Platelet activation, epinephrine, and blood pressure in obstructive sleep apnea syndrome

Atherothrombotic brain infarcts often occur during sleep in the early morning hours.¹ Snoring has been suggested to be a risk factor for brain infarction.² Heavy snoring is associated with obstructive sleep apnea, which may have several harmful effects on the cardiovascular system.^3-5 In healthy men, alteration in platelet function is related to an increased risk of cardiovascular death,⁶ and there is evidence that platelet activation is pathophysiologically involved in ischemic heart disease.^7,8 Epinephrine can activate platelets⁹ via binding to alpha₂ receptors on the platelet surface,¹⁰ and was shown to be increased in patients with obstructive sleep apnea syndrome(OSAS).¹¹ Thus, epinephrine-induced platelet activation may be involved in the increase of cerebrovascular and cardiovascular morbidity and mortality in patients with OSAS.

The purpose of our study was to investigate the relation among OSAS, morning and evening plasma epinephrine levels, blood pressure, and platelet activation as measured by P-selectin expression.

Methods. Study population. Seventy-five patients referred for polysomnograms (PSGs) to the New England Medical Center's sleep laboratory because of a suspected sleep disorder were recruited consecutively and prospectively over a period of 4 months. Patients were enrolled in the study if they were at least 18 years old and had given their informed written consent. Patients were referred by neurologists (approximately 50%), pulmonologists (approximately 8%), psychiatrists (approximately 6%), general internists (approximately 10%), otorhinolaryngologists (approximately 6%), and obstetricians and other specialists (approximately 20%). About one-third of the patients in our laboratory were diagnosed as having OSAS. Most of the remaining patients had periodic limb movement disorder or restless legs syndrome. Patients were rarely found to have parasomnias. Three of our staff were also included as true non-OSAS control subjects. Among the 75 recruited patients, four patients were reexamined for nasal continuous positive airway pressure (CPAP) titration (only their non-CPAP nights were used), one patient left the sleep laboratory before the end of the sleep study, and one patient's morning blood sample was clotted. These last two patients were excluded, leaving 76 participants, with a mean age of 42.7 years (range, 24 to 74 years), in the study population. Fifty patients did not meet the criteria for OSAS (defined as fewer than five obstructive apneas or hypopneas per hour of total sleep time [TST]), 19 patients had an apnea hypopnea index(AHI) of five or more but less than 50 per hour of TST (mild-to-moderate OSAS), and seven patients had an AHI of more than 50 per hour of TST (severe OSAS). All participants were surveyed for sleep characteristics, health history, and sociodemographics by questionnaires administered at the time of their PSG.

Study protocol. The protocol was approved by the Human Subjects Committee at the New England Medical Center. Printed information requesting subjects to refrain from taking aspirin, ibuprofen, or naproxen for at least 7 days and caffeine, nicotine, or alcohol for 24 hours before the scheduled sleep study was mailed to prospective subjects in advance of their appointment in the sleep laboratory. Subjects arrived in the laboratory at approximately 8 PM. Informed written consent was obtained from each patient. A careful history of sleep characteristics as well as sociodemographics was obtained by oral questions and a written questionnaire. Blood samples were taken and blood pressure measured around 9 PM after a minimum of 15 minutes of supine rest, and at 6 AM before the patient assumed an upright position. We performed a pretest with seven normal subjects who had venipunctures in a vein that had been accessed 4 hours earlier, and in a noninjured vein to evaluate whether morning venipuncture of the same vein punctured the evening before resulted in increased platelet activation.

Sleep studies. All participants were included in the overnight diagnostic PSG investigations in the sleep laboratory. A semiautomated Sleep Scoring System (Telefactor Corporation, West Conshohocken, PA) was used for registration and scoring of the PSG. The investigation was initiated around 10 PM and terminated at 6 AM. The PSG included six channels of EEG, electro-oculogram (EOG), and mental and submental electromyogram (EMG). The surface EMGs of both anterior tibial muscles were recorded as described by Coleman.¹² The oronasal airflow was also recorded via thermistors mounted over the nose and mouth, and the thoracic and abdominal respiratory movements were recorded by impedance plethysmography. Arterial oxygen saturation was measured continuously via a noninvasive infrared finger probe. The EKG was recorded continuously between the forearms. Sleep staging followed the recommendations of Rechtschaffen and Kales.¹³ Apnea was defined as a cessation of airflow lasting at least 10 seconds, accompanied by a drop of saturated oxygen (SaO₂) by more than 2% below the immediately preceding baseline. For hypopnea scoring we applied a combination of criteria widely used in American Sleep Disorders Association(ASDA)-accredited sleep laboratories in the United States according to Moser et al.¹⁴ Hypopnea was defined as a decrease of the airflow by 50% or more below the waking baseline, accompanied by a drop of SaO₂ by more than 2% below the immediately preceding baseline. Arousal scoring followed the guidelines suggested by the Atlas Task Force of the ASDA.¹⁵

Blood pressure recordings. Because we added blood pressure measurements after the study had already begun, we did not have blood pressure values for the first 11 patients. Blood pressure was measured with a blood pressure cuff and stethoscope after at least 15 minutes of supine rest the evening before the sleep study was started (9 PM) and in the morning before the subject got up (6 AM).

Blood sampling. Blood samples were obtained after the subject had rested for 15 minutes at 9 PM prior to the onset of the PSG, and at 6 AM before the patient rose from a horizontal position. Blood was obtained by venipuncture with a 21-gauge butterfly needle as follows: The first 3 mL of blood was discarded, the next 10 mL was put on ice immediately after being drawn into ethylenediaminetetraacetic acid (EDTA) for epinephrine analysis, and 3 mL was drawn for cholesterol analysis. Then, 0.9 mL of blood was drawn into a 1-mL tuberculin syringe and put immediately into an Eppendorf tube containing 100 µL of 3.8% sodium citrate solution for flow cytometric analysis. Three microliters of whole blood in citrate solution was pipetted quickly into Eppendorf tubes, each containing 47 µL of a mixture of N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES) Tyrode buffer, glycoprotein Ib (GPIb), and P-selectin antibodies, plus 0, 0.5, 1, 2, or 5 µM adenosine 5′-diphosphate (ADP). The blood samples of the first 11 patients (six non-OSAS and five mild-to-moderate OSAS patients) were not stimulated with 5 µM ADP. After incubation for 15 minutes at room temperature, 50 µL of 1% paraformaldehyde in HEPES Tyrode buffer was added to each of the tubes. After another 45 minutes at room temperature, these tubes were refrigerated at 4 °C and processed within the next 2 days. Samples for epinephrine analysis were spun at 1,000g at 6°C for 10 minutes and cholesterol blood was spun at 1,000g at 6 °C for 15 minutes. After the EDTA plasma was pipetted into 5 mL-biofreeze vials, it was stored at -70 °C until processed within the next 3 months. The serum for cholesterol analysis was stored at -40 °C until processed within the next 4 weeks.

Flow-cytometric assay. Rationale. P-selectin is considered a stable and reproducible platelet surface marker for determining the percentage of platelets that are activated at any given time.

Reagents. The platelet population was identified by using a monoclonal antibody (mAb) against GPIb coupled to fluorescein isothiocyanate(FITC) from AMAC (Westbrook, ME), clone SZ2, P-selectin expression was determined with a CD62 mAb, conjugated to phycoerythrin (PE) from Becton Dickinson (San Jose, CA), clone AC 1.2. Negative controls were FITC and PE-labeled immunoglobulin G (IgG) isotype (opticlone) from AMAC. All antibodies were used at optimum concentrations that gave maximal fluorescence with minimal nonspecific binding. The dilution buffer was always HEPES-buffered saline (129 mmol/L NaCl, 8.9 mmol/L NaHCO₃, 0.8 mmol/L KH₂PO₄, 0.8 mmol/L MgCl₂, 5.6 mmol/L dextrose, and 10 mmol/L HEPES; pH, 7.4) that had been passed through a 0.22-µm filter to remove dust particles. ADP was from BIO/DATA Corporation (Horsham, PA).

Flow-cytometric analysis procedure. The flow-cytometric analysis of platelets from whole blood is based on the methods of Shattil et al.¹⁶ and Warkentin et al.¹⁷ Within 2 minutes of collection, 3 µL of whole blood was added to 47µL of filtered HEPES-buffered saline with appropriate dilutions of GPIb-mAb, CD62-mAb, and ADP (final ADP concentrations of 0, 0.5, 1, 2, or 5µmol/L). The samples were incubated without stirring for 15 minutes at 22 to 26 °C, then mixed with 50 µL of 1% formyl saline (1% weight-per-volume paraformaldehyde in HEPES Tyrode buffer) to stop further activation. After another 45 minutes of incubation at room temperature, they were refrigerated at 4 °C. Immediately before analysis the fixed cells were suspended gently with 2 mL of 62-µm filtered sheath fluid(FACSFlow containing sodium chloride, potassium chloride, disodium EDTA, and sodium fluoride; Becton Dickinson, San Jose, CA). The fixed samples were analyzed in a Becton Dickinson FACScan System within 2 days of collection. The platelet population was identified by its forward and side light scatter characteristics, and enclosed in an electronic bit map. Two thousand platelets per sample were analyzed as they passed through a 488-nm focused laser beam one at a time. Samples from each subject were labeled with CD42b-FITC antibody to confirm that more than 98% of the analyzed particles in each were GPIb positive. Compensation was adjusted so that a P-selectin positive control platelet population labeled with only PE-ACI.2 measured the same percent-positive platelets when it was additionally labeled with FITC-SZ2 antibody.

Other measurements Epinephrine concentrations in venous plasma were determined by high-performance cation-exchange chromatography with electrochemical detection.¹⁸ Cholesterol was analyzed using the Synchron CX system (Beckman Instruments Inc., Fullerton, CA), which measures the cholesterol concentration by a timed end point method.¹⁹

Data presentation and statistics. Data were analyzed using the SPSS for Windows 5.0.2 statistical package (SPSS, Inc., Chicago, IL) and are presented as mean and SD of the mean if not stated otherwise. The data were tested for normal distribution with the Kolmogorov-Smirnov test. Data points between the three study groups (AHI < 5/hr, 5/hr ≤ AHI < 50/hr, AHI≥ 50/hr) were compared using the chi-square test, Kruskal-Wallis test, Mann-Whitney U-test, and Wilcoxon's rank sum test. Data within one study group were analyzed using Friedmann's two-way ANOVA. Data were analyzed for correlation with Pearson's correlation coefficient if distributed normally, and with Spearman's correlation coefficient if not distributed normally. The association of AHI and epinephrine levels or platelet activation was examined by multiple stepwise forward regression analysis, controlling for age, sex, body mass index (BMI), cholesterol, and total arousal index (TAI). Multiple stepwise forward regression analysis controlling for sociodemographics and correlating factors was further used to identify independent influences on blood pressure values. A p value of less than 0.05 was considered significant.

Results. Clinical characteristics. The true control subjects did not differ from the patients included in the non-OSAS group in sleep architecture, blood pressure, and platelet activation, but had lower platelet aggregability values on stimulation with ADP (mean difference between groups, 15.86 ± 4.76%; p ≤ 0.029). All patients with OSAS had poor sleep quality and excessive daytime somnolence. Patients with OSAS were older, more obese, and were more frequently men than patients without sleep apnea (table 1). There was a positive correlation between AHI and age (Spearman's correlation coefficient, 0.350; p = 0.002) and BMI (Spearman's correlation coefficient, 0.460; p < 0.0001). Four patients without sleep apnea and four patients with mild-to-moderate OSAS smoked. Six non-OSAS patients (AHI < 5/hr), five patients in the mild-to-moderate OSAS group (5/hr ⩽ AHI < 50/hr), and two patients of the severe OSAS group (AHI ≥ 50/hr) took platelet antiaggregating agents (ibuprofen, acetylsalicylic acid, or naproxen) within 5 days prior to the sleep study (the number of patients taking antiaggregating medications was not different between the groups). Patients taking antiaggregating agents did not have significantly different platelet activation values as measured by P-selection expression. Eight non-OSAS patients, nine patients with mild-to-moderate OSAS, and four patients with severe OSAS were being treated for systemic hypertension at the time of the study. Thus, more OSAS than non-OSAS patients took antihypertensive medications (p = 0.006). A variety of medications had been prescribed to these patients, but analysis showed no correlation between platelet- or hypertension-specific drugs and any value.

Sleep analysis. Analysis of breathing. The non-OSAS group showed a mean AHI that was lower than the mean AHI of the mild-to-moderate group and that of the severe group (severe OSAS, 60.4± 8.7/hr TST; mild-to-moderate OSAS, 13.8 ± 8.1/hr TST; non-OSAS, 1.9 ± 1.4/hr TST; p < 0.0001). There was no difference in baseline O₂ saturation between the groups, whereas the mild-to-moderate group and the severe group each showed a significantly lower average desaturation (severe OSAS, 83.2 ± 5.7%; mild-to-moderate OSAS, 89.5 ± 2.8%; non-OSAS, 91.8 ± 2.6%) and significantly more time spent in apnea or hypopnea (severe OSAS, 155.8 ± 53.6 minutes; mild-to-moderate OSAS, 30.4 ± 25.7 minutes; non-OSAS, 2.4 ± 3.6 minutes) than the non-OSAS patients. There was a significant negative correlation of the baseline O₂ saturation with the time spent in apnea or hypopnea (Pearson's correlation coefficient, -0.259; p = 0.024). The correlation of the time spent in apnea or hypopnea with the AHI was positive (Pearson's correlation coefficient, 0.942; p < 0.0001).

Sleep architecture. The TAI and the amount of stage 1 were higher in the severe group than in the non-OSAS group (severe OSAS: TAI, 40.9± 20.8/hr TST; stage 1, 49.9 ± 30% of TST; non-OSAS: TAI, 8.2 ± 8.1/hr TST; stage 1, 23.2 ± 10.5% of TST;p ⩽ 0.036), which had spent more time in stages 3 and 4(slow-wave sleep) (severe OSAS: 3.3 ± 5.8% of TST; non-OSAS, 22.7± 9.7% of TST; p = 0.001). TST was not significantly different between the three groups. AHI correlated positively with percentage of sleep stage 1 (Pearson's correlation coefficient, 0.424;p< 0.0001) and minutes of sleep stage 1 (Pearson's correlation coefficient, 0.554; p < 0.0001), and correlated negatively with percentage of slow-wave sleep (Pearson's correlation coefficient,-0.251; p = 0.029).

EMG analysis of the anterior tibial muscles. The severe OSAS group had a significantly lower index of periodic limb movements per hour of TST than the non-OSAS group (severe OSAS, 1.7 ± 3.3/hr TST; non-OSAS, 14.2 ± 16.2/hr TST). The majority of non-OSAS patients had periodic limb movements as documented by PSG.

Blood sample analysis. Epinephrine. The evening and morning epinephrine analyses of one mild-to-moderate OSAS patient were excluded because of extraordinary high values (9 PM, 598.1 pg/mL; 6 AM, 598.4 pg/mL). Furthermore, exact morning epinephrine analyses of three non-OSAS patients were not possible because standard and epinephrine peaks overlapped in the chromatography.

The 6 AM epinephrine values correlated positively with AHI (Pearson's correlation coefficient, 0.252; p = 0.033), 9 PM platelet activation (Pearson's correlation coefficient, 0.448; p < 0.0001), and aggregability on stimulation with 0.5 µm ADP (Spearman's correlation coefficient, 0.249; p = 0.035). Multiple linear regression analysis controlling for age, sex, BMI, cholesterol, and TAI showed that AHI was associated independently with 6 AM epinephrine. The positive correlation between 6 AM epinephrine and AHI was mainly related to high 6 AM epinephrine levels in the severe OSAS group (table 2).

Platelet activation. Blood samples were taken from different veins at 9 PM and 6 AM in 64 of the 76 patients. The morning platelet activation of the 12 patients with double access of the same vein did not differ significantly from that of the 64 patients in whom different veins were accessed. Exclusion of these 12 patients from statistical analysis did not change our results. Furthermore our pretest showed no significant differences in platelet activation at different times or after venipuncture of different veins: platelet activation 11 AM, 4.49 ± 2.44%; 3 PM(same vein like 11 AM), 4.01 ± 2.79%; 3 PM (opposite arm), 4.44± 3.08 (p = not significant). The evening blood sample containing 5 µM ADP of one mild-to-moderate OSAS patient was clotted and therefore could not be analyzed by flow cytometry.

The 9 PM platelet activation correlated positively with AHI (Pearson's correlation coefficient, 0.296; p = 0.009), the time spent in apnea or hypopnea (Pearson's correlation coefficient, 0.245;p = 0.033), and 6 AM epinephrine (Pearson's correlation coefficient, 0.448; p < 0.0001). The positive correlation between 9 PM platelet activation and AHI was mainly related to high 9 PM platelet activation levels in the severe OSAS group (see table 2). The 6 AM platelet activation correlated negatively with the baseline O₂ saturation (Spearman's correlation coefficient, -0.235; p = 0.041). Multiple linear regression analysis controlling for age, sex, BMI, cholesterol, TAI, AHI, and 6 AM epinephrine showed that AHI and 6 AM epinephrine were associated independently with 9 PM platelet activation. Although not statistically significant, morning and evening GPIb expression was consistently lower in the severe OSAS group. Morning and evening platelet aggregability was highest in the severe OSAS group after stimulation with different concentrations of ADP.

Blood pressure. AHI was correlated positively with increased systolic blood pressure values at 9 PM (Spearman's correlation coefficient, 0.604; p < 0.0001) and 6 AM (Spearman's correlation coefficient, 0.580; p < 0.0001), and with increased diastolic blood pressure values at 9 PM (Spearman's correlation coefficient, 0.336; p = 0.006) and 6 AM (Spearman's correlation coefficient, 0.406; p = 0.001). Increased morning diastolic blood pressure also correlated positively with 6 AM epinephrine (Spearman's correlation coefficient, 0.264; p= 0.038). Patients with OSAS had higher systolic blood pressures at 9 PM and 6 AM, and patients with severe OSAS had higher diastolic blood pressures at 6 AM than non-OSAS patients (systolic at 9 PM: severe OSAS, 155 ± 28.7 mm Hg; mild-to-moderate OSAS, 140 ± 13 mm Hg; non-OSAS, 123 ± 18 mm Hg; systolic at 6 AM: severe OSAS, 155 ± 23.8 mm Hg; mild-to-moderate OSAS, 142.1 ± 12.4 mm Hg; non-OSAS, 122.8± 18.7 mm Hg; p < 0.0001; diastolic at 6 AM: severe OSAS, 100 ± 22 mm Hg; non-OSAS, 83.4 ± 12.7 mm Hg; p = 0.029). A significant correlation between increased blood pressure values and platelet activation levels was not found.

The multiple regression analysis controlling for sociodemographic characteristics and factors correlating significantly with blood pressure values showed that the AHI was an independent risk factor for elevated systolic blood pressure values. Furthermore, age was identified as an independent risk factor for elevated morning and evening systolic blood pressure. The male sex was also considered an independent risk factor for elevated morning systolic and diastolic blood pressure. Other independent factors were high BMI, which increased morning and evening diastolic blood pressure and 6 AM systolic blood pressure, and high TAI, which elevated evening diastolic blood pressure values.

Discussion. The current study demonstrates an association among OSAS, morning epinephrine levels, and evening platelet activation levels. Levels of morning plasma epinephrine and evening P-selectin expression were increased in patients with severe OSAS. To our knowledge, this is the first study to investigate the relation of platelet activation and plasma catecholamines in OSAS. Our findings suggest that increased epinephrine levels and associated elevated platelet activation levels may play a role in the increased cardiovascular and cerebrovascular morbidity of patients with OSAS.

Epinephrine and platelet activation. The elevated mortality rate associated with OSAS has been attributed to cardiovascular mortality.²⁰ OSAS is also overrepresented in cross-sectional populations that survive myocardial infarction.²¹ Snoring, which almost invariably accompanies OSAS, was reported to be a risk factor for stroke.²²

Besides hypertension, increased BMI, male sex, and advanced age-very common factors in OSAS-other pathophysiologic mechanisms possibly involved in high cardiovascular and cerebrovascular morbidity in OSAS are increased risk for atherosclerosis due to apnea-induced hypoxemia²³ and altered fibrinolytic²⁴ and platelet function.^25,26 In a preliminary study of six patients, Bokinsky et al.²⁶ recently found that platelet aggregability (measured as P-selectin expression) was increased in patients with OSAS and decreased after treatment with CPAP. This suggests that the change in platelet function is secondary to OSAS. The reason for this OSAS-related change is unknown, but it may at least partially be related to the reduced catecholamine levels after CPAP therapy.²⁷

There were neither significant differences between the three groups in regard to platelet aggregability to ADP nor correlations between AHI with platelet aggregability to ADP, although epinephrine, which was highest in the severe OSAS group, can increase the platelet-activating effect of low concentrations of ADP.²⁸ Figures et al.²⁸ assumed an increase in ADP-receptor binding affinity caused by epinephrine, but they also showed that an increase of in vitro ADP decreases the influence of epinephrine on ADP-receptor binding affinity. This could explain the decreasing differences between OSAS and non-OSAS in platelet aggregability and the weak correlations between AHI and platelet aggregability with increasing concentrations of ADP.

The platelet activation levels of our patients who took platelet-antiaggregating agents 5 days prior to their PSG did not differ from those of the other participants. This is to be expected because ibuprofen, acetylsalicylic acid, and naproxen do not affect P-selectin expression primarily.

The patients with severe OSAS had the highest platelet activation and epinephrine values at 6 AM, suggesting a greater vulnerability to increased platelet activation during subsequent morning hours, when thrombotic stroke and cardiovascular events are most common. Angoli et al.¹ found that cerebrovascular accidents occurred more frequently between 6 AM and 2 PM. When Mitler et al.²⁹ specifically investigated deaths due to ischemic heart disease, they found an increase between 6 AM and 10 AM. We found that platelet activation values increased in the evening compared with morning values, and multiple regression analysis defined morning epinephrine and AHI as independent factors for the increased evening platelet activation. Platelet survival half-life is much higher than epinephrine survival half-life.^30,31 Therefore, platelet activation caused by morning epinephrine could still be observed at least partially in the evening.

In brief, there is an association between respiratory events (apneas and hypopneas) and evening and morning P-selectin expression or morning epinephrine levels. Increased daytime ventilation with lowered sympathicoadrenal activity due to decreased chemoreflex activity³² could be one reason for the evening epinephrine decrease in OSAS patients.

The platelet activation values at 6 AM were decreased from nighttime values in all groups. This finding is consistent with the results of some studies³³ but contradicts those of previous studies^34,26 that measured platelet aggregability in small numbers of normal subjects and OSAS patients. One reason for the different results could be the different method of measuring platelet aggregation or activation. Only Bokinsky et al.²⁶ measured P-selectin expression in OSAS patients and normal subjects. However, they found much higher activation levels than we did, probably because they drew blood samples through tubing.

Stimulation of platelets not only causes P-selectin expression but also reduces the amount of GPIb receptors.³⁵ High AHIs were associated with decreased GPIb, but this result was not significant statistically. However, activation-dependent changes in binding of different monoclonal antibodies do not have to be uniform.³⁶

Influence of position. The regression analysis showed that a high AHI increased morning epinephrine levels independently. Upright posture also increases epinephrine.³⁷ It is to be expected that platelet activation in OSAS patients is more pronounced than platelet activation in non-OSAS patients during the subsequent hours, because the proaggregatory effect of epinephrine persists for at least 3 hours.³¹

Furthermore, upright posture and daily activity have an additional effect not only on epinephrine but also on platelet aggregability.³⁷ Accordingly, patients with severe OSAS had significantly higher platelet activation levels at 9 PM than non-OSAS patients. Morning platelet activation was highest in the severe OSAS group, but was not statistically significant. Lande et al.³⁸ found that patients with various cardiac diseases required extremely high levels of plasma epinephrine (approximately 29 nM) to elevate platelet activation. They also found that lower levels of epinephrine can increase platelet activation in hypertensive subjects (approximately 2.5 nM).³⁹ Our severe OSAS patients were diagnosed to have hypertension, and their highest epinephrine levels occurred in the morning; however, their levels were still lower than the previously mentioned values (approximately 0.2 nM). To what degree upright posture and daily activity additionally increase plasma epinephrine in hypertensives requires further investigation.

When a patient lies in a supine position, plasma catecholamines should fall to baseline levels within 2 hours.³¹ The OSAS and non-OSAS patients in our study were recumbent for at least 15 minutes before blood was drawn at 9 PM. The evening epinephrine values were equally low in OSAS and non-OSAS patients, whereas OSAS patients had an increase in morning plasma epinephrine, possibly caused by stressful nightly events such as apneas. Arousals, which are associated with increased sympathetic activity and elevated blood pressure,⁴⁰ were not identified as independent risk factors for elevated morning epinephrine and evening platelet activation. This suggests the importance of apnea-related aspects other than increased sympathetic tone for increased platelet activation(e.g., O₂ desaturation or hypertension). Wedzicha et al.⁴¹ found that platelets of patients with chronic hypoxemia and lung disease had an increased tendency to aggregate under conditions of hypoxemia and hypercapnia. We found an association between 6 AM platelet activation and baseline O₂ saturation, which correlated significantly and negatively with the time spent in apnea or hypopnea. The time spent in apnea or hypopnea was correlated significantly and positively with the AHI.

Blood pressure. Our OSAS patients had higher systolic blood pressures at 9 PM and 6 AM, and higher diastolic blood pressure at 6 AM than non-OSAS patients. Increased systolic and diastolic morning and evening blood pressures correlated significantly and positively with the AHI. In OSAS patients, 9 PM diastolic blood pressure was lower than 6 AM diastolic blood pressure. Thus the statistically significant differences between groups in 6 AM diastolic blood pressure no longer persisted at 9 PM. These findings confirm the previously reported association between OSAS and hypertension.^5,42 However, to our knowledge, the decrease of high morning diastolic blood pressure during the day in severe OSAS patients has not been described previously. Elevated morning catecholamines, especially norepinephrine, could contribute to high morning diastolic blood pressure values in OSAS. Increased daytime ventilation with reduced sympathoadrenal activity due to decreased chemoreflex activation³² could contribute to the decrease of diastolic blood pressure in OSAS patients over the day. This hypothesis, however, is contradicted by Marrone et al.,²⁷ who found an increased secretion of norepinephrine in OSAS patients during the day. We measured plasma epinephrine at two points during the day and found a tendency for epinephrine to be decreased in the evening. Marrone et al.²⁷ measured catecholamines in collected urine. The daytime decrease in diastolic blood pressure could also be due to the intake of antihypertensive medications, which were significantly more common in severe OSAS patients than in non-OSAS patients. The mechanism by which elevated systolic blood pressure is sustained during the day is not known. The repeated increase in sympathetic tone,⁴³ repeated hemodynamic oscillations caused by frequent apneic episodes⁴⁴ with consecutive neurohumoral or vascular changes,⁴⁵ and increased catecholamines during wakefulness²⁷ may be partially responsible for the development of diurnal hypertension in patients with sleep apnea.

Our study identified AHI as an independent risk factor for increased systolic blood pressure values. Diastolic blood pressure values are more likely to be increased by sociodemographic factors, such as high BMI, advanced age, and male sex. BMI and age themselves, however, correlated significantly and positively with AHI. Hypertension was recently shown to be associated with platelet activation.⁴⁶ However, we did not find a significant correlation between blood pressure values and platelet activation values.

Sleep architecture in OSAS. We found a reduction of slow-wave sleep and an increase of sleep stage 1 and arousals in severe OSAS. These results agree with those of previous studies.⁴⁷ The changed sleep architecture in OSAS is secondary to airway obstruction. At high altitudes, for example, normal subjects show an increase of nocturnal arousals and a decrease in slow-wave sleep.⁴⁸

Our severe OSAS group had significantly less periodic leg movements and significantly more arousals in sleep than the non-OSAS group. The small number of periodic leg movements in severe OSAS might be due to scoring criteria that do not allow a score for a periodic leg movement occurring during an arousal.⁴⁹ Our non-OSAS patients had a mild increase in periodic leg movements and arousals. This might be the result of a first-night effect⁵⁰ or the fact that our non-OSAS group did not consist of only true control subjects but consisted mainly of patients who reported daytime sleepiness and sleep disturbances.

We found that OSAS was associated with increased plasma epinephrine, platelet activation, blood pressure levels, and sleep impairment. If confirmed, our results suggest that drugs that decrease platelet activation or the influence of epinephrine on platelet function have a possibly protective effect in patients with OSAS. However, to identify OSAS as the cause of increased plasma epinephrine and platelet activation, the findings of the current study require confirmation from a larger investigation that also includes the influence of CPAP treatment. Reduction of plasma catecholamines and platelet activation after successful CPAP treatment would suggest that OSAS is the primary cause of epinephrine-associated platelet activation.

Acknowledgments

The skillful technical assistance of Ms. Cindy Boothby and Ms. Paula Abbott is gratefully acknowledged.