corresponding Ppa elevation with a delay that seemed to be strongly related to circulatory plus instrumental factors, and therefore we measured -basal and -peak corresponding to Sp_O2-basal and Sp_O2-lowest, respectively. Awake and asleep and Sp_O2 alterations are shown in Figure 2. In , Ppa-peak during REM showed a significantly higher value than the other values (p < 0.05). In Sp_O2, Sp_O2-lowest during REM was significantly lower than the values of Sp_O2 during wakefulness, Sp_O2-lowest during NREM, and Sp_O2-basal during REM (p < 0.05).

View larger version (36K): [in this window] [in a new window]   Figure 1.   A representative polysomnographic tracing in a patient with OSAS with daytime PH. During REM sleep, Ppa elevation is more exaggerated compared with NREM sleep, with SpO2 decrease in every apneic period. Furthermore, the rapid rise of Ppa in association with the appearance of phasic REM and its recovery to the initial level immediately after the disappearance of REM are evident.

View larger version (28K): [in this window] [in a new window]   Figure 2.   Ppa and SpO2 alterations during wakefulness and each sleep stage in all patients with OSAS. In Ppa, Ppa-peak values during REM show significantly higher values than the other Ppa values. In SpO2, SpO2-lowest values during REM are also significantly lower than SpO2 during wakefulness, SpO2-lowest during NREM, and SpO2-basal during REM. *p < 0.05 with one-way ANOVA followed by Fisher's PLSD test.

Ppa Response Curve to Hypoxia During Sleep

In each patient, the augmentation of in association with the decrease in Sp_O2 during each apnea was analyzed by a linear regression method. Figure 3 shows a typical Ppa response to hypoxia, revealing linear increases with Sp_O2 depression in both stages. Mean response curves to hypoxia during both NREM and REM sleep for the PH() and PH(+) groups are shown in Figure 4. In both PH() and PH(+) groups, the slopes in response to hypoxia, i.e., the slopes of the regression line, were almost the same between NREM and REM sleep: 0.36 ± 0.09 and 0.40 ± 0.15 in PH(), and 0.74 ± 0.27 and 0.71 ± 0.30 mm Hg/% in PH(+), respectively. For evaluating the magnitudes of of the mean response line, the values at the same Sp_O2 level were calculated. The value at the same Sp_O2 level was significantly higher in REM than in NREM only in the PH(+) group but not in the PH() group; in the PH(+) group, at Sp_O2 75% was 41.3 ± 8.2 mm Hg in NREM versus 47.8 ± 12.0 mm Hg in REM (p < 0.01); in the PH() group, at Sp_O2 85% was 14.9 ± 5.8 in NREM versus 16.4 ± 6.9 in REM (p = 0.074) (Figure 5). Thus, in the PH(+) group, the response curve was shifted upwards in REM compared with NREM.

View larger version (11K): [in this window] [in a new window]   Figure 3.   Ppa response to hypoxia during NREM and REM sleep in a patient with daytime PH. In both sleep stages, linear increases in Ppa are elicited by decreases in SpO2. Subject: T.N.

View larger version (15K): [in this window] [in a new window]   Figure 4.   Relationship between SpO2 and Ppa during sleep in all patients with OSAS with and without daytime PH. Mean response curves to hypoxia during NREM and REM sleep are shown in the two groups of patients with and without daytime PH, PH(+), and PH(), respectively. The slopes of the regression lines are almost the same between NREM and REM in both PH() and PH(+) patients. In the PH(+) group, Ppa at SpO2 75% is significantly higher in REM than NREM (*p < 0.01 versus NREM).

View larger version (18K): [in this window] [in a new window]   Figure 5.   Ppa evaluated at the same level of SpO2 during NREM and REM sleep in patients with and without daytime PH. Because both baseline and lowest SpO2 during apnea were lower in patients with daytime PH than without PH, comparison of Ppa was made at different SpO2 levels, i.e., SpO2 75% in patients with PH and 85% without PH. Only in patients with daytime PH, Ppa at SpO2 75% was significantly higher in REM than in NREM sleep (*p < 0.01 versus NREM).

Blood Catecholamine Concentration During Sleep

In four patients of the PH(+) group, we measured blood catecholamine concentration during NREM and REM sleep. Norepinephrine and epinephrine concentrations were widely divergent in the respective patients; norepinephrine in NREM and REM, 268.8 ± 147.0 versus 297.5 ± 100.1 pg/ml; epinephrine in NREM and REM, 14.3 ± 7.8 versus 27.5 ± 35.0 pg/ml, respectively.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The main findings obtained in the present study were that the Ppa response line to oxygen desaturation was shifted upwards during REM, contrary to that seen in NREM sleep, and that this Ppa elevation was confirmed to exist even before falling asleep. This result may signify that the vascular tone of the pulmonary artery could have been increased more in REM than in NREM irrespective of the change in the degree of hypoxia. Transient Ppa elevation during apnea has been reported by several investigators (11). Hypoxic pulmonary vasoconstriction is considered to be responsible for the periodic Ppa elevation that is observed concomitantly with apnea.

Several factors responsible for these Ppa changes may be considered. One candidate is the accumulated CO₂, as it could have modified hypoxic pulmonary vasoconstriction because more CO₂ retention during prolonged apnea in REM would enhance this condition. Second, some humoral or neural control mechanisms may be involved. These mechanisms may further affect pulmonary vascular tone, thus additionally affecting hypoxic pulmonary vasoconstriction. The effect of CO₂-induced blood hydrogen ion concentration on hypoxic pulmonary vasoconstriction has already been reported. Lloyd demonstrated that decreased blood pH could increase pressor response to hypoxia in pulmonary artery in adult dogs (22). In our study, the Sp_O2 level was depressed more during REM than NREM sleep. Consequently, CO₂ retention and decreased blood pH could occur (23), followed by pulmonary vascular resistance and Ppa elevation. The effect on hypoxic pulmonary vasoconstriction by gaseous CO₂ itself may certainly be involved in the Ppa and pulmonary vascular resistance responses to hypoxia. However, if it is supposed that CO₂ is the only factor to influence hypoxic pulmonary vasoconstriction, the slope of the Ppa response to hypoxia should be more augmented in REM than in NREM sleep. Actually, Laks and coworkers demonstrated that Ppa elevation was more augmented under hypoxic condition with hypercapnia than with eucapnia in awake OSAS patients (24). In the present study, the slopes of hypoxic Ppa response were not different between REM and NREM sleep, but its response curve shifted upwards more in REM than in NREM in patients with daytime PH. Taken these findings together, it is suggested that some factors other than CO₂ retention could have increased the pulmonary vessel tone during REM sleep.

Thus, other factors were also considered as possibly inducing an enhanced hypoxic pulmonary vasoconstriction. Catecholamines, histamine, and serotonin can be enumerated as humoral factors that potentially modify hypoxic pulmonary vasoconstriction (25). It has been reported that blood catecholamine concentrations increased during REM compared with NREM sleep (28), but in our study, blood catecholamine concentrations did not show any difference between the two sleep stages. The reason why catecholamine concentrations changed little might be that sampling time was not adequate to obtain fully developed REM-associated effects. Nevertheless, it can be argued that this REM-specific Ppa elevation was not directly related to blood catecholamine because the augmented Ppa response to hypoxia was already observed at the time of blood collection. Figure 1 indicates that the rapid rise of Ppa is observed right after the appearance of phasic REM, its rise is somewhat attenuated during tonic REM, and also that it returned to the initial level immediately after the REM sleep. These observations suggest that Ppa responds promptly to REM sleep. This rapid Ppa response strongly suggests the contribution of neural mechanisms rather than catecholamines, because the pressor effect of the latter takes more than a few minutes (29).

We hypothesized, therefore, that the REM-specific Ppa elevation in patients with OSAS was induced by direct neural mechanisms. It was reported that muscular sympathetic activity increased during REM sleep in healthy humans (30). Many previous studies claimed that the sympathetic nerve system played few roles in impinging on pulmonary vessels and in hypoxic pulmonary vasoconstriction (31), whereas Shirai and coworkers reported increased efferent pulmonary sympathetic activities during systemic hypoxia, and also that the Ppa level was elevated in association with increased sympathetic activity in cats (34). In patients with OSAS, systemic blood pressure was reported to increase during REM sleep (19, 20), but there have been few studies on the Ppa alteration simultaneously analyzed with sleep stages. Marrone and coworkers reported augmented Ppa levels during REM, but they concluded that diminished oxygen saturation was responsible for this (35). We speculate that direct neural control can modulate hypoxic pulmonary vasoconstriction and elevate Ppa during REM sleep independently of the effect by hypoxia, hypercapnia, or humoral factors in patients with OSAS.

In patients with OSAS, it has not been resolved whether nocturnal desaturation caused by sleep apnea could result in daytime PH, whereas in an animal model, it was established that repetitive hypoxia can cause chronic PH in conjunction with vascular remodeling (36, 37). In this study, the patients with daytime PH revealed higher cardiac output and higher pulmonary vascular resistance than patients without PH (Table 2). Recent studies reported increased pulmonary vascular resistance in patients with OSAS with resting PH (4, 7), which is in good agreement with our data. Increased cardiac output and pulmonary vascular resistance could partially account for daytime blood gas abnormality, but taking the insignificant difference in daytime Pa_O2 between the PH(+) and PH() groups into consideration, those increases could be affected by some other factors. It has been suggested that daytime PH in OSAS can be elicited by vascular remodeling together with both nocturnal and daytime oxygen desaturation (38). Such pulmonary vascular remodeling in patients with daytime PH might induce greater sympathoadrenal tone or greater sensitivity to sympathoadrenal stimulation.

In conclusion, the present study demonstrated that the response curve of Ppa to oxygen desaturation was significantly shifted upwards in REM compared with NREM sleep in patients with daytime PH. As this is not the case in patients without daytime PH, it is suggested that these state-specific elevations in Ppa occur in patients whose vascular tone has already increased before falling asleep, and/or that vascular remodeling, at least in part, may have already existed as a facilitating component.