Abstract
Previous studies have reported the effects of obstructive sleep apnea (OSA) and cardiometabolic disorders on cardiovascular disease (CVD), but associations between cardiometabolic biomarkers and two cardinal features of OSA (chronic intermittent hypoxia and sleep fragmentation) and their interactions on CVD in OSA populations remain unclear. A total of 1727 subjects were included in this observational study. Data on overnight polysomnography parameters, biochemical biomarkers, and anthropometric measurements were collected. Metabolic syndrome (MS), including blood pressure, waist circumference (WC), fasting glucose, triglycerides (TG), and high‐density lipoprotein cholesterol (HDL‐C), was diagnosed based on modified criteria of the Adult Treatment Panel III. WC, mean arterial pressure, TG and low‐density lipoprotein cholesterol (LDL‐C) were independently associated with apnea‐hypopnea index (AHI) after adjustment for confounding factors (β = 0.578, P = 0.000; β = 0.157, P = 0.001; β = 1.003, P = 0.019; and β = 4.067, P = 0.0005, respectively). Furthermore, the interaction analysis revealed joint effects between hypertension, obesity, hyperglycemia, and LDL‐C dyslipidemia and AHI on CVD. The relative excess risks of CVD due to the interactions with OSA were 2.06, 1.02, 0.48, and 1.42, respectively (all P < 0.05). In contrast, we found no independent effect of the microarousal index (MAI) on CVD. However, LDL‐C level and some MS components (WC, TG) were associated with MAI. Our findings indicate that hypoxemia and cardiometabolic disorders in OSA may potentiate their unfavorable effects on CVD. Sleep fragmentation may indirectly predispose patients with OSA to an increased risk of CVD. Thus, cardiometabolic disorders and OSA synergistically influence cardiometabolic risk patterns.
Keywords: cardiovascular disease, dyslipidemia, metabolic syndrome, obstructive sleep apnea, sleep fragmentation
1. INTRODUCTION
Obstructive sleep apnea (OSA) is a common sleep disorder, affecting approximately 2%‐4% of all middle‐aged adults. OSA is characterized by recurrent episodes of upper airway collapse that lead to two cardinal features, including chronic intermittent hypoxia (CIH) and sleep fragmentation.1 Metabolic syndrome (MS) has a strong relationship with OSA and excessive cardiovascular risk.2, 3 OSA is associated with an increased likelihood of cardiovascular diseases (CVD).4, 5, 6 Some investigators believe that OSA should be classified under MS, which should be called syndrome Z.7 Although increased serum atherogenic lipids, such as low‐density lipoprotein cholesterol (LDL‐C) and total cholesterol, are not a specific component in MS, previous work demonstrated that OSA is associated with highly oxidized LDL‐C.8 Thus, we defined MS components and atherogenic LDL‐C as cardiometabolic disorders in this study.
Recent studies have demonstrated that OSA is related to cardiometabolic disorders due to CIH.9, 10 Furthermore, sleep fragmentation is another common pathophysiology of OSA. We are aware that OSA in epidemiological studies is also associated with an increase in overall CVD mortality and morbidity.11 However, the simultaneous independent relationships between the cardiometabolic disorders and two cardinal features of OSA remain unclear. OSA and cardiometabolic disorders develop in a very complex manner, and long‐term CVD pathology may be influenced by their interactions. Previous clinical studies showed that OSA could modify the effect of the hypertension on chronic heart failure, particularly in patients with concurrent coronary heart disease (CHD).12, 13 However, none of these works analyzed the interactive effects on CVD between cardiometabolic disorders, including MS components, serum atherogenic lipid, and CIH or sleep fragmentation in OSA. On the other hand, continuous positive airway pressure (CPAP) is the gold standard therapy for OSA and provides a substantial therapeutic benefit to CVD.14 However, reports on its effects on cardiometabolic disorders are conflicting.15 In addition, the effect of CPAP on BP varies, most likely owing to the multifactorial nature of hypertension (which involves factors related to metabolism, such as body weight).16 Long‐term weight loss could reduce OSA severity in obese patients with diabetes.17 Clinicians of overweight patients with OSA should consider the advantages of incorporating weight loss and dietary management into OSA treatment programs.18 Considering the increased risk of CVD in some patients with cardiometabolic disorders, it is important to consider the relatively independent and joint effect on CVD in OSA.
To our knowledge, this is the first report identifying the independent and joint associations of cardiometabolic disorders and OSA with the prevalence of multiple CVD in an OSA population, which may help us to improve our understanding of the pathophysiology of these disorders and yield more effective treatment strategies for OSA.
2. METHODS
2.1. Study population
This large‐scale cross‐sectional study consists of 2022 consecutive participants who were referred to the sleep laboratory of Shanghai Jiao Tong University Affiliated Sixth People’s Hospital for suspected OSA between January 2012 and January 2017. We excluded 295 patients based on the following reasons: (a) aged <18 years (n = 25); (b) previously treated for OSA (n = 97); (c) the presence of another systemic disease, such as respiratory, endocrine disease, cancer, psychiatric diseases, and pregnancy (n = 84); 4) and missing data (n = 89). Ultimately, we used the data of 1727 participants to explore risk factors for and their interactive effects on CVD. This study was approved by the Internal Review Board of the Institutional Ethics Committee of Shanghai Jiao Tong University Affiliated Sixth Hospital and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all subjects before they participated in this study; each participant first received a one‐page leaflet containing information on the informed consent procedure.
2.2. Anthropometric and biochemical measurements
Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared (kg/m2). Waist circumference (WC) was measured midway between the lower costal margin and the iliac crest while the subject was standing. Daytime BP was measured after at least 5 minutes of rest in a sitting position using a mercury sphygmomanometer following the American Society of Hypertension Guidelines, and the mean of three measurements was recorded. Mean arterial pressure (MAP) was calculated as [(systolic BP (SBP) + 2 * diastolic BP(DBP)]/3. Our thresholds for the MS definition are general not population specific. The blood pressure threshold for hypertension is 140/90 mm Hg. The other MS components were defined as follows: (a) abdominal obesity: WC ≥102 cm in males and 88 cm in females; (b) hyperglycemia: fasting glucose (FG) ≥110 mg/dL (6.1 mmol/L) or a history of diabetes or current use of anti‐diabetic medication; and (c) dyslipidemia: triglycerides (TG) ≥150 mg/dL (1.7 mmol/L) or high‐density lipoprotein cholesterol (HDL‐C) <40 mg/dL (1.03 mmol/L) in males and <50 mg/dL (1.29 mmol/L) in females or current use of a lipid‐lowering drug. MS was defined as the presence of three or more of these components according to the diagnostic criteria of the United States National Cholesterol Education Program Adult Treatment Panel III (NCEPIII).19 Other atherogenic dyslipidemia was based on the following definition: total cholesterol (TC) ≥5.17 mmol/L, low‐density lipoprotein cholesterol (LDL‐C) ≥3.33 mmol/L, apoA‐I <1.2 g/L, apoB >1.1 g/L and apoE levels >0.05 or <0.03 g/L or current use of a lipid‐lowering drug according to the diagnostic criteria of the NCEPIII.19 Insulin resistance was defined as an insulin resistance index ≥2.5 of the homeostasis model assessment, which was calculated as fasting serum insulin (μU/mL) × FG (mmol/L)/22.5.20 A fasting blood sample was collected from the antecubital vein of all participants on the morning after polysomnography (PSG) monitoring. Serum lipid, glucose, and insulin levels were measured in the hospital laboratory using routine procedures.
2.3. Overnight PSG parameters
Respiratory events were recorded using a laboratory‐based PSG instrument (Alice 4 or 5; Respironics, Pittsburgh, PA, USA). The oxygen desaturation index (ODI) was defined as the number of times per hour of sleep that the blood oxygen level dropped by ≥4% from baseline. The microarousal index (MAI) was defined as the number of arousals per hour of sleep. The lowest pulse oxygen saturation (LSpO2) was the lowest oxygen saturation value recorded during sleep. We defined the fourth quartile of MAI ≥38.5 as the high MAI in the total of sample. The apnea‐hypopnea index (AHI) was the number of apnea and hypopnea events per hour during sleep. OSA severity was defined according to current standards and based on the criteria of the American Academy of Sleep Medicine as follows: AHI <5, non‐OSA; mild OSA, 5 ≤ AHI < 15; moderate OSA, 15 ≤ AHI < 30; and severe OSA, AHI ≥30.21
2.4. CVD outcome definition and information collection
Cardiovascular disease comorbidities, including CHD, heart failure, myocardial infarction, stroke, arrhythmias and transient ischemic attack, were classified based on the International Statistical Classification of Diseases (ICD)‐10.4, 5, 6 All CVD outcomes were based on the symptoms and some objective biochemical measurements and laboratory tests according to the Chinese population diagnostic criteria. All CVD diagnostic information were obtained from our hospital or other hospitals and physician office records through the medical Internet system in China or was collected by self‐reported. In addition, a study technician interviewed the participants using a standardized questionnaire and collected information on medical history and health‐related characteristics. In brief, the interviewer asked each participant if a doctor had ever diagnosed him/her with any of the aforementioned conditions and if he/she had ever undergone the treatments for CVD (ie, coronary bypass surgery or coronary angioplasty). An “unsure” response was permitted for all items. We defined prevalent CVD as a “yes” response to one or more of the conditions or procedures listed above. Participants who responded “no” to all questions were considered CVD‐free. The remainder (ie, those giving a combination of “no” and “unsure” responses) were classified as “unknown” disease status and excluded from the analyses. Smoking status and alcohol consumption were also ascertained. Smoking status and alcohol consumption were classified as current or noncurrent smoker and current or noncurrent drinker, respectively.
2.5. Statistics
Data are presented as the means (standard deviation, SD), medians (interquartile range), or the number (%) depending on whether data were normally distributed, skewed, or categorical, respectively. Differences in baseline characteristics among the four groups were examined using the Kruskal‐Wallis H test, one‐way analysis of variance, Fisher’s exact test, or the χ 2 test according to the distribution of the data. A post hoc analysis was used to compare differences between the two groups. P‐values for linear trends across quartile groups were calculated using the polynomial linear trend test for continuous variables. After integrating the clinical MS components definition and the close relation among serum lipids and metabolic biomarkers (TC vs LDL‐C vs apoB; TG vs apoE; HDL‐C vs apoA‐I; SBP vs DBP vs MAP),19 we selected MS components (TG, HDL‐C, MAP, and FG) and LDL‐C as the cardiometabolic biomarkers in our further analysis. The Spearman test or Pearson test analysis confirmed the close correlation (all correlation coefficients >0.6; Table S1). Stepwise multivariate linear regression analyses were performed on the cardiometabolic predictor with the cardinal features of OSA. In epidemiological research, an interaction refers to a situation where the effect of one risk factor (A) on a certain disease outcome is different across strata of another risk factor (B), or vice versa. If the combined or interaction effect of A and B is larger (or smaller) than the sum of the effects of A and B, the interaction should be the focus. Multiplicative interaction effects were evaluated with binary logistic regression including four interaction groups [non‐OSA * non‐MS (reference), non‐OSA * MS, OSA * non‐MS, and OSA * MS],12, 13 and the odds ratio (OR) of the multiplicative interaction between OSA and metabolic disorder was adjusted for age, sex, BMI, and lifestyle factors using CVD as the dependent variable. The additive interaction was evaluated using three indices in Excel sheet written by Andersson: relative excess risk due to interaction (RERI) = OR11 − OR10 − OR01 + 1, and the attributable proportion due to the interaction (AP) was <0, and a synergy index (S) of <1 were defined as no additive interaction.22, 23 The analyses were performed using SPSS software (ver. 20.0; SPSS Inc, Chicago, IL, USA). P‐values <0.05 were considered significant.
3. RESULTS
3.1. Patient baseline characteristics
The baseline characteristics of the 1727 patients according to OSA severity are presented in Table 1. Patients with OSA were the older than without OSA, and OSA patients also exhibited increased values for BMI, serum lipids, FG, insulin, sleep parameters, and current smoking and drinking prevalence. In total, 81.5% of all subjects had OSA. In addition, 17.7% of all patients had reported at least one manifestation of CVD; of these, 52.94% had CHD. The subjects were divided into four subgroups according to the presence or absence of MS and OSA (non‐OSA without MS; non‐OSA with MS; OSA without MS; and OSA with MS). The prevalence rates of CVD in these subgroups were 4.7%, 14.3%, 25.9%, and 35.9%, respectively.
Table 1.
Characteristics of the groups classified according to obstructive sleep apnea (OSA) severity using current clinical definitions
| Non‐OSA (n = 319) | Mild OSA (n = 282) | Moderate OSA (n = 320) | Severe OSA (n = 806) | P value | |
|---|---|---|---|---|---|
| Demographics | |||||
| Age, y | 37 (30, 47) | 41 (33, 51) | 43 (35, 54) | 43 (35, 53) | <0.001 |
| Male, N (%) | 426 (59.2) | 456 (74) | 488 (80.7) | 1706 (88.2) | <0.001 |
| BMI, kg/m2 | 23.78 (21.67, 25.71) | 25.15 (23.31, 27.43) | 25.991 (23.98, 28.40) | 27.68 (25.64, 30.02) | <0.001 |
| NC, cm | 36 (34, 39) | 38 (36, 40) | 39 (37, 41) | 41 (39, 43) | <0.001 |
| WC, cm | 86 (80, 92.5) | 92 (86, 97) | 95 (90, 100) | 99 (94, 105) | <0.001 |
| HC, cm | 96 (92, 101) | 99 (95, 103) | 100 (96, 105) | 103 (99, 108) | <0.001 |
| SBP | 120 (112, 126) | 120 (115, 132) | 125 (116, 135) | 127 (120, 138) | <0.001 |
| DBP | 78 (70, 81) | 79 (71, 85) | 80 (72, 86) | 80 (76, 90) | <0.001 |
| Biochemistry assays | |||||
| TC, mmol/L | 4.34 ± 0.93 | 4.71 ± 0.87 | 4.65 (4.11,5.28) | 4.79 (4.21, 5,5.4) | <0.001 |
| TG, mmol/L | 1.12 (0.74, 1.65) | 1.42 (1.00, 2.08) | 1.61 (1.14, 2.275) | 1.73 (1.26, 2.52) | <0.001 |
| HDL‐C, mmol/L | 1.11 (0.95,1.3) | 1.05 (0.92,1.23) | 1.02 (0.91,1.185) | 1.01 (0.89,1.15) | <0.001 |
| LDL‐C, mmol/L | 2.66 ± 0.82 | 2.99 ± 0.78 | 2.94 (2.47, 5,3.5) | 3.05 (2.53, 3.56) | <0.001 |
| apoA‐I, g/L | 1.11 (0.97, 1.28) | 1.08 (0.96, 1.23) | 1.07 (0.96, 1.23) | 1.08 (0.97, 1.21) | 0.011 |
| apoB, g/L | 0.725 (0.61, 0.85) | 0.83 ± 0.18 | 0.81 (0.71, 0.95) | 0.85 (0.75, 0.98) | <0.001 |
| apoE, mg/L | 3.93 (3.23,4.74) | 4.1 (3.44,5.02) | 4.315 (3.5575, 5.3225) | 4.47 (3.68,5.63) | <0.001 |
| Lp(a), g/L | 7.8 (4.3, 16.6) | 7.65 (3.9, 17.2) | 7.2 (3.7, 14.29) | 7 (3.6, 14.3) | 0.002 |
| Fasting glucose, mmol/L | 4.99 (4.65, 5.30) | 5.13 (4.85, 5.56) | 5.25 (4.86, 5.79) | 5.41 (5.02, 6.02) | <0.001 |
| Insulin, μU/L | 7.4 (5.07, 10) | 8.9 (6.2, 13) | 10 (7.27, 15) | 12 (8.8, 18) | <0.001 |
| HOMA‐IR | 0.50 ± 0.67 | 0.74 ± 0.74 | 0.92 (0.51, 1.34) | 1.13 (0.72, 1.54) | <0.001 |
| Sleep parameters | |||||
| AHI | 1.7 (0.6, 3.1) | 9.2 (7, 12.1) | 21.6 (18.2, 25.85) | 58.3 (44.3, 70.725) | <0.001 |
| MAI | 12.8 (9.4, 20.2) | 18 (12.6, 26.1) | 23 (14.85, 31.15) | 38.4 (23.8, 55.42) | <0.001 |
| LSpO2 | 93.5 (91, 95) | 87 (84, 91) | 83 (78, 87) | 72 (63, 79) | <0.001 |
| ODI | 1 (0, 3) | 9 (6, 12) | 21 (17, 27) | 56 (42, 71) | <0.001 |
| ESS | 4 (0, 8) | 6 (3, 10.5) | 6 (3, 11) | 10 (6, 14.5) | <0.001 |
| Medical and lifestyle history | |||||
| Diabetes, N (%) | 45 (6.3) | 59 (9.5) | 77 (12.7) | 271 (14) | <0.001 |
| Obesity, N (%) | 73 (10.3) | 130 (21.2) | 175 (29.4) | 876 (45.6) | <0.001 |
| Hypertension, N (%) | 120 (18.9) | 177 (31.2) | 222 (39.9) | 924 (51.4) | <0.001 |
| Dyslipidemia, N (%) | 199 (62.4) | 217 (77) | 272 (85) | 697 (86.5) | <0.001 |
| MS, N (%) | 55 (17.24) | 50 (17.7) | 79 (24.7) | 264 (32.8) | <0.001 |
| CVD, N (%) | 21 (6.6) | 28 (9.9) | 45 (14.1) | 212 (26.3) | <0.001 |
| CHD, N (%) | 9 (2.8) | 13 (4.6) | 23 (7.2) | 117 (14.5) | <0.001 |
| Myocardial infarction, N (%) | 2 (0.6) | 3 (1) | 3 (0.9) | 27 (3.3) | <0.001 |
| Heart failure, N (%) | 6 (1.8) | 8 (2.8) | 10 (3.1) | 45 (14) | <0.001 |
| Stroke, N (%) | 1 (0.3) | 5 (1.7) | 9 (2.8) | 41 (5) | <0.001 |
| Arrhythmias, N (%) | 12 (3.7) | 13 (4.6) | 11 (3.4) | 41 (5) | 0.061 |
| Transient ischemic attack, N (%) | 7 (2.1) | 6 (2.1) | 9 (2.8) | 22 (2.7) | 0.045 |
| Current smoking %, N (%) | 143 (19.9) | 206 (33.3) | 231 (38.2) | 917 (47.4) | <0.001 |
| Alcohol consumption, N (%) | 207 (28.8) | 225 (36.3) | 245 (40.5) | 932 (48.2) | <0.001 |
AHI, apnea‐hypopnea index; apo, apolipoprotein; BMI, body mass index; CHD, coronary heart disease; CVD, cardiovascular disease; DBP, diastolic blood pressure; ESS, Epworth Sleepiness Scale; HC, hip circumference; HDL‐C, high‐density lipoprotein cholesterol; HOMA‐IR, insulin resistance index calculated by the homeostasis model assessment; LDL‐C, low‐density lipoprotein cholesterol; Lp(a), lipoprotein(a); LSpO2, lowest oxygen saturation; MAI, microarousal index; MS, metabolic syndrome; NC, neck circumference; ODI, oxygen desaturation index; SBP, systolic blood pressure; TC, total cholesterol; TG, triglycerides; WC, waist circumference.
The data are presented as means ± SD; skewed data are presented as the median (IQR), and categorical data as the number (%). Differences in the baseline characteristics among the four groups were examined using a Kruskal‐Wallis H test or χ 2 tests according to the characteristics of the data distribution.
3.2. Relationship between cardiometabolic biomarkers and the cardinal features of OSA
The relationship between metabolism and the cardinal features of OSA was examined. Figure 1 depicts the observed association between AHI or MAI and metabolic biomarkers levels without any adjustments for confounders. Positive dose‐effect relationships were identified between AHI or MAI quartiles and MS biomarkers, such as FG, WC, blood TG, LDL‐C level, and MAP. A negative dose‐response relationship was observed between AHI or MAI and HDL‐C levels. Potential associated factors of two OSA cardinal features are provided in Table 2. After BMI adjustment of Model 2, multivariate linear regression revealed that metabolic biomarkers, including WC (β = 0.578, P = 0.000), MAP (β = 0.157, P = 0.001), fasting LDL‐C level (β = 4.067, P = 0.0005), and TG (β = 1.003, P = 0.019) were significantly associated with the OSA severity indexed by AHI (model 3). In terms of MAI, similar metabolic biomarkers without MAP exhibited significant relationships with MAI in model 3.
Figure 1.
The tendency of metabolic biomarker levels across AHI and MAI quartiles. The trend was tested by the polynomial linear trend test for continuous variables. Positive linear trends among the AHI and MAI quartiles were significant for the FG, MAP, LDL‐C, WC, and TG levels; negative trend was found for HDL‐C levels. Abbreviations: AHI, apnea‐hypopnea index; FG, fasting blood glucose; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; MAI, microarousal index; MAP, mean arterial pressure; Q1‐Q4, the lowest to the highest quartile; TG, triglycerides; WC, waist circumference
Table 2.
Stepwise multivariate linear regression model for AHI and MAI
| Predictors | AHI | MAI | ||||
|---|---|---|---|---|---|---|
| Β | SE[β] | P‐value | Β | SE[β] | P‐value | |
| Model 1 | ||||||
| BMI | 1.952 | 0.157 | 0.000 | 0.785 | 0.130 | 0.000 |
| TG | 0.960 | 0.437 | 0.028 | 1.067 | 0.372 | 0.004 |
| HDL‐C | ||||||
| LDL‐C | 4.178 | 0.693 | 0.000 | 4.509 | 0.598 | 0.000 |
| Fasting glucose | 0.784 | 0.319 | 0.014 | |||
| MAP (mm Hg) | 0.170 | 0.048 | 0.000 | |||
| Model 2 | ||||||
| WC (cm) | 0.786 | 0.055 | 0.000 | 0.324 | 0.046 | 0.000 |
| TG | 1.042 | 0.428 | 0.015 | 0.991 | 0.373 | 0.008 |
| HDL‐C | ||||||
| LDL‐C | 4.086 | 0.693 | 0.000 | 4.444 | 0.601 | 0.000 |
| Fasting glucose | ||||||
| MAP (mm Hg) | 0.168 | 0.048 | 0.000 | |||
| Model 3 | ||||||
| BMI | 0.708 | 0.266 | 0.008 | |||
| WC (cm) | 0.578 | 0.095 | 0.000 | 0.324 | 0.046 | 0.000 |
| TG | 1.003 | 0.428 | 0.019 | 0.991 | 0.373 | 0.008 |
| HDL‐C | ||||||
| LDL‐C | 4.067 | 0.692 | 0.000 | 4.444 | 0.601 | 0.000 |
| Fasting glucose | ||||||
| MAP (mm Hg) | 0.157 | 0.048 | 0.001 | |||
AHI, apnea‐hypopnea index; BMI, body mass index; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; MAI, microarousal index; MAP, mean arterial pressure; TG, triglycerides; WC, waist circumference.
Model 1 was adjusted by age, sex, dyslipidemia or diabetes treatment history and drinking and smoking status for BMI, TG, HDL‐C, BP, and fasting glucose.
Model 2 was adjusted using the variables listed above in model 1 for waist circumference, TG, HDL‐C, BP, and fasting glucose.
Model 3 was adjusted by the above variables in model 1 for BMI, waist circumference, TG, HDL‐C, BP, and fasting glucose.
3.3. Independent associations of OSA, dyslipidemia, hyperglycemia, and abdominal obesity with CVD
We next explored the independent associations of cardiometabolic disorders and OSA with CVD. Binary forward logistic regression models were used to determine CVD risk factors, including TG, LDL‐C, and HDL‐C dyslipidemia, hypertension, abdominal obesity, and hyperglycemia in model 1. Model 1 was adjusted for age, sex, BMI, smoking, and drinking status; model 2 was further adjusted for OSA (AHI ≥5 events per hour) and MAI (Table 3). The risk of CVD increased significantly in those with HDL‐C dyslipidemia, hyperglycemia, hypertension, and abdominal obesity in model 1. After further adjustment for MAI and OSA, all risk factors remained significant in model 1 and independent of OSA in model 2. In contrast, MAI exhibited no relationship with CVD in either model. Logistic regression for CHD exhibited similar results to the CVD models (data not shown).
Table 3.
Binary logistic regression model for the risk of CVD
| OR (95% CI) | |
|---|---|
| Model 1 | |
| TG dyslipidemia | |
| HDL‐C dyslipidemia | 1.562 (1.141, 2.138)b |
| Hyperglycemia | 1.721 (1.274, 2.325)c |
| Abdominal obesity | 2.212 (1.674, 2.923)c |
| Hypertension | 2.725 (2.009, 3.696)c |
| LDL‐C dyslipidemia | 1.615 (1.222,2.133)b |
| Model 2 | |
| TG dyslipidemia | |
| HDL‐C dyslipidemia | 1.425 (1.034, 1.963)a |
| Hyperglycemia | 1.700 (1.258, 2.297)b |
| Abdominal obesity | 2.095 (1.582, 2.774)c |
| Hypertension | 2.678 (1.973, 3.634)c |
| LDL‐C dyslipidemia | 1.533 (1.158,2.030)b |
| OSA (AHI >5 events/h) | 2.155 (1.318, 3.523)b |
| MAI | |
BMI, body mass index; CVD, cardiovascular disease; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; OSA, obstructive sleep apnea; TG, triglycerides.
The data are presented as the odds ratio (OR) (95% CI). We performed forward binary logistic regression. Models were adjusted for age, sex, BMI and smoking and drinking status. Age, BMI and MAI were as continuous variables, and the others as categorized variables in the model.
P < 0.05.
P < 0.01.
P < 0.001.
3.4. Interaction effects of OSA cardinal feature and dyslipidemia, hyperglycemia, abdominal obesity, and hypertension on CVD
Non‐OSA (41.8%) was most frequently observed in non‐MS subjects, whereas severe OSA (63%) was most frequently observed in patients with at least three MS components. Four‐level multiplicative interactions between the presence or absence of OSA (AHI ≥5 or MAI ≥38.5) and cardiometabolic disorders (MS components and LDL‐C dyslipidemia) were quantified (Figure 2). For example, we adjusted for age, sex, BMI, abdominal obesity, hyperglycemia, TG, LDL‐C, HDL‐C dyslipidemia and drinking and smoking status. Then, hypertension and OSA (AHI ≥5) and the product of hypertension and OSA are entered as independent variables, and CVD was entered as the dependent variable in the logistic regression model. Participants with OSA (AHI ≥5) and hypertension exhibited increased odds of CVD compared with the other three groups of patients, that is, those with either OSA or hypertension alone or neither (OR, 9.06; 95% confidence interval [CI]: 3.42, 18.85, P = 0.000). The addictive effects between AHI or MAI and cardiometabolic biomarkers on CVD are presented in Table 4. The RERI value was 2.06 (95% CI: 0.76, 4.23, P = 0.000), which was 2.06‐fold increased as a result of the additive interaction between AHI ≥5 and hypertension compared with their independent effects for CVD. AP was 0.43 (95% CI: 0.0169, 0.9632, P = 0.008), representing 43% of CVD cases that were attributed to the additive interaction among participants exposed to both OSA and hypertension. S was 2.68 (95% CI: 0.9468, 3.5410, P = 0.000), suggesting it had a positive effect on the interaction (S > 1 indicates a positive effect, whereas S < 1 indicates a negative interaction). All three measures of the additive interaction indicate biologically joint effects on the CVD between AHI ≥5 and hypertension. In addition, the effects were greater than the sum of the estimated effects of OSA alone and hypertension alone. Specially, the joint effects on CVD were noted between high MAI and LDL‐C dyslipidemia rather than other metabolic components (Table 4). Overall, OSA exhibited significant interactions with four components, namely, hypertension, high glucose or diabetes, high LDL‐C dyslipidemia, and abdominal obesity. When OSA was defined by an of AHI ≥10, the results remained unchanged (data not shown). Table S2 showed that additive interactive effects of OSA and cardiometabolic biomarkers on CVD based on the MS of Chinese diabetes society.
Figure 2.
The multiplicative interaction effects between OSA (AHI ≥5 or MAI ≥38.5) and metabolism dysfunction on CVD risk. The interactions were evaluated with binary logistic regression including four interaction groups (non‐OSA * non‐MS [reference], non‐OSA * MS, OSA * non‐MS, and OSA * MS), after adjusting for age, sex, BMI, and drinking and smoking status in the logistic regression model. Abbreviations: MAI, microarousal index; MS, metabolic syndrome; OSA, obstructive sleep apnea
Table 4.
Additive interactive effects of obstructive sleep apnea (OSA) and cardiometabolic biomarkers on CVD
| AHI + cardiometabolic biomarkers | MAI + cardiometabolic biomarkers | |||||
|---|---|---|---|---|---|---|
| Estimate | 95% CI | P value | Estimate | 95% CI | P value | |
| Hypertension | ||||||
| RERI | 2.0652 | 0.7643, 4.2335 | 0.0008 | −0.3211 | −0.4512, 0.4587 | 0.0746 |
| AP | 0.4328 | 0.0169, 0.9632 | 0.0087 | −0.0.565 | −0.3211, 0.2398 | 0.0687 |
| S | 2.6801 | 0.9468, 3.5410 | 0.0008 | 0.8741 | 0.5132, 1.5820 | 0.0653 |
| Abdominal obesity | ||||||
| RERI | 1.0231 | −1.1770, 3.2232 | 0.0133 | −0.4571 | −1.5030, 0.5883 | 0.3814 |
| AP | 0.2177 | −0.2477, 0.6832 | 0.0192 | −0.2349 | −0.8210, 0.3511 | 0.4320 |
| S | 1.3823 | 0.6042, 3.1623 | 0.0124 | 0.6742 | 0.2735, 1.6620 | 0.3917 |
| Hyperglycemia | ||||||
| RERI | 0.4827 | −1.6965, 2.6619 | 0.0246 | −0.1072 | −0.9902, 1.2045 | 0.8482 |
| AP | 0.1272 | −0.4451, 0.6995 | 0.0122 | −0.0558 | −0.4956, 0.6071 | 0.7528 |
| S | 1.2088 | 0.4711, 3.1022 | 0.0453 | 0.9315 | 0.3233, 1.0598 | 0.3867 |
| HDL‐C dyslipidemia | ||||||
| RERI | 0.3135 | −1.1777, 1.8047 | 0.0577 | −0.1407 | −0.8413, 1.1227 | 0.0786 |
| AP | 0.0873 | −0.3494, 0.5241 | 0.0621 | −0.0659 | −0.4889, 0.5208 | 0.7601 |
| S | 1.1378 | 0.5619, 2.3037 | 0.0603 | 0.7416 | 0.4319, 1.0175 | 0.0564 |
| TG dyslipidemia | ||||||
| RERI | −0.0903 | −1.2753, 1.0947 | 0.6151 | −0.3350 | −0.9952, 0.3251 | 0.3198 |
| AP | −0.0441 | −0.6150, 0.5267 | 0.6955 | −0.3762 | −1.1725, 0.4200 | 0.3565 |
| S | 0.9205 | 0.3334, 2.5415 | 0.7996 | −0.4860 | −1.5421, 0.5742 | 0.5410 |
| LDL‐C dyslipidemia | ||||||
| RERI | 1.4257 | −1.0727, 2.3240 | 0.0097 | 0.6479 | −0.4530, 1.7489 | 0.0243 |
| AP | 0.3987 | −0.0069, 1.004 | 0.0141 | 0.2182 | −0.3147, 0.8510 | 0.0190 |
| S | 1.3704 | 1.2513, 2.5842 | 0.0083 | 1.8894 | 0.7218, 3.0731 | 0.0286 |
AHI, apnea‐hypopnea index; AP, attributable proportion due to interaction; CVD, cardiovascular disease; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; MS, metabolic syndrome; RERI, the relative excess risk due to interaction; S, the synergy index; TG, triglycerides.
The additive interaction was evaluated using the index in Excel sheet written by Andersson. The RERI value is >0 as a result of the additive risk for CVD, and AP represents the percentage of CVD cases that are attributed to the additive interaction among participants exposed to both OSA and metabolic dysfunction. S > 1 indicates that there was positive effect, while S < 1 with negative interaction.
4. DISCUSSION
The present study showed that both AHI and MAI were associated with metabolic dysfunction, including MS components and atherogenic dyslipidemia. OSA apparently augmented the adverse effect of hypertension, abdominal obesity, dyslipidemia and abnormal glucose status on the cardiovascular system. Our study provided a cross‐sectional profile of the interactive effects between different components of cardiometabolic disorders and OSA on CVD risk in an OSA population. In our work, the OSA and related terms definition were provided in table S3.
All MS components except high TG dyslipidemia were independently associated with CVD in the OSA population. We observed no significant association between high TG levels and an increased risk for CVD, which was not consistent with a previous report in a community‐based sample24; however, the independent role of TG as a CVD risk factor has been controversial.25 The reason for this finding may be that the association between high TG and CVD is mediated by other MS components, such as an abnormal HDL‐C in the general population26 and diabetes in the OSA population.27 In this study, independent associations among other MS components were consistent with the results of previous prospective reports.3
In the present study, the multiplicative interaction and additive effect analyses suggested a combined effect of high glucose, hypertension, abdominal obesity, high LDL‐C and OSA on CVD. The underlying mechanisms explaining the associations of OSA and metabolic dysfunction with CVD are not entirely understood; however, several mechanisms related to CIH have been proposed. For example, the coexistence of severe OSA and MS can exacerbate left ventricular concentric hypertrophy and diastolic dysfunction.28 Ivani reported that patients with MS comorbid for OSA exhibited increased sympathetic activity and reduced baroreflex sensitivity compared with MS patients without OSA.29 OSA also results in persistently increased sympathetic activity, even during the day and in the absence of any comorbidities.30 Oxidative stress is suspected to be involved in the pathophysiology of OSA and has been linked to oxidized LDL‐C and MS8 Furthermore, MS impairs endothelial function, reduces the ability of baroreceptors in blood vessels to activate the sympathetic nervous system, decreases the distensibility of large elastic arteries, and disrupts BP regulation.31
On the other hand, previous work reported that fragmented sleep was positively associated with TC and LDL‐C levels, heart rate, blood pressure, and cortisol levels in a general population.32 In our work, sleep fragmentation plays an important role in lipid homeostasis in addition to nocturnal hypoxia in OSA patients. Sleep fragmentation causes systemic inflammation and increases in adrenocorticotropic hormones and cortisol, which can result in changes in lipid metabolism.33, 34 Although evidence is lacking to evaluate whether sleep deprivation is directly associated with CVD, previous studies demonstrated a significant association between sleep fragmentation and endothelial dysfunction.35 Thus, it is biologically plausible that OSA combined with cardiometabolic disorders may predispose an individual to CVD in part due to a common pathophysiological mechanism related to sleep disorder.
From a clinical perspective, our results suggest that the association between CVD and OSA may be partly explained by metabolic dysfunction, and this view is further supported by a randomized trial demonstrating that CPAP monotherapy was not more cost‐effective with respect to treating insulin sensitivity, dyslipidemia, BP, or obesity vs combined therapy.36 To effectively manage OSA, it is crucial to identify the relationship between clusters of metabolic risk factors and OSA.37 Treatment strategies for OSA that are tailored according to the complications that are present could lead to improved CVD outcomes in patients with OSA.18 However, this hypothesis is currently speculative, and more studies are warranted to precisely explore the pathophysiology of CVD in OSA patients.
Some limitations to our study should be discussed. First, because this study was cross‐sectional, temporality was unclear and observational. Thus, causality cannot be inferred. Second, despite adjusting for several common confounders, other more complex factors were not considered, such as lifestyle factors of exercise and dietary habits. Third, the study sample largely comprised of young and middle‐aged men with severe OSA. Thus, the results are mainly applicable to these populations given that the metabolic profile of men and women may be different. Fourth, the first night effect may prevent the PSG from reflecting the real sleep condition at home. Finally, diagnostic information from the hospital and physician’s office was collected and self‐reported. We could not further explore the relation between the diagnostic information, such as echocardiographic parameters, and OSA to confirm our conclusion. However, these limitations did not compromise the value of our work. Further studies on the effect of the interaction between OSA and cardiometabolic disorders are required.
5. CONCLUSIONS
In this study, cardiometabolic disorders and OSA were synergistically associated with an increased risk of CVD, possibly through the common pathophysiology related to the hypoxemia and sleep fragmentation. These results suggest that cardiometabolic disorders and OSA may interact via the mechanisms partly related to hypoxia and sleep fragmentation, thus leading to CVD. OSA therapy for CVD should be tailored based on metabolic comorbidities and OSA.
CONFLICT OF INTEREST
The authors declare no competing financial interests.
AUTHORS' CONTRIBUTIONS
Prof. Shankai Yin, Jian Guan and HongliangYi had full access to all of the data in the study and took responsibility for the integrity of the data and the accuracy of the data analysis. Study design: Xiaolong Zhao, Jian Guan, Hongliang Yi and Shankai Yin; Data collection: Huajun Xu, Xinyi Li, Yingjun Qian; Statistical analysis: Huajun Xu; Manuscript draft: Xiaolong Zhao, Jian Guan, Hongliang Yi and Shankai Yin.
Supporting information
ACKNOWLEDGMENTS
This study was supported by grants‐in‐aid from National Key R&D Program of China (2017YFC0112500); National Natural Science Foundation of China (81770987, 81700896, 81701306, 81770988); Innovation Program of Shanghai Municipal Education Commission (2017‐01‐07‐00‐02‐E00047); multi‐center clinical research project from school of medicine, Shanghai Jiao Tong University (DLY201502) and Shanghai Shen‐Kang Hospital Management Center Project (SHDC12015101).
Zhao X, Li X, Xu H, et al. Relationships between cardiometabolic disorders and obstructive sleep apnea: Implications for cardiovascular disease risk. J Clin Hypertens. 2019;21:280–290. 10.1111/jch.13473
Xiaolong Zhao and Xinyi Li contributed equally to this paper.
Contributor Information
Jian Guan, Email: guanjian0606@sina.com.
Shan kai Yin, Email: skyin@sjtu.edu.cn.
REFERENCES
- 1. Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The occurrence of sleep‐disordered breathing among middle‐aged adults. N Engl J Med. 1993;328:1230‐1235. [DOI] [PubMed] [Google Scholar]
- 2. Coughlin SR, Mawdsley L, Mugarza JA, Calverley PM, Wilding JP. Obstructive sleep apnoea is independently associated with an increased prevalence of metabolic syndrome. Eur Heart J. 2004;25:735‐741. [DOI] [PubMed] [Google Scholar]
- 3. Lakka HM, Laaksonen DE, Lakka TA, et al. The metabolic syndrome and total and cardiovascular disease mortality in middle aged men. JAMA. 2002;288:2709‐2716. [DOI] [PubMed] [Google Scholar]
- 4. Gottlieb DJ, Yenokyan G, Newman AB, et al. Prospective study of obstructive sleep apnea and incident coronary heart disease and heart failure: the Sleep Heart Health Study. Circulation. 2010;122:352‐360. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Shahar E, Whitney CW, Redline S, et al. Sleep‐disordered breathing and cardiovascular disease: cross‐sectional results of the Sleep Heart Health Study. Am J Respir Crit Care Med. 2001;163:19‐25. [DOI] [PubMed] [Google Scholar]
- 6. Valenza MC, Baranchuk A, Valenza‐Demet G, Muñoz‐Casaubon T, Martin‐Navajas JA, Healey J. Prevalence of risk factors for atrial fibrillation and stroke among 1210 patients with sleep disordered breathing. Int J Cardiol. 2014;174:73‐76. [DOI] [PubMed] [Google Scholar]
- 7. Wilcox I, McNamara SG, Collins FL, Grunstein RR, Sullivan CE. Syndrome Z: the interaction of sleep apnoea, vascular risk factors and heart disease. Thorax. 1998;53:S25‐S28. [PMC free article] [PubMed] [Google Scholar]
- 8. Feres MC, Fonseca FA, Cintra FD, et al. An assessment of oxidized LDL in the lipid profiles of patients with obstructive sleep apnea and its association with both hypertension and dyslipidemia, and the impact of treatment with CPAP. Atherosclerosis. 2015;241:342‐349. [DOI] [PubMed] [Google Scholar]
- 9. Murphy AM, Thomas A, Crinion SJ, et al. Intermittent hypoxia in obstructive sleep apnoea mediates insulin resistance through adipose tissue inflammation. Eur Respir J. 2017;4:pii: 1601731. [DOI] [PubMed] [Google Scholar]
- 10. Trzepizur W, Le Vaillant M, Meslier N, et al. Independent association between nocturnal intermittent hypoxemia and metabolic dyslipidemia. Chest. 2013;143:1584‐1589. [DOI] [PubMed] [Google Scholar]
- 11. Levy P, Tamisier R, Arnaud C, et al. Sleep deprivation, sleep apnea and cardiovascular diseases. Front Biosci (Elite Ed). 2012;4:2007‐2021. [DOI] [PubMed] [Google Scholar]
- 12. Cai A, Zhang J, Wang R, et al. Joint effects of obstructive sleep apnea and resistant hypertension on chronic heart failure: a cross sectional study. Int J Cardiol. 2018;257:125‐130. [DOI] [PubMed] [Google Scholar]
- 13. Wang L, Cai A, Zhang J, et al. Association of obstructive sleep apnea plus hypertension and prevalent cardiovascular diseases: a cross‐sectional study. Medicine (Baltimore). 2016;95:e4691. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Abuzaid AS, Al Ashry HS, Elbadawi A, et al. Meta‐analysis of cardiovascular outcomes with continuous positive airway pressure therapy in patients with obstructive sleep apnea. Am J Cardiol. 2017;120:693‐699. [DOI] [PubMed] [Google Scholar]
- 15. Coughlin SR, Mawdsley L, Mugarza JA, Wilding JP, Calverley PM. Cardiovascular and metabolic effects of CPAP in obese males with OSA. Eur Respir J. 2007;29:720‐727. [DOI] [PubMed] [Google Scholar]
- 16. Robinson GV, Langford BA, Smith DM, Stradling JR. Predictors of blood pressure fall with continuous positive airway pressure (CPAP) treatment of obstructive sleep apnoea (OSA). Thorax. 2008;63:855‐859. [DOI] [PubMed] [Google Scholar]
- 17. Kuna ST, Reboussin DM, Borradaile KE, et al. Long‐term effect of weight loss on obstructive sleep apnea severity in obese patients with type 2 diabetes. Sleep. 2013;36:641‐649. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Hudgel DW. Critical review: CPAP and weight management of obstructive sleep apnea cardiovascular co‐morbidities. Sleep Med Rev. 2018;37:14‐23. [DOI] [PubMed] [Google Scholar]
- 19. National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) . Third Report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III) final report. Circulation 2002;106:3143‐3421. [PubMed] [Google Scholar]
- 20. Bonora E, Targher G, Alberiche M, et al. Homeostasis model assessment closely mirrors the glucose clamp technique in the assessment of insulin sensitivity: studies in subjects with various degrees of glucose tolerance and insulin sensitivity. Diabetes Care. 2000;23:57‐63. [DOI] [PubMed] [Google Scholar]
- 21. Berry RB, Budhiraja R, Gottlieb DJ, et al. Rules for scoring respiratory events in sleep: update of the 2007 aasm manual for the scoring of sleep and associated events. Deliberations of the sleep apnea definitions task force of the American Academy of Sleep Medicine. J Clin Sleep Med. 2012;8:597‐619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Andersson T, Alfredsson L, Källberg H, Zdravkovic S, Ahlbom A. Calculating measures of biological interaction. Eur J Epidemiol. 2005;20:575‐579. [DOI] [PubMed] [Google Scholar]
- 23. Nagaya T, Yoshida H, Takahashi H, Kawai M. Resting heart rate and blood pressure, independent of each other, proportionally raise the risk for type‐2 diabetes mellitus. Int J Epidemiol. 2010;39:215‐222. [DOI] [PubMed] [Google Scholar]
- 24. Nordestgaard BG, Benn M, Schnohr P, Tybjaerg‐Hansen A. Nonfasting triglycerides and risk of myocardial infarction, ischemic heart disease, and death in men and women. JAMA. 2007;298:299‐308. [DOI] [PubMed] [Google Scholar]
- 25. Durrington PN. Triglycerides are more important in atherosclerosis than epidemiology has suggested. Atherosclerosis. 1998;141:S57‐S62. [DOI] [PubMed] [Google Scholar]
- 26. Quispe R, Manalac RJ, Faridi KF, et al. Relationship of the triglyceride to high‐density lipoprotein cholesterol (TG/HDL‐C) ratio to the remainder of the lipid profile: the Very Large Database of Lipids‐4 (VLDL‐4) study. Atherosclerosis. 2015;242:243‐250. [DOI] [PubMed] [Google Scholar]
- 27. Liu A, Cardell J, Ariel D, et al. Abnormalities of lipoprotein concentrations in obstructive sleep apnea are related to insulin resistance. Sleep. 2015;38:793‐799. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Usui Y, Takata Y, Inoue Y, et al. Coexistence of obstructive sleep apnoea and metabolic syndrome is independently associated with left ventricular hypertrophy and diastolic dysfunction. Sleep Breath. 2012;16:677‐684. [DOI] [PubMed] [Google Scholar]
- 29. Trombetta IC, Somers VK, Maki‐Nunes C, et al. Consequences of comorbid sleep apnea in the metabolic syndrome–implications for cardiovascular risk. Sleep. 2010;33:1193‐1199. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Henderson LA, Macefield VG. Obstructive sleep apnoea and hypertension: the role of the central nervous system. Curr Hypertens Rep. 2016;18:59. [DOI] [PubMed] [Google Scholar]
- 31. Jean‐Louis G, Zizi F, Clark LT, Brown CD, McFarlane SI. Obstructive sleep apnea and cardiovascular disease: role of the metabolic syndrome and its components. J Clin Sleep Med. 2008;4:261‐272. [PMC free article] [PubMed] [Google Scholar]
- 32. Ekstedt M, Akerstedt T, Söderström M. Microarousals during sleep are associated with increased levels of lipids, cortisol, and blood pressure. Psychosom Med. 2004;66:925‐931. [DOI] [PubMed] [Google Scholar]
- 33. Trammell RA, Verhulst S, Toth LA. Effects of sleep fragmentation on sleep and markers of inflammation in mice. Comp Med. 2014;64:13‐24. [PMC free article] [PubMed] [Google Scholar]
- 34. Späth‐Schwalbe E, Gofferje M, Kern W, Born J, Fehm HL. Sleep disruption alters nocturnal ACTH and cortisol secretory patterns. Biol Psychiatry. 1991;29:575‐584. [DOI] [PubMed] [Google Scholar]
- 35. Kohansieh M, Makaryus AN. Sleep deficiency and deprivation leading to cardiovascular disease. Int J Hypertens. 2015;2015:615681. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Chirinos JA, Gurubhagavatula I, Teff K, et al. CPAP, weight loss, or both for obstructive sleep apnea. N Engl J Med. 2014;370:2265‐2275. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Joosten SA, Hamilton GS, Naughton MT. Impact of weight loss management in OSA. Chest. 2017;152:194‐203. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials