The Prevalence of Obstructive Sleep Apnea in ... - ATS Journals

17 downloads 73 Views 47KB Size Report
TREFOR OWEN MORGAN, ADRIANNE ILA ANDERSON, and ROBERT J. PIERCE ..... have shown repetitive acute rises in BP with each apnea in pa- tients with ...
The Prevalence of Obstructive Sleep Apnea in Hypertensives CHRISTOPHER JOHN WORSNOP, MATTHEW THOMAS NAUGHTON, COLIN EDWIN BARTER, TREFOR OWEN MORGAN, ADRIANNE ILA ANDERSON, and ROBERT J. PIERCE Department of Respiratory Medicine and Hypertension Clinic, Austin and Repatriation Medical Centre, Heidelberg, Victoria, Australia

This study was designed to measure the prevalence of obstructive sleep apnea in untreated and treated hypertensive patients by comparing them with normotensive subjects, taking into account the possible confounding variables body mass index, age, sex, and alcohol consumption. Subjects with no known sleep disorders were recruited, had full polysomnography, and had their blood pressure assessed with a 24-h ambulatory monitor. Subjects with a mean 24-h blood pressure greater than 140/90, and receiving no treatment for, or with no history of, hypertension were classified as untreated hypertensives; those receiving antihypertension medication were classified as treated hypertensives; those with a mean 24-h blood pressure less than 140/90 and no history of hypertension were classified as normotensives. Thirty-eight percent of the 34 untreated and 38% of the 34 treated hypertensives, and 4% of the 25 normotensives had apnea-hypopnea indexes greater than 5. Logistic regression analysis showed that body mass index (p 5 0.001), age (p 5 0.07), sex (p 5 0.07), treated hypertension (p 5 0.05), and untreated hypertension (p 5 0.06) were associated with the presence of sleep apnea, but that alcohol consumption (p 5 0.82) was not. It is concluded that there is a relationship between sleep apnea and hypertension that, although partially explained by the confounding variables body mass index, age, and sex, persists when these are allowed for. Worsnop CJ, Naughton MT, Barter CE, Morgan TO, Anderson AI, Pierce RJ. The prevalence of obstructive sleep apnea in hypertensives. AM J RESPIR CRIT CARE MED 1998;157:111–115.

Obstructive sleep apnea (OSA) (1) and systemic hypertension (HT) are common conditions affecting middle-aged and elderly adults. Both conditions are associated with significant morbidity and mortality (2, 3). An association between OSA and HT has been documented, and the reported prevalence of HT in those with OSA ranges from 15 to 56% (2, 4–9). The wide variation between studies is due to differences in the populations studied as well as varying definitions of OSA and HT. There is controversy as to whether this association might simply be due to a confounding factor such as obesity, which commonly occurs along with both OSA and HT, or whether there is a direct causal link between them. There is experimental evidence to support a causal link as blood pressure (BP) has been shown to rise acutely with each apnea during the night (10, 11). Also associated with apneas are repetitive arousals, hypoxia, and rises in catecholamines and sympathetic nervous system activity, all of which can lead to daytime systemic HT. The prevalence of OSA among patients who are hypertensive (HTs) has also been studied and ranges from 12 to 83% (12–20), varying for reasons similar to those above. Lavie and

(Received in original form September 11, 1996 and in revised form August 15, 1997) Supported by a grant from the Department of Veterans Affairs, Australia. Correspondence and requests for reprints should be addressed to Dr. Christopher John Worsnop, Department of Respiratory Medicine, Austin and Repatriation Medical Centre, Studley Road, Heidelberg, Victoria 3084, Australia. Am J Respir Crit Care Med Vol 157. pp 111–115, 1998

coworkers (12) screened 50 patients with HT using a questionnaire, and 16 who had a sleep study; 13 (26%) had an apnea index greater than 5 per hour. Kales and coworkers (13) used polysomnography (PSG) to diagnose OSA in 50 patients from a HT clinic, all receiving treatment; 30% had an apnea index greater than 30, and 64% had an apnea index greater than 3, compared with zero and 24%, respectively, in a control group of 50 normotensive (NT) control subjects. The two groups were matched for age and sex, but not body mass index (BMI). Fletcher and coworkers (14) compared 46 male HTs taken off medication with 34 male NTs and found that 30% of the HTs and 9% of the NTs had an apnea index greater than 10 determined with PSG. The two groups were matched for age and weight. Williams and colleagues (15) performed partial daytime nap studies on 23 male HTs receiving anti-HT medication and on eight male NTs and found that 48 and 12%, respectively, had more than five apneas per hour. The two groups had the same average weight and age. Hirschowitz and coworkers (16) performed sleep studies in five groups of men. OSA was defined using PSG as having more than five apneas per hour of sleep. The prevalences of OSA in the five groups were: HTs receiving treatment but with persistently high BP and erectile dysfunction, 51%; HTs successfully treated but with erectile dysfunction, 33%; untreated HTs, 37%; those with only erectile dysfunction, 42%; and a normal control group, 23%. The apnea index adjusted for weight and age was significantly higher in the treated HTs but with elevated BP than in the successfully treated HTs. In a group of patients referred to a sleep clinic, Escourrou and coworkers (17) found a

112

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

higher apnea-hypopnea index (AHI) in a group of HTs, most of whom were treated, than in a group of NTs. There were no significant differences in BMI, age, or sex between the NTs and the HTs. Isaksson and Svanborg (18) found that 56% of treated but uncontrolled HTs and 19% of treated and controlled HTs had OSA using a Static Charge Sensitive Bed. The 16 subjects in each group were matched for BMI, age, and sex. Kiselak and coworkers (19) studied 19 obese subjects with a portable sleep monitor in their homes and found that five of the six HTs and two of the 13 NTs had respiratory disturbance indices greater than 10. The HT status was based on a clinic BP measurement greater than 160/95 or if the subject was receiving anti-HT medication. Gleadhill and coworkers matched 34 male HTs with 34 male NTs according to BMI and age and found that 12% of the HTs and none of the NTs had more than five 4% oxygen saturation dips per hour. Six of the above studies found a higher prevalence of OSA among HTs than among NTs, but as there was a wide range in the prevalences of OSA in the HTs we decided to reexamine the issue further by designing a study in which OSA and HT were both objectively assessed. To obtain a more valid estimate of the prevalence we took steps to ensure that our subjects were not biased towards having OSA. We used a NT control group, and to allow for the effects of anti-HT medication we used both treated and untreated HTs. Regression analysis was used to allow for the effects of potential confounding variables. Aim

This study was designed to measure the prevalence of OSA in untreated HTs, treated HTs, and NTs taking into account sex, age, BMI, and alcohol consumption.

METHODS Subjects Subjects were recruited from local service clubs, shopping centers, and the Austin and Repatriation Medical Centre’s HT Clinic. They were initially invited to participate in a HT study so that subjects being enrolled were not biased towards having a sleep-related problem. Once a subject expressed interest in participating, the protocol was fully explained; then he or she gave informed consent. Three potential subjects withdrew when they learned that they had to have an overnight sleep study. If it was thought that anyone sought enrollment in the study because he or she knew it was a sleep study, he or she was excluded. The Austin and Repatriation Medical Centre’s Human Ethics Committee approved the study.

VOL 157

1998

whose mean 24-h BP was greater than 140/90 formed the untreated HT group. Those who had a past history of HT, but who were not currently being treated, and whose mean 24-h BP was less than 140/90 were excluded. Also excluded were subjects with diabetes, renal disease, or any cardiovascular disease.

Confounding Variables All subjects had their weight and height measured, and daily alcohol consumption was assessed using a standardized questionnaire (21), which took into account their past as well as their present drinking habits.

Polysomnography All subjects underwent full PSG (22), during which the electroencephalogram, submental electromyogram, and bilateral electrooculograms were recorded for assessing sleep state. Arterial oxygen saturation (Oxy Pulse Oximeter; Radiometer Pacific, Denmark), abdominal and thoracic movements with body plethysmography (Department of Biomechanical Engineering, Austin and Repatriation Medical Centre), nasal and oral airflow via a thermistor, and body position were all recorded using the Sleepwatch system (Compumedics, Melbourne, Australia). Each PSG study was manually staged in 20-s epochs using standard criteria (23). The scorer was blinded to the subjects’ BP and HT classification. Hypopneas and apneas were defined as reductions in the thermistor signal to less than 50% and less than 20% of baseline, respectively, in association with a desaturation of at least 4%. The AHI, that is, the number of apneas and hypopneas per hour of sleep, was used in this study to define OSA. However, it can be influenced by the subject’s sleeping position (24) and the amount of rapid eye movement (REM) sleep (10), so the AHI during sleep when supine (back AHI) and the AHI during REM sleep (REM-AHI) were also calculated. Two measures of the degree of oxygen desaturation during sleep were also made, the average desaturation with each apnea or hypopnea (avdesat) and the lowest oxygen saturation during sleep (lowsat), to see if they could help clarify the association between OSA and HT. Subjects were also asked about snoring using the standard question, “Do you snore?” to which each could reply “never,” “rarely,” “sometimes,” “often,” or “only after alcohol.” Some subjects were unable to answer this question.

Statistics Chi-square tests and logistic regression analyses were carried out using SPSS for Windows 6.0 (SPSS, Chicago, IL).

RESULTS Ninety-three subjects completed the study. There were 34 treated HTs, 34 untreated HTs, and 25 NTs. The descriptive data for each group are shown in Table 1. The subjects tended

BP Assessment All subjects were asked about a history of HT, medications used, and other medical conditions. All had their BP measured with a 24-h ambulatory monitor (Accutracker II; Suntech Medical Instruments, Raleigh, NC). The device was attached in the morning and removed the next morning. It uses a microphone placed over the brachial artery to detect Korotokov sounds. Subjects were advised to carry out their normal activities, but to avoid getting the device wet. Blood pressure was measured every 30 min 6 5 min from 6:00 A.M. to 10:00 P.M. (daytime BP), and every 60 min 6 10 min from 10:00 P.M. to 6:00 A.M. (nighttime BP). The exact time of BP recorded was determined randomly so that subjects could not anticipate cuff inflation. The first inflation of the 24-h period was to 250 mm Hg, and subsequent inflations were to 30 mm Hg above the previous systolic level. After each inflation there was a 2-s pause, and then the cuff was deflated at 3 mm Hg/s. Subjects who had no history of HT, were not receiving any antiHT medication, and whose mean 24-h BP was less than 140/90 formed the NT group. Those who were receiving anti-HT medication formed the treated HT group, irrespective of their 24-h BP readings. Those who had no history of HT, were not receiving any medications, and

TABLE 1 PATIENT CHARACTERISTICS FOR UNTREATED AND TREATED HYPERTENSIVES AND FOR NORMOTENSIVES*

Total patients, n Age, yr Male:female BMI, kg/m2 Alcohol, g/d Mean 24-h systolic BP Mean 24-h diastolic BP Daytime systolic BP Daytime diastolic BP Nighttime systolic BP Nighttime diastolic BP

Untreated HT

Treated HT

NT

34 58.0 (2.1) 33:1 28.7 (9.6) 18.4 (4.8) 151.2 (1.9) 91.1 (1.7) 155.7 (2.2) 92.2 (1.8) 132.1 (2.2) 81.9 (1.8)

34 60.9 (2.0) 26:8 28.9 (0.8) 8.1 (2.7) 139.3 (2.9) 83.2 (2.5) 140.5 (3.2) 83.1 (2.3) 128.2 (3.6) 76.0 (2.7)

25 54.6 (2.6) 22:3 25.9 (0.4) 11.4 (3.2) 119.0 (2.1) 72.8 (1.2) 121.0 (2.2) 74.5 (1.3) 108.6 (2.1) 64.8 (1.4)

Definition of abbreviations: HT 5 hypertensives; NT 5 normotensives; BMI 5 body mass index; BP 5 blood pressure. * Values are means with SEM shown in parentheses.

113

Worsnop, Naughton, Barter, et al.: Sleep Apnea in Hypertensives

to be middle-aged men as they were recruited in a Veterans’ hospital and from service clubs. As expected, the HT groups had higher BP readings than did the NTs. The untreated HTs consumed more alcohol than the other two groups; there was no apparent reason for this. The descriptive data for the various indices of OSA calculated are shown in Table 2. In both HT groups, the proportion of subjects with AHI . 5 was significantly greater than in the NT group (p 5 0.005, x2 analysis). The proportions with AHI . 10 (p 5 0.003) and AHI . 15 (treated, p 5 0.04; untreated, p 5 0.009) were also significantly greater in each HT group than in the NT group (x2 ). Using REM-AHI increased the proportion in each group that could be considered to have OSA, and the proportion with REM-AHI . 5 was greater in the untreated HT group than in the NT group (p 5 0.03), but not in the treated HT group (p 5 0.09, x2). Two thirds of the HT subjects slept on their backs, and in those, AHI indices were no greater when supine than when in other sleeping positions. The oxygen saturation data showed that most subjects did not have profound desaturations. To allow for the differences in age, BMI, sex, and alcohol consumption between the groups, logistic regression was used with AHI . 5 as the dichotomous outcome variable. Comparisons between untreated HTs and NTs, and between treated HTs and NTs showed that BMI (p 5 0.001), age (p 5 0.07), sex (p 5 0.07), treated HT (p 5 0.05) and untreated HT (p 5 0.06) were associated with OSA, but that alcohol consumption (p 5 0.82) was not. If the dependent variable was AHI . 10, the association with HT was seen again (p 5 0.08 for both treated and untreated HT); and if the dependent variable was AHI . 15 the association was again present (p 5 0.08 for both treated and untreated HT). Using the mean nighttime BP or mean daytime BP to classify HT-NT status rather than mean 24-h BP made no difference in the results of the analysis. The odds ratios showed that untreated HTs were eight (95% CI, 0.9 to 72) times more likely, and treated HTs were nine (95% CI, 1.0 to 85) times more likely to have AHI . 5 than NTs when the confounding variables were taken into account. When the relationship was reexamined using BP as a continuous variable and with AHI . 5 as the dependent variable and all confounders included, the association between OSA and BP persisted for both mean 24-h systolic (p 5 0.07) and diastolic BP (p 5 0.05). This relationship was present for daytime BP (systolic, p 5 0.06; diastolic, p 5 0.02), but not for nocturnal BP (systolic, p 5 0.23; diastolic, p 5 0.28). The odds ratios for these variables, based on a difference in BP of 30 mm Hg, are shown in Table 3. They indicate that someone with a sys-

TABLE 2

TABLE 3 THE ODDS RATIOS AND 95% CONFIDENCE INTERVALS FOR HAVING AN AHI GREATER THAN 5 IF THE 24-h SYSTOLIC BP, 24-h DIASTOLIC BP, DAYTIME DIASTOLIC BP, NIGHTTIME SYSTOLIC BP, AND NIGHTTIME DIASTOLIC BP ARE 30 mm Hg GREATER*

Systolic BP over 24 h Diastolic BP over 24 h Daytime systolic BP Daytime diastolic BP Nighttime systolic BP Nighttime diastolic BP

Odds Ratios

95% Confidence Intervals

2.4 3.5 2.2 4.8 1.8 2.0

0.9–6.0 1.0–12.5 1.0–5.3 1.3–18.5 0.7–4.6 0.6–6.8

* These are derived from a logistic regression taking into account the effects of BMI, age, sex, and alcohol consumption.

tolic BP 30 mm Hg higher than someone else would be nearly 2.5 times more likely to have an AHI greater than 5, and if the diastolic BP was 30 mm Hg higher he or she would be 3.5 times more likely to have an AHI greater than 5. The other indices for OSA did not show significant relationships with HT status when multiple logistic regression (including the confounding variables) was used: REM-AHI (p 5 0.78), back AHI (p 5 0.98), lowsat (p 5 0.65), avdesat (p 5 0.25). When BP was used as a continuous variable, multiple logistic regression (including the confounders) showed a relationship between lowsat and mean 24-h BP (systolic, p 5 0.05; diastolic, p 5 0.05) and between lowsat and daytime BP (systolic p 5 0.04), (diastolic, p 5 0.03), but not between lowsat and nighttime BP (systolic p 5 0.20, diastolic, p 5 0.19). The percentages of subjects in each BP category responding to the snoring question are shown in Table 4. Chi-squared analysis showed that the untreated HTs were more likely to report snoring (p 5 0.04). However, logistic regression (including the confounders) showed that this relationship was mainly due to BMI (p 5 0.04) with untreated HTs no longer having a significant relationship with snoring (p 5 0.30).

DISCUSSION Although limited by small sample size, this study has demonstrated an association between OSA and HT. More specifically it has shown that treated or untreated HTs are more likely to have AHI . 5 than are NTs. BMI is a major confounding variable in this association, but age and sex also contribute. The association between OSA and HT persists when allowance is made for the confounding variables. A detailed examination of the data demonstrated that the BP level itself is quantitatively involved in the relationship. Also daytime BP

THE SLEEP STUDY RESULTS FOR THE HT AND THE NT GROUPS

AHI* REM-AHI* Back AHI* Av. desat.* Low sat.* Number (%) with AHI . 5 Number (%) with AHI . 10 Number (%) with AHI . 15

Untreated NT

Treated HT

NT

TABLE 4

8.3 (2.5) 16.5 (3.6) 6.8 (3.9) 91.4 (0.8) 86.0 (1.4) 13 (38)† 8 (23)† 5 (15)†

9.0 (2.3) 12.4 (3.0) 11.8 (4.9) 90.3 (0.8) 87.3 (1.2) 13 (38)† 8 (23)† 8 (23)†

1.2 (0.3) 4.0 (1.1) 3.7 (1.4) 92.5 (0.5) 90.7 (0.5) 1 (4) 0 0

THE PERCENTAGE OF UNTREATED AND TREATED HYPERTENSIVES AND NORMOTENSIVES, MAKING EACH RESPONSE CHOICE TO THE QUESTION ABOUT SNORING

Definition of abbreviations: AHI 5 apnea-hypopnea index; REM-AHI 5 AHI during REM sleep only; Back AHI 5 AHI while sleeping supine; Av. desat. 5 average arterial oxygen desaturation per apnea/hypopnea; Low sat. 5 lowest arterial oxygen saturation for the night. For other definitions, see Table 1. * Values are means with SEM shown in parentheses. † p , 0.05 (χ2) compared with NT.

Snoring unknown Never snores Snores rarely Snores sometimes Snores often Only snores after alcohol

Untreated HT (n 5 34)

Treated HT (n 5 34)

NT (n 5 25)

0 9 3 41 44 3

18 21 15 26 21 0

4 24 12 36 24 0

For definition of abbreviations, see Table 1.

114

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

is more important than nighttime BP. The HTs in this study had lower nighttime BP than daytime BP, a phenomenon that has been previously described (25). There appears to be no advantage in using REM-AHI, back AHI, lowsat, or avdesat over using AHI when considering group data in this relationship between OSA and HT. Others have also found that there is no difference in lowsat between obese HTs and NTs (19) and that lowsat is not an independent predictor of BP because it has a high correlation with BMI . 25. However, in an individual patient, it may still be important to look beyond the AHI to gain a more complete picture of his or her OSA. The important features of this study are, first, the subjects were not biased in any way towards having sleep-related problems. Studies using subjects referred to a sleep clinic have had a potential bias of overestimating the prevalence of OSA. This study avoided that problem. This would explain why the prevalence of OSA in our NT group was lower than that reported in other studies. Second, in this study allowance was made for possible confounding variables so that their contribution to the relationship could be evaluated. Third, measurement of BP over 24 h allowed for a more reliable diagnosis of HT (26– 28). Fourth, assessment of OSA was thorough so that its estimated prevalence was more accurate. Fifth, a control group was used so that any peculiar biases in the acquisition of the data could be seen in the control group as well as in the HT groups. Separating the HTs into treated and untreated groups addressed the issue of whether anti-HT medication itself may predispose to OSA. As both HT groups had the same prevalence of OSA, anti-HT medication cannot explain the association. There is a concern that 24-h ambulatory BP monitoring may artificially elevate BP, particularly at night, by producing arousals. Hla and coworkers (4) found that during sleep the Accutracker II produces arousals and a BP rise for as long as 10 s after 60% of cuff inflation. However, Davies and coworkers (29) found that automatic cuff inflation produced arousal and BP elevation for as long as 50 s, although it was not made clear how frequently this happened. If BP was artificially elevated by the ambulatory monitor, then it may mean that some of the untreated HTs were in fact NTs. If this was the case the relationship between HT and OSA may have been artificially weakened in this study. Nighttime BP was lower than daytime BP in the three groups in this study, so that if arousals were having an influence on the nighttime BP is was not to a great degree. The results of this study have indicated that the relationship between HT and OSA is partly due to confounding by the common risk factors, BMI, age, and sex and partly due to an association independent of these risk factors. There is insufficient evidence at present to attribute causality to the association, but studies with intra-arterial measurement of BP (11) have shown repetitive acute rises in BP with each apnea in patients with OSA. This study complements the study of Hla and coworkers (4), which examined the relationship from the reverse direction by assessing the prevalence of HT among nonsnorers, snorers, and apneics selected from a general population. The relationship persisted when allowance was made for the confounding variables BMI, age, and sex. Grunstein and coworkers (30) also found that in subjects with OSA, AHI was a determinant of BP independent of obesity. Carlson and coworkers (6) found that OSA, age, and obesity were independent risk factors for HT. An increased prevalence of OSA in HTs compared with that in NTs matched for age and percentage of ideal body weight has also been found. Defining OSA using

VOL 157

1998

the number of desaturations per hour, Gleadhill and coworkers (20) found that after allowing for BMI, a group of treated HTs had 1.9 times the desaturation index compared with a NT group. Even though an association between OSA and HT has been demonstrated to be largely (although not completely) due to BMI, further conclusions about cause and effect cannot be drawn from this type of study. It is known that obesity can lead to both OSA (1, 24) and HT (31), and this has been suggested as an adequate explanation of the OSA-HT link. Weight loss has also been shown to reduce HT and OSA in obese subjects (9, 19). Our data do not support this as there is some relationship persisting when allowance is made for BMI. Even when the relationship is mainly due to elevated BMI, it is still possible that there is a direct causal link between OSA and HT in overweight subjects. A number of studies have achieved a reduction in BP by relieving OSA with nasal continuous positive airway pressure (32, 33), suggesting that there is a direct causal link. It is also possible that obesity may predispose to HT, secondarily causing OSA, or obesity may predispose to OSA, secondarily causing HT. There is evidence although not conclusive, to support both possibilities (29). It has been postulated that in essential HT there is increased drive from arterial chemoreceptors, which produce ventilatory instability and sleep apnea. Repetitive apneas can cause a sustained rise in BP through the night (10, 11), but it is not clear if this is a cause of HT. Another issue this study has helped to address is whether the relationship between OSA and HT is due to the BP level itself or to the disease HT. If the level of BP is important it would be expected that at least nighttime BP would be associated with OSA, but in this study the reverse was found, with daytime BP being associated with OSA but not nighttime BP. Thus, having the disease HT seems to be the factor related to OSA rather than just an elevated BP. Hoffstein and Mateika (34) found that in nonapneic snorers and in those with mild OSA the evening BP was higher than the morning BP, which is consistent with our data, although in severe OSA the morning BP was higher than the evening BP, and the difference was eliminated when subjects were matched for age and weight. A relationship between snoring and HT was not convincingly demonstrated in this study, and this is not unexpected (4, 19) as it is recognized that subjects’ self-reports of snoring are not particularly accurate. As an objective measurement of snoring can be difficult to obtain and as this was not a primary aim of the study, the issue was not pursued further. It is possible that there is not a true relationship between simple snoring (without OSA) and HT, thus weakening the association between snoring overall and HT, although it has been shown in a large sample that there is an increased risk of snoring in HTs compared with that in NTs after adjusting for age and BMI (27). In summary, this study has shown that there is a relationship between OSA and HT that, although partially explained by the confounding variables BMI, age, and sex, persists when these are controlled for. Acknowledgment : The writers appreciate the assistance given by Denise Bertram and Olive Morgan with the ambulatory blood pressure monitoring, and also the statistical advice provided by Ted Byrt of the Biostatistics Department of the Royal Children’s Hospital, Melbourne.

References 1. Young, T., M. Palta, J. Dempsey, J. Skatrud, S. Weber, and S. Badr. 1993. The occurrence of sleep-disordered breathing among middleaged adults. N. Engl. J. Med. 328:1230–1235. 2. Partinen, M., and C. Guilleminault. 1990. Daytime sleepiness and vascu-

Worsnop, Naughton, Barter, et al.: Sleep Apnea in Hypertensives

3.

4.

5. 6.

7. 8.

9.

10.

11.

12.

13.

14.

15.

16.

17. 18.

lar morbidity at seven-year follow-up in obstructive sleep apnea patients. Chest 97:27–32. Rossner, S., L. Lagerstrand, H. E. Persson, and C. Sachs. 1991. The sleep apnoea syndrome in obesity: risk of sudden death. J. Intern. Med. 230: 135–141. Hla, K. M., T. B. Young, T. Bidwell, M. Palta, J. B. Skatrud, and J. Dempsey. 1994. Sleep apnea and hypertension: a population-based study. Ann. Intern. Med. 120:382–388. Rauscher, H., W. Popp, and H. Zwick. 1992. Systemic hypertension in snorers with and without sleep apnea. Chest 102:367–371. Carlson, J. T., J. A. Hedner, H. Ejnell, and L. Peterson. 1994. High prevalence of hypertension in sleep apnea patients independent of obesity. Am. J. Respir. Crit. Care Med. 150:72–77. Mendelson, W. B. 1992. Sleepiness and hypertension in obstructive sleep apnea. Chest 101:903–909. Millman, R. P., S. Redline, C. C. Carlisle, A. R. Assaf, and P. D. Levinson. 1991. Daytime hypertension in obstructive sleep apnea. Chest 99:861–866. Rajala, R., M. Partinen, T. Sane, R. Pelkonen, K. Huikuri, and A. M. Seppalainen. 1991. Obstructive sleep apnoea syndrome in morbidly obese patients. J. Intern. Med. 230:125–129. Stoohs, R., and C. Guilleminault. 1992. Cardiovascular changes associated with obstructive sleep apnea syndrome. J. Appl. Physiol. 72:583– 589. Tilkian, A. G., C. Guilleminault, J. S. Schroeder, K. L. Lehrman, F. B. Simmons, and W. C. Dement. 1976. Hemodynamics in sleep-induced apnea. Ann. Intern. Med. 85:714–719. Lavie, P., R. Ben-Yosef, and A. E. Rubin. 1984. Prevalence of sleep apnea syndrome among patients with essential hypertension. Am. Heart J. 108:373–376. Kales, A., E. O. Bixler, R. J. Cadieux, D. W. Schneck, L. C. Shaw, T. W. Locke, A. Vela-Bueno, and C. R. Soldatos. 1984. Sleep apnoea in a hypertensive population. Lancet ii:1005–1008. Fletcher, E. C., R. D. DeBehnke, M. S. Lovoi, and A. B. Gorin. 1985. Undiagnosed sleep apnea in patients with essential hypertension. Ann. Intern. Med. 103:190–195. Williams, A. J., D. Houston, S. Finberg, C. Lam, J. L. Kinney, and S. Santiago. 1985. Sleep apnea syndrome and essential hypertension. Am. J. Cardiol. 55:1019–1022. Hirschowitz, M., I. Karacan, A. Gurakar, and R. L. Williams. 1989. Hypertension, erectile dysfunction, and occult sleep apnea. Sleep 12:223– 232. Escourou, P., A. Jirani, H. Nedelcoux, P. Duroux, and C. Gaultier. 1990. Systemic hypertension in sleep apnea syndrome. Chest 98:1362–1365. Isaksson, H., and E. Svanborg. 1991. Obstructive sleep apnea syndrome in male hypertensives, refractory to drug therapy. Nocturnal automatic blood pressure measurements: an aid to diagnosis? Clin. Exp. Hypertens. A. 6&7:1195–1212.

115 19. Kiselak, J., M. Clark, V. Pera, C. Rosenberg, and S. Redline. 1993. The association between hypertension and sleep apnea in obese patients. Chest 103:775–780. 20. Gleadhill, I. C., E. E. McCrum, C. C. Patterson, G. D. Johnston, A. Evans, and J. MacMahon. 1993. Sleep related hypoxaemia in hypertensive and normotensive men. Thorax 48:534–536. 21. Sobell, L. C., S. A. Maisto, M. B. Sobell, and A. M. Cooper. 1979. Reliability of alcohol abusers’ self-reports of drinking behaviour. Behav. Res. Ther. 17:157–160. 22. Martin, R., A. J. Block, M. A. Cohn, W. A. Conway, D. W. Hudgel, A. C. P. Powles, M. H. Sanders, and P. L. Smith. 1985. Indications and standards for cardiopulmonary sleep studies. Sleep 8:371–379. 23. Rechtschaffen, A., and A. Kales, editors. 1968. A Manual of Standardized Terminology, Techniques and Snoring System for Sleep Stages of Human Subjects. Brain Information Service/Brain Research Institute, UCLA, Los Angeles, CA. 24. Stradling, J. R., and J. H. Crosby. 1991. Predictors and prevalence of obstructive sleep apnoea and snoring in 1001 middle aged men. Thorax 46:85–90. 25. Wilcox, I., R. R. Grunstein, F. Collins, J. Doyle, D. Kelly, and C. Sullivan. 1992. 24-hour blood pressure in patients with obstructive sleep apnoea: “dippers” vs “nondippers.” J. Sleep Res. 1(S):254. 26. Weber, M. A. 1988. Whole-day blood pressure. Hypertension 11:288– 298. 27. Irving, J. B., H. M. Brash, F. Kerr, and B. J. Kirby. 1976. The value of ambulatory monitoring in borderline and established hypertension. Postgrad. Med. J. 52(Suppl. 7):137–139. 28. Perloff, D., M. Sokolow, and R. Cowan. 1983. The prognostic value of ambulatory blood pressures. J.A.M.A. 249:2792–2798. 29. Davies, R. J. O., P. J. Belt, S. J. Roberts, N. J. Ali, and J. R. Stradling. 1993. Arterial blood pressure responses to graded transient arousal from sleep in normal humans. J. Appl. Physiol. 74:1123–1130. 30. Grunstein, R. R., I. Wilcox, T. S. Yang, J. A. Hedner, Y. Gould, and M. J. Dodd. 1992. Sleep apnoea: a confounding factor in the relationship between pattern of obesity and systemic hypertension? J. Sleep Res. 1(S):87. 31. Havlik, R. J., H. B. Hubert, R. R. Fabsitz, and M. Feinleib. 1983. Weight and hypertension. Ann. Intern. Med. 98:855–859. 32. Wilcox, I., R. R. Grunstein, J. A. Hedner, J. Doyle, F. L. Collins, P. J. Fletcher, D. T. Kelly, and C. E. Sullivan. 1993. Effect of nasal continuous positive airway pressure during sleep on 24-hour blood pressure in obstructive sleep apnea. Sleep 16:539–544. 33. Naughton, M., and R. J. Pierce. 1991. Effects of continuous positive airway pressure on blood pressure and body mass index in obstructive sleep apnoea. Aust. N.Z. J. Med. 21:917–919. 34. Hoffstein, V., and J. Mateika. 1992. Evening-to-morning blood pressure variations in snoring patients with and without obstructive sleep apnea. Chest 101:379–384.