Author + information
- Received May 13, 2005
- Revision received August 4, 2005
- Accepted September 8, 2005
- Published online January 17, 2006.
- Andrew D. Krahn, MD⁎,⁎ (, )
- Raymond Yee, MD⁎,
- Mark K. Erickson, BSc†,
- Toby Markowitz, BSEE†,
- Lorne J. Gula, MD⁎,
- George J. Klein, MD⁎,
- Allan C. Skanes, MD⁎,
- Charles F.P. George, MD⁎ and
- Kathleen A. Ferguson, MD⁎
- ↵⁎Reprint requests and correspondence:
Dr. Andrew D. Krahn, London Health Sciences Center, University Campus, 339 Windermere Road, London, Ontario N6A 5A5, Canada.
Objectives This study was designed to assess the impact of prevention of bradycardia with physiologic pacing on the severity of obstructive sleep apnea.
Background Apneic episodes during sleep are associated with slowing of the heart rate during apnea and tachycardia with subsequent arousal. Patients with permanent pacemakers may have reduced episodes of sleep apnea when their pacemaker rate is set faster than their spontaneous nocturnal heart rate.
Methods We conducted a prospective, randomized, single-blind crossover trial of temporary atrial pacing in obstructive sleep apnea to reduce the apnea hypopnea index (AHI). Fifteen patients (age 60 ± 13 years, 12 men) with moderate to severe obstructive sleep apnea (AHI 34 ± 14) underwent insertion of an externalized atrial permanent pacing system via the left subclavian vein. Patients underwent overnight respiratory sleep studies in hospital, during atrial pacing at 75 beats/min, and with pacing turned off. The order of pacing mode was randomized, with crossover the subsequent night to the other mode. Patients were blinded to pacing mode, and the analysis of sleep recordings was blind to pacing mode.
Results Pacing was tolerated without complications in all patients. Overnight physiologic pacing did not affect the AHI (pacing 39 ± 21/h vs. control 42 ± 21/h, p = 0.23, 95% confidence interval −9.3 to 2.5 for difference), desaturation time (pacing 3.8 ± 6.0% vs. control 3.5 ± 4.3%, p = 0.70), or the minimum SaO2(pacing 75 ± 10% vs. control 77 ± 11%, p = 0.38). There was a borderline significant reduction in circulatory time with pacing (pacing 23.4 ± 3.2 s vs. control 25.5 ± 4.4 s, p = 0.09).
Conclusions Temporary atrial pacing does not appear to improve respiratory manifestations of obstructive sleep apnea. Permanent atrial pacing in this patient population does not appear to be justified.
Sleep apnea syndrome is a disorder characterized by recurrent episodes of apnea during sleep, typically producing excessive daytime sleepiness (1). The disorder is associated with significant morbidity and mortality and is significantly underdiagnosed and undertreated (2–4). Sleep apnea is diagnosed when the apnea-hypopnea index (AHI) is at least 5 and symptoms are present. Based upon these criteria, 4% of middle-age men and 2% of middle-age women are affected (1).
Apneic episodes during sleep are generally associated with profound bradycardia with cyclical variation of the heart rate during apnea and resumption of breathing (5). Garrigue et al. (6) reported that nocturnal atrial overdrive pacing resulted in an improvement in the frequency of sleep apnea events in patients. Fifteen subjects with dual-chamber pacemakers and mild to moderate sleep apnea were randomized, in a crossover design, to DDD pacing at 15 beats/min above the spontaneous nocturnal heart rate or spontaneous rhythm with VVI backup pacing at 40 beats/min. Overnight polysomnography demonstrated a 60% reduction in AHI. Nocturnal atrial overdrive pacing reduced both central and obstructive apnea and hypopnea. Importantly, the patients were derived from a population of pacemaker patients with symptoms suggesting sleep apnea, with subsequent testing suggesting a mixed population of obstructive sleep apnea (OSA) and central sleep apnea.
Patients with pacemakers represent a small percentage of people with sleep apnea. Our goal was to examine the effect of temporary nocturnal atrial overdrive pacing in subjects with a primary diagnosis of obstructive sleep apnea syndrome who did not have a conventional indication for pacing. We hypothesized that an increase in physiologic heart rate would reduce the number and severity of apneic events. If temporary pacing reduced apnea severity, permanent atrial pacing might represent a potential treatment option for patients with OSA, particularly patients in whom nasal continuous positive airway pressure (CPAP) had failed.
Subjects with previously diagnosed sleep apnea without standard indications for cardiac pacing were recruited to undergo two consecutive night sleep studies with temporary pacing randomized to no pacing or AAI at a lower rate of 75 beats/min. Subjects were recruited from a local sleep clinic with new or previously diagnosed OSA. Patients had moderate to severe sleep apnea (AHI 10 to 50) on overnight polysomnography. The baseline diagnostic sleep study montage included electroencephalogram (C3/A2, C4/A1, O2/A1), electrooculogram, submental electromyogram (EMG), left and right anterior tibialis EMG, electrocardiogram (ECG), thoraco-abdominal motion, oronasal airflow (nasal pressure), and arterial oxygen saturation with pulse oximetry. The studies were scored manually, and the total AHI was calculated for the night. Obstructive apneas were defined as the cessation of airflow for at least 10 s accompanied by ongoing respiratory effort. Obstructive hypopnea was defined as a reduction in airflow of at least 50% for at least 10 s accompanied by a reduction in respiratory effort and by an arousal or an arterial oxygen desaturation of at least 3%.
Subjects had refused or failed CPAP, defined as discontinuation because of inefficacy and/or intolerance and/or refusal to continue to use CPAP. Subjects were excluded from the study if they were <18 years of age, had a diagnosis of another sleep disorder in addition to sleep apnea (including central sleep apnea), were taking sedative or hypnotic medications, were taking warfarin, had New York Heart Association functional class IV heart failure, had a previous myocardial infarction or cardiac bypass surgery, had features suggesting cor pulmonale (hematocrit >50%, awake arterial PCO2>50 mm Hg or severe hypoxemia during an overnight polysomnogram, SaO2<70%), were pregnant, or were unable or unwilling to attend follow-up or provide informed consent. Patients that had severe excessive daytime sleepiness (Epworth Sleepiness Scale [ESS] ≥15) were excluded because we did not want untreated patients with severe sleepiness to be delayed in getting known effective therapy (CPAP) or stopping CPAP for an unproven therapy. Patients with a body mass index >40 kg/m2were excluded because they tend to have severe OSA that is most effectively treated with CPAP, and the protocol mandated that they be untreated for at least two weeks before testing (to wash out any effect of CPAP). The potential risks and benefits of atrial pacing and central venous cannulation were explained in detail. All subjects provided written informed consent. The Health Sciences Research Ethics Board of the University of Western Ontario approved the protocol.
Pacing and sleep study
The study was a single-blind randomized crossover design with a control and treatment arm. Subjects presented for insertion of a temporary pacing system after fasting for a minimum of 6 h. Using sterile technique and local anesthetic, subjects underwent cannulation of the left subclavian vein. An active fixation, bipolar, permanent pacing lead (Model 5076, Medtronic Inc., Minneapolis, Minnesota) was introduced though a 5-mm cutaneous incision and advanced to the right atrial appendage under fluoroscopic guidance. After adequate pacing and sensing thresholds were verified, the pacing lead was connected to an external permanent pacing generator (Model KSR 903, Medtronic Inc.) and fixed to the skin with a suture to anchor the lead, and a temporary adhesive dressing was applied over the generator and lead. Output amplitude was set to 0.4 ms and 3.5 V or two times threshold, whichever was higher. A chest X-ray was performed to verify lead position and exclude pneumothorax.
Subjects received AAI pacing at 75 beats/min or OAO pacing (no pacing) on two consecutive nights in a randomized, crossover single-blind fashion. The pacemaker was connected to the subject for both nights. Overnight respiratory recording was performed using the Stardust Respiratory Disorder Diagnostic Device (Stardust, Respironics Inc., Murrysville, Pennsylvania). The standard recordings included measures of oral/nasal airflow (nasal pressure), respiratory effort, arterial oxygen saturation, heart rate, and body position. The parameters from these recordings provided an assessment of sleep-disordered breathing severity. A DR190 digital Holter monitor (NorthEast Monitoring, Maynard, Massachusetts) capable of recording ECG and continuous pacemaker telemetry was worn both nights.
Following the first night of sleep, the study coordinator downloaded recorded data from the pacemaker, the Holter monitor, and the Stardust recorder; reset the recording devices; verified normal pacing function; and programmed the pacemaker to the crossover mode. The following morning, similar data were retrieved from the devices, and the lead was removed under fluoroscopic guidance and an adhesive bandage as applied to the insertion site.
Data collection and analysis
Baseline measurements included age, gender, weight, body mass index (kg/m2), neck circumference, and blood pressure. A research coordinator administered the ESS; the Sleep Apnea Subscale of the Sleep Disorders Questionnaire; and a detailed questionnaire about snoring, sleepiness, and associated features of SAS (Table 1).The primary outcome of the study was the difference in AHI during AAI pacing compared with control patients. We estimated that 13 patients were required in a crossover design to provide 80% power to detect a 50% reduction in AHI. Secondary outcomes included a comparison of treatment and control nights for total study time, mean oxygen saturation (SaO2), lowest SaO2, percent of sleep time with SaO2<90%, mean apnea-hypopnea duration (s), and circulatory time (time from end of apnea to nadir of SaO2). The studies were scored using the automated analysis protocol of the Stardust software. An experienced scorer who was blinded to the treatment status of the subject then reviewed the studies manually on an epoch-by-epoch basis. Responders to AAI pacing were defined as subjects who improved by at least one American Academy of Sleep Medicine classification; severe to moderate, moderate to mild, or mild to normal (7). Continuous variables are presented as mean ± SD. Continuous data were compared using a paired Student ttest. A p value at the 0.05 level or lower was considered significant. Data are presented as mean ± SD.
The 15 subjects enrolled in the trial were predominantly overweight middle-aged men with borderline hypertension (Table 1). Their average AHI level from diagnostic polysomnography was classified as severe at 34 ± 14/h. Twelve of the 15 subjects had used CPAP in the past. Nine of the 15 subjects were identified as high risk for sleep apnea by the Sleep Apnea Subscale of the Sleep Disorders Questionnaire (established cutoffs >32 for women and >36 for men), but only 5 of 15 were categorized as sleepy using the ESS (≥10).
The AHI values during control and AAI pacing for individual subjects are illustrated in Figure 1.Mean AHI showed no evidence of benefit from pacing (38.6 ± 20.5/h control vs. 42.1 ± 20.7/h pacing, p = 0.23). By American Academy of Sleep Medicine criteria (7), only one subject was classified as a responder (Subject #10, AHI reduced from 34 to 18).
The pacing intervention successfully eliminated nocturnal bradycardia with the expected rise in mean heart rate (Fig. 1). All subjects increased heart rate with AAI pacing. Subject #6 had extensive noise on the Holter ECG and the Stardust pulse rate, making reliable heart rate calculations impossible. By design, mean heart rate increased from 65 ± 7 beats/min during the control night to 77 ± 1 beats/min during AAI pacing (p < 0.0001). There were no complications from the lead implant or pacing. Two subjects complained of minor tenderness at the lead insertion site.
There was no difference in other indices of sleep apnea burden, including oxygen saturation, duration of desaturation, and apnea-hypopnea duration (Table 2).There was a trend toward a reduction in circulatory time during AAI pacing (25.5 ± 4.4 s control vs. 23.4 ± 3.2 s AAI, p = 0.09), likely reflecting increased cardiac output attributable to increased heart rate with AAI pacing.
The current study does not support a role for cardiac pacing in the treatment of obstructive sleep apnea. The apparent benefit from permanent pacing in patients with both forms of sleep apnea previously reported was not replicated in patients that did not require pacing for a conventional bradycardia indication (6). Garrigue et al. (6) focused on patients with established bradycardia who had a stable permanent pacemaker with ongoing findings suggesting sleep apnea during nominal pacing programming. They found a benefit from permanent pacing in subjects with both central and obstructive apnea that we could not confirm in subjects who did not require pacing for a conventional bradycardia indication. Our population was somewhat older than is typical for a sleep apnea population, but was similar in age to the Garrigue et al. study (6). We sought to investigate a role for permanent pacing by testing blinded physiologic pacing in patients who would be ideal candidates for an alternate approach to improving their sleep apnea but who did not have an indication for pacing. Two other studies have focused on patients with permanent pacemakers (8,9). Luthje et al. (8) and Pepin et al. (9) studied 20 and 15 patients with previously implanted devices and did not detect any change in the AHI with nocturnal pacing.
Our study demonstrated no effect in this clinically relevant population, suggesting that a large enough clinical effect on AHI that would be needed to justify an invasive strategy using permanent pacing is unlikely to have been missed. The 95% confidence limits of the effect on AHI (−9.3 to 2.5) do not include a sufficient effect to justify an invasive strategy that would fail to normalize or markedly improve the patient’s sleep apnea. Although a reduction in AHI as much as 9 points appears to represent potential clinical benefit, the risk of hypertension and coronary artery disease associated with sleep apnea is increased above an AHI of 5 (10–13). In a single small study that found a reduction in cardiovascular risk in patients treated with CPAP for sleep apnea, the mean AHI fell from 33.7 ± 16.8/h at baseline to 3.9 ± 2.9/h (10). This target is well below the residual AHI burden of disease of 25 that represents the optimal estimate of effect from pacing in the current study, suggesting ongoing moderate to severe sleep apnea. Additional parameters that gauge severity of sleep disruption, physiologic response to apnea, and patient symptoms did not suggest any evidence of benefit. One postulate for the previous beneficial observation is that the utility of pacing may be restricted to improving central sleep apnea. The concept that atrial pacing would improve obstructive sleep apnea is a speculative one. The Garrigue et al. (6) study showed a greater reduction in apnea index in central sleep apnea patients (13 ± 17 to 6 ± 7, p = 0.007) when compared with obstructive apnea patients (6 ± 4 to 3 ± 1, p = 0.03). The central patients had more apnea events and had a more dramatic response (14). The improvement in central sleep apnea with pacing is more consistent with our understanding of the physiology of central apnea in heart failure. The reversal of bradycardia in the autonomic and neurohormonal milieu of heart failure seems more likely to have an impact on central sleep apnea and cardiac function than does reversal of the secondary bradycardia-reducing OSA (15).
The current findings are limited to the small population studied. Further investigations to explore the role of pacing in central sleep apnea seem warranted on the basis of the current study and previous work. Because patients had failed or refused CPAP, failure of standard therapy with CPAP may have made it more likely for other treatments to fail. However, the patients were blinded to the pacing component of the study. One strength of the study was the crossover design, which minimized the possibility of a period effect by maintaining stable sleeping conditions as best as feasible under the circumstances. The same design did not allow for a washout period between nights. There is no evidence that overnight backup pacing should have led to persistent effects; this view is supported by the nonsignificant test for interaction between the first and second night (Wilcoxon rank-sum test, p > 0.10). Given the unproven and invasive nature of atrial pacing as a treatment, we felt it would not be appropriate to recruit patients with moderate to severe OSA who had not previously been offered or treated with CPAP. A full sleep study may have been more sensitive to detect subtle benefits with more comprehensive sleep data, though the overall primary end point was clearly negative. The Stardust recorder does not directly measure sleep time, but rather breathing and airflow. Thus, we estimated sleep time from time in bed, but did not measure it directly, potentially introducing error in the calculation of the AHI. Invasive pacing was not feasible in the outpatient environment, where full sleep studies are performed in our institution. For safety reasons, the study was conducted on the cardiology floor with ambulatory sleep assessment.
The ideal form of pacing that may reduce the frequency of apneic episodes is unknown. The current approach was chosen because it represented typical physiologic pacing that is used to prevent low heart rates in other conditions, such as bradycardia-mediated atrial fibrillation (16). A different heart rate may have influenced results, though no trend was evident with the current heart rate that clearly increased nocturnal rates.
Temporary atrial pacing does not appear to reduce sleep apnea in subjects with obstructive sleep apnea. Permanent pacing in the OSA patient population does not appear to be justified at this time.
The authors would like to thank Steve Saunders and Kim Heighway for their tireless efforts in patient recruitment and data collection.
This study was funded by Medtronic Inc., Minneapolis, Minnesota. Drs. Krahn, Yee, and Klein have received consulting fees from Medtronic and have been paid speakers for Medtronic Inc. Mr. Erickson and Mr. Markowitz are employees of Medtronic Inc. Dr. George has received consulting fees and is on the medical advisory board for SleepTech LLC (Kinnelon, New Jersey). Dr. Ferguson has received grant support and has been a paid speaker for Respironics Inc.
- Abbreviations and Acronyms
- apnea hypopnea index
- continuous positive airway pressure
- Epworth Sleepiness Scale
- obstructive sleep apnea
- oxygen saturation
- Received May 13, 2005.
- Revision received August 4, 2005.
- Accepted September 8, 2005.
- American College of Cardiology Foundation
- Pepin J.L.,
- Defaye P.,
- Garrigue S.,
- et al.
- Milleron O.,
- Pilliere R.,
- Foucher A.,
- et al.
- Gillis A.M.,
- Connolly S.J.,
- Lacombe P.,
- et al.,
- The atrial pacing peri-ablation for paroxysmal atrial fibrillation (PA (3)) study investigators