INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

ARTERIAL BLOOD PRESSURE (BP), which reflects sympathetic outflow (24, 32, 33), is in part dependent on peripheral arterial chemoreceptor (22, 32, 33) and arterial baroreceptor (8, 31) afferent input. In awake normal humans the peripheral chemoreceptor reflex appears to be sympathoexcitatory (32, 33) and the arterial baroreceptor reflex sympathoinhibitory (34). A barostimulation-induced inhibition of the peripheral chemoreceptor and a chemostimulation-induced inhibition of the baroreceptor have also been postulated (11, 34).

Sleep has been associated with increased arterial baroreceptor sensitivity (31) and decreased peripheral chemoreceptor sensitivity (2, 11). Each of these effects could contribute to normally observed decreases in sympathetic outflow (30, 32) and arterial BP during sleep (6, 30). However, normal responses of transient stimulation of these receptors during sleep have not been well studied in humans. Furthermore, although the increased prevalence of diurnal and nocturnal hypertension (17, 29a) as well as sleep-disordered breathing in African-Americans at young ages (29, 29a) would suggest that race may be an important determinant of chemo- and baroreceptor function, the effects of race on the interactions of chemo- and baroreceptor activity have not been studied.

We thus designed this study to investigate normal ventilatory and chronotropic baroreceptor responses to transient chemo- and barostimulation in these subjects. We hypothesized that sleep may differentially modulate physiological responses to arterial chemo- and barostimulation in young African-Americans and Caucasians in a manner that might contribute to the increased diathesis for development of chronic systemic hypertension in African-Americans.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Subjects. Twenty-six normal, nonsnoring young subjects (6 African-American men, 7 African-American women, 7 Caucasian men, 6 Caucasian women), 20-36 yr of age, were recruited from the university staff and community. All were naive regarding the study hypothesis, and none had prior experience with similar experiments. All subjects were screened by a physician investigator before inclusion in the study and were without cardiorespiratory disease or symptoms suggestive of obstructive sleep apnea or other primary sleep disorder or circadian rhythm abnormality. Subjects were free of any medications at the time of enrollment and study and were selected only if they were within normal limits of body mass index (23). Absence of hypertension was determined by lack of any history of hypertension and normal BP readings obtained by the physician investigators on at least two successive occasions. The women were in the middle of their menstrual cycle, except one Caucasian, who was in the 1st wk. Three of the women were taking oral contraceptives. Informed written consent was obtained from each subject according to the guidelines of the World Medical Association Declaration of Helsinki and with approval of the Institutional Review Board of the University of Illinois at Chicago.

Measurements. Subjects were studied supine, wearing a tight-fitting silicone face mask (Bird), awake and asleep. Subjects who could not consolidate sleep in the supine position were allowed to shift to a semisupine position with the head and mask apparatus adjusted to the side in the later portion of the night. The face mask was fitted to a two-way valve to separate inspiratory and expiratory ports. The inspiratory port was connected to a pneumotachograph (Fleisch no. 3) to measure inspiratory airflow and coupled to a respiratory integrator (Validyne) to obtain tidal volume (VT). The expiratory port was connected to an infrared PCO₂ analyzer (model 5200, Ohmeda) to measure end-tidal PCO₂. Arterial oxyhemoglobin saturation (Sa_O2) was continuously monitored by finger pulse oximeter (model N-200, Nellcor, Hayward, CA). Digital arterial beat-to-beat systolic and diastolic BP and heart rate were recorded noninvasively from the third finger of the left hand with an infrared plethysmographic volume clamp (Finapres, Ohmeda, Englewood, CO) as described elsewhere (9, 26, 27). The hand with the Finapres was restrained at midchest level. A precordial electrocardiogram was also monitored. A 20-gauge intravenous catheter was placed in the hand contralateral to the Finapres using sterile technique and wrapped on a hand board; a solution of 0.9% saline was run at a keep-open rate. All data were collected simultaneously on a polygraph recorder (model 78D, Grass) and a Bio-logic Sleepscan system (Bio-logic Systems, Mundelein, IL). Digitized data were archived in raw form on a magnetooptical disk connected to the system computer. Each record was scored by hand by the physician investigators. Sleep was staged with central and occipital referential electroencephalogram (EEG), right and left electrooculogram, and submental electromyogram. Low-pass filters were set at 35 Hz and high-pass filters at 0.3 Hz.

Protocol. After signing informed consent, subjects were prepared for standard polysomnography, and an intravenous line was introduced in a dorsal vein of the hand. The face mask was applied and attached to the respiratory apparatus suspended above the subject. Air leak was checked with a CO₂ probe and was checked in the same way throughout the night. Subjects lay quietly while awake. Data collection during wakefulness was begun between 9 and 11 PM. In random order, stimuli to induce transient hypoxemia, hypertension, and concomitant hypertension and hypoxemia were administered. Hypoxemia was induced with three to eight breaths of pure N₂ titrated to cause Sa_O2 nadir between 75 and 89%. The gas was administered using a two-way valve fitted to a manifold to allow the inspirate to be selected from the N₂-containing reservoir bag or from room air (13). The hypertensive stimulus was 25-37.5 µg of phenylephrine (PE) in 0.9% saline as a 0.5-ml bolus flushed with 2 ml of saline (1). Concomitant hypertensive and hypoxemic stimulation (N₂ + PE) was achieved with N₂ given immediately after the PE bolus and flush. A minimum of 3 min was required between stimuli. The goal was to obtain three stimulations each of N₂, PE, and N₂ + PE.

Data analysis. For each stimulus (N₂, PE, and N₂ + PE), the 20 consecutive cardiac cycles preceding the stimulus were scored as the baseline segment, and the 40 consecutive cardiac cycles beginning at beat 21 after the administration of the stimulus were scored as the response segment. The decision to score and analyze cardiac cycles 20-60 after the stimulus as the response segment was made after inspection identified this region as the region invariably associated with the onset and resolution of measurable heart rate, BP, and ventilatory responses. BP, minute ventilation (I), and Sa_O2 of each of these pre- and poststimulus cardiac beats were scored; however, only data measured during expiration were analyzed (12, 31). To correct for the lag of the Sa_O2 display in assigning Sa_O2 values for hypoxemic stimuli (N₂ and N₂ + PE), the Sa_O2 curve was manually advanced such that the nadir of the curve was located directly below the area of maximal I. All poststimulus data were normalized as change from the averaged 20 cardiac cycles preceding the stimulus. Baroresponse was calculated as the slope of arterial BP plotted against the pulse (R-R) interval of the subsequent cardiac cycle (pulse interval/BP, in ms/mmHg, where  is change) (31). Systolic and mean BP were used for this analysis. There were no significant differences between results obtained with systolic BP and mean BP; results using systolic BP are reported. The slope of the ventilatory response was calculated as I/Sa_O2. I was calculated according to the following formula: I = VT/TI × TI/TT, where TI is inspiratory time and TT is total respiratory cycle duration (5).

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Baroresponse slopes (pulse interval/systolic blood pressure, where  is change) for African-American and Caucasian subjects during wakefulness (Wake) and non-rapid-eye-movement (NREM) and rapid-eye-movement (REM) sleep during hypoxemia (N2), hypertension (phenylephrine, PE), and concomitant hypoxemia and hypertension (N2 + PE). Values are means ± SE for 40 successive cardiac cycles after stimulus. Baroresponse (PE) is increased from wakefulness to NREM sleep for all subjects (P = 0.007) and attenuated with concomitant hypoxemia (N2 + PE, P < 0.002). Baroresponse is significantly greater in Caucasians than in African-Americans (P = 0.02). Baroresponse attenuation during N2 + PE is significantly greater in African-Americans than in Caucasians in NREM sleep (P = 0.03). No statistically significant gender differences are present. There were insufficient data for Caucasian women during N2 and N2 + PE to be included in ANOVA.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Anthropometric data for the subjects are displayed in Table 1. A total of 625 stimulations were analyzed among the 26 subjects (307 in African-Americans, 318 in Caucasians). The overall degree of chemostimulation (as measured by degree of Sa_O2 decrease) and barostimulation (as measured by systolic BP increase) was similar in the African-American and Caucasian subjects, the only statistically significant difference in barostimulation being a greater increase in BP during N₂ + PE in wakefulness and REM sleep in the African-Americans (Table 2). The only statistically significant difference in degree of induced hypoxemia was a greater decrease in Sa_O2 in Caucasians during N₂ in NREM sleep (Table 3). All subjects became transiently hypocapnic during hypoxemic stimulation; the range of CO₂ decrease was 0.4-1.6%. ANOVA did not find a statistically significant main effect of individual subject on responses.

Heart rate and BP responses. Baroresponses for all states and stimulus types are shown in Fig. 1 for all subjects. During wakefulness and sleep, with the hypoxemic stimulus (N₂) alone, the baroresponse slope was negative; that is, heart period decreased as BP increased. With the hypertensive stimulus (PE) alone, baroresponse slope was positive; that is, heart period increased as BP increased. Baroresponse was significantly increased in NREM sleep compared with wakefulness for all subjects (P = 0.007). A similar trend to increased baroresponse in REM sleep did not reach statistical significance (P = 0.19). With concomitant hypoxemia and hypertension (N₂ + PE), baroresponse remained positive but was attenuated compared with PE alone (P  0.002). Polysomnographic tracings showing baroresponse during PE and N₂ + PE in a representative Caucasian man are displayed in Figs. 2 and 3, respectively.

View larger version (30K): [in this window] [in a new window]   Fig. 2.   Polysomnographic tracing during NREM sleep in a representative Caucasian man after a 25-µg bolus of PE. Cardiac baroresponse is apparent in blood pressure (BP) tracing. Inspiration is up and expiration down for flow and volume tracings. C3-A2 and O2-A1, central and occipital referential electroencephalogram, respectively; LOC and ROC, left and right electrooculogram, respectively; SaO2, arterial O2 saturation; EMGChin, submental EMG. Segment begins 9 cardiac cycles after PE bolus.

View larger version (32K): [in this window] [in a new window]   Fig. 3.   Polysomnographic tracing during NREM sleep in subject in Fig. 2 after simultaneous 25-µg bolus of PE and inhalation of 100% N2. SaO2 nadir is 85%. Cardiac baroresponse is attenuated compared with response to PE alone. Segment begins 8 cardiac cycles after PE bolus. See Fig. 2 legend for definition of abbreviations.

Ventilatory responses. For all subjects, mean ventilatory response slopes (I/Sa_O2) to the hypoxemic challenge (N₂) were not significantly different between wakefulness and sleep. African-Americans demonstrated a greater (more negative) slope than Caucasians during wakefulness and sleep. This difference was statistically significant only for NREM sleep (P = 0.01). These data are displayed in Fig. 4.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Ventilatory response slopes (I/SaO2, where I is minute ventilation) for African-American and Caucasian subjects during wakefulness and NREM and REM sleep during transient hypoxemia (N2) and concomitant hypoxemia and hypertension (N2 + PE). Values are means ± SE for 40 successive cardiac cycles after stimulus. I/SaO2 during N2 is not significantly different between wakefulness and sleep. African-Americans demonstrate a greater (more negative) slope overall than Caucasians during wakefulness and sleep; ANOVA was significantly different only in NREM sleep (P = 0.01). Transient hypertension induced along with hypoxemia (N2 + PE) shows a statistically significant attenuation of hypoxemic ventilatory response in African-American and Caucasian men (P = 0.04) in NREM and REM sleep but not in women. There were insufficient data for Caucasian women during N2 to be included in ANOVA.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

These data in normal young African-Americans and Caucasians indicate that sleep is associated with a greater enhancement of baroresponse to transient BP elevation in normal young Caucasians than in African-Americans. Chemoresponsiveness in NREM sleep appeared to be greater in the African-Americans than in the Caucasians.

It must be emphasized that the number of subjects in this protocol is relatively small, and thus these physiological findings, specifically those indicating a difference between African-American and Caucasian responses, may not be representative of young African-American and Caucasian populations overall. We attempted to control for variability in subject selection by limiting these studies to young, healthy, nonobese, nonapneic, normotensive subjects. Furthermore, we collected and analyzed multiple runs of interventions from all subjects, as well as a large number of data points for each intervention, spanning 20 cardiac cycles of baseline and 40 cardiac cycles of response. Each significant main effect of ANOVA, including those that indicated a difference in responses between the groups of African-American and Caucasian subjects, was also subjected to nonparametric analysis, objectively confirming that no individual subject deviated greatly from the group behavior within each race. The possibilities of type 1 error also pertain to the use of multiple tests and variables in a relatively small group of subjects. The positive findings we report were hypothesis driven rather than outcomes of testing for all possible interactions. Furthermore, the multiple individual contrasts were controlled by protected least significant difference at levels that would be unlikely to allow false positives. At the same time, type 2 error in this study cannot be ruled out, and statistical similarities reported between African-Americans and Caucasians and between men and women in this protocol might in fact have become significantly different with more subjects (7). Also there were many fewer REM than NREM data in this protocol, and REM data were not obtained in all subjects. Thus, for baroresponse, N₂ REM data included three African-American men, three Caucasian men, three African-American women, and no Caucasian women; N₂ + PE data included four African-American men, six Caucasian men, five African-American women, and one Caucasian woman. For ventilatory N₂ response, REM data included three African-American men, three Caucasian men, three African-American women, and no Caucasian women; N₂ + PE data included four African-American men, six African-American men, five African-American women, and one Caucasian woman.

African-Americans appear to have an increased prevalence of diurnal and nocturnal hypertension compared with Caucasians at virtually all ages (29a) and increased sleep-disordered breathing compared with Caucasians at a relatively early age (29). African-American women as a group appear to have the highest prevalence of systemic hypertension from middle to old age (29a). In this study the African-American subjects showed decreased baroresponses with and without concomitant hypoxemia compared with similar Caucasian subjects, and this was most specific to NREM sleep. Decreased baroresponses have similarly been found in patients with obstructive sleep apnea (4), older subjects (16), and young subjects with borderline hypertension or a strong family history of hypertension (36, 37). In the present study, African-American subjects also demonstrated relatively increased ventilatory responses to hypoxemia. Increased peripheral chemosensitivity (increased ventilatory drive) has been found in young men who demonstrate mild essential hypertension (35). The net effect of decreased baroresponse and increased chemoresponse may result in increased exposure during sleep to sympathetic stimulation in the young African-American subjects, particularly if the baroreflex normally decreases the reflex vascular effects of chemostimulation (22).

The accuracy of finger pulse oximetry in dark-pigmented vs. lighter skinned subjects has been questioned by studies in which greater variability between measured finger pulse oximetry and direct measurement of Sa_O2 was found in "black" than in "white" patients undergoing mechanical ventilation (20). Although there are no similar data for healthy young populations, we cannot exclude the possibility that there was a greater degree of hypoxemia among African-American subjects in this protocol. This would not be likely to influence comparison of the slopes of the hypoxemic ventilatory response but could have resulted in occultly greater depression of baroresponse with combined stimuli in the African-Americans.

Although we did not observe consistent changes in airflow or end-tidal CO₂ in these subjects suggestive of increasing upper airway resistance during sleep, we did note that some subjects (African-Americans and Caucasians) snored intermittently during REM sleep. Our methodology cannot rule out the possibility of a systematic difference in upper airway resistance between groups, as might occur, for example, during PE challenges (14), which thus could have affected the measurement of ventilatory and baroresponses.

Although there was an expected decrease in baseline ventilation (before stimulations) from wakefulness to sleep in this study, we found, somewhat unexpectedly, that the hypoxemic ventilatory response was not attenuated with sleep. Sleep has been shown to decrease ventilatory response to progressive isocapnic hypoxemia in humans (2, 10). However, hypocapnic hypoxemia, similar to the present study, has not been found to produce consistent ventilatory changes in NREM sleep compared with wakefulness, possibly because hypocapnia may cause a selectively decreased ventilatory response during wakefulness (15). It is also likely that a transient hypoxemic stimulus is modulated differently from progressive hypoxemia during sleep.

Using a bolus of PE, Abdel-Rahman and colleagues (1) found decreased baroreceptor responses in wakefulness in young normal women compared with men (race not specified). We, however, found that responses during wakefulness were similar in men and women of both races. The women in the present study displayed baroresponses to 25 µg of PE quantitatively similar to those in the study of Abdel-Rahman et al.; the men in the present study showed less baroresponse than did the men in the earlier study (1). We cannot specifically explain this difference, but we would note that the state of wakefulness itself may involve different levels of apprehension and sympathetic stimulation during such experiments; neither study clearly differentiates such change in arousal within the state of wakefulness.

Although the duration of baseline and response interval varied slightly among subjects, stimuli, and states, we do not believe that such variation affected results, since all analyzed stimulations utilized a stable and reasonably lengthy and representative baseline, whereas each response interval captured the beginning and resolution of changes in ventilation, heart rate, and BP. Furthermore, there was no difference in the number of response beats among subjects in the final data analysis, since these data were averaged to produce one observation for each cell (sleep and stimulus type) for each subject in the ANOVA.

We did not employ "sham" stimuli; therefore, we cannot be certain whether the methods used (gas inhalation, bolus of 2 ml of liquid) themselves contributed to responses. In particular, Caucasian women tended to arouse more frequently than the other groups during N₂ inhalation, and this protocol cannot differentiate whether this tendency to arousal was due to the methodology of gas inhalation employed rather than the hypoxemia itself. We believe it is unlikely that the methodology of PE infusion was responsible for the results we found. Subjects did not show a particular tendency to arouse with PE alone, and, using a similar bolus of 2 ml of saline in normal subjects, Abdel-Rahman et al. (1) did not demonstrate an effect on heart rate or BP.

There was considerable variation in the degree of hypoxemic challenge we employed, which could have affected between-group results. We aimed from the outset for an Sa_O2 range of 75-89% for hypoxemic challenges, comparable to the range of Sa_O2 in previous studies of hypoxemic challenge during sleep (10). The number of N₂ breaths was determined for each subject as the number that reliably brought the Sa_O2 down to that range and was adjusted throughout each subject's study as necessary. This number tended to vary with state as well as with subject. We found, for example, that wakefulness tended to be more resistant than REM sleep to Sa_O2 decrease. However, there were no systematic trends in that regard, and the range of breaths of N₂, as well as the range of Sa_O2 decrease, was similar among subjects. Decreases in Sa_O2 from baseline were similar in African-Americans and Caucasians for all stimuli and states (Table 3).

We used a noninvasive method of BP measurement rather than an intra-arterial catheter to maximize subject safety and minimize discomfort. Arterial BP measured in this manner has correlated well with pressure measured by arterial catheter in awake and sleep states during similarly changing conditions of BP (9, 26, 27). Furthermore, the baroresponse slopes we obtained with hypertension are similar to those obtained in normal subjects using intra-arterial catheters under similar conditions (33).

We excluded data that coincided with or followed EEG arousal, inasmuch as even minimal abrupt EEG changes, below the minimal threshold for scoring of arousal by standard guidelines, may be associated with evidence of autonomic arousal, including increased blood pressure (9). Furthermore, we excluded any data associated with an abrupt BP rise with or just after the stimulation, whether or not an EEG arousal was present. However, we cannot exclude arousal-related autonomic discharge not discernible on EEG, since arousal stimuli unassociated with clear EEG change may cause autonomic discharge (9, 24).

The race-related differences in chemo- and baroresponses in this group of normal young subjects during sleep may have implications concerning diathesis for the development of systemic hypertension, particularly since African-Americans have also been found to have elevated BP during sleep compared with Caucasians as early as adolescence (17). Further studies of race-related chemo- and baroresponses in normal young humans might help lead to improved strategies of screening and therapy of individuals with a diathesis for diurnal hypertension and sleep-disordered breathing.