INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Hypertension is a common complication of obstructive sleep apnea in adults. It may result in increased cardiovascular morbidity and mortality. Although the obstructive sleep apnea syndrome (OSAS) is common in children, affecting 1-3% of the preschool population (1, 2), little is known about hypertension in these patients. Several case reports have shown that hypertension does occur in some children with OSAS (3). However, the cases described were all severe and do not represent the spectrum of childhood OSAS seen today.

We hypothesized that children with OSAS would have an increased prevalence of high blood pressure. We therefore measured blood pressure during polysomnography in children with OSAS.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Design

Pediatric patients with suspected OSAS underwent serial blood pressure measurements during polysomnography. Polysomnography results were then used to subdivide subjects into those with OSAS and those with primary snoring (PS), which is snoring without obstructive apnea (6). Informed consent was obtained from the parents/legal guardians of each child, and assent was obtained from the child if older than 5 yr of age. The study was approved by the Institutional Review Board of Johns Hopkins University.

Study Group

Patients were recruited sequentially from those referred to the Pediatric Sleep Disorders Clinic at Johns Hopkins Hospital for evaluation of snoring and difficulty breathing during sleep. Children less than 1 yr of age, or those with cardiac disease, renal disease, or other diseases that predisposed them to hypertension were excluded from the study. Patients using medications that could potentially affect blood pressure (including beta-agonists) at the time of study were excluded.

A history was obtained and physical examination performed by one of the investigators. Growth percentiles were obtained using standard growth charts (National Center for Health Statistics, adapted by Ross Laboratories). Children were considered obese if their weight was greater than 120% of their ideal weight for height (7). The body mass index (BMI) was defined as the weight in kilograms divided by the square of the height in meters.

Polysomnography

Polysomnographic studies were performed overnight. No sedation or sleep deprivation was used. Children were accompanied by a parent throughout the night. During polysomnography, the following parameters were measured and recorded continuously on a computerized system (Alice 3; Healthdyne, Marietta, GA): electroencephalogram (C3/A2, O1/A2); right and left electrooculogram; submental electromyogram (EMG); tibial EMG; electrocardiogram; chest and abdominal wall motion (piezoelectric transducers), oronasal airflow (three-pronged thermistor); end-tidal PCO₂, measured at the nose by infrared capnometry (Nellcor N-1000; Van Nuys, CA); Sa_O2 by pulse oximetry (Nellcor N-1000); oximeter pulse waveform, and systolic and diastolic blood pressure (see below). Children were also monitored and recorded on videotape, using an infrared video camera, and were continuously observed by a polysomnography technician. The following parameters were measured:

1. Sleep architecture: Assessed by standard techniques (8). Arousals were defined as recommended by the American Sleep Disorders Association (9).
2. Obstructive apneas: Defined as the presence of chest/abdominal wall motion in the absence of airflow. As children have a higher respiratory frequency than adults and frequently desaturate even with short apneas, all obstructive apneas greater than or equal to two breaths duration were counted (10, 11). The obstructive apnea index was defined as the number of obstructive apneas per hour of sleep. Mixed apneas (apneas with both central and obstructive components) were included in the apnea index.
3. Hypopneas: Defined as a reduction in airflow of 50%, in the presence of chest/abdominal wall motion, associated with desaturation greater than or equal to 4%. The apnea hypopnea index (AHI) was defined as the number of obstructive apneas and hypopneas per hour of sleep.
4. Sa_O2: The Sa_O2 nadir and mean Sa_O2 were determined. Arterial oxygen saturation measurements associated with a poor pulse waveform were discounted. The percentage of total sleep time during which Sa_O2 was less than 92% was measured.
5. End-tidal carbon dioxide tension (PET_CO2): The mean and peak PET_CO2 were determined. The percentage of total sleep time during which PET_CO2 was greater than or equal to 50 mm Hg was measured.

Children were diagnosed with OSAS if they had an obstructive apnea index greater than 1/h, Sa_O2 nadir less than 90% associated with obstructive apnea, peak PET_CO2 greater or equal to 53 mm Hg or hypoventilation greater or equal to 10% of total sleep time (10). Most subjects with OSAS met multiple criteria. Subjects were classified as having PS if they did not meet any of the above criteria. Two children who had normal sleep studies except for transient desaturation to 89% (one in conjunction with central apnea, and one during REM sleep not associated with apnea) were included in the PS group.

Measurement of Blood Pressure

Systolic and diastolic blood pressure (BP) were measured every 15 min during polysomnography, using an automated system (Dinamap; Johnson & Johnson Medical Inc., Tampa, FL) and appropriately sized arm cuffs. Technicians were carefully trained to use cuff size as recommended by the 1987 Task Force on Blood Pressure Control in Children (12). Technicians were instructed to discontinue BP recordings if the measurements disturbed the patient's sleep. Mean systolic and diastolic pressures were calculated for each patient over the course of the night. Data obtained during periods of wakefulness during the night were analyzed separately. Blood pressure values were compared to recent guidelines on age-appropriate normative data (13).

In children, BP changes with growth and development. Therefore, normative BP values are based on gender, age, and height percentiles (13). As we studied children of differing ages and height, BP values were normalized by deriving a BP index. This consisted of the difference between the subject's mean BP and the BP at the 95th percentile for that subject's age, gender, and height, based on established normative data (13). Blood pressure indices were calculated separately for systole and diastole. Patients were considered to be hypertensive if mean systolic BP, diastolic BP, or both were greater than the 95th percentile (13).

Statistical Analysis

Demographic and polysomnographic differences between OSAS and PS subjects were compared using the unpaired t test (continuous variables) and chi-square analysis (categorical variables). All results are expressed as mean ± SD unless otherwise indicated. Mean indexed systolic and diastolic BP, both awake and asleep, were compared between groups using the unpaired t test. For the remainder of the analysis, OSAS and PS groups were collapsed to study OSAS as a continuous spectrum of severity, thus avoiding an arbitrary cutoff between normal and abnormal. Using paired t tests, the mean intra-individual change in BP was analyzed for the following state changes: wake to non-rapid eye movement (non-REM), wake to rapid eye movement (REM), and non-REM to REM. Multiple linear regression was used to identify demographic, morphometric, and polysomnographic factors that might predict BP. Log transformation of continuous predictor variables was performed when indicated to reduce the influence of outlying predictor variables.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

A total of 75 children were studied. Three were excluded. One of these had taken beta-agonists on the day of study for an asthma exacerbation, and another was undergoing postoperative evaluation. The third patient, who was obese, was receiving treatment for hypertension. Five studies were terminated. In one case, the parents decided to terminate the entire sleep study because the child had difficulty sleeping; in one case the BP equipment malfunctioned; and in three cases, the blood pressure readings interfered with the child's sleep, and thus the BP portion of the study was discontinued. Of the remaining 67 patients, 26 had normal sleep studies and were classified as PS, and 41 had studies demonstrating OSAS.

Subject characteristics are shown in Table 1. Patients with OSAS were slightly younger than patients with PS, and were more likely to be male. There were no significant differences in height, BMI, or percentage of obese subjects between the two groups. Most patients were healthy other than for OSAS secondary to adenotonsillar hypertrophy. Eleven patients (nine with OSAS) had a history of reactive airways disease, but required either no treatment or cromolyn alone at the time of study. Three children (two with OSAS) had a history of prematurity. Of the children with OSAS, one child had achondroplasia, one had Down syndrome and one had a translocation of chromosome 18. One child with PS had Pierre Robin sequence. Two patients, both obese, had persistent OSAS following previous tonsillectomy and adenoidectomy; one child with PS had undergone previous tonsillectomy and adenoidectomy secondary to recurrent tonsillitis.

Polysomnography

Polysomnography results are shown in Table 2. Sleep architecture and efficiency were within normal limits for our laboratory and were similar between the two groups. The degree of OSAS ranged from mild to severe, but on average was moderate by pediatric standards. Many children with OSAS demonstrated obstructive hypoventilation (i.e., snoring associated with retractions, paradoxical respiration, and hypercapnia), rather than complete obstructive apneas, and thus had a low apnea index. This is a pattern characteristic of pediatric OSAS (11, 14).

Blood Pressure

Patients with OSAS had significantly higher diastolic BP during both wakefulness and sleep, compared with subjects with PS. During sleep, the diastolic BP index was 15 ± 8 mm Hg for OSAS versus 23 ± 8 mm Hg for PS (p < 0.001). During wakefulness, the diastolic BP index was 8 ± 13 mm Hg for OSAS versus 16 ± 9 mm Hg for PS (p < 0.005). There were no significant differences in the systolic BP indices during either sleep or wakefulness between the two groups (sleep: 19 ± 11 versus 20 ± 22 mm Hg; wake: 12 ± 14 versus 13 ± 15 mm Hg).

Compared with normative data, the subjects with OSAS and PS had relatively high blood pressures. Five (12%) of OSAS patients had BP greater than 95th percentile during sleep, compared with one (4%) PS subject. Thirteen (32%) OSAS patients had BP values (systolic and/or diastolic) greater than 95th percentile during sleep and/or wakefulness, compared with 5 (19%) PS subjects (NS).

For the group as a whole, BP correlated with both the apnea index (Figure 1) and AHI (diastolic BP: r = 0.59, p < 0.001 for AI, and r = 0.54, p < 0.001 for AHI; systolic BP: r = 0.28, p < 0.02 for AI, and r = 0.21, NS for AHI). We elected to use the apnea index, since normative data for hypopneas in children have not been established. Simple regression showed significant correlations between the BP indices and the apnea index and BMI, respectively (Figures 1 and 2). Therefore, multiple linear regression was performed. In addition to both BMI and AI, a significant linear effect was noticed between indexed diastolic BP and age. The latter finding suggested that age was incompletely controlled by the indexing process, and therefore should be included in the multiple regression analysis. Using multiple linear regression models that included age, BMI, and AI as predictor variables (all three log-transformed), 58% of the variability in indexed diastolic BP and 19% of the variability in indexed systolic BP was predicted. When holding the other predictor variables constant, age, BMI, and AI were all independent predictors of indexed diastolic BP. In contrast, only BMI was significantly related to indexed systolic BP. When additional predictor variables were added individually to the basic model, none was statistically significant. These variables included gender, race, and the polysomnographic variables noted in Table 3.

View larger version (16K): [in this window] [in a new window]   Figure 1.   Blood pressure index (the difference between the subject's mean BP during sleep and the BP at the 95th percentile for age and height) is plotted against the apnea index for diastole (panel A) and systole (panel B). There was a significant correlation between the BP index and apnea index for both systole and diastole.

View larger version (17K): [in this window] [in a new window]   Figure 2.   Systolic (panel A) and diastolic (panel B) BP indices are shown for OSAS and PS patients. Non-obese patients are represented by filled circles, and obese patients by open circles. Group means and standard deviations are shown.

There was a significant drop in BP during sleep compared with wakefulness (Figure 3). Both systolic and diastolic BP values were 7 ± 10 mm Hg lower during sleep (p < 0.001). Most subjects had lower BP during sleep; in fact, all the subjects with elevated BP values showed a dip in BP during sleep compared with wakefulness. There was no significant correlation between the indices of OSAS severity (i.e., apnea index and Sa_O2 nadir) and the size of the dip in BP from wakefulness to sleep. When BP during wakefulness was compared with non-REM and REM sleep, it was found that awake BP was higher than both non-REM and REM sleep (p < 0.0001 for both systole and diastole), but non-REM and REM did not differ significantly from each other.

View larger version (23K): [in this window] [in a new window]   Figure 3.   The changes in systolic (panel A) and diastolic (panel B) BP from wakefulness to sleep are shown for individual subjects. OSAS subjects are depicted as filled circles, and PS subjects as open circles. Group means and standard deviations are represented by the error bars. There was a significant decrease in BP during sleep (p < 0.001 for both systole and diastole).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study has shown that children with OSAS have higher diastolic blood pressures than children with PS. The degree of BP elevation was related to the severity of obstructive sleep apnea and the degree of obesity.

To the best of our knowledge, this is the first study to prospectively evaluate BP in a cohort of children with OSAS. However, a number of cases of children with hypertension related to OSAS have been reported in the literature (3, 15). Most of these reports are from the older literature and describe children with severe OSAS resulting in respiratory failure, heart failure, or coma. In addition, Guilleminault and colleagues (5) reported in 1976 that five of eight children with OSAS were hypertensive during wakefulness. These children, while not in extremis, had severe OSAS with high apnea indices (5). Childhood OSAS is now more widely recognized, and children are usually diagnosed with milder illness, before severe complications occur. Nevertheless, the current study shows that high blood pressure occurs even among milder cases. Although the degree of BP elevation in this study was mild, the children were very young and had only moderate OSAS. It is therefore conceivable that children who do not receive treatment for OSAS will develop worsening hypertension over the years.

The etiology of OSAS-related hypertension has been studied extensively in adults (16). It is thought to be due to a number of factors, particularly sympathetic nervous system activation secondary to arousal, and to a lesser degree, hypoxemia (17). Changes in cardiac output secondary to intrathoracic pressure swings may also play a role. The mechanisms for hypertension in children are probably similar to those for adults. Although cortical arousals at the termination of obstructive apneas are less common in children than adults (18), children may manifest signs of subcortical arousal, including autonomic changes such as tachycardia (19). It is therefore possible that these subcortical arousals are associated with elevations of BP. In the current study, we found a correlation between the frequency of obstructive apnea and BP, but no correlation between Sa_O2 and BP. This suggests that arousal rather than hypoxemia may be a major determinant of BP elevation in children. Because BP was averaged every 15 min rather than being measured on a beat-to-beat basis, we could not further address the pathophysiology of the hypertension. However, Shiomi and coworkers (20) found that an 11-yr-old child with OSAS had the greatest increase in BP upon arousal following obstructive apneas. These facts suggest that arousal (either cortical or subcortical) plays a role in BP elevation in children with OSAS.

In the current study, we found a significant difference in diastolic BP between children with OSAS and PS, despite the fact that many children in the OSAS group had very mild disease. We used an AI of 1 to distinguish between the groups, as this level of apnea index has been shown to be statistically abnormal in children (10). Because diastolic BP was proportional to the AI, a more severe group of patients may have shown an even greater increase in BP.

Diastolic BP increased with increasing apnea index. Both systolic and diastolic BP also increased with the degree of obesity, as measured by the BMI. However, the children with OSAS in this study had a higher diastolic BP than those with PS, although the BMI and the percentage of obese subjects were comparable between the two groups. Multiple linear regression techniques demonstrated that BMI was a major determinant of BP, but that apnea index was important, independently, in determining diastolic BP. These findings are similar to findings in adults with OSAS. Coy and colleagues (21) studied adults with OSAS and found that the respiratory disturbance index independently predicted diastolic BP, whereas the BMI independently predicted systolic BP.

Although the children with OSAS had a higher BP than those with PS, most subjects still had negative BP indices. This is to be expected, as we were comparing data from sleeping subjects with data obtained in awake individuals. In order to compare children over a large age and height range, we normalized BP values by deriving a BP index based on normative data obtained in a large sample (> 56,000) of children (13). The normative data were obtained from awake subjects. Blood pressure during sleep is lower than during wakefulness; studies in children have shown an average decline of 5-13 mm Hg during systole and 4-29 mm Hg during diastole (22). Therefore, the presence of an elevated BP during sleep is strongly suggestive of hypertension.

In this study, both the children with OSAS and PS had a higher level of BP than the general pediatric population, where the prevalence is 1% (13). This is not surprising, as both primary snoring (25) and upper airway resistance syndrome (26) can potentially affect blood pressure. It is probable that normal, nonsnoring children would have lower BP during sleep than the children enrolled in this study.

We found that both the AI and the AHI correlated significantly with the diastolic BP. We chose to use the apnea index, because there is a wide variability in hypopnea definitions among different centers (27, 28), and because normative data for hypopneas in children have not been established (11). In our laboratory, hypopneas are not scored unless there is associated desaturation. This is a common practice (28), based on the fact that thermisters do not provide quantitative measurements of airflow.

In the current study, the subjects with OSAS were more likely to be male. This was expected. In adults, OSAS is more common in males than in females. The gender distribution for childhood OSAS has not been studied formally. However, we and others (Carol Rosen, personal communication) have noticed a male preponderance. The subjects with OSAS were also younger than those with PS. Use of the BP indices normalized the data in regard to both age and gender, thereby minimizing these differences between the two groups.

Most subjects tolerated the BP measurements well, without arousing. In only a few cases was the study discontinued because BP measurements disrupted sleep. The sleep efficiency and architecture during this study were similar to that normally seen in our laboratory in children with OSAS and similar to that described in the literature (29). Children tend to have a high threshold for arousal from sleep (30), and previous studies have shown that children tolerate BP measurements with arm cuffs well during sleep (31). Hla and colleagues (32) measured BP simultaneously with arm cuffs and beat-to-beat plethysmography in sleeping adults, and showed that cuff inflation can cause microarousals associated with transient elevation of plethysmographically measured BP, but that these fluctuations do not last long enough to affect the cuff measurements. Although we cannot exclude the possibility that cuff inflation resulted in autonomic changes in our subjects, this effect would be expected to be similar between those with OSAS and PS.

We found that BP was lower during sleep than wakefulness. There was no difference in BP between REM and non- REM sleep. Blood pressure is known to have a circadian variation. In normal adults and children, BP is lower during sleep than during wakefulness (22, 33, 34), and REM values are variable (33, 34). In contrast to normal subjects, nocturnal BP is variable in adults with OSAS. Some adult patients with OSAS have a decline in BP with sleep. However, others show no decline or even an elevation in BP (35). In the current study, most children, including all those with BP indices greater than the 95th percentile, showed a decline in BP from wakefulness to sleep.

The diagnosis of hypertension cannot be made solely on the basis of BP measurements on a single occasion. However, the high prevalence of elevated BP in this study indicates the need for longitudinal studies evaluating the relationship between childhood OSAS and hypertension. In addition, the effect of treatment of OSAS on BP should be studied.

In conclusion, we have shown that children with moderate OSAS have higher diastolic blood pressure than control subjects during wakefulness and/or sleep. We recommend that blood pressure be measured in all children diagnosed with obstructive sleep apnea. Further studies are required to elucidate the pathophysiology of hypertension in this group, and the effect of treatment of OSAS on blood pressure.