Frequent cyclic variation of heart rate is associated with left ventricular diastolic dysfunction in patients without ischemia

Abstract Background Cyclic variation of heart rate (CVHR) associated with sleep‐disordered breathing reflects cardiac autonomic responses to apneic/hypoxic stress. However, the association of CVHR with cardiac function is unclear. Methods We investigated a total of 181 patients who underwent both 24‐hour Holter electrocardiography (ECG) and quantitative gated single‐photon emission computed tomography (SPECT) myocardial functional imaging, excluding patients with atrial fibrillation, myocardial infarction, structural heart disease, and implantable devices, from January 2017 to July 2018. The number of CVHR per hour (CVHR index) in sleeping‐time Holter ECG was compared with the parameters of left ventricular (LV) systolic and diastolic functions assessed by cardiac SPECT functional imaging, peak filling rate (PFR), first‐third mean filling rate (1/3 MFR), and time to peak filling rate (TTPF). Results In all patients, the CVHR index was not associated with any parameters of cardiac functions. However, in a propensity score–matched subgroup of patients without ischemia (N = 39), the CVHR index was negatively correlated with PFR (r = −0.35, P < .05) and 1/3 MFR (r = −0.37, P < .05) but positively correlated with TTPF (r = 0.43, P < .01) and was not correlated with LV ejection fraction. Multivariate linear regression analysis revealed that high CVHR index was independently associated with LV diastolic dysfunction, even after adjusting for the relative wall thickness and LV mass index assessed by echocardiography. Conclusion These results indicate that the high frequency of CVHR in sleeping time is associated with LV diastolic dysfunction in nonischemic patients, irrespective of LV geometry.


| INTRODUCTION
Sleep-related breathing disorders are associated with cardiac diastolic dysfunction. Previous studies demonstrated that apnea-hypopnea index (AHI) in patients with obstructive sleep apnea (OSA) was correlated with left ventricular (LV) diastolic dysfunction evaluated by echocardiography. 1,2 However, polysomnography, which is the gold standard for diagnosis, is rather troublesome because the test includes electroencephalography that requires a one-night stay in a medical facility. Episodes of OSA are accompanied by a characteristic heart rate alteration that consists of bradycardia during apnea, followed by abrupt tachycardia on its cessation, which is known as cyclic variation of heart rate (CVHR). [3][4][5] CVHR creates dips in the R-R interval series. Recently, CVHR has been available by analyzing 24-hour Holter electrocardiography (ECG). CVHR index, which is the CVHR per hour in sleeping-time Holter ECG, has been closely correlated with AHI. [5][6][7] The automated detection of CVHR index from Holter ECG provides a powerful screening tool for OSA, even in older patients and in those with dementia and cardiac autonomic dysfunction. However, it is unclear whether CVHR is associated with cardiac function. This study aimed to investigate whether CVHR is associated with LV systolic or diastolic function.
Furthermore, myocardial ischemia itself causes not only LV systolic dysfunction but only LV diastolic dysfunction. [8][9][10][11][12] Therefore, the comorbidity of myocardial ischemia was expected to mask the relationship between CVHR and LV diastolic dysfunction. To address this issue in this study, we also investigated the subgroup analysis in patients with and without myocardial ischemia.

| Study population and ethical considerations
We retrospectively investigated a total of 181 hospitalized patients who underwent Holter ECG, echocardiography, and quantitative gated singlephoton emission computed tomography (SPECT) from January 2017 to July 2018, after excluding patients who met the following criteria: age <20 years, atrial fibrillation and complete atrioventricular block at night, implantable devices, myocardial infarction, and structural heart disease.
Since the LV diastolic function is impaired in an ischemic heart, to examine the impact of ischemia on cardiac function, we divided the 181 patients into the ischemic heart disease (IHD) group and nonischemic heart disease (non-IHD) group ( Figure 1). IHD group included 142 patients with either significant coronary stenosis confirmed by coronary angiography or multidetector computed tomography or history of coronary revascularization, such as percutaneous coronary intervention and coronary artery bypass graft, whereas other patients were assigned to non-IHD group. This study and the retrospective data used were approved by the Bioethical Committee of Noto General Hospital and conducted in accordance with the ethical standards in the 1964 Declaration of Helsinki and its later amendments.

| Assessment of CVHR
ECG signals were obtained via the Holter LS-300 system at a sampling frequency of 125 Hz. A 24-hour Holter automatic arrhythmia analysis system (SCM-850S, Fukuda Denshi) identified all R-wave positions and excluded abnormal beats, such as ventricular and supraventricular ectopic complexes and artifacts. The algorithm to detect CVHR automatically was prepared as previously reported. 6 The CVHR index was calculated as the mean number of CVHR per hour in bed, which was obtained from the behavior records of subjects.

| Evaluation of quantitative gated SPECT findings
ECG-gated single-photon emission computed tomography (SPECT) was performed at rest and during sinus rhythm using thallium-201 chloride (111 MBq) or technetium-99 m tetrofosmin (740 MBq) as a tracer. At 60 minutes after intravenous tracer injection, ECG-gated SPECT data were obtained for 60 seconds per projection from 30 projections during a 180 rotation using a two-headed gamma camera with a low-energy, high-resolution parallel-hole collimator. A cardiac cycle was divided into 16 frames. The data were stored in a 64 Â 64-word matrix nuclear computer system, and no attenuation or scatter correction was applied to this protocol. For the data analysis, the quantitative gated SPECT program (Cedars-Sinai Medical Center, Los Angeles, California) was applied to process short-axis tomograms to determine the LV end diastolic volume, LV end-systolic volume, and LV ejection fraction (LVEF). Peak filling rate (PFR), first-third mean filling rate (1/3 MFR), and time to peak filling rate (TTPF) were also obtained as the diastolic functional parameters. 14,15

| Statistical analysis
Normally distributed continuous variables were presented as mean ± standard deviation (SD) for each group unless stated and analyzed by unpaired t-test for two-group comparisons. Categorical variables were compared using the chi-square test. Pearson's correlation analysis was used to evaluate the correlations between the parameters.
The multivariable linear regression analysis was employed to identify the independent parameter for impaired LV diastolic function. A propensity score was used to match the subgroups in a 1:1 ratio and T A B L E 1 Clinical characteristics of total and a propensity score-matched comparable subgroups Propensity score-matched subgroup

| RESULTS
The flowchart of this study is shown in Figure 1. The clinical characteristics of the 181 patients of this study are presented in Table 1 and Table S1. 38% of patients had diabetes, and 88% had hypertension.
The Holter ECG showed that the mean CVHR index was 15 per hour.
The average values of LVMI and RWT were within the normal range.
In these subjects, the CVHR index was associated with E/A ratio but not associated with DcT, left atrial diameter (LAD), LVEF, PFR, 1/3 MFR, or TTPF ( Figure 2).
There were statistical differences in some parameters of clinical backgrounds between the non-IHD group and IHD groups (Table S1).
To avoid selection bias, we used propensity score matching to select comparable groups of age, sex, BMI, hypertension, and diabetes. Thus, each of the 39 patients was selected for further analysis. There was no statistical difference in any parameters of clinical backgrounds between the non-IHD group and IHD groups (Table 1). In the propensity score-matched subgroup of patients without IHD, the CVHR index was negatively correlated with E/A ratio, PFR, and 1/3 MFR but positively correlated with TTPF and was not associated with DcT, Lad, and LVEF ( Figure 3). Meanwhile, in the subgroup of patients with IHD, the CVHR index was not associated with E/A ratio, DcT, LAd, LVEF, PFR, 1/3 MFR, or TTPF ( Figure 4). F I G U R E 2 Relationship between CVHR index and LV systolic and diastolic function in all subjects. CVHR, cyclic variation of heart rate; DcT, declaration time; E/A, early to late (atrial) diastolic transmural flow velocity; LVEF, left ventricular ejection fraction; 1/3 MFR, first-third mean filling rate; PFR, peak filling rate; r, correlation coefficient; TTPF, time to peak filling rate revealed that only CVHR index was a significant independent parameter that was associated with decreased PFR and 1/3 MFR or increased TTPF after adjustment for LVMI and RWT.

| DISCUSSION
The major findings of this retrospective study are as follows: Heart failure with preserved ejection fraction (HFpEF) is a clinical syndrome in which patients have symptoms and signs of heart failure with normal or near-normal LVEF. Now, it accounts for 56% of patients with heart failure, and its prevalence is increasing. 17 LV diastolic dysfunction is one of the primary pathophysiological characteristics that underlie HFpEF.
OSA has been associated with diastolic dysfunction and increased LVMI, which is the result from episodes of repetitive hypoxia during sleep and increases in afterload. 18 Holter ECG was reported to be closely correlated with AHI in full polysomnography. [5][6][7] (A) (B) P P P P P P P F I G U R E 4 Relationship between CVHR index and LV systolic and diastolic function in the subgroup of ischemic patients. CVHR, cyclic variation of heart rate; DcT, declaration time; E/A, early to late (atrial) diastolic transmural flow velocity; IHD, ischemic heart disease; LAd, left atrial diameter; LVEF, left ventricular ejection fraction; 1/3 MFR, first-third mean filling rate; PFR, peak filling rate; r, correlation coefficient; TTPF, time to peak filling rate T A B L E 2 Univariable and multivariable linear regression analyses for E/A, PFR, 1/3 MFR, and TTPF in patients without IHD The LV diastolic dysfunction of the local myocardium exposed to ischemia due to organic coronary artery stenosis, now on, in the past, is thought to persist regardless of the presence of OSA. 8,11,12 Previous studies demonstrating the relationship between AHI and LV diastolic dysfunction were excluded patients with a history of cardiovascular disease. 1,2 Therefore, the subjects in the previous studies are similar to those in the non-IHD group of our study. On the other hand, hypertension and hyperglycemia are individually important risk factors for LV hypertrophy. 25 Another report showed that severe OSA and metabolic syndrome could cause LV diastolic dysfunction individually and synergistically. 26 No articles have directly been associated with AHI in patients with ischemic heart disease, but myocardial ischemia might be considered an influential factor enough to mask the contribution of CVHR/OSA to LV diastolic function.
Echocardiography is noninvasive and frequently used for evaluation of LV function; however, the intra-and interoperator variation in measurement tend to be larger. In contrast, PFR, 1/3 MFR, and TTPF obtained from gated SPECT have been highly correlated with LV end diastolic pressure (LVEDP), which was measured using the invasive method of cardiac catheterization. 27

| CONCLUSION
High frequency of CVHR in sleeping time is associated with LV diastolic function in nonischemic patients, irrespective of LV geometry.
CVHR determined by Holter ECG may be a useful screening index that reflects early LV diastolic dysfunction in patients without structural and ischemic heart disease.

CONFLICT OF INTEREST
The authors have no conflict of interest.
Writing-review and editing: Manabu Nakano.
All authors have read and approved the final version of the manuscript.
Dr Takanori Yaegashi had full access to all of the data in this study and takes complete responsibility for the integrity of the data and the accuracy of the data analysis.

TRANSPARENCY STATEMENT
Takanori Yaegashi affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.