Atrial Overdrive Pacing for Sleep Apnea: A Door Now Closed?

Sleep apnea remains a common clinical problem, with adverse cardiovascular implications, yet few treatment options. For obstructive sleep apnea (OSA), the cornerstone of therapy is continuous positive airway pressure (CPAP), which eliminates upper airway obstruction, apneas, hypoxia, sympathetic activation, and arousals from sleep, while lowering blood pressure and heart rate (1, 2). In patients with heart failure and coexisting OSA, CPAP also increases left ventricular ejection fraction (3). Although OSA is common in patients with and without heart failure, most individuals with Cheyne-Stokes respiration and central sleep apnea (CSA) have heart failure. In trials involving patients with heart failure, treatment of CSA by CPAP stabilized breathing, lowered norepinephrine concentrations, and increased left ventricular ejection fraction (4). However, because of problems with bed-partner acceptance, nasal congestion and nasofacial discomfort, or claustrophobia, CPAP is not universally applicable for sleep apnea. Even with good compliance, CSA may be alleviated, but not abolished (4).

Thus, the search for alternate therapies continues. At present, there is no effective drug therapy for OSA, and oral appliances or surgery are not as effective as CPAP (5, 6). Electrical stimulation of the upper airway dilator muscles has been attempted, but has not proved viable in the long term (7). The apnea-hypopnea index (AHI) of patients with heart failure with CSA can be improved by nocturnal oxygen therapy (8) and adaptive servoventilation (9), but these effects have only been demonstrated in small, brief trials, in which cardiovascular function was not evaluated.

Given this background, there was considerable excitement when Garrigue and colleagues (10) reported that atrial overdrive pacing, at a rate of 15 beats/minute faster than the mean nocturnal heart rate, reduced the AHI by 60% in patients who had received pacemakers for symptomatic sinus bradycardia or tachyor bradyarrhythmias. Could this be a novel therapy for sleep apnea, and a new indication for cardiac pacing?

Garrigue and coworkers (10) subjected 15 patients with an AHI of greater than 15/hour, and with a relatively equal distribution of CSA and OSA, to this nocturnal pacing protocol. Obstructive apneas fell from 6 to 3/hour, central apneas from 13 to 6/hour, and the overall AHI from 28 to 11 events/hour. They proposed that surges in parasympathetic tone caused marked nocturnal heart rate variability, and suggested, without providing evidence, that atrial pacing might stabilize respiration by preventing this nocturnal vagotonia. However, as the authors admitted, it would be difficult to explain the observed reduction in obstructive events on this basis.

Another possibility, which we favor, is that an improvement in cardiac output with atrial overdrive pacing led to reductions in lung to chemoreceptor circulation time and left ventricular filling pressure, and thereby stabilized breathing by reducing loop gain and preventing the hyperventilation that initiates CSA (11). If so, then one would expect a greater effect on CSA than OSA where respiratory control system instability plays less of a pathophysiologic role.

In the present issue (pp. 118-122), L

Does this study put to rest the concept that atrial overdrive pacing can be applied for treatment of sleep-related breathing disorders? On the surface, this would appear to be the case. However, the results may not be generalizable because pacemakers were not implanted for therapy of sleep apnea, but rather for management of diverse cardiac arrhythmias. Differences in the nature of these arrhythmias and the type of sleep apnea may explain differences between the present study and that of Garrigue and others (10).

First, implant indications in the two studies differed. In Garrigue and coworkers (10), the primary indication for cardiac pacing was bradycardia, whereas in the present study by Lüthje and others (12), only 50% had bradycardia. Consequently, their nocturnal heart rates were faster, and cardiac outputs were probably higher. Second, whereas OSA and CSA were of approximately equal prevalence in Garrigue and colleagues, OSA predominated in Lüthje and coworkers. Thus, if our assumptions about the effects of increasing cardiac output through overdrive pacing are correct, the potential for alleviating sleep apnea would have been less in Lüthje and coworkers' patients than in those of Garrigue and others, both because of their higher baseline heart rate and their predominance of OSA. However, neither group measured the hemodynamic effects of atrial overdrive pacing so that we do not know whether the increase in pacing rate of 7 and 15 beats/minute increased cardiac output and reduced left ventricular filling pressures sufficiently to alter the latter stimulus to hyperventilation (13).

Where might cardiac pacing fit in the therapy of sleep apnea? Garrigue and coworkers and Lüthje and colleagues studied patients without left ventricular systolic dysfunction (mean left ventricular ejection fractions were 54 and 47%, respectively). Thus, neither study addresses this important question in the population of greatest interest -namely, those with sleep apnea and heart failure. In such subjects, the potential role of respiratory control system instability related to circulatory delay and hyperventilation is greater than in the subjects included by these two groups.

Thus, the door remains open to future studies of nocturnal cardiac pacing in heart failure. However, because a subsequent study by Garrigue's group (14) failed to demonstrate any effect of atrial overdrive pacing on OSA, future clinical trials should focus mainly on patients with CSA. In our view, biventricular resynchronization pacing, which has already been demonstrated to reduce the composite endpoint of death and hospitalization (15), holds more promise than atrial overdrive pacing because of its more pronounced hemodynamic benefits. In a small, nonrandomized study of patients with heart failure (mean left ventricular ejection fraction of 24%) with left bundle branch block and CSA, cardiac resynchronization therapy lowered the AHI from 19.2 to 4.6/hour (16). Randomized trials evaluating the impact of this pacing modality, particularly with respect to CSA, are eagerly awaited. Whether sleep apnea could become an independent indication for some form of cardiac pacing remains an open question.

Conflict of Interest Statement: J.S.F. is a coinvestigator on a multicenter Canadian Institutes of Health research university/industry clinical trial of CPAP for patients with heart failure with CSA, with Respironics, ResMed, and Tyco as industrial partners. The grant ran from 1998-2004, and approximately $1,500,000 was the contribution by these three partners over this time period; T.D.B. is the principal investigator on a multicenter Canadian Institutes of Health research university/industry clinical trial of CPAP for patients with heart failure with CSA, with Respironics, ResMed, and Tyco as industrial partners. The grant ran from 1998-2004, and approximately $1,500,000 was the contribution by these three partners over this time period. He consulted for Medtronic and received $6,000 in 2004.