Predictor factors of sleep-disordered breathing in heart failure

Background Heart failure (HF) is characterized by its high mortality, frequent hospitalizations, and reduced quality of life. Sleep-disordered breathing (SDB), one of the common comorbidities, accelerates the progression of HF. Objectives The objectives of the study were (a) to investigate the prevalence and type of SDB in HF patients and (b) to determine the predictors of SDB. Materials and methods In a cross-sectional analytic study, all eligible patients of Assiut Chest and Cardiology Department admitted (100 patients) during the period from August 2015 to March 2016 were included in this study. Clinical assessment, full-night attended polysomnography, and echocardiography were recorded and compared between patients with (SDB) (85 patients) and those without SDB (15 patients). Results SDB was found in 85% of patients [53% had obstructive sleep apnea (OSA) and 32% had central sleep apnea (CSA)]. OSA patients are characterized by higher BMI and neck and waist circumference. There was a higher prevalence of hypertension, as well as mean blood pressure, systolic blood pressure, diastolic blood pressure, in OSA patients. Loud snoring was the only clinical symptom associated with OSA as compared with CSA. CSA patients had a significant reduction in PaCO2. OSA patients showed a significant increase in desaturation index and time spent with oxygen saturation less than 90%. Maximum heart rate and brady/tachy index were significantly increased in OSA. Cycle length was significantly increased in CSA. Conclusion The prevalence of sleep apnea was high in patients with stable HF (85%). OSA was the predominant type (53%). The predictors of SDB were BMI (≥30), systemic hypertension, neck circumference more than 40 cm, waist circumference more than 110 cm, and ejection fraction (left ventricular ejection fraction) (≤45%).


Introduction
Heart failure (HF) is a highly prevalent syndrome with an estimated annual incidence of half a million. Despite therapeutic advances in pharmacology, HF remains a leading cause of morbidity, as well as frequent hospitalizations and a reduced quality of life [1]. Sleep-disordered breathing (SDB), one of the most common comorbidities in HF, occurs in 50-80% of patients, and is known to accelerate the progression of the condition, thus increasing morbidity and mortality [2,3].
The present study aimed (a) to investigate the prevalence and type of SDB in HF patients and (b) to determine the predictors of SDB in HF patients.

Materials and methods
This cross-sectional analytic study was carried out in the Chest and Cardiology Department, Faculty of Medicine, Assiut University Hospital, during the period from August 2015 to March 2016.
The study design was approved by the Scientific Ethics Committee of Faculty of Medicine of Assiut University.
Informed consent was obtained from the patients.
Among 100 patients who were admitted during this period and fulfilled the inclusion criteria, only 85 patients had SDB. Patients were classified into two groups: group 1, the SDB group (85 patients), and group 2, the without SDB (no SDB) group (15 patients). Group 1 patients were further subclassified into those with central sleep apnea (CSA) and those with obstructive sleep apnea (OSA) based on the predominant (>50%) type of sleep apnea.

Exclusion criteria
Patients with recent unstable angina or myocardial infarction within 3 months of the study, those with a history of chronic lung disease (i.e. obstructive pulmonary disease), pregnant woman, those with a history of stroke or clinical signs of peripheral or central nervous system disorders were excluded from the study.
All patients were subjected to a comprehensive clinical assessment, an overnight-attended polysomnography, and echocardiography.

Clinical assessment and sleep history
Clinical data of patients were recorded, including age, sex, anthropometric measures (height, weight, BMI, and neck and waist circumference), history of diabetes mellitus, and history of hypertension. New York Heart Association classification (NYHA) functional classification of HF was determined. The standard HF therapy was noted for each patient (angiotensin-converting enzyme inhibitors, β-blockers, diuretics, and amiodarone). Sleep history was as follows: daytime symptoms included morning headache and excessive daytime sleepiness and nighttime symptoms included loud snoring, disturbed sleep rhythm, witnessed apnea, bad dreams, and frequent nocturea). Epworth Sleepiness Score was used in our study. Results of at least 10 are considered to represent significant excessive daytime sleepiness [4].

Echocardiography
All participants underwent transthoracic echocardiography (Vivid S5; GE Healthcare, United States), which was performed in the left lateral decubitus position. Images were obtained in different views (parasternal long-axis as well as apical four chambers, two chambers, and three chambers) according to current standards, by personnel blinded to the results of polysomnography [5].
Left atrial diameter, left ventricular (LV) end-diastolic internal dimension, LV end-systolic internal dimension, interventricular septal thickness, and LV posterior wall thickness were determined from M-mode measurements. In addition, left ventricular ejection fraction (LVEF) was measured according to the modified Simpson's method using apical four and two chamber views. Right ventricle diameter and pulmonary artery systolic pressure (PASP) were also measured. PASP can be determined using the modified Bernoulli equation: (PASP)=4×(peak TR velocity) 2+RAP.
The maximum velocity of the tricuspid valve regurgitant jet (v) is measured using continuous wave Doppler and added to the estimated right atrial pressure (RAP) [6].

Polysomnography
All patients underwent full-night attended polysomnography (Somnstar 4100; Sensor-Medics Co., Yorba Linda, California, USA) in the sleep laboratory of the Assiut university hospital. The polysomnogram systematically monitors the electroencephalogram (C3-A2, C4-A1), electro-oculogram, electromyogram of the chin, electrocardiogram, body positions, nasal and oral airflow, thoracic and abdominal effort, limb movements, pulse oximetry, and snoring sound level. The polysomnograms were all scored manually according to the American academy of sleep medicine [7].

Cheyne-Stokes respiration
Cheyne-Stokes respiration is defined as three consecutive central apneas and/or central hypopneas separated by a crescendo and decrescendo change in breathing amplitude with a cycle length of at least 40 s and the central apnea index of at least 5 per hour [7]. Cycle length is the duration of the central apnea plus the duration of a respiratory phase. If central hypopneas occur, the cycle length is defined as the time from the zenith in the respiratory phase preceding the central hypopnea to the zenith of the next respiratory phase [7]. Minimum and maximum heart rate was recorded. Brady/tachy index, which is the average number of bradycardia and tachycardia events per hour, was also recorded.

Statistical analysis
Data were analyzed using SPSS (Statistical Package for Social Science), version 16 (IBM Inc., Armonk, NewYork, USA) software was used for analysis of results. The results in this study were expressed as mean±SD or number and percentage. Comparison between two groups was made using the t-test and the one-way analysis of variance test for comparison between more than two groups. The difference was considered significant when P value of less than 0.05. Logistic regression analysis was performed to detect predictors of SDB in HF. Table 1 shows demographic data and clinical examination in HF patients with SDB and those no SDB. Group 1 patients with SDB had significantly higher age group, BMI, neck and waist circumference, and serum cholesterol level. Moreover, the prevalence of hypertension, as well as mean blood pressure (MBP), systolic blood pressure (SBP), and diastolic blood pressure (DBP), was significantly increased in patients with SDB.

Results
As shown in Table 2, patients with OSA had a significant increase in BMI and the prevalence of hypertension (MBP, SBP and DBP) as compared with the CSA and the no SDB group. However, cases with CSA recorded a statistically significant difference in age, neck and waist circumference, and serum cholesterol level as compared with the no SDB group. Table 3, CSA was significantly higher in NYHA functional class (III and IV) as compared with the OSA and the no SDB group. It was observed that the Epworth Sleepiness Scale was significantly higher in OSA patients. OSA patients showed a significant increase in daytime and night-time symptoms of sleep apnea compared with no SDB patients. The prevalence of loud snoring was significantly increased in patients with OSA as compared with CSA patients. Table 4 shows that CSA patients had a significant respiratory alkalosis and a decrease in the mean value of PaCO 2 associated with marked atrial fibrillation (AF) as compared with OSA and no SDB.

As shown in
As regards polysomnographic parameters (Table 5), the respiratory disturbance index (RDI) was significantly increased in OSA patients. Patients with CSA had a significant increase in the Cheyne-Stokes respiration index as well as cycle length when compared with the OSA and the no SDB group.
As regards respiratory oxygenation during sleep, patients with OSA had a significantly more severe arterial  oxyhemoglobin desaturation and time spent with oxygen saturation below 90%. Moreover, a significant reduction in minimum and average oxygen saturation was observed in OSA patients when compared with CSA and no SDB patients. As regards heart rate recording during sleep, OSA patients showed a significant reduction in minimum heart rate and a significant increase in maximum heart rate and brady/tachy index.
As shown in Table 6, LVEF was significantly reduced in patients with CSA as compared with OSA and no SDB patients. Patients with HF reduced ejection fraction (≤45%) had a significantly higher prevalence of CSA, whereas HF patients with preserved ejection fraction (>45%) had a significantly higher prevalence of OSA.
As regards LV dimensions, CSA patients had a significant increase in LV end diastolic and LV end systolic diameter, whereas in patients with OSA, both posterior wall diameter and interventricular septal diameter were significantly increased. There was no significant difference in PASP between both OSA and CSA. OSA was significantly higher in patients with hypertensive heart disease, whereas CSA had a significantly higher prevalence in ischemic heart disease and dilated cardiomyopathy. Table 7 shows that the possible predictive factors of SDB in HF were BMI of at least 30, systemic hypertension, neck circumference more than 40 cm, waist circumference more than 110 cm, LVEF (≤45%), NYHA functional class III and IV, AF, male sex, and age more than 50 years.

Discussion
In the current study, we have studied the prevalence of SDB in HF, and found that 85% of patients with HF had SDB; OSA accounted for 53%, whereas CSA sleep apnea accounted for (32%). Our results are consistent    reported a lower prevalence of SDB, with a clear predominance of CSA over OSA. Javaheri and colleagues [11] reported that the prevalence of SDB in 81 outpatients with stable HF was 51% (10% had OSA and 40% had CSA). Moreover, Vazir and colleagues [12] demonstrated that the prevalence of SDB was 53% (15% OSA and 38% CSA). This variability in the reported prevalence of SDB may be attributed to the small number of patients involved, or the use of different monitoring techniques to measure airflow. SDB in those studies was evaluated with cardiorespiratory polygraphy instead of standard polysomonography.
As regards patient characteristics of SDB in HF, we observed that patients with SDB were characterized by old age (56.35±7.23vs. 51.13±4.32 years in no SDB) and male sex (63.5 vs. 36.5%). Similarly, Paulino and colleagues [8] demonstrated that HF patents with SDB had significantly older age group (mean age; 60± 13 vs. 55±14 years in no SDB) and male sex (86 vs.67%). Arzt and colleagues [13] also reported a higher prevalence of SDB in men than in women (49 vs. 36%) and also found that the rate of SDB increased as age increased (the prevalence in patients≤50 years was 31%, whereas in those>80 years the prevalence was 59%). The sex-related differences may be attributed to differences in sleep characteristics between the two sexes; men have less stable sleep architecture compared with women, with a greater number of sleep-wake transitions and shorter slow wave sleep, which may predispose to respiratory control system instability and central apneas [14].  Our results revealed that neck circumference, which is an indirect measure of fat deposition around the neck has been shown to be an important and independent predictor of SDB among HF patients (42.07±2.50 vs. 39.87±1.51 cm in no SDB). Similarly, Ferreira and colleagues [9], who studied 103 patients with HF to identify the possible risk factors for sleep apnea in HF patients reported that an increased neck circumference has been associated with an increased risk for SDB in HF (41.0 vs. 38.0 cm in no sleep apnea, P=0.003). Moreover, Loo and colleagues [18], who studied 332 HF patients (mean age: 62±10 years) to determine a possible predictor of SDB, demonstrated that neck circumference is one of the predictors of SDB among HF patients (OR=1.06; 95% CI 1.02-1.10; P=0.001). This can be attributed to the fact that increased fat around the neck, related to an android pattern of adiposity, may contribute to pharyngeal narrowing and collapsibility. Another additional mechanism, the redistribution of excess fluid from the legs to the neck during sleep, resulting in increased peripharyngeal edema and cervical congestion, compresses the upper airway and obstructs airflow [19].
Waist circumference is another predictor, which has been found as a risk factor for SDB among HF patients. Our findings are similar to those of Loo and colleagues [18], who found that waist circumference (OR=1.06; 95% CI 1.01-1.13, P=0.001) is one of the independent predictors of SDB among HF patients.
In our study, we observed that the prevalence of systemic hypertension was significantly higher in patients with OSA (62.3 vs. 20.0% in no SDB), with a significant increase in the mean value of SBP, DBP, and MBP. This is in accordance with the findings of Logan and colleagues [20], who demonstrated that, in medically treated patients with HF, the prevalence of systolic hypertension was significantly increased in patients with OSA (83%), with a significant increase in daytime SBP (190.43±43.12 vs. 140.54±23.11 mmHg in no SDB). OSA plays a linking role between hypertension and HF. Intermittent apnea-related hypoxia, as well as the intrathoracic pressure swings and arousals associated with respiratory events, augment sympathetic nervous system activity, which may contribute to elevated blood pressure [21].
On comparing clinical symptoms suggestive of sleep apnea between CSA and OSA, our study revealed that loud snoring was the prevalent night-time symptom in patients with OSA (56 vs. 33.3% as compared with CSA). Similarly, Javaheri and colleagues [21] found that 78% of patients with OSA had nocturnal snoring, and only 28% of patients with CSA had nocturnal snoring. Moreover, Bitter and colleagues [22] investigated more than 1500 HF (LVEF ≤45%, NYHA ≥II) patients for possible symptoms of sleep apnea and found that snoring was the only symptom independently associated with OSA (66 vs. 25%). In

Conclusion
The prevalence of sleep apnea was high in patients with stable HF despite optimized therapy (85%). OSA was the predominant type (53%). The possible predictive factors of SDB in HF were BMI of at least 30, systemic hypertension, neck circumference more than 40 cm, waist circumference (>110 cm), and LVEF (≤45%).