Sleep disordered breathing (SDB) appears to be common

Transcription

1 The Current Prevalence of Sleep Disordered Breathing in Congestive Heart Failure Patients Treated with Beta-Blockers Mary Macdonald, RPSGT; James Fang, M.D.; Steven D. Pittman, MSBME; David P. White, M.D.; Atul Malhotra, M.D. Divisions of Cardiovascular Diseases, Sleep Medicine, and Pulmonary and Critical Care Medicine, Brigham and Women s Hospital and Harvard Medical School, Boston, MA Study Objectives: Although sleep disordered breathing is thought to be common in patients with systolic heart failure, prior studies are difficult to interpret due to a variety of factors including small sample sizes, referral bias to sleep laboratories among participants, lack of modern medical therapy for congestive heart failure, and the failure to use modern techniques to assess breathing such as nasal pressure. Our objective was to determine the current prevalence of sleep disordered breathing in a state-of-the-art congestive heart failure clinic. Methods: We conducted a prospective study of consecutive patients who visited our heart failure clinic to assess the prevalence of sleep apnea in all eligible patients on maximal medical therapy. We used 4-channel recording equipment and modified Chicago criteria for scoring respiratory events (using heart rate response as a surrogate for arousal from sleep). Results: We observed that among the 108 participants, 61% had some form of sleep disordered breathing (31% central apnea with Cheyne Stokes respiration and 30% obstructive sleep apnea). Sleep Scientific investigations disordered breathing was significantly associated with atrial fibrillation (OR = 11.56, p = 0.02) and worse functional heart failure class (OR = 2.77, p = 0.02), after adjusting for male sex, age over 60 years, body mass index, and left ventricular ejection fraction. Conclusions: We conclude that both obstructive and central sleep apnea remain common in congestive heart failure patients despite advances in medical therapy, and that the previously reported high prevalence values are unlikely to be explained by referral bias or participation bias in prior studies. These data have important clinical implications for practitioners providing CHF therapy. Keywords: Apnea, sleep, congestive heart failure, obstructive, central, outcome, lung Citation: Macdonald M; Fang J; Pittman SD; White DP; Malhotra A. The current prevalence of sleep disordered breathing in congestive heart failure patients treated with beta-blockers. J Clin Sleep Med 2008;4(1): Sleep disordered breathing (SDB) appears to be common among patients with congestive heart failure (CHF). 1,2 Prior studies have suggested that 40%-50% of patients with CHF and left ventricular systolic dysfunction will have some form of SDB, either obstructive or central sleep apnea (CSA). 1-3 However, important limitations exist in these previous reports. Many were based in sleep laboratories and therefore subject to referral and participatory bias. 2 To our knowledge, only one study, using nonconsecutively enrolled subjects, has prospectively Disclosure Statement This study was supported in part by Respironics. Dr. White is employed by Respironics and has received research and consulting fees from Itamar Medical and consulting fees from Aspire Medical. Dr. Malhotra has received research support from Respironics, Inspiration Medical, NMT Medical, Restore Medical, Pfizer, and Cephalon. Dr. Fang and Ms. Macdonald have indicated no financial conflicts of interest. Mr. Pittman is an employee of Respironics. Submitted for publication June, 2007 Accepted for publication October, 2007 Address correspondence to: Atul Malhotra, Sleep Disorders Research BIDMC, Brigham and Women s Hospital, 75 Francis Street, Boston, MA, 02115; Tel: (617) ; Fax: (617) ; amalhotra1@partners.org evaluated the prevalence of SDB in outpatients in a heart failure clinic. 4 Other studies have been limited by technical issues such as the use of respiratory monitoring systems that do not employ state-of-the-art sensors for ventilation i.e., nasal pressure to assess hypopneas. 5 Previous studies have also been difficult to reconcile due to variable definitions of CSA, and a failure to distinguish between CSA in general and Cheyne Stokes respiration (CSR) in particular. In many cases, it can be difficult to distinguish obstructive from central apnea type in CHF patients, leading some authors to suggest that different metrics for sleep disordered breathing severity may be desirable. 6 Perhaps most importantly, many previous studies have not included study populations with currently optimal medical management including adequately dosed β-blockers, angiotensin converting enzyme inhibitors (ACEI), angiotensin II receptor antagonists (ARBs) and aldosterone antagonists. 7 This lack of optimized medical therapy may have particular relevance since such therapies may have an impact on the prevalence of SDB such as CSR. 8,9 For example, β-blockers are known to influence hypoxic chemosensitivity which would be predicted to influence the development of CSR. 10,11 Similarly, ACEI and diuretic therapy can both contribute to lower intracardiac filling pressures, which could importantly influence the occurrence of CSR. 8,12 The goals of this study were to (1) prospectively evaluate the prevalence of SDB in consecutive medically optimized outpa- Journal of Clinical Sleep Medicine, Vol. 4, No. 1,

2 Table 1 Baseline Clinical Characteristics of Study Participants and Non-Participants* Participants Non-Participants p value (n=108) (n=45) Female n,% 16 (15) 14 (31) 0.04 Age (yr) 57 ± ± BMI (kg/m 2 ) 26.8 ± ± LVEF, % 20 (15-30) 20 (15-25) 0.89 Class II 67 (62) 27 (60) Class III-IV 41 (38) 18 (40) 0.96 Ischemic 47 (44) 26 (58) Idiopathic 61 (66) 19 (42) 0.15 β-blockers n,% 39 (82) 39 (86) 0.68 Other Medications, % Loop diuretics 88 ** ACE inhibitor 83 ** Digitalis 63 ** Anticoagulant 48 ** Spironolactone 36 ** Statin 35 ** Antiplatelet 34 ** Antiarrhythmic 30 ** Nitrate 25 ** ARB 13 ** AHI (/h) 24.7 ±17.2 ** CSR time, % 24 ±30 ** Atrial fibrillation n, % 15 (14) ** *Unless indicated, values expressed as mean ±SD or median (95% CI range) ** not compared tients in a heart failure clinic; (2) examine risk factors for SDB and specifically CSR in this population; and (3) explore alternative metrics for CSR. METHODS One hundred and eight subjects out of 153 consecutively eligible patients from the Heart Failure Clinic of the Brigham & Women s Hospital agreed to participate. Prior to enrollment, subjects were not asked about symptoms or risk factors for SDB. Subjects were included if they had clinically stable heart failure; were between the ages of 18 and 85 years; had New York Heart Association (NYHA) functional class II, III, or IV; and had a left ventricular ejection fraction (LVEF) of <40%. Subjects were excluded if they had primary valvular heart disease, primary diastolic failure (i.e., congestive heart failure with normal systolic function), a history of major lung disease (i.e., obstructive pulmonary disease), a history of pneumothorax in the prior 6 months, current or previous use of positive airway pressure noninvasive ventilation, current use of supplemental oxygen, or the presence of an artificial airway. Subjects who had been hospitalized or had medication changes within the past 30 days were also excluded. All subjects provided written informed consent. This study was approved by the Partners Institutional Review Board. Subjects underwent echocardiographic assessment of LVEF 12 months prior to performing a sleep study, during which time LVEF was unlikely to significantly change, Sleep Apnea in Heart Failure Table 2 Clinical Characteristics of Subjects with and without Sleep Disordered Breathing* no SDB SDB p value (n=42) (n=66) Prevalence 39% 61% Female n,% 7 (17) 9 (14) 0.88 Age (yr) 56 ±9 58 ± BMI (kg/m 2 ) 26.8 ± ± AHI (/h) 9.5 ( ) 32.9 ( ) <0.001 CSR time, % 0 (0-7) 34 (9-72) <0.001 LVEF, % 20 (15-25) 20 (15-30) 0.29 Class II 31 (74) 36 (55) Class III-IV 11 (26) 30 (45) 0.07 Ischemic 15 (36) 32 (49) Idiopathic 27 (64) 34 (51) 0.27 β-blockers n, % 39 (93) 50 (77) 0.06 Atrial fibrillation n, % 1 (2) 14 (21) 0.01 *Unless indicated, values expressed as mean ±SD or median (95% CI range) since a repeat echocardiogram is performed if any important change in clinical status occurs. All echocardiograms were read by board certified cardiologists with advanced training in echocardiography. The Teichholz formula is routinely used to assess LVEF, which must be agreed upon by the overreading echocardiographer. Subjects underwent an in-home, unattended overnight study with a standard 4-channel recording device (StarDust, Respironics, Inc, Murrysville, PA). This device records nasal pressure, thoracic excursion (as measured by a piezoelectric crystal), body position, pulse oximetry, and heart rate derived from pulse oximetry. The data were downloaded and scored by an experienced scorer. A modified version of the 1999 American Academy of Sleep Medicine (AASM) criteria for scoring respiratory events was used. 13,14 An obstructive apnea was defined as 10-sec cessation of airflow as measured by nasal pressure associated with the continuation of thoracic effort. A central apnea was defined as 10-sec cessation of airflow without thoracic effort. A hypopnea was defined as a 50% reduction in airflow for 10 sec or a discernable change in airflow with either a 3% oxyhemoglobin desaturation or an arousal, as defined by a 10% increase in heart rate. 15 Arousal was defined based on a 10% change in heart rate because EEG was not available using the StarDust system. We believe that a relative rather than absolute change in heart rate is more reliable for comparing patients with and without β-blockade. The apnea-hypopnea index (AHI) and central apnea index (CAI) were calculated based on total number of events per hour of total recording time (TRT). SDB was defined by AHI of 15 events per hour. Another metric of CSR was also used for exploratory purposes. In this case, CSR was defined as a symmetrical crescendo-decrescendo respiratory pattern with a >50% difference between peak and nadir nasal pressure or respiratory effort amplitude, occurring within a 30 to 90-second period. Primary CSR was defined as a study with a CSR time (as a percentage of total recording time) 33%. Primary OSA was predefined as an AHI 15 events per hour and a CSR time of <33%. Journal of Clinical Sleep Medicine, Vol. 4, No. 1,

3 M Macdonald, J Fang, S Pittman et al Table 3 Clinical Characteristics of Subjects with and without Cheyne Stokes Respiration* no CSR CSR p value (n=74) (n=34) Prevalence 69% 31% Female n, % 13 (18) 3 (9) 0.37 Age (yr) 56 ± ± BMI (kg/m 2 ) 27.0 ± ± AHI (/h) 13.6 ( ) 37.8 ( ) <0.001 CSR time, % 3.1 ( ) 68.6 ( ) <0.001 LVEF, % 24 (15-30) 15 (13-25) 0.02 Class II 52 (70) 15 (44) Class III-IV 22 (30) 19 (56) 0.02 Ischemic 31 (66) 16 (47) Idiopathic 43 (34) 18 (53) 0.77 β-blockers n,% 66 (89) 23 (68) 0.49 Atrial fibrillation n, % 6 (8) 9 (26) 0.02 *Values expressed as mean ±SD or median (95% CI range) Differences between groups were assessed using the unpaired Student s t-test for parametric results and the Mann-Whitney rank sum test for nonparametric data. χ 2 was used to compare proportions between groups. Logistic regressions were used to assess the odds ratios of SDB, and CSR, conferred by various independent variables. A value of p < 0.05 was considered statistically significant, with results being reported as the mean ± SD or median and 95% confidence intervals. Candidate independent variables were first established using a univariate model. The final model included only those variables with a significant effect on the dependent variable as measured by the likelihood ratio test statistic (p < 0.05). Multiple logistic regression was then used to establish the multivariate variables that remained independently predictive of SDB and CSR. The same regression model was also used to determine if the self-reported use of β-blockers independently predicted the presence or absence of SDB and CSR Odds ratios and 95% confidence intervals were calculated. Calculations were performed in SigmaStat 3.0 software. RESULTS 108 subjects (92 males and 16 females) enrolled in the study (participation rate of 71%). The primary reason for refusal to participate was the perceived inconvenience of the in-home sleep study. Other than a higher proportion of females among the nonenrolled, there were no significant baseline clinical differences between those who enrolled and those who did not (see Table 1). Sleep Disordered Breathing The prevalence of SDB in this cohort was 61% (57 males and 9 females). 31% of subjects had primary CSR (31 males and 3 females); 30% had primary OSA (26 males and 6 females). Between the groups with and without SDB (Table 2), the only significant difference was the presence of atrial fibrillation in the SDB group (21% vs. 2%, p = 0.01). There were no significant differences in baseline demographics, NYHA class, LVEF, CHF etiology, or β-blocker use. Figure 1 The prevalence of Cheyne Stokes Respirations (CSR) varies with the definition used. With a rising threshold for percentage of CSR time, there is a falling prevalence of this breathing abnormality. At a cutoff of 33% CSR time, almost one third of participants had this disease. After adjusting for male sex, age over 60 years, BMI, and LVEF, subjects with SDB had a nearly 12-fold increased odds for atrial fibrillation (OR = 11.56, 95% CI , p = 0.02), and a significantly greater odds for a worse functional class of heart failure (OR = 2.77, 95% CI , p = 0.02). Cheyne Stokes Respiration The group with CSR time 33% had a mean CSR time of 68.6% ( ); the group without CSR had mean CSR time 3.1% ( ) (Table 3, Figure 1).The CSR group had significantly more impaired functional capacity than the group without CSR (NYHA class III-IV 44% vs 30% class II, p = 0.02), a lower LVEF (15% [ ] vs. 25% [ ], p = 0.02), and had more subjects with atrial fibrillation (26% vs. 8%, p = 0.02). There were no significant differences in baseline demographics, CHF etiology, or β-blocker use. After adjusting for male sex, age over 60 years, BMI, and LVEF, subjects with CSR had a nearly 6-fold increased odds for atrial fibrillation (OR = % CI , p = 0.01), and a significantly greater odds for a worse functional class of heart failure (OR = % CI , p = 0.01). Using a CAI cut-off of 5/h and 15/h respectively (consistent with definitions of primary CSA used in previous studies), 10,16 the prevalence of CSA in our group is 28% and 17%. NYHA functional class and the presence of atrial fibrillation still emerge as significant differences between those with CAI 15 and those with a CAI <15/h. (NHYA: 61% Class III-IV vs. 33%, p = 0.05; a. fib: 39% vs. 9% p = 0.003). β-blocker use was not significantly different between groups. β-blocker Usage Eighty-two percent of all subjects were using β-blockers (Table 4). The group not using β-blockers was significantly more functionally impaired than the group that was (63% Class III-IV vs. 33% Class III-IV, p = 0.03). There were no other significant differences in baseline demographics, severity of SDB, CHF eti- Journal of Clinical Sleep Medicine, Vol. 4, No. 1,

4 Table 4 Clinical Characteristics of Subjects on and off Blockers* Β-Blockers no β-blockers p value (n=89) (n=19) Prevalence 82% 18% Female n, % 11 (12) 5 (26) 0.23 Age (yr) 57 ± ± BMI (kg/m 2 ) 28.1 ± ± AHI (/h) 19.4 ( ) 31.8 ( ) 0.06 CSR time, % 9.0 ( ) 17.1 ( ) 0.36 LVEF, % 20 (15-30) 25 (15-30) 0.56 Class II 60 (67) 7 (37) Class III-IV 29 (33) 12 (63) 0.03 Ischemic 38 (43) 9 (47) Idiopathic 51 (57) 10 (53) 0.91 Atrial fibrillation n, % 10 (11) 5 (26) 0.17 * Values expressed as mean ±SD or median (95% CI range) ology, or the presence of atrial fibrillation; 77% of subjects with SDB and 68% with CSR were using β-blockers. After adjusting for male sex, age over 60, BMI, and LVEF, β-blocker use did not independently predict the presence or absence of SDB or CSR. Discussion In this heart failure clinic-based study of SDB in a cohort of unselected consecutive patients with CHF who were receiving optimally dosed medical management with both β-blockade and renin-angiotensin system antagonists, a high prevalence of SDB persisted (61% overall; 31% CSR and 30% OSA) supporting previous work suggesting that SDB is common in CHF. Our data add to the existing literature by showing that the previously reported prevalence rates are not likely a product of selection bias or a result of inadequate or antiquated medical therapy. Although β-blocker therapy in heart failure is associated with improvements in LVEF, functional class, and survival, it did not appear to affect the prevalence of SDB in our cohort, which was almost uniformly treated with these agents. In fact, the prevalence of SDB in this study is similar to the rates previously reported in other heart failure cohorts, which range from 41% to 75%. In these previous studies, β-blocker use was either not reported or limited to a minority of participants. In one of the only recent true prevalence studies where β-blocker therapy was reported, only 30% of heart failure patients were on these agents. 4 Furthermore, this prior study was limited to 53 subjects referred to a sleep laboratory with less advanced heart failure (NYHA I/II 75%, mean LVEF 34%). Thus, it would appear that newer therapies for CHF do not have a clinically important impact on overall breathing instability. However, because different criteria have been used to define the apnea hypopnea index in each of these studies, further work is clearly needed to determine if variations in prevalence reported in the literature are methodological or biological. Patients with both SDB (61%) and CSR (31%) were both more likely to have atrial fibrillation and poorer NYHA functional class than those without SDB or CSR. These differences were not explained by differences in demographic variables, medical therapy, or CHF etiology. Although previous studies have identified the association of atrial fibrillation with SDB Sleep Apnea in Heart Failure Table 5 Clinical Characteristics of Subjects with Cheyne Stokes Respiration vs. Obstructive Sleep Apnea* CSR OSA p value (n=34) (n=32) Prevalence 31% 30% Female n, % 3 (9) 6 (6) 0.3 Age (yr) 59 ± ± BMI (kg/m 2 ) 28.5 ± ± AHI (/h) 37.8 ( ) 26.0 ( ) <0.001 CSR time, % 68.6 ( ) 8.9 ( ) <0.001 LVEF, % 15 (13-25) 30 (20-32) Class II 15 (44) 21 (58) Class III-IV 19 (56) 11 (34) 0.13 Ischemic 16 (47) 16 (50) Idiopathic 18 (53) 16 (50) 0.99 β-blockers n,% 23 (68) 24 (75) 0.89 Atrial fibrillation n,% 9 (26) 5 (16) 0.44 *Values expressed as mean ±SD or median (95% CI range) and CSR, an association between NYHA functional class and CSR has not been previously reported to our knowledge. Although this finding would support the notion that CSR is a consequence of progressive heart failure, and in fact patients in our cohort with CSR had significantly lower LVEFs than those without CSR, the lack of a β-blocker interaction in this study may suggest otherwise. Another interpretation of these findings is that aggressive medical management of CHF may improve LVEF in some patients, whereas a persistently low LVEF may be a marker of a particularly impaired cardiovascular system. This impairment could manifest as an elevated filling pressure which could lead to CSR. 9 In addition, as seen in Table 5, there are some possible differences between those patients with CSR compared with those with OSA. OSA patients had lower AHI as well as higher LVEF as compared to those with CSR. However, β-blocker use did not differ between these groups. One limitation of our study is the use of a four-channel recording device that does not measure sleep. Thus, our measurements of AHI, CAI, and CSR time are based on total recording time rather than total sleep time (TST). This issue may be relevant in heart failure since sleep efficiency is often low. However, TRT may also be useful (in addition to TST), since many subjects experience CSR while awake. Furthermore, although it is recognized that sleep efficiency is low in heart failure, patients were studied at home (using a minimally invasive device) rather than in the sleep laboratory and therefore sleep efficiency may have been better than previous reports would indicate. Because our goal was to assess a large group of patients in a CHF clinic, we believe that our use of home monitoring allowed us access to the largest and most representative group possible. We have previously tried to refer all CHF patients for overnight inlaboratory polysomnography, but only the most motivated patients were willing to undergo this test. Thus, although our use of home monitoring is a potential limitation, it may have helped us access a more generalizable sample of heart failure patients. Sensitivity in our study was also optimized by using nasal pressure rather than thermistry as in some of the prior reports. In conclusion, SDB and CSR remain highly prevalent in subjects with CHF despite optimization of medical therapy. This Journal of Clinical Sleep Medicine, Vol. 4, No. 1,

5 M Macdonald, J Fang, S Pittman et al observation suggests that the relationship between SDB and heart failure is still poorly understood since effective therapies for heart failure appear to have no major effect on SDB. Because available evidence suggests that CSR may contribute to the progression of heart failure, 20 CSR may still represent a potential target of therapy for the persistently symptomatic patient on optimal medical therapy REFERENCES 1. Javaheri S, Parker TJ, Wexler L, et al. Occult sleep-disordered breathing in stable congestive heart failure [published erratum appears in Ann Intern Med 1995 Jul 1;123(1):77]. Ann Intern Med 1995;122: Sin D, Fitzgerald F, Parker J. et al. Risk factors for central and obstructive sleep apnea in 450 men and women with congestive heart failure. Am. J. Respir. Crit. Care Med. 1999;160: Christ M, Sharkova Y, Fenske H, et al. Brain natriuretic peptide for prediction of Cheyne-Stokes respiration in heart failure patients. Int J Cardiol 2007;116: Ferrier K, Campbell A, Yee B, et al. Sleep-disordered breathing occurs frequently in stable outpatients with congestive heart failure. Chest 2005;128: Hosselet JJ, Norman RG, Ayappa I, Rapoport DM. Detection of flow limitation with a nasal cannula/pressure transducer. Am J Respir Crit Care Med 1998;157: Pack AI, Goldberg LR. Routine polysomnography is not indicated in congestive heart failure. Con. J Clin Sleep Med 2005;1: Sin D, Logan A, Fitzgerald F, et al. Effects of continuous positive airway pressure on cardiovascular outcomes in heart failure patients with and without Cheyne-Stokes respiration. Circulation 2000;102: Solin P, Bergin P, Richardson M, Kaye DM, Walters EH, Naughton MT. Influence of pulmonary capillary wedge pressure on central apnea in heart failure. Circulation 1999;99: Lloyd TC, Jr. Effect of increased left atrial pressure on breathing frequency in anesthetized dog. J Appl Physiol 1990;69: Witte KK, Thackray S, Nikitin NP, Cleland JG, Clark AL. The effects of long-term betablockade on the ventilatory responses to exercise in chronic heart failure. Eur J Heart Fail 2005;7: Warner MM, Mitchell GS. Role of catecholamines and beta-receptors in ventilatory response during hypoxic exercise. Respir Physiol 1991;85: Solin P, Roebuck T, Johns DP, Walters EH, Naughton MT. Peripheral and central ventilatory responses in central sleep apnea with and without congestive heart failure. Am. J. Respir. Crit. Care Med. 2000;162: Rechtschaffen A, Kales A. A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. Los Angeles: Brain Information Service/Brain Research Institute, UCLA, AASM. Sleep-related breathing disorders in adults: Recommendations for syndrome definition and measurement techniques in adults. Sleep 1999;22: Pitson DJ, Stradling JR. Autonomic markers of arousal during sleep in patients undergoing investigation for obstructive sleep apnoea, their relationship to EEG arousals, respiratory events and subjective sleepiness. J Sleep Res 1998;7: Malhotra A, Berry R, White D. Central sleep apnea. In: Carney PR, Berry RB, Geyer JD, Clinical Sleep Disorders. Philadelphia: Lippincott Williams & Wilkins, 2005: Krum H, Roecker EB, Mohacsi P, et al. Effects of initiating carvedilol in patients with severe chronic heart failure: results from the COPERNICUS Study. JAMA 2003;289: Krum H, Mohacsi P, Katus HA, et al. Are beta-blockers need- ed in patients receiving spironolactone for severe chronic heart failure? An analysis of the COPERNICUS study. Am Heart J 2006;151: Packer M, Coats AJ, Fowler MB, et al. Effect of carvedilol on survival in severe chronic heart failure. N Engl J Med 2001;344: Lanfranchi PA, Braghiroli A, Bosimini E, et al. Prognostic value of nocturnal Cheyne-Stokes respiration in chronic heart failure. Circulation 1999;99: Bradley TD, Logan AG, Kimoff RJ, et al. Continuous positive airway pressure for central sleep apnea and heart failure. N Engl J Med 2005;353: Eckert DJ, Jordan AS, Merchia P, Malhotra A. Central sleep apnea: Pathophysiology and treatment. Chest 2007; 131: Javaheri S. Acetazolamide improves central sleep apnea in heart failure: a double-blind, prospective study. Am J Respir Crit Care Med 2006;173: Javaheri S. Effects of continuous positive airway pressure on sleep apnea and ventricular irritability in patients with heart failure. Circulation 2000;101: Malhotra A, Muse VV, Mark EJ. Case records of the Massachusetts General Hospital. Weekly clinicopathological exercises. Case An 82-year-old man with dyspnea and pulmonary abnormalities. New England Journal of Medicine 2003;348: Journal of Clinical Sleep Medicine, Vol. 4, No. 1,