Hyperlipoproteinemia

Study objectives: To determine whether traditional risk factors for cardiovascular disease (CVD) and regional fat distribution, especially the central obesity type and increased parapharyngeal fat pads, are associated with the degree of obstructive sleep apnea (OSA). To determine whether there are interrelationships between body fat, serum leptin levels, and the degree of OSA.

Design and setting: Prospective mono-center cross-sectional study in a university hospital in Germany.

Patients: Eighty-five consecutive male patients who were referred for evaluation of suspected OSA.

Measurements and results: The major dependent outcome variable was the apnea-hypopnea index (AHI), the average number of apneas and hypopneas per hour of sleep, determined by overnight polysomnography. Independent measures were anthropometric data, body composition analysis (bioelectrical impedance analysis [BIA]), cardiovascular risk factor evaluation (smoking, hypertension, serum lipoproteins, diabetes or impaired glucose tolerance, uric acid, fibrinogen), and leptin. Adipose tissue quantification of the abdominal and neck regions was performed by nuclear MRI (NMR). Significant linear relationships of AHI with fasting blood glucose, uric acid, fibrinogen, body weight, body mass index (BMI), sum of fat skin folds, and percentage of body fat could be established, whereas there was no correlation with age. The presence of OSA was independent of smoking, hypertension, and lipoproteins. NMR scans showed that AHI was significantly correlated with intra-abdominal fat and subcutaneous abdominal fat, whereas subcutaneous fat in the neck region and parapharyngeal fat in the airway vicinity were not correlated. Leptin concentrations correlated with AHI and with biochemical markers of the metabolic syndrome (lipoproteins, glucose) but were not dependent on AHI. Logistic regression analysis found percentage of body fat (BIA) and BMI as good predictors of AHI > 10 with a sensitivity of 95.5% but a low specificity (46.2%). Multiple regression analysis identified the sum of fat skin folds, body weight, and BMI as good predictors for the degree of OSA.

Conclusions: We conclude that OSA is independent from most traditional risk factors for CVD. Regional body fat distribution predicts the presence and degree of OSA, but fat accumulation in the neck and parapharyngeal region are of minor importance. Leptin concentrations when controlled for body fat are not related to the degree of OSA.

Key words: body fat; cardiovascular risk factors; leptin; nuclear MRI; obesity; obstructive sleep apnea; polysomnography; regional fat distribution

Abbreviations: AHI = apnea-hypopnea index; BIA = bioelectrical impedance analysis; BMI = body mass index; CI = confidence interval; CVD = cardiovascular disease; FIRI = fasting insulin resistance index; HDL = high-density lipoprotein; LDL = low-density lipoprotein; nCPAP = nasal continuous positive airway pressure; NMR = nuclear MRI; OSA = obstructive sleep apnea; VLDL = very low-density lipoprotein

**********

Obstructive sleep apnea (OSA) is characterized by repeated collapse of the pharynx during sleep, which leads to oxygen desaturation, fragmentation of sleep, and often daytime sleepiness, (1) the latter leading to an increased risk of traffic accidents. (2) A high prevalence of the condition, affecting approximately 2 to 4% of middle-aged adults, was found in an epidemiologic study. (3) Retrospective studies (4,5) indicate that there is an association of OSA with morbidity and mortality caused by cardiovascular and cerebrovascular causes, but the precise effects of OSA on health are still under debate. (6,7) The association of OSA with commonly known cardiovascular risk factors remains uncertain. It had been suggested that the morbidity associated with OSA was particularly caused by obesity. (8) Obesity and age were found to be strong predictors for OSA. (9) An increase in body mass index (BMI) of 1 SD is associated with a fourfold increase in risk for OSA (defined as an apnea-hypopnea index [AHI] > 5). (3) The fact that neck circumference was the most powerful predictor of OSA among all anthropometric variables studied suggests that the upper body or central obesity, rather than a more generalized distribution of body fat, may be important for the development of OSA. (3) It is presumed that increased fat deposition in the neck region or adjacent to the upper airway is responsible for the development of OSA in obese subjects, (10,11) but increased fat deposition next to the upper airway can also be found in nonobese subjects with OSA. (12)

Obesity in general is a major cause of morbidity and mortality through the development of insulin resistance, dyslipoproteinemia, and hypertension leading to cardiovascular disease (CVD). A clear relation between body weight and development of CVD has been established in epidemiologic studies. (13,14) Moreover, it has become clear that especially the central fat distribution is associated with increased risk for CVD. (15,16) Numerous hormones modulate fat storage through their effects on energy balance. Leptin is a circulating hormone synthesized and expressed in adipose tissue and a potential afferent signal of body fat stores. (17) Plasma leptin concentrations are increased in people who are obese in direct proportion to body fat mass. (18) The question remains whether the regional fat distribution has an impact on OSA and is linked with leptin levels. Thus, the purpose of this study was to examine prospectively a series of patients with suspected OSA and determine the relation to traditional cardiovascular risk factors, and in particular, to analyze the relation between body fat distribution, circulating leptin levels, and sleep-related breathing disorder.

MATERIALS AND METHODS

Study Design

The trial was designed as a cross-sectional study. All patients gave written informed consent to participate in the trial. The study protocol was approved by the university ethics committee, and the study was performed in accordance with the guidelines of the Declaration of Helsinki and its current revision.

Patients

Eighty-five consecutive male patients who had been referred to our hospital by general practitioners or pulmonologists for evaluation of suspected OSA were included in the study between June 1997 and May 1998. Three patients were later excluded from analysis because central apnea was diagnosed; one patient was excluded because a growth hormone-secreting microadenoma had been detected and related metabolic disorders could not be excluded. The remaining 81 patients (mean age, 55 years; range, 27 to 75 years) could be evaluated. Twelve patients (15%) had a history of coronary heart disease, 3 patients (4%) had a history of cerebrovascular ischemic insults, 42 patients (52%) had a history of hypertension, 8 patients (10%) had a history of diabetes, 9 patients (11%) had a history of thyroid disease, and 35 patients (43%) had a history of hyperuricemia. Twenty-eight patients (35%) had excessive daytime sleepiness.

Study Procedures

An overnight polysomnography in the sleep laboratory was performed in all 81 patients included in the evaluation of the study. Blood was drawn at 8 AM after an overnight fast before and 120 min after an oral glucose tolerance test for the determination of multiple clinical chemistry parameters (at 120 min only glucose, insulin, and C-peptide). Anthropometric data were measured and recorded, a bioelectrical impedance analysis (BIA) was performed, and nuclear MRI (NMR) of the neck and the abdomen regions were obtained. Complete NMR scans could be evaluated in only 63 patients (78%; see "Discussion").

Determination of Risk Factors and Laboratory Investigations

The prevalent cardiovascular risk factors and concomitant diseases were recorded in all patients according to the following criteria:

Cigarette Smoking: Patients were classified into three groups: those who currently smoked or had quit smoking < 1 year before the study were considered smokers; nonsmokers were subclassified into ex-smokers who had quit > 1 year before and individuals who had never smoked. In smokers and ex-smokers, the number of pack-years was calculated.

Hypertension: A previous diagnosis of hypertension was recorded. Systolic BP [greater than or equal to] 160 mm Hg and/or diastolic BP [greater than or equal to] 95 mm Hg or known hypertension treated with antihypertensive medication was considered positive. Mild hypertension was assumed if systolic BP was [greater than or equal to] 140 mm Hg and/or diastolic BP was [greater than or equal to] 90 mm Hg.

Hyperlipoproteinemia: Cholesterol and triglyceride concentrations in serum and lipoprotein fractions were determined enzymatically using standard laboratory procedures (Lipid Research Clinic method). For analysis of the lipoprotein risk profile, low-density lipoprotein (LDL)-C was categorized as follows: LDL-C > 130 mg/dL indicated mild risk, > 160 mg/dL indicated moderate risk, and > 190 mg/dL indicated severe risk. Further subgroups were defined as patients with high-density lipoprotein (HDL)-C < 30 mg/dL and triglycerides > 150 mg/dL.

Diabetes Mellitus/Glucose Tolerance: The diagnosis of diabetes mellitus was established according to the World Health Organization criteria as fasting glucose of > 7.8 mmol/L. (19,20) Impaired glucose tolerance was assumed when a 2-h postchallenge (75 g glucose) plasma glucose was > 11.1 mmol/L. The empirical fasting insulin resistance index (FIRI; fasting glucose x fasting insulin/25) was used to estimate insulin sensitivity. (21)

Obesity: The classification of obesity was performed according to Bray (22): BMI 20 to 25, class 0; BMI 25 to 30, class I; BMI 30 to 35, class 2; BMI 35 to 40, class 3; and BMI > 40, class IV.

Hyperuricemia: Uric acid was determined in serum and elevations were defined as concentrations > 6.8 mg/dL (mild) or > 7.5 mg/dL (moderate).

Fibrinogen: Fibrinogen was determined in plasma and risk classes were defined as values > 300 mg/dL, > 350 mg/dL, or > 400 mg/dL.

Lipoprotein(a): Lipoprotein(a) was determined in serum and elevation was defined as levels > 30 mg/dL.

Leptin: Leptin was determined in plasma using the DSL-10-23100 ACTIVE Human Leptin ELISA-kit (Diagnostic System Laboratories; Webster, TX). The sensitivity of this assay was 0.5 ng/mL, and the interassay coefficient of variation was 4.6%.

Sleep Studies

Polysomnography included the following signals: two leads of EEG ([C.sub.4][A.sub.1]; [C.sub.3][A.sub.2]), two leads of electro-oculogram, and a submental electromyogram were continuously registered by surface electrodes. A one-channel ECG was registered continuously. Leg movements were detected by an anterior tibialis electromyogram. Breathing sounds were monitored by a microphone placed at the jugular vein. Respiration was continuously recorded: airflow was monitored by combined oronasal thermistors, abdominal and chest wall movements were monitored by inductive plethysmography, and arterial oxygen saturation was measured by pulse oximetry. All variables were recorded by a computerized system approved for OSA diagnostics (NewMedics; Oehringen, Germany). Sleep stages were determined manually according to standard criteria. (23) Central, obstructive, and mixed apneas were defined according to the usual criteria. (24) We defined hypopnea as a > 50% reduction in airflow from its value during quiet wakefulness for at least 10 s followed by either oxygen desaturation of at least 4 percentage points or by an episode of arousal. The total number of apnea and hypopnea episodes were divided by total sleep time to calculate the apnea index and the hypopnea index. The sum of the apnea index and the hypopnea index then became the apnea-hypopnea index (AHI), which was used as the primary study parameter. Sleep apnea was considered to be present if AHI was > 10. OSA was classified as mild with AHI > 10 and [less than or equal to] 20, moderate with AHI > 20 and [less than or equal to] 40, and severe with AHI > 40.

Anthropometry and BIA

Anthropometric measures were performed using standard procedures. Broca index, as a measure of obesity, was defined as height (in centimeters)--weight (in kilograms), normal value < 100. Skin fat folds were measured twice on the patient's right side using a Holtain caliper (Crymych; Dyfed, UK). The percentage of body fat and fat-free mass were calculated on the basis of four skinfold thickness measurements. (25)

BIA was performed using a multifrequency, phase-sensitive device (Model 2000-M; Data Input; Frankfurt, Germany). The data were analyzed using the manufacturer's software (Nutri 4; Data Input).

NMR

NMR was performed on a 1.5 T scanner (Philips ACS S15; Philips Medizinsysteme; Hamburg, Germany) using a conventional head-neck coil and body coil for transmission and reception of the signal. Transversal T1 weighted spin-echo sequences were performed for quantification of the fat distribution of the neck and abdomen. The imaging protocol included the following sequences with scan parameters: (1) transversal T1 weighted spin-echo sequence of the neck parallel to the hard palate (field of view, 220; rectangular field of view, 90%; time of repetition, 426 ms; time of echo, 15 ms; numbers of acquisition, 4; matrix 128 x 256; slice thickness, 5 mm/gap 0.5 mm 15 slices; scan time, 3.22 min); (2) transversal T1 weighted spin-echo sequence of the abdomen centered at the level of lumbar vertebra L2/L3 (field of view, 500; rectangular field of view, 100%; time of repetition, 300 ms; time of echo, 20 ms; numbers of acquisition, 2; matrix 128 x 256; slice thickness, 10 mm/gap 1.0 mm, 8 slices; scan time, 1.27 min).

Image analysis for quantification of fat distribution of the neck and abdomen was performed by computerized seeded-region of interest analysis using the standard software on a separate work station (EasyVision 4; Philips Medizinsysteme). Fat volume was measured in each slice by mapping the adipose tissue compartments. The fat tissue of the neck region was distinguished and separated into parapharyngeal and subcutaneous fat compartments. Abdominal fat was separated into intra-abdominal and subcutaneous fat tissue. A pixel-by-pixel analysis of all slices was performed, and the number of pixels containing fat counted in each compartment was converted into a volume to quantify the proportions of the fat compartments.

In 21 of 81 patients, NMR could not be performed, mostly due to known claustrophobia. Thus, a total of 60 patients participated in NMR analysis.

Data Analysis and Statistics

Continuous variables are expressed as mean values [+ or -] SD unless otherwise stated. The primary study parameter, the combined AHI, was used as the dependent variable in all analyses, either as a continuous variable or as a categorical variable as described above. Relations between continuous variables were examined by simple linear regression analyses. Not normally distributed data were log-transformed before analysis. Differences between the groups described by categorical variables were analyzed by [chi square] statistics to test for an association between conditions and the presence of OSA. Odds ratios and exact 95% confidence intervals (CIs) were determined for proportions. Multiple regression analysis and logistic analysis were used for the further examination of risk factors that were significantly associated with OSA. Additional analyses were performed with leptin as the dependent parameter. All reported p values were tested two-sided, and a p value of < 0.05 was considered to indicate statistical significance. All calculations were performed using StatView 5.0. (SAS Institute; Cary, NC) on a personal computer with Windows-95 (Microsoft; Redmond, WA).

RESULTS

Sleep Studies

Polysomnographic protocols could be evaluated in all patients who entered the study. Fifteen patients (19%) were found to have an AHI [less than or equal to] 5, 11 patients (14%) had an AHI > 5 and [less than or equal to] 10, 8 patients (10%) had an AHI > 10 and [less than or equal to] 15, 4 patients (5%) had an AHI > 15 and [less than or equal to] 20, 22 patients (27%) had an AHI > 20 and [less than or equal to] 40, and 21 patients (26%) had an AHI > 40. However, if an AHI > 10 is considered diagnostic for OSA, then approximately two thirds of the patients received positive diagnoses. The low r value between age and AHI indicates that there is no apparent relationship when examined as a linear relationship (Table 1).

Smoking

Twenty-nine individuals (36%) had never smoked and 29 individuals (36%) were ex-smokers, whereas 23 patients (28%) were currently smoking. Linear regression analysis between the number of pack-years and AHI was not significant (r = 0.11, p = 0.39; Table 1). [chi square] statistics did not show significance either, although the smoking status tended to be significantly associated with the presence of severe OSA (Table 2).

Arterial Hypertension

Forty-two patients (52%) had a history of arterial hypertension. Systolic BP [greater than or equal to] 160 mm Hg was measured in 7 patients (9%) and diastolic BP [greater than or equal to] 95 mm Hg was measured in 13 patients (16%). Mild hypertension (systolic BP [greater than or equal to] 140 mm Hg, diastolic BP [greater than or equal to] 90 mm Hg) was determined in 41 patients (51%) and 28 patients (35%), respectively. Forty-four patients (54%) had either previously known hypertension or currently elevated BP values. Linear regression analysis between systolic or diastolic BP and AHI was not significant (Table 1). [chi square] statistics between the presence of hypertension (or mild hypertension) and OSA did not show any significance either (Table 2).

Lipoproteins

Sixty-six patients (82%) had LDL cholesterol concentrations [greater than or equal to] 130 mg/dL, 39 patients (48%) had LDL-C [greater than or equal to] 160 mg/dL, and 16 patients (20%) had LDL-C [greater than or equal to] 190 mg/dL. Twelve patients (15%) had low HDL-C levels (< 30 mg/dL), and 33 patients (41%) had elevated triglycerides (> 150 mg/dL). There was no correlation between these data and AHI in linear regression analyses (Table 1) or [chi square] statistics (Table 2). Analyses were not different when patients with lipid lowering medications (n = 8) were excluded (data not shown). The mean ([+ or -] SD) concentration of lipoprotein(a) was 19.1 [+ or -] 26.4 mg/dL. Neither linear regression or [chi square] analysis revealed any significant relations with OSA (Tables 1, 2).

Glucose Metabolism

The mean ([+ or -] SD) fasting or postchallenge blood glucose levels were 89 [+ or -] 22 mg/dL or 109 [+ or -] 41 mg/dL, insulin levels were 14 [+ or -] 8 mIU/L or 62 [+ or -] 45 mIU/L, and C-peptide levels were 4.2 [+ or -] 1.5 mIU/L or 11.8 [+ or -] 3.9 mIU/L, respectively. FIRI was calculated as 2.9 [+ or -] 1.9, indicating a high degree of insulin resistance in the group as a whole. Eight patients (10%) had a history of diabetes, but only two patients received oral antidiabetic medication. None of the patients were treated with insulin. Four patients (5%) had elevated fasting blood glucose levels (> 7.0 mmol/L), and two patients had values > 7.8 mmol/L. All of these patients were among the group with previously known diabetes. Five patients (6%) had a pathologic glucose tolerance test finding, all of whom were also among the aforementioned group. Linear regression analysis between parameters describing glucose metabolism (Table 1) revealed a weak but significant positive correlation only between fasting blood glucose (r = 0.24, p = 0.029) and a tendency toward significance for FIRI (r = 0.2, p = 0.08). [chi square] statistics did not show any significant correlations between the presence of diabetes or impaired glucose tolerance and OSA (Table 2).

Hyperuricemia

Uric acid concentrations were 6.4 [+ or -] 1.2 mg/dL in the whole cohort after exclusion of all patients receiving uric acid-lowering medications. There were 36 patients with previously known elevations of uric acid. Patients with previously known hyperuricemia had significantly higher uric acid concentrations than persons with no history of hyperuricemia (7.6 [+ or -] 0.8 mg/dL vs 5.6 [+ or -] 0.7 mg/dL, analysis for variance). Linear regression showed a highly significant correlation between untreated uric acid concentrations and AHI (Table 1). The significance of this relation was independent of uric acid-lowering medications and use of diuretics (data not shown). Also [chi square] statistics showed a significantly increased likelihood of having OSA when uric acid concentrations are elevated (Table 2). Uric acid concentrations showed significant colinearities with a number of other parameters.

Fibrinogen

The mean ([+ or -] SD) fibrinogen concentration was 298 [+ or -] 64 mg/dL. There was a highly significant positive correlation between fibrinogen concentrations and AHI (Table 1). [chi square] statistics showed that having a fibrinogen concentration above the 50th percentile (300 mg/dL) tended to be associated with a higher incidence of mild OSA (AHI > 5) and was significantly associated with severe OSA (AHI > 40). Fibrinogen concentrations > 350 mg/dL were also associated with a higher incidence of severe OSA (AHI > 40; Table 2).

Anthropometric Measures

The patients' mean ([+ or -] SD) body weight was 95.2 [+ or -] 18.2 kg, height was 178 [+ or -] 6.6 cm, BMI was 30.0 [+ or -] 5.0, and Broca index was 122 [+ or -] 20. Their mean waist-to-hip ratio was 1.02 [+ or -] 0.07. Using the above-described obesity classification according to Bray, (22) 10 patients (12%) were in class 0, 38 patients (47%) were in class 1, 23 patients (28%) were in class 2, 5 patients (6%) were in class 3, and 5 patients (6%) were in class 4. Linear regression analysis revealed highly significant correlations between AHI and body weight, BMI, Broca index, sum of fat skin folds, and percentage of body fat (Table 1), whereas the relationship with the waist-to-hip ratio tended to be significant. [chi square] statistics showed an increased likelihood for AHI to be associated with all classes of obesity (Table 2).

Regional Body Fat Distribution

Intra-abdominal fat volume was 211 [+ or -] 106 [cm.sup.3] (range, 20 to 525 [cm.sup.3]); subcutaneous fat volume was 169 [+ or -] 83 [cm.sup.3] (range, 32 to 427 [cm.sup.3]). The subcutaneous fat volume in the neck region was 57 [+ or -] 21 [cm.sup.3] (range, 20 to 145 [cm.sup.3]), and the volume of the parapharyngeal fat pads was 6.3 [+ or -] 2.5 [cm.sup.3] (range, 2.1 to 16 [cm.sup.3]). All NMR parameters were positively skewed and had to be logarithmically transformed before analyses. To exclude major interobserver variability in NMR measurements of fat tissue volumes, we performed validation of the analyses between the two researchers who analyzed the data. Interobserver variability was 0.6 [+ or -] 0.6% in abdominal (subcutaneous and intra-abdominal) tissue, 1.2 [+ or -] 1.6% in neck subcutaneous tissue, and 4.4 [+ or -] 3.3% in parapharyngeal fat pads.

Linear regression analysis showed a significant correlation between the amount of intra-abdominal fat (r = 0.29, p = 0.025), whereas the correlation with subcutaneous abdominal fat and subcutaneous fat of the neck region was weaker and not significant (r = 0.23, p = 0.077; and r = 0.20, p = 0.12, respectively). Surprisingly, there was no association of AHI with the size of the parapharyngeal fat pads (r = 0.044, p = 0.73; Table 1).

Serum Leptin Concentrations

Serum leptin concentrations (log-transformed) were significantly correlated with AHI (r = 0.39, p = 0.0003). There were multiple colinearities between serum leptin concentrations and parameters describing body composition, MRI analyses, lipoproteins, parameters describing glucose metabolism, and other biochemical parameters (Table 3). To identify whether leptin levels and AHI are correlated with the effect of percentage of body fat removed, we performed partial correlation analysis. The correlation of leptin with AHI becomes weaker (r = 0.25), suggesting no independent influence of leptin. Significant correlations could be established with biochemical markers that are characteristic of the metabolic syndrome, ie, elevated very low-density lipoprotein (VLDL) cholesterol and triglycerides, total triglycerides, low HDL cholesterol, fasting and postchallenge glucose, insulin and C-peptide, and the resulting increase in FIRI.

To analyze whether subjects with similar BMI (or percentage of body fat) have higher leptin levels if they have higher degrees of OSA, we ranked all subjects according to BMI (or percentage of body fat) and allocated the subject with the respective higher AHI of neighboring pairs to a higher-AHI group, and subjects with a lower AHI to a lower-AHI group. Unpaired comparisons of the leptin levels between higher AHI and lower AHI did not reveal significant differences (BMI ranking, 8.5 [+ or -] 7.9 ng/mL vs 9.1 [+ or -] 9.2 ng/mL, p = 0.74; percentage of body fat ranking, 10.1 [+ or -] 10.3 ng/mL vs 7.6 [+ or -] 6.2 ng/mL, p = 0.19).

In stepwise multiple regression analysis with serum leptin levels as the dependent parameter, only percentage body fat, waist-to-hip ratio, and subcutaneous neck fat were predictors of serum leptin, accounting together for 74% of its variability. AHI was not found to be significant.

Multivariate Analyses With AHI as the Dependent Parameter

To identify parameters that would best predict the presence of OSA, defined as AHI > 10, we calculated a logistic regression model using parameters with significant correlations in univariate analyses: BMI, percentage of body fat, fasting blood glucose, fibrinogen, uric acid, leptin (log transformed), and NMR intra-abdominal fat (log transformed). The only significant predictors were percentage of body fat as determined by BIA (p = 0.008) and BMI (p = 0.046). These parameters were followed by fasting glucose (p = 0.074), leptin concentrations (p = 0.088), and intra-abdominal fat (p = 0.13). All other parameters did not allow significant predictions. The sensitivity was 95.5%; the specificity, however, was low (46.2%).

Using a multiple regression model with AHI as the dependent parameter and simple-to-determine measurements as independent parameters, a highly significant (p < 0.0001) prediction of AHI could be made. The following parameters contributed to 61% of the variability of AHI: sum of skin fat folds (p = 0.003), body weight (0.005), BMI (p = 0.013), fasting blood glucose (p = 0.58), percentage of body fat (BIA) [p = 0.68], and waist-to-hip ratio (p = 0.97).

DISCUSSION

In this study, we found a highly significant correlation between the degree of sleep-related breathing disorder (ie, combined apnea and hypopnea index) and anthropometric measures of obesity (ie, body weight and BMI) in patients presenting with suspected OSA. Moreover, regional body fat distribution analyzed by MRI showed significant correlations to the amount of intra-abdominal fat and borderline significant correlation with subcutaneous abdominal fat, whereas no correlation was found to parapharyngeal fat pads or subcutaneous fat of the neck region. Serum leptin concentrations were significantly correlated with AHI and, in addition, there were multiple colinearities between serum leptin concentrations and parameters describing body composition, lipoproteins, and glucose metabolism. Leptin levels were dependent on percentage of body fat, waist-to-hip ratio, and subcutaneous neck fat, but not on AHI.

According to a previous study, (26) the prevalence of OSA is increased fourfold in patients with obesity. Obesity plays a major part in the development of the metabolic syndrome, which consists of insulin resistance, diabetes or impaired glucose tolerance, hypertension, and dyslipoproteinemia. It has been recognized that the type of regional fat distribution (abdominal-visceral vs gluteal-femoral) plays an important role in the development of the metabolic syndrome. (27-29) Not only increased body weight but fat distribution plays a major role in the development of OSA. Visceral (central) obesity has been recognized to be associated more often with OSA than other forms of obesity. (30) A special role in the development of OSA is attributed to the fat deposition in the neck region, especially in the parapharyngeal region. An early study using NMR observed that six patients with OSA had a larger cross-sectional area of adipose tissue adjacent to the airway than five weight-matched control subjects. (11) In another study (10) with NMR in 21 male subjects with OSA and 9 subjects without OSA, the authors found that the volume of the adipose tissue adjacent to the upper airway correlated with the AHI. These results are in contrast to the findings of our study, in which we found a correlation between AHI and BMI as well as AHI and intra-abdominal fat but not between AHI and parapharyngeal fat or neck subcutaneous fat. The discrepancies may arise from different study populations with selection of subjects with a disparity in the relationship between obesity and adipose tissue volume or from small sample sizes in general. Parapharyngeal fat pad sizes show a large interindividual variability and the distribution is skewed. Thus, a large sample number is needed to determine true differences, and differences found in small samples may be caused by patient selection bias. In our study, which contains the largest study population in the literature with NMR scans, all patients were included consecutively. The differences to the first study cited are most likely caused by the method of fat volume quantification. We performed a computer-based pixel-by-pixel analysis of the whole parapharyngeal fat pad, which is a more precise technique than the evaluation of fat pad diameters. Our results do not exclude the possibility that in certain cases increased parapharyngeal fat may predispose to upper airway narrowing and OSA, because Mortimore et al (12) demonstrated that nonobese patients with OSA had substantially greater deposits of fat anterolaterally to the upper airway than BMI- and neck circumference-matched control subjects. Nevertheless, as our data indicate, the amount of intra-abdominal fat seems to be of much greater importance for the presence of OSA.

Adipose tissue secretes a hormonal substance into the circulation, recognized as leptin, which participates in the regulation of multiple metabolic pathways. (31) Leptin alters neuroendocrine functions as well as energy intake and expenditure by binding to specific receptors in the hypothalamus. Leptin levels are increased exponentially with increasing fat mass, and most obese humans have increased leptin levels, (32) indicating that obesity in most obese individuals is a leptin-resistant state. Leptin levels seem to be higher in OSA patients than in similarly obese control subjects without sleep apnea independent of body fat. (33) We found highly significant correlations of leptin levels with parameters describing whole body fat composition (BMI, percentage of body fat, skin folds) as well as regional fat distributions (NMR: subcutaneous fat of the neck region, parapharyngeal fat pads, subcutaneous, and intra-abdominal fat), findings that are in agreement with those of other investigators. (34) The correlation between AHI and leptin levels found in this study may thus represent the link between body fat mass and the development of OSA. One explanation might be that the AHI can be considered an indicator of the degree of physiologic stress caused by sleep-disordered breathing. Stress may lead to increased leptin levels either through an increase in leptin secretion or indirectly. (35) Furthermore, the suggestion that leptin may be causally related to OSA has been suggested in the literature. (36) In this context, it has been demonstrated in animal studies that leptin can prevent respiratory depression in obesity, implying that a deficiency in CNS leptin levels or activity may induce hypoventilation in some obese subjects, (37) suggesting that both obesity and OSA may be caused by leptin resistance. However, according to our data, leptin levels are not dependent on AHI and they are not correlated with AHI when controlled for BMI or body fat, as shown by others. (33) Thus, the causality remains controversial between our findings and the literature. The discrepancy between the latter study and our data may be explained by the fact that Phillips et al (33) investigated subjects with newly diagnosed OSA and without any other diseases, as opposed to a large number of comorbidities in our study.

It is of interest that in recent studies it has been shown that hyperleptinemia in OSA could be abrogated after treatment with nasal continuous positive airway pressure (nCPAP) with no significant change in the anthropometric parameters measured (34); whereas Chin et al, (38) in addition to demonstrating a fall of leptin levels with nCPAP, demonstrated significant reductions in visceral fat accumulation. These findings suggest that other mechanisms apart from fat mass could contribute to the increased leptin levels in OSA subjects and raises the intriguing possibility that treatment of OSA with nCPAP, as well as exogenous leptin agonists or agents that restore leptin receptor sensitivity, may play a role in the reversal of the primary pathophysiologic abnormalities of the OSA syndrome. In this context, it has been reported that in OSA patients there is evidence of cerebral arousal and increased sympathetic nerve discharge induced by nocturnal events of asphyxiation. (39) Therefore, it has been suggested that alleviating the neuroexcitation with nCPAP treatment may result in a decrease in leptin expression. (34) However, our data performing unpaired comparisons of leptin levels in subjects with comparable percentage of body fat but high or low degree of OSA (ie, high or low AHI) revealed no differences in leptin levels.

Obesity is linked to the metabolic syndrome in humans. The role of leptin in regulating neuroendocrine functions in patients with the metabolic syndrome is not clear, although insulin resistance has been associated with increased leptin levels in one study in humans. (40) We found a correlation between serum leptin and parameters describing glucose metabolism, especially the parameters indicating insulin resistance (FIRI), suggesting an association between obesity and the metabolic syndrome in OSA patients. Apart from contributing to the regulation of fat metabolism, leptin may also regulate insulin release in the form of an adipo-insular feedback, (41) while insulin resistance has been shown to cause increased secretion of leptin independent of fat mass. (42,43) Moreover, leptin concentrations were closely correlated to the changes in lipoproteins typically found in insulin resistance states, ie, elevated VLDL and low HDL levels. It should be pointed out that leptin has been associated with an increased risk of myocardial infarction. (44)

The repetitive increase in sympathetic tone may be responsible for the development of diurnal hypertension, (45) and patients with OSA have an increased risk of diurnal hypertension independent of obesity and age. (46,47) However, in our study we were not able to determine an association between hypertension and the degree of AHI, as were others. (48,49) According to data from the largest cross-sectional study to examine this association (Sleep Heart Health Study, > 6,000 patients (47)), an odds ratio of 1.37 (95% CI, 1.03 to 1.83) was found for the presence of hypertension between the lowest and highest OSA category. We believe that our study was underpowered to determine an association of OSA with hypertension.

An increasing prevalence of OSA with age is known from epidemiologic studies, (3,9) establishing the diagnosis usually in the fifth through the seventh decade. The absence of a consistent relationship between AHI and age in the present study is explained by the fact that patients with suspected sleep-related breathing disorders were admitted and, thus, even younger patients with a higher degree of OSA are frequently identified in sleep laboratories.

There was no association between the lipoprotein parameters and the presence of OSA, even after correction for lipid-lowering medications. A recent study (50) reporting an association between the number of apolipoprotein E4 alleles, a gene leading to higher LDL cholesterol levels, and the severity of OSA suggests increased prevalence of high LDL in patients with OSA. Again, the present study was underpowered to detect such an association because the apoE4 allele frequency is only approximately 15% in the population and carriers of this allele have, on average, only 5% higher LDL levels. (50) The clinical consequences of the synergistic effects on the cardiovascular risk of high LDL and OSA in apoE genotype carriers may be substantial.

Uric acid levels were positively correlated with AHI, and elevated uric acid levels were significantly associated with the presence of severe OSA (AHI > 40). There were, however, multiple colinearities of uric acid, eg, with body weight and body fat stores. The role of uric acid as an independent cardiovascular risk factor has been questioned, (51) although others argue that it is independently associated with cardiovascular events. (52)

Fibrinogen, an acute phase protein and clotting factor, has been found to be an independent risk factor for CVD. (53) Among the underlying mechanisms, fibrinogen appears to directly enhance atherogenesis by its conversion to fibrin, which binds LDL and stimulates proliferation of vascular smooth muscle. In our study, a correlation was found between serum fibrinogen levels and AHI, independent of smoking. Moreover, fibrinogen was a strong risk factor for severe OSA (AHI > 40), together with other cardiovascular risk factors.

The degree of OSA was independent of smoking status. Cigarette smoking as a potential risk factor for sleep-disordered breathing has been suggested. Data from the Wisconsin cohort showed that current cigarette smokers are at a greater risk for sleep-disordered breathing than nonsmokers. (54) In our study, only 28% were current smokers, so a strong association could be missed easily in our study.

Several potential shortcomings of this study have to be addressed. This study was not designed to include matched control subjects. Because the intention was to determine the degree of dependence of OSA on obesity and on multiple cardiovascular risk factors, we included all patients who were consecutively referred for suspected OSA to exclude a primary selection bias. We believe, however, that the diagnosis of OSA (as AHI > 10) in two thirds of the patients and the absence of OSA (as AHI < 10) in one third of the patients renders in the investigated cohort a sufficient balance to draw meaningful conclusions about patients referred for suspected OSA. Drawing conclusions about relationships between the independent variable AHI and diseases that can be attributable to OSA rather than other confounding comorbidities may be limited.

We were not able to perform MRI for all patients, because due to technical limitations the most overweight patients did not fit in the NMR unit. Thus, it is not possible to ultimately clarify the role of the regional body fat composition and especially the amount of parapharyngeal fat in the most obese patients, which may limit the conclusions drawn from our study. However, of the 18 patients excluded from NMR, only 3 patients were excluded because of excessive weight. The other 15 patients were excluded because of claustrophobia (n = 9), unavailability of the scanner at the scheduled time point (n = 4), history of mammary artery bypass surgery (n = 1), and a metal fragment in the head (n = 1). Thus, the excessive weight bias is limited to a total of < 5% of the patients.

There might be confounding of the cardiovascular parameters due to the use of medications that could alter the independent parameters of the study. These confounding effects are small because for hypertension the presence or absence rather than the absolute BP values were used, the number of patients receiving lipid-lowering or uric acid-lowering medications were very low. However, the study was, in general, insufficiently powered to clearly confirm or reject the relations between hypertension and hyperlipoproteinemia. One of the most important questions, "Is the degree of OSA dependent on leptin levels, or are leptin levels dependent on OSA?" could not be resolved; however, the evidence increases the likelihood that leptin plays a central role in OSA.

CONCLUSION

Taken together, this study shows a highly significant correlation between the degree of sleep-related breathing disorder and anthropometric measures of obesity, especially visceral fat accumulation. With respect to regional fat distribution, the amount of intra-abdominal fat is associated with OSA but not with the size of parapharyngeal fat. In addition, there were multiple colinearities between serum leptin concentrations and parameters describing body fat composition, lipoprotein, and glucose metabolism. Leptin levels were dependent on only percentage of body fat, subcutaneous fat in the neck region, and waist-to-hip ratio, but not on AHI. Most of the known risk factors for CVD, especially hypertension, smoking, and dyslipoproteinemia, however, were not associated with higher degree of OSA, indicating the potential major role of the percentage of body fat and regional fat distribution and its metabolic consequences for cardiovascular events in patients with OSA. For any given BMI, mortality is higher if fat is distributed centrally (visceral adiposity) compared with a more generalized pattern of distribution. (55) Whether OSA contributes to mortality in this respect remains to be established.

ACKNOWLEDGMENT: The authors thank Niklas Nolden, MD, and Gregor Sadowski, MD, for help in performing the study.

REFERENCES

(1) Guilleminault C. Clinical features and evaluation of obstructive sleep apnea. In: Kryger MH, Roth T, Dement WC, eds. Principles and practice of sleep medicine. Philadelphia, PA: W.B. Saunders, 1994; 667-677

(2) Teran-Santos J, Jiminez-Gomez A, Cordero-Guevara J, et al. The association between sleep apnea and the risk of traffic accidents. N Engl J Med 1999; 340:847-851

(3) Young T, Palta M, Dempsey J, et al. The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med 1993; 328:1230-1235

(4) Partinen M, Jamieson A, Guilleminault C. Long-term outcome for obstructive sleep apnea syndrome patients. Chest 1988; 94:1200-1204

(5) He J, Kryger MH, Zorick FJ, et al. Mortality and apnea index in obstructive sleep apnea. Chest 1988; 94:9-14

(6) American Thoracic Society/American Sleep Disorders Association. Statement on health outcomes research in sleep apnea. Am J Respir Crit Care Med 1998; 157:335-341

(7) Gibson GJ. Public health aspects of obstructive sleep apnoea. Thorax 1998; 53:408-409

(8) Wright J, Johns R, Watt I, et al. Health effects of obstructive sleep apnoea and the effectiveness of continuous positive airways pressure: a systematic review of the research evidence. BMJ 1997; 314:851-860

(9) Strohl KP, Redline S. Recognition of obstructive sleep apnea. Am J Respir Crit Care Med 1996; 154:279-289

(10) Shelton KE, Woodson H, Gay S, et al. Parapharyngeal fat in obstructive sleep apnea. Am Rev Respir Dis 1993; 148:462-466

(11) Horner RL, Mohiaddin RH, Lowell DG, et al. Sites and sizes of fat deposits around the pharynx in obese patients with obstructive sleep apnoea and weight matched controls. Eur Respir J 1989; 2:613-622

(12) Mortimore IL, Marshall I, Wraith PK, et al. Neck and total body fat deposition in nonobese and obese patients with sleep apnea compared with that in control subjects. Am J Respir Crit Care Med 1998; 157:280-283

(13) Hubert HB, Feinleib M, McNamara PM, et al. Obesity as an independent risk factor for cardiovascular disease: a 26-year follow-up of participants in the Framingham Heart Study. Circulation 1983; 67:968-977

(14) Rabkin SW, Mathewson FA, Hsu PH. Relation of body weight to development of ischemic heart disease in a cohort of young North American men after a 26 year observation period: the Manitoba Study. Am J Cardiol 1977; 39:452-458

(15) Larsson B, Svardsudd K, Welin L, et al. Abdominal adipose tissue distribution, obesity, and risk of cardiovascular disease and death: a 13 year follow up of participants in the study of men born in 1913. BMJ 1984; 288:1401-1404

(16) Donahue RP, Abbott RD, Bloom E, et al. Central obesity and coronary heart disease in men. Lancet 1987; 1:821-824

(17) Auwerx J, Staels B. Leptin. Lancet 1998; 351:737-742

(18) Rosenbaum M, Nicolson M, Hirsch J, et al. Effects of gender, body composition and menopause on plasma concentrations of leptin. J Clin Endocrinol Metab 1996; 81:3424-3427

(19) Word Health Organization Study Group. Diabetes mellitus. WHO Tech Rep Ser 1985; No. 727

(20) Alberti KG, Zimmet PZ. Definition, diagnosis and classification of diabetes mellitus and its complications: Part 1. Diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med 1998; 15:539-553

(21) Duncan MH, Singh BM, Wise PH, et al. A simple measure of insulin resistance. Lancet 1995; 346:120-121

(22) Bray GA. Pathophysiology of obesity. Am J Clin Nutr 1992; 55:4883-4943

(23) Rechtschaffen A, Kales AA, eds. A manual of standardized terminology, techniques and scoring systems for sleep stages of human subjects. Washington, DC: Government Printing Office, 1968; National Institutes of Health publication No. 204

(24) Guilleminault C, Tilkian A, Dement WC. The sleep apnea syndromes. Annu Rev Med 1976; 86:293-296

(25) Durnin JV, Womersley J. Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged from 16 to 72 years. Br J Nutr 1974; 32:77-97

(26) Kopelman PG. Altered respiratory function in obesity: sleep disordered breathing and the Pickwickian syndrome. In: Bjorntorp P, Brodoff BN, eds. Obesity. Philadelphia, PA: Lippincott, 1992; 568-575

(27) Bjorntorp P. The regulation of adipose tissue distribution in humans. Int J Obes Relat Metab Disord 1996; 20:291-320

(28) Kissebah AH, Krakower GR. Regional adiposity and morbidity. Physiol Rev 1994; 74:761-811

(29) Pouliot MC, Despres JP, Nadeau A, et al. Visceral obesity in man: associations with glucose tolerance, plasma insulin, and lipoprotein levels. Diabetes 1992; 41:826-834

(30) Grunstein K, Wilcox I, Yang T-S, et al. Snoring and sleep apnoea in men: association with central obesity and hypertension. Int J Obes 1993; 17:533-540

(31) Spiegelman BM, Flier JS. Adipogenesis and obesity: rounding out the big picture. Cell 1996; 87:377-389

(32) Considine RV, Sinha MK, Heimann ML, et al. Serum immunoreactive-leptin concentrations in normal weight and obese humans. N Engl J Med 1996; 334:292-295

(33) Phillips BG, Kato M, Narkiewicz K, et al. Increases in leptin levels, sympathetic drive, and weight gain in obstructive sleep apnea. Am J Physiol Heart Circ Physiol 2000; 279:H234-H237

(34) Ip MS, Lam KS, Ho C-M, et al. Serum leptin and vascular risk factors in obstructive sleep apnea. Chest 2000; 118:580-586

(35) Heiman MI, Ahima LS, Craft B, et al. Leptin inhibition of the hypothalamic-pituitary-adrenal axis in response to stress. Endocrinology 1997; 138:3859-3863

(36) O'Donnell CP, Tankersley CG, Polotsky VP, et al. Leptin, obesity, and respiratory function. Respir Physiol 2000; 119: 163-170

(37) O'Donnell CP, Schaub CD, Haines AS, et al. Leptin prevents respiratory depression in obesity. Am J Respir Crit Care Med 1999; 159:1477-1484

(38) Chin K, Shimizu K, Nakamura T, et al. Changes in intra-abdominal visceral fat and serum leptin levels in patients with obstructive sleep apnea syndrome following nasal continuous positive airway pressure therapy. Circulation 1999; 100:706-712

(39) Somers VK, Dyken ME, Clary MP, et al. Sympathetic neural mechanisms in obstructive sleep apnea. J Clin Invest 1995; 96:1897-1904

(40) Segal KR, Landt M, Klein S. Relationship between insulin sensitivity and plasma leptin concentration in lean and obese men. Diabetes 1996; 45:987-991

(41) Kieffer TJ, Heller RS, Habener JF. Leptin receptors expressed on pancreatic beta-cells. Biochem Biophys Res Commun 1996; 224:522-527

(42) Kolaczynski JW, Nyce MR, Considine RV, et al. Acute and chronic effects of insulin on leptin production in humans: studies in vivo and in vitro. Diabetes 1996; 45:699-701

(43) Dagogo-Jack S, Fanelli C, Paramore D, et al. Plasma leptin and insulin relationships in obese and nonobese humans. Diabetes 1996; 45:695-698

(44) Soderberg S, Ahren B, Jansson JH, et al. Leptin is associated with increased risk of myocardial infarction. J Intern Med 1999; 246:409-418

(45) Fletcher EC. The relationship between systemic hypertension and obstructive sleep apnea: facts and theory. Am J Med 1995; 98:118-128

(46) Carlson JT, Hedner JA, Ejnell H, et al. High prevalence of hypertension in sleep apnea patients independent of obesity. Am J Respir Crit Care Med 1994; 150:72-77

(47) Nieto FJ, Young TB, Lind BK, et al. Association of sleep-disordered breathing, sleep apnea, and hypertension in a large community-based study. JAMA 2000; 283:1829-1836

(48) Stradling JR, Crosby JH. Relation between systemic hypertension and sleep hypoxaemia or snoring. BMJ 1990; 300: 75-78

(49) Jennum P, Sjol A. Snoring, sleep apnea and cardiovascular risk factors. Int J Epidemiol 1993; 22:439-444

(50) Kadotani H, Kadotani T, Young T, et al. Association between apolipoprotein E epsilon 4 and sleep-disordered breathing in adults. JAMA 2001; 285:2888-2890

(51) Culleton BF, Larson MG, Kannel WB, et al. Serum uric acid and risk for cardiovascular disease and death: the Framingham Study. Ann Intern Med 1999; 131:7-13

(52) Fang J, Alderman MH. Serum uric acid and cardiovascular mortality. The NHANES I epidemiologic follow-up study, 1971-1992. JAMA 2000; 283:2404-2410

(53) Ma J, Hennekens CH, Ridker PM, et al. A prospective study of fibrinogen and risk of myocardial infarction in the physicians' health study. J Am Coll Cardiol 1999; 33:1347-1352

(54) Wetter DW, Young TB, Bidwell TR, et al. Smoking as a risk factor for sleep-disordered breathing. Arch Intern Med 1994; 154:2219-2224

(55) Bjorntorp P. Obesity. Lancet 1997; 350:423-426

* From the Department of Medicine II (Drs. Schafer and Ewig), the Department of Radiology (Dr. Pauleit), the Department of Clinical Pharmacology (Drs. Sudhop and Berthold), and the Medical Policlinic (Dr. Gouni-Berthold), University of Bonn, Bonn, Germany.

Manuscript received May 8, 2001; revision accepted April 9, 2002.

Correspondence to: Harald Schafer, MD, Department of Pulmonary Medicine, Klinikum Saarbruecken, PO Box 102629, 66026 Saarbruecken, Germany; e-mail: h.schaefer@klinikum-saarbruecken.de

COPYRIGHT 2002 American College of Chest Physicians
COPYRIGHT 2003 Gale Group