Hypoxia-associated Component of RR-interval Fluctuations in Patients with Obstructive Sleep Apnea Syndrome

An RR-interval fluctuation has been used as an important clinical tool for identifying a patient at risk in cardiovascular disease. However, the role of hypoxia on RR-interval fluctuations has not been determined. Methods and Results. We performed ambulatory ECGs monitoring and measured overnight arterial oxygen saturation (SpO2) in 26 patients with obstructive sleep apnea syndrome in whom sever hypoxia occurred during sleep. By means of maximal entropy method, time series of ECG-RR intervals were transformed into frequencies. The minimal SpO2 during sleep were compared with spectrum powers of the following frequency ranges; 1) 0.0001 0.05 Hz, 2) 0.05 0.1 Hz, 3) 0.1 0.15 Hz, 4) 0.15 0.2 Hz, 5) 0.2 0.25 Hz, 6) 0.25 0.3 Hz, and 7) 0.3 0.5 Hz. Among the seven analyzed frequencies, the increase in the 0.05 0.1 Hz power of RR-interval fluctuations was linearly correlated with the minimal SpO2 during sleep (r=.80, p<.001). Conclusion. This study revealed that hypoxia contributes to 0.05 0.1 Hz RR-interval fluctuations in patients with obstructive sleep apnea syndrome. RR-interval fluctuations will provide important information about hypoxia.


Introduction
RR-interval fluctuations have provided important clinical information [1][2][3][4][5] for identifying patients at risk in cardiovascular disease. The DIAMOND study showed that the analysis of heart rate dynamics has significant prognostic value that is independent of the clinical risk factors 6 in patients with depressed left ventricular function after an acute myocardial infarction. However, the underling mechanisms of the RR-interval fluctuation remain unclear. The heart is obligate anaerobe and the heart function is sensitive to hypoxia. Therefore, hypoxia may be responsible for RR-interval fluctuations. We hypothesized that hypoxia induces RR-interval fluctuations. Obstructive sleep apnea causes severe hypoxia attributable to obstruction of the respiratory air way during sleep, while no such abnormality occurs during daytime. Obstructive sleep apnea may be an ideal model when investigating effects of hypoxia on RR-intervals. The goal of this study was to elucidate the contribution of hypoxia to RR-interval fluctuations in patients with obstructive sleep apnea syndrome.

Subjects
This study included 26 patients with obstructive sleep apnea syndrome who fulfilled the standard criteria for obstructive sleep apnea syndrome 7 . Patients were enrolled between 1 December 2012 and 30 November 2013 at Takahata Public Hospital, Takahata, Japan. Exclusion criteria were the presence of atrial fibrillation, atrioventricular block, intraventricular conduction block, pacemaker rhythm, hypotension, because these conditions may interfere appropriate and precise evaluations of RR-intervals or oxygen saturation. This study was approved by the Ethics Committee of Yamagata Prefectural University of Health Science, and the Ethics Committee of Takahata Public Hospital. Written informed consent was obtained from all patients.

Measurements
ECGs were recorded using NASA and CM5 leads for 24 hours. An ambulatory digital ECG recording device 8 (FM180; Fukuda Denshi, Tokyo, Japan) with a sampling rate of 1,000/sec was used with an automatic measurement system (SCM8000; Fukuda Denshi, Tokyo, Japan), and was transferred to a personal computer on which the MemCalc system [9][10][11] had been installed (MemCalc/Chiram; Suwa Trust, Tokyo, Japan). Data recording and processing were performed according to the recommendations of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology 12 . The MemCalc system was used to transform the RR-interval fluctuations into frequencies. This method allowed us to ascertain how power distributes as a function of frequency (power spectral density estimation) 13 . The analyzed spectral components were 1) 0.0001-0.05 Hz, 2) 0.05-0.1 Hz, 3) 0.1-0.15 Hz, 4) 0.15-0.2 Hz, 5) 0.2-0.25 Hz, 6) 0.25-0.3 Hz, and 7) 0.3-0.5 Hz. Each spectral measure was computed as amplitudes i.e. areas under the power spectral density. We defined this parameter as 'spectrum power', and is presented in square milliseconds. We conducted separate analyses during awake and sleep conditions and calculated the sleep/awake ratios as 'spectrum power' during sleep divided by 'spectrum power' during awake. Sleep and awake times were determined by the monitor diary that patients wrote. Further details of the MemCalc system are described elsewhere 9 .
Arterial oxygen saturation (SpO 2 ) was measured in all subjects using an LS-300 device (Fukuda Denshi, Tokyo, Japan) 14,15 , that incorporated a finger pulse oximeter (Envitec, Wismar, Germany). The overnight recordings were analyzed using SCM8000 System (Fukuda Denshi, Tokyo, Japan). The lowest oxygen saturation during sleep was defined as the minimal SpO 2 .

Statistical analysis
Statistical analyses were performed using Stat View version 5.0 (SAS Institute Inc., Cary, NC, USA). Pearson's correlation was applied to regression analysis. Statistical significance was inferred for p < .05.

Clinical characteristics of study subjects
Clinical characteristics of study subjects are presented in table 1. Twenty six subjects with sleep apnea syndrome were studied. Eight out of 26 patients were treated with continuous positive airway pressure (CPAP) device during sleep. The apnea-hypopnea index (AHI) of study subjects were 5.1-66.0 per hour. The minimal oxygen saturation during sleep were 53-90%. Twenty one subjects were also treated as hypertension, 5 as hypercholesterolemia, 3 as diabetes mellitus and one as previous myocardial infarction. Two patients received beta blockade (one hypertension and one previous myocardial infarction). One patient with previous myocardial infarction has reduced left ventricular function with 39% of ejection fraction by echocardiography. In the other 25 patients, ejection fractions were 57-83%. Comparison with the degree of oxygen desaturation Figure 1 shows the power spectrum density of RR-interval time series recorded at the condition with 71% or 96% of SpO 2 in a patient with obstructive sleep apnea syndrome. The spectral powers during hypoxia were shifted upward to those during normoxia, mainly on the 0.1 Hz or lower spectral range. This suggests that hypoxia evoked fluctuation of RR intervals. In the same case, the relations of 0.05-01 Hz powers and SpO 2 were plotted by every five-minute (Fig. 2). Increases in 0.05-0.1 Hz powers in this case were correlated proportionally to decreases in oxygen saturation. The spectral powers during hypoxia were shifted upward to those during normoxia, mainly on the 0.1 Hz or lower spectral range. PSD, power spectrum density; SpO 2: arterial oxygen saturation.

Discussion
Results showed that hypoxia evokes RR-interval fluctuations in obstructive sleep apnea syndrome. Among the seven analyzed frequencies, we found that the increase in the 0.05-0.1 Hz power correlated linearly with the minimal SpO 2 during sleep. This evidence indicates that the RR-interval fluctuations, especially 0.05-0.1 Hz fluctuations will provide important information related to hypoxia.

Hypoxia and RR-interval fluctuations
Previously, chronic effects of hypoxia were reported by Cornolo et al. 16 in subjects staying at high altitude. They found that 0.04-0.15Hz components (low-frequency components) of RR-interval fluctuations decreased after a six-day stay at the high-altitude mountain. However, acute effects of hypoxia were not clear. Several studies have examined that in subjects staying at the hypoxic chamber [17][18][19] , where oxygen contents were low. However, these studies failed to detect the relation between hypoxia and RR-interval fluctuations. Differences in study design from the present study were 1) degree of hypoxia and 2) analyzed spectrum bands. Patients with obstructive sleep apnea syndrome are exposed by severe hypoxia during sleep, therefore their sleep are interrupted. If the air way is obstructed completely, then the oxygen supply stops completely. Hypoxia caused by obstructive sleep apnea might be more severe than that caused by reduction of air oxygen content. Moreover, we analyzed seven narrow spectra of RR-interval fluctuations separately. These analytical methods helped us determine the relations sharply and specifically. Finally, the results clearly revealed the presence of hypoxia-related components within the RR-interval fluctuations.

Origin of RR-interval fluctuations
The spectrum of hypoxia-related RR-interval oscillation was categorized into low frequency in heart rate variability analysis. Oscillations of this category were explained to be associated with baroreflex activation and to be jointly regulated by sympathetic and parasympathetic interaction 20 . However, a recent study showed that the baroreflex sensitivity was suppressed in patients with sleep apnea syndrome 21 . Moreover, several studies suggested that the altered heart rate variability is associated with various pathologic conditions such as hemorrhagic shock 22 and septic shock 23 . We supposed that the previous explanation does not cover all of RR-interval fluctuations; the other mechanisms might influence pathological fluctuations. Hypoxia may be a new candidate origin of pathological fluctuations. However, interactions between hypoxia and RR fluctuation remain unclear. Possible mechanisms were oscillatory changes in 1) pacemaker potentials in sinus node, 2) disturbed electrical conductions in conduction system and 3) mechanical stretch in ventricular muscle cells. Further studies will be necessary to determine the mechanism of this phenomenon.

Clinical implication and limitation
This study revealed that RR-interval fluctuations were associated with hypoxia. This might be a key to the understanding of a prognostic value of RR-interval fluctuation in cardiovascular disease 5 . Previously, we reported in an experimental model that the depolarization fluctuations directly triggered ventricular fibrillation on the heart with high-dose loading of sodium channel blocker 24 . Since depolarization fluctuations can be seen as RR interval fluctuations in a conventional ECG recording, hypoxia-associated fluctuation may trigger life-threatening events. With consideration of these relations, more attention should be paid to the RR-interval fluctuations, especially 0.05-0.1 Hz fluctuations.
The criticisms in this study are study size and population. The study size was small, and the study population was limited to patients with obstructive sleep apnea syndrome. Additional research conducted with a large study size and in other pathological conditions associated with hypoxia is required.