Methods

This is a longitudinal observational study of adult patients residing in Ohio and diagnosed with SDB by a sleep physician from April 2002 to July 2011 as a result of a minimum 1 overnight polysomnogram performed within the Cleveland Clinic Foundation. Inclusion criteria were as follows: (1) uninterrupted care within the Cleveland Clinic Foundation at least 6 months before and during the follow-up period after the polysomnogram and (2) completion of an echocardiographic examination within 1 month of the baseline polysomnogram recorded in the Cleveland Clinic prospective echocardiogram registry. Exclusion criteria included EF of <50%, patients with a history of HF, patients with moderate or more valvular regurgitation, those residing outside Ohio, and patients without an echocardiographic examination within 1 month of the polysomnogram evaluation. The entry date in the study was set at the date of polysomnogram, and all patients were followed up until incident HF or death occurred or until the last date of a physical examination performed by a physician within the Cleveland Clinic Foundation system up to September 2013. Institutional review board approval was obtained, and data were collected using multiple sources. Sleep evaluation included a sleep history and a routine overnight polysomnogram attended by a sleep technologist, reviewed by a board certified sleep physician. Apnea/hypopnea was defined by a minimum 50% decrease from the baseline airflow pressure signal excursions, lasting at least 10 seconds and with at least 3% desaturations from preevent baseline or/and association with an arousal (minimum 3-second awaking). Apnea/Hypopnea Index (AHI) was defined as the total number of respiratory events per hour of sleep, and SDB was defined based on an AHI ≥5 events/hour of sleep.

Comprehensive baseline transthoracic echocardiograms were performed with commercially available systems. Measurements performed for clinical purposes and available in the registry included left ventricular EF (LVEF), LV, and left atrial (LA) volumes (biplane Simpson method). Standard echocardiographic measurements of the right-sided heart function were made in accordance with current guidelines. Right ventricular (RV) fractional area change was defined using the formula ([end-diastolic area − end-systolic area]/end-diastolic area) × 100. Tricuspid annular plane systolic excursion was measured as the distance of systolic movement of the junction between the tricuspid valve and the RV free wall. Systolic pulmonary artery pressure was estimated from the maximal continuous-wave Doppler velocity of the tricuspid regurgitation jet using systolic transtricuspid pressure gradient calculated by the modified Bernoulli equation and the addition of estimated right atrial pressure. An index of pulmonary vascular resistance (PVR) was derived by dividing the maximal velocity of the tricuspid regurgitation jet by the RV outflow tract velocity − time integral. All measurements were performed and averaged over 3 cardiac cycles.

Baseline and follow-up clinical data were collected using multiple sources, and in conformity with definitions and recommendations published by the American College of Cardiology/American Heart Association in 2013: data elements and definitions and/or with 2010 American College of Cardiology/American Heart Association HF Performance Measurement Set. Clinical data available in the prospective echocardiographic registry were merged with data obtained during the baseline polysomnogram evaluation, by manual abstraction from comprehensive physician sleep study reports, and with socioeconomic and clinical characteristics recorded in Epic through intrainstitutional automated electronic data fields capturing all the aspects of care at any time before and during the follow-up period of the enrolled patients. Comprehensive assessments of co-morbidities at baseline were used as independent covariates for calculating the Charlson Comorbidity Index.

A composite outcome was calculated using incident HF or death. HF was collected from follow-up clinical data, and mortality was determined using telephone visits and the Social Security Death Index obtained from the Ohio Department of Health. HF was defined as meeting stage C or D criteria of classification of HF.

Standard statistical software packages (SAS version 9.3; SAS Institute, Inc., Cary, North Carolina and SPSS version 20.0; SPSS Inc., Chicago) were used for data analysis. “Time zero” was date of polysomnogram. The Kolmogorov–Smirnov test was performed to test the normality of continuous baseline cohort characteristics further summarized as mean ± SD or medians (interquartile range). Frequencies and percentages were calculated for the categorical data. Missing values (<15% and only covariables) were imputed using fivefold multiple imputations with the Markov chain Monte Carlo technique. Bootstrap bagging for variable selection and regression coefficients and their variance-covariance matrix for the final model were subsequently estimated for each imputed data set and further combined using the method of Rubin et al to produce the final estimates. Crude cumulative incidence function estimates of composite HF or death risk were plotted as a step function using the method of Fine and Gray. Cox proportional univariate and multivariate prognostic hazard models predicting HF and death were built. Sequential Cox models were performed to determine the incremental prognostic benefit of LA volume index (LAVi) and PVR over clinical data and LVEF, with incremental prognostic value being defined by a significant increase in global chi-square. Median values of LAVi and PVR were used to divide patients into 2 equal groups for Kaplan–Meier analysis, with survival compared using a 2-sided log-rank test.

Results

Of 375 consecutive patients, identified from billing records, with an echocardiographic examination done within 1 month of an overnight in-lab sleep study from April 2002 to July 2011, 285 patients were diagnosed by a sleep physician with SDB. Of those, 70 patients were excluded due to an HF co-morbidity, and an additional 13 patients were excluded because of reduced LVEF, more than mild valvular disease, or an insufficient follow-up period (<6 months). Baseline characteristics of the final cohort (n = 202) are included in Table 1 . They show an expected prevalence of risk factors and preserved LV function. Eighty-one patients (40%) had severe SDB (AHI ≥30 events/hour). Table 2 demonstrates the correlation between SDB severity (assessed by AHI) and echocardiographic characteristics of interest. Of note, there was no significant correlation between AHI and echocardiographic characteristics except for LV mass index which had a weak correlation with AHI ( r = 0.19, p = 0.007).

All 202 patients were followed up over a period of 3.1 ± 2.4 years (range, 0.5 to 8.4 years). Thirty-four patients (16.8%) reached the composite end point defined as HF or death. There were no documented noncardiac deaths during follow-up. Incident HF occurred in 27 patients (13.4%), and 7 patients died (3.5%) during the study period. In our cohort, event-free survival rate was 98% (4 events total) and 97.5% (5 events in 2 years) at 1 and 2 years, respectively. Hazard ratios (HRs) of the relevant parameters before and after adjustment for age are listed in Table 3 . Interestingly, there was no association between AHI/use of continuous positive airway pressure and clinical outcome shown on Cox proportional hazards models. In multivariable Cox proportional hazard models (after adjustment for age and AHI, Table 4 ), RV end-systolic area (HR 1.3, p = 0.038) and PVR (HR 1.4, p = 0.005) were associated with the composite end point, independent of age and AHI.

Table 3

Univariable associations of event (heart failure/mortality composite outcome) during follow-up in sleep-disordered breathing patients without heart failure at baseline

Variables	Univariate Hazard ratio (95% CI)	p	Age ∗ adjusted Hazard Ratio (95% CI)	p
Apnea hypo index	1.0 (0.99-1.01)	0.787	1.0 (0.99-1.01)	0.713
Use of continuous positive airway pressure	1.9 (0.9-3.7)	0.075	1.8 (0.9-3.6)	0.102
Left ventricular ejection fraction	0.8 (0.6-1.1)	0.110	0.6 † (0.5-0.9)	0.010
Left atrial volume index	1.8 (1.4-2.4)	<0.001	1.7(1.3- 2.3)	<0.001
Left ventricular mass index	1.5 (1.1-2.1)	0.007	1.5(1.1- 2.0)	0.016
Left ventricular end-systolic volume index,	1.0 (0.7-1.4)	0.912	1.1 (0.8- 1.5)	0.625
Left ventricular end-diastolic volume index,	1.1 (0.8-1.5)	0.467	1.2 (0.9-1.7)	0.219
Trans-mitral early diastolic wave,	1.5(1.2-2.0)	0.002	1.5(1.1-2.0)	0.005
Trans-mitral atrial filling wave,	0.7(0.5-1.1)	0.128	0.7(0.5- 0.99)	0.041
Early diastolic / atrial filling ratio	1.5 (1.2- 1.8)	<0.001	1.4 (1.2- 1.7)	<0.001
Trans-mitral early diastolic deceleration time	1.1( 0.8- 1.5)	0.685	1.0 ( 0.7- 1.5)	0.875
Right ventricular end systolic area	1.2 (0.99 -1.5)	0.059	1.3 (1.02- 1.6)	0.035
Right ventricular end diastolic area	1.2 ( 0.9-1.6)	0.142	1.3 (0.97- 1.7)	0.084
Right ventricular fractional area change	0.8 (0.6- 1.04)	0.095	0.8 (0.6- 1.1)	0.118
Tricuspid annular plane systolic excursion	0.8 ( 0.5-1.1)	0.098	0.7 ( 0.5-1.1)	0.092
Pulmonary vascular resistance,	1.3(1.1- 1.6)	0.006	1.4 (1.1-1.7)	0.005

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Effect of Smoking on Infarct Size and Major Adverse Cardiac Events in Patients With Large Anterior ST-Elevation Myocardial Infarction (from the INFUSE-AMI Trial)
2. Risk Factor Optimization and Guideline-Directed Medical Therapy in US Veterans With Peripheral Arterial and Ischemic Cerebrovascular Disease Compared to Veterans With Coronary Heart Disease
3. Frequency of Stent Thrombosis Risk at 5 Years in Women Versus Men With Zotarolimus-Eluting Compared With Sirolimus-Eluting Stent
4. Comparison of Characteristics and Outcomes of Heart Failure Patients With Preserved Versus Reduced Ejection Fraction in a Multiethnic Southeast Asian Cohort

Stay updated, free articles. Join our Telegram channel

Join