Differing Effects of Obstructive and Central Sleep ... - ATS Journals

Download PDF
9 downloads 0 Views 176KB Size Report
Comment
Feb 15, 2013 - Institute and Toronto General Hospital, Toronto, Ontario, Canada; 2Centre for Sleep Medicine and Circadian Biology, University of Toronto,.
Differing Effects of Obstructive and Central Sleep Apneas on Stroke Volume in Patients with Heart Failure Dai Yumino1,2, Takatoshi Kasai1,2, Derek Kimmerly1, Vinoban Amirthalingam1, John S. Floras1,3, and T. Douglas Bradley1,2 1 Sleep Research Laboratory, Centre for Sleep Health and Research, and Department of Medicine, University Health Network Toronto Rehabilitation Institute and Toronto General Hospital, Toronto, Ontario, Canada; 2Centre for Sleep Medicine and Circadian Biology, University of Toronto, Toronto, Ontario, Canada; and 3Department of Medicine, Mount Sinai Hospital, Toronto, Ontario, Canada

Rationale: Obstructive sleep apnea and central sleep apnea increase risk of mortality in patients with heart failure (HF), possibly because of hemodynamic compromise during sleep. However, beat-to-beat stroke volume (SV) has not been assessed in response to obstructive and central events during sleep in patients with HF. Because obstructive events generate negative intrathoracic pressure that reduces left ventricular (LV) preload and increases afterload, but central events do not, obstructive events should lead to greater hemodynamic compromise than central events. Objectives: To determine the effects of obstructive and central apneas and hypopneas during sleep on SV in patients with HF. Methods: Patients with systolic HF (LV ejection fraction < 45%) and sleep apnea underwent beat-to-beat measurement of SV by digital photoplethysmography during polysomnography. Change in SV from before to the end of obstructive and central respiratory events was calculated and compared between these types of events. Measurements and Main Results: Changes in SV were assessed during 252 obstructive and 148 central respiratory events in 40 patients with HF. Whereas SV decreased by 6.8 (68.7)% during obstructive events, it increased by 2.6 (65.4)% during central events (P , 0.001 for difference). For obstructive events, reduction in SV was associated independently with LV ejection fraction, duration of respiratory events, and degree of oxygen desaturation. Conclusions: In patients with HF, obstructive and central respiratory events have opposite hemodynamic effects: whereas obstructive sleep apnea appears to have an adverse effect on SV, central sleep apnea appears to have little or slightly positive effects on SV. These observations may have implications for therapeutic approaches to these two breathing disturbances. Keywords: stroke volume; central sleep apnea; obstructive sleep apnea

(Received in original form May 17, 2012; accepted in final form November 27, 2012) Supported by Canadian Institutes of Health Research operating grant MOP82731, Fuji-Respironics Inc. unrestricted research fellowships (D.Y. and T.K.), a Canadian Institutes of Health Research postdoctoral fellowship (D.K.), and by a Heart and Stroke Foundation of Ontario Career Investigator award and the Canada Research Chair in Integrative Cardiovascular Biology (J.S.F.). Author Contributions: D.Y. performed experiments, collected and analyzed data, wrote first draft, and contributed to revisions, including final draft of the manuscript; T.K. collected and analyzed data and contributed to revisions of the manuscript; D.K. assisted in experimental design, performed experiments, analyzed data, and contributed to revisions of the manuscript; V.A. collected data and contributed to revisions of the manuscript; J.S.F. assisted in experimental design and contributed to subject recruitment and revisions of the manuscript; T.D.B. funded and supervised the project, assisted in experimental design, contributed to subject recruitment, and wrote revisions and final draft of the manuscript. Correspondence and requests for reprints should be addressed to T. Douglas Bradley, M.D., University Health Network Toronto General Hospital, 9N-943, 200 Elizabeth Street, Toronto, ON, M5G 2C4, Canada. E-mail: [email protected] This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org Am J Respir Crit Care Med Vol 187, Iss. 4, pp 433–438, Feb 15, 2013 Copyright ª 2013 by the American Thoracic Society Originally Published in Press as DOI: 10.1164/rccm.201205-0894OC on December 13, 2012 Internet address: www.atsjournals.org

AT A GLANCE COMMENTARY Scientific Knowledge on the Subject

In patients with heart failure (HF), obstructive respiratory events generate negative intrathoracic pressure that reduces left ventricular (LV) preload and increases afterload, but central respiratory events do not. Therefore, obstructive events should lead to greater hemodynamic compromise than central events. However, beat-to-beat stroke volume (SV) has not been assessed in response to obstructive and central events during sleep in patients with HF. What This Study Adds to the Field

This study demonstrates that obstructive and central events have differing effects on SV: obstructive events caused a decrease, and central events caused a slight increase. For obstructive events, reductions in SV were associated independently with LV ejection fraction, duration of respiratory events, and degree of oxygen desaturation. Therefore, in patients with HF, obstructive sleep apnea appears to have adverse hemodynamic effects, whereas central sleep apnea appears to have little or slightly positive hemodynamic effects.

Central sleep apnea (CSA) and obstructive sleep apnea (OSA) are common in patients with heart failure (HF), and are associated with increased mortality (1, 2). Factors contributing to this excess mortality are not well understood. One possibility is that the combined effects of intermittent hypoxia, arousals from sleep, and excessive sympathetic nervous activity arising from CSA and OSA, plus negative intrathoracic pressure generation during obstructive apneas (OAs), could reduce stroke volume (SV) and impair tissue perfusion (3, 4). During OAs, negative inspiratory intrathoracic pressure generated against the occluded pharynx increases left ventricular (LV) transmural pressure, and hence afterload (5). Exposure to increased afterload causes much greater reductions in SV and cardiac output (CO) in the failing than in the normal LV (5, 6). Negative intrathoracic pressure also increases venous return, augmenting right ventricular preload, whereas OSA-induced hypoxic pulmonary vasoconstriction increases right ventricular afterload (7). Consequent right ventricular distension and leftward septal displacement during diastole impairs LV filling (8). During Mueller maneuvers that simulate OAs, this combination of increased LV afterload and diminished preload reduces SV and CO more in patients with HF than in healthy subjects (5). Moreover, whereas SV recovers abruptly to baseline in healthy subjects at apnea termination, in patients with HF, SV remains depressed for several seconds before recovering. However, these data were acquired in awake subjects, and may not reflect the influence of OAs on SV and CO during sleep.

434

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

In contrast to OAs, during central apneas (CAs), inspiratory efforts are absent and, therefore, no negative intrathoracic pressure is generated. As such, LV afterload should not increase, and LV preload should not decrease to the same extent as they would during obstructive events. Consequently, one would anticipate that central events would have less of a negative effect on SV and CO than obstructive events (9). We therefore hypothesized that, in patients with HF, OAs and obstructive hypopneas (OHs) will reduce SV and CO more during sleep than CAs and central hypopneas (CHs). Some aspects of this study have been published in abstract form (10).

METHODS For more detailed description of the methods, please see the online supplement.

Validation of Digital Photoplethysmographic Determination of SV To test the validity of noninvasive, beat-by-beat measures of SV by digital photoplethysmography (DPP) (Portapres; Finapres Medical Systems BV, Amsterdam, The Netherlands) during respiratory maneuvers, we compared changes in SV measured by DPP and echocardiographic-Doppler (echo-Doppler) during various respiratory maneuvers. These studies were performed on five healthy male subjects without any history of cardiovascular or respiratory diseases on no medications. The arm was placed in a splint to keep it straight during studies. The DPP has a height correction unit to adjust for differences between finger level and heart level to obtain comparable measurements of blood pressure and SV, irrespective of finger level (11). Breathing maneuvers were performed in the following order: an end-expiratory apnea (i.e., mouth pressure ¼ 0 mm Hg), Valsalva maneuvers at mouth pressures of 115 and 130 mm Hg, and Mueller maneuvers at mouth pressures of 215 and 230 mm Hg. Baseline SVs were computed as the mean values obtained during the 5-second period of normal breathing before each breathing maneuver and all changes in SV during each maneuver were calculated as differences from baseline. An ordinary least-products regression analysis was used to examine differences in the absolute measure as well as breathing maneuver–mediated changes in SV indices between the DPP and echo-Doppler techniques.

SV during Sleep in Patients with HF Subjects. Inclusion criteria were: (1) men and women aged 18 years or older; (2) HF due to ischemic or nonischemic dilated cardiomyopathy for 6 months or longer; (3) LV ejection fraction of 45% or less; (4) New York Heart Association class I–III; (5) sleep apnea defined as 10 apneas and hypopneas or more per hour of sleep (apnea–hypopnea index), and with at least 10 apneas and hypopneas during stage 2 sleep. Exclusion criteria were: (1) treated sleep apnea; (2) unstable angina, myocardial infarction, or cardiac surgery within the previous 3 months; (3) cardiac pacing; and (4) atrial fibrillation. The protocol was approved by the research ethics board of the Toronto Rehabilitation Institute, and all subjects provided written consent before participation. Demographic characteristics, medical history, and medication use were recorded. LV ejection fraction was measured by echocardiography within 3 months before polysomnography. Polysomnography. Subjects underwent overnight polysomnography (12, 13). Thoracoabdominal motion was monitored by respiratory inductance plethysmography, and nasal airflow by nasal pressure cannulae. Arterial oxygen saturation was monitored by oximetry on the ear. CA was defined as a 90% or greater reduction in tidal volume for at least 10 seconds without thoracoabdominal motion, and CH as 50–90% reduction in tidal volume from baseline for at least 10 seconds with in-phase thoracoabdominal motion and without airflow limitation on nasal pressure. OAs and hypopneas were similarly defined, except that they had to be accompanied by out-of-phase thoracoabdominal motion, or airflow limitation on nasal pressure (14). SV and CO were measured continuously by DPP, as described previously here, with the arm placed in a splint to keep it straight during studies. Subjects slept supine throughout the night to maintain constant position of the DPP.

VOL 187

2013

Statistical Analysis We calculated the mean difference in SV, heart rate, and CO, expressed as percent change, between the average values for the 5 seconds immediately before (baseline) and the last 5 seconds at the end of 10 randomly selected respiratory events of each type in stage 2 non–rapid eye movement sleep. We chose to make measurements of SV and CO during the 5 seconds before event onset because breathing is unobstructed during hyperpneas before either obstructive or central events, so that the loading conditions on the heart should be similar. Therefore, this is a reasonable period to act as a baseline to assess the effects of obstructive and central events on SV and CO. The duration of obstructive and central events was measured, and the degree of O2 desaturation associated with these events was calculated as the difference between the peak SaO2 just before or at the onset of respiratory events (i.e., baseline) and the nadir SaO2 during or just after termination of respiratory events. Comparisons between the obstructive and central events were performed by Student’s t test or and by Mann-Whitney U test. In addition, one-way ANOVA was used to test for overall differences in percent change of SV, heart rate, and CO among the different event types, followed by the Tukey-Kramer method. To evaluate factors associated with change in SV, we employed multivariate generalized logistic modeling with step-wise variable selection. Data are presented as means (6SD), or frequencies. A P value of less than 0.05 was considered statistically significant. All analyses were performed using SPSS 13.0.1 (SPSS Inc., Chicago, IL).

RESULTS Validation of DPP SV

Subjects’ mean (6SD) age and body mass index (BMI) were 28.0 (67.6) years and 24.6 (62.7) kg/m2, respectively. Table 1 provides summary data for the ordinary least-products regression analyses performed on the absolute (ml) and relative (%) breathing maneuver–related changes in SV derived from DPP and echoDoppler. These results indicate that both fixed and proportional biases were observed when comparing absolute SV changes between DPP and echo-Doppler. The 95% confidence interval for the ordinary least-products y intercept (a9) did not include 0 (267 to 250), which suggests that the DPP method of measurement provides consistently larger SV values than echo-Doppler. Furthermore, the 95% confidence interval for the coefficient of ordinary least-products slope (b9) did not include 1 (1.26–1.42), indicating that the dispersion of SV measures between DPP and echo-Doppler is proportional to the magnitude of SV change. However, when changes in SV were normalized to the SV during baseline normal breathing before breathing maneuvers, no systematic differences were observed between DPP and echoDoppler (Table 1). In addition, changes in SV derived from DPP correlated very strongly with those derived from echoDoppler during respiratory maneuvers (P , 0.001; Figure 1). SV during Sleep in Patients with HF

Of the 40 patients that we studied, 30 had predominantly OSA (i.e., >50% obstructive events), of whom 80% were men. Their mean age was 56.5 (610.3) years, BMI was 29.8 (66.5) kg/m2, and left ventricular ejection fraction (LVEF) was 33.4 (69.2)%. A total of 10 patients had predominantly CSA (i.e., .50% central events), of whom 80% were men. Their mean age was 59.5 (614.6) years, BMI was 28.5 (66.5) kg/m2, and LVEF was 29.0 (67.6)%. As shown in Tables 2 and 3, we studied 400 respiratory events during sleep in these 40 patients with HF. Of these events, 252 were obstructive and 148 were central. Obstructive events were accompanied by a significantly greater reduction in SV and CO than central events (26.8 6 8.7 versus 2.6 6 5.4%, P , 0.001; and 25.0 6 9.1 versus 2.5 6 7.4%, P , 0.001, respectively). Heart rate increased significantly more during obstructive events than during central events (2.2 6 7.3 versus 0.0 6

Yumino, Kasai, Kimmerly, et al.: Stroke Volume and Sleep Apnea

435

TABLE 1. RELATIONSHIP BETWEEN DIGITAL PHOTOPLETHYSMOGRAPHIC AND ECHOCARDIOGRAPHIC-DOPPLER–DERIVED CHANGES IN STROKE VOLUME DURING RESPIRATORY MANEUVERS IN HEALTHY SUBJECTS Variable

r

a9 (95% CI)

Fixed Bias*

b9 (95% CI)

Proportional Bias†

DSV (ml) DSV (%)

0.44 0.90

258 (267 to 250) 20.81 (22.62 to 2.11)

Yes No

1.34 (1.26 to 1.42) 1.03 (0.64 to 1.67)

Yes No

Definition of abbreviations: a9 ¼ y intercept in ordinary least-products regression model; b9 ¼ slope in ordinary leastproducts regression model; CI ¼ confidence interval; r ¼ product-moment correlation coefficient; SV ¼ stroke volume. a9 and b9 are coefficients in the ordinary least-products regression model (digital photoplethysmography–SV) ¼ a9 1 b9 (Echo-SV). * Fixed bias, if 95% CI for a9 does not include “0.” y Proportional bias, if 95% CI for b9 does not include “1.”

6.7%; P ¼ 0.003). However, there were no differences between obstructive and central events in terms of event duration (23.1 6 8.4 s versus 22.0 6 8.4 s, respectively; P ¼ 0.212), baseline SaO2 (93.7 6 2.4% versus 94.1 6 2.4%, respectively; P ¼ 0.080), nor the degree of oxygen desaturation (25.9 6 4.9% and 26.1 6 4.3%, respectively; P ¼ 0.967). Nevertheless, there was a graded reduction in percent change in SV and CO from CA to CH to OH to OA (P , 0.001; Figure 2). Degrees of oxygen desaturation in response to CA, CH, OH, and OA were 27.2 (64.7)%, 24.0 (62.5)%, 24.9 (64.1)%, and 27.8 (66.0)%, respectively (P , 0.001 for ANOVA; CA versus CH, P , 0.001; OH versus OA, P , 0.001; CA versus OA, P ¼ 0.855; CH versus OH, P ¼ 0.555). In addition, we compared the effects of respiratory events within the 10 patients who had all types of respiratory events, but in whom the numbers of each type of event differed. A total of 10 events were analyzed for each subject, for a total of 100 events, but the number of each type of event varied among subjects. For example, one subject had one OA, three OHs, two CAs, and four CHs (total ¼ 10), whereas another subject had five OAs, two OHs, two CAs, and one CH (total ¼ 10). The number of each of these types of events analyzed was, however, similar (27 OAs, 27 OHs, 27 CAs, and 19 CHs). Obstructive events were accompanied by a significantly greater reduction in SV and CO than were central events (24.1 6 7.4 versus 3.1 6 5.7%, P , 0.001; 23.4 6 8.8 versus 2.9 6 7.3%, P , 0.001, respectively). There was also a graded reduction in the change in SV and CO from CAs to CHs to OHs to OAs (from 3.2 6 6.7% to 2.9 6 3.9% to 23.4 6 6.1% to 24.7 6 8.5%, P , 0.001; and from 3.9 6 8.8% to 1.5 6 4.2% to 22.2 6 6.3% to 24.5 6 10.9%, P , 0.001). Change in heart rate was not different between obstructive and central events (0.7 6 5.8 versus 0.0 6 7.5%, P ¼ 0.562). In 252 obstructive events, multivariable analysis revealed that reduction in SV was independently associated with LV ejection fraction, duration of respiratory events, and degree of oxygen desaturation (Table 4). However, age, sex, etiology of HF, b-blocker use, presence of apnea versus hypopnea, and arousal were not independently associated with reduction in SV.

the same night, the effects of obstructive and central events on SV and CO were in the same direction and of similar magnitude as in the entire group, several of whom did not have all types of respiratory events on the same night. We can therefore conclude that the distribution of obstructive and central events among subjects had little or no effect on the SV and CO responses to them, thus minimizing the potential for confounding. Upper airway occlusion during OAs triggers four key pathophysiological effects: hypoxia; arousal from sleep at termination of apnea; sympatho-excitation; and exaggerated negative intrathoracic pressure. During OA, PaO2 falls and PaCO2 rises, leading to increased, but futile, inspiratory efforts that continue until the apnea is terminated by an arousal. Exaggerated negative intrathoracic pressure is generated during each inspiratory effort. The increase in right ventricular preload and LV afterload caused by the generation of negative intrathoracic pressure is a hemodynamic disturbance unique to OSA. Increased negative intrathoracic pressure increases venous return to the right ventricle, increasing its preload, whereas hypoxia causes pulmonary vasoconstriction and an increase in right ventricular afterload. If, as a result, right ventricular diastolic pressure exceeds LV diastolic pressure, the intraventricular septum will be displaced leftward and impede LV diastolic filling (15, 16). Exaggerated negative intrathoracic pressure also increases the pressure gradient between LV intracavitary and extracavitary (i.e., intrathoracic or pericardial) pressure, thus increasing end-systolic LV transmural pressure (5, 17, 18), a principal determinant of LV afterload. In patients with coronary artery

DISCUSSION Our study has given rise to several novel findings that provide insights into the effects of OSA and CSA on SV and CO during sleep in patients with HF. First, and most importantly, we found, in patients with HF with reduced LV ejection fraction, that obstructive events reduced SV and CO, whereas central events did not. In addition, there was a graded reduction in the change of SV and CO from CA to CH to OH to OA. Second, during obstructive events, the reduction in SV was independently associated with a lower LV ejection fraction, longer duration of respiratory events, and greater oxygen desaturation. A particular strength of our study is that, in a subset of the subjects who had all types of respiratory events on

Figure 1. This plot demonstrates a very close relationship between change in stroke volume assessed by digital photoplethysmography (DPP) versus that measured by echocardiographic-Doppler (echoDoppler; % from baseline) during respiratory maneuvers.

436

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

TABLE 2. CHARACTERISTICS OF THE PATIENTS WITH HEART FAILURE Variables Age, yr Male, % Body mass index, kg/m2 New York Heart Association class Ischemic etiology, % Left ventricular ejection fraction, % Thiazide or loop diuretics, % Angiotensin-converting enzyme inhibitor and/or angiotensin-2 antagonist, % b-blocker, % Total sleep time, min Apnea–hypopnea index, no./h of sleep Obstructive events, no. (%) Central events, no. (%)

Values (n ¼ 40) 57.2 6 11.3 80 29.5 6 6.4 2.1 6 0.5 35 32.3 6 8.8 70 100

280 29.2 83.7 16.3

6 6 6 6

83 63 19.1 14.4 (72) 14.4 (28)

Values are expressed as means 6 SD, unless otherwise indicated.

disease who also have OSA, these increases in afterload, in conjunction with declining oxygen saturation, can trigger ischemic changes in the electrocardiogram (19). Apnea-induced hypoxia can also reduce the rate of LV relaxation, thereby increasing the impedance to LV filling and increasing LV filling pressures (20). By one or more of these mechanisms, OAs can induce marked reductions in SV and CO that are proportional to the degree of developed negative intrathoracic pressure (5). Nevertheless, ours is the first study to measure beat-by-beat SV and CO during sleep in patients with HF with OSA, and to demonstrate that, in such patients, SV and CO decrease progressively during OAs and hypopneas. These observations provide insights into how OSA can trigger nocturnal cardiac ischemia and acute pulmonary edema (19, 21), impair myocardial contractility, and contribute to the development and progression of HF. They also help to explain the increased mortality risk associated with the presence of OSA in patients with HF (22). In addition, acute alleviation of OSA by continuous positive airway pressure (CPAP) eliminates apnea-related intermittent hypoxia and abolishes negative intrathoracic pressure swings, thereby reducing right ventricular preload and LV afterload (23). It also reduces blood pressure that further reduces LV afterload. These unloading effects are probably a mechanism by which chronic CPAP therapy led to a 9% improvement in LV ejection fraction in a 1-month randomized trial involving patients with HF with severe OSA (24). In contrast to OAs, no inspiratory efforts are made during CAs, so that no negative intrathoracic pressure is generated. Similarly, because there is no airflow limitation during CHs, the degree of negative inspiratory intrathoracic pressure generation is lower than during OHs (3, 5, 9). Because intrathoracic pressure becomes less negative during the apneic phase than during the hyperpneic phase of the periodic breathing cycle, it stands to reason that the potential adverse effects of negative intrathoracic pressure would be reduced during CAs and hypopneas compared with hyperpneas. Consequently, one would anticipate that adverse hemodynamic effects related to generation of negative intrathoracic pressure would be less pronounced in patients with HF with CSA than in those with OSA. This is indeed what we observed. During obstructive events, SV and CO fell significantly, whereas during central events they actually increased modestly compared with the ventilatory period. Differences in SV and CO responses to obstructive and central events, however, could not be attributed to differences in duration of events, baseline SaO2, or degree of oxygen desaturation during events, because these factors did not differ between obstructive and central events.

VOL 187

2013

These observations are consistent with those of our previous studies in which we compared the effects of simulated OA (i.e., Mueller maneuvers) and CAs (i.e., breath holds) during wakefulness on SV and CO in patients with HF under normoxic conditions. During Mueller maneuvers in which 230 cm H2O of intrathoracic pressure was generated, SV and CO decreased significantly, whereas during breath holds in which negative intrathoracic pressure was not generated, neither SV nor CO changed (5, 9). However, superficially, our observations vary from those of some other investigators. For example, regarding hemodynamic effects of central events, Maze and colleagues (25) purported to measure LV inflow tract velocity as an index of SV in HF with Cheyne-Stokes respiration, and reported that flow–velocity rose at the end of hyperpnea and fell at the end of apnea. However, Cheyne-Stokes respiration was not defined, they had no means of precisely timing flow–velocity measurements with the apnea–hyperpnea cycle, and could not distinguish between obstructive and central events, because they did not monitor respiration objectively. Similarly, Chen and Scharf (26) and Tarasiuk and Scharf (27) reported that, in sedated animals with normal cardiac function, CAs induced a decrease in SV. However, in both cases, apneas were induced during sedation by pharmacologically paralyzing the respiratory muscles and turning off the ventilator. Thus, apneas were not central, because PCO2 remained above the apnea threshold, so that central drive to the respiratory muscles would have persisted, but the paralyzed muscles could not respond. Therefore, these apneas did not duplicate the pathophysiology of CAs in HF. Thus, seeming discrepancies between these studies and ours are most likely attributable to differing experimental conditions. In view of our finding, because of the greater unloading of the LV by nocturnal CPAP in patients with HF with OSA than in those with CSA, one would expect a greater degree of improvement in LV function during chronic CPAP treatment of OSA than of CSA. In fact, this appears to be the case: treatment of OSA by CPAP for 1–3 months has been reported to improve LV ejection fraction by 5–12% (24, 28, 29), whereas treatment of CSA by CPAP for similar periods was reported to increase LV ejection fraction by 2.5–8% (30, 31). However, there have been no long-term randomized trials evaluating whether treating OSA by CPAP in patients with HF improves morbidity and mortality. In the case of CSA, evidence from one multicenter randomized trial demonstrated that CPAP attenuated CSA in association with a modest improvement in LV ejection fraction, a decrease in sympathetic nervous activity, and an increase in exercise performance (30). However, these improvements were not associated with any improvement in heart transplant–free survival or a reduction in hospitalizations. Another interesting finding was that, during obstructive events, the lower the LV ejection fraction, the longer the respiratory event, and the greater the degree of O2 desaturation, the TABLE 3. CHARACTERISTICS OF 400 RESPIRATORY EVENTS IN THE 40 PATIENTS Variables Respiratory events, no. Obstructive hypopneas, no. Obstructive apneas, no. Central hypopneas, no. Central apneas, no. Duration of respiratory events, s Degree of oxygen desaturation, % Presence of arousals, % Values are expressed as means 6 SD, unless otherwise indicated.

Values 400 168 84 53 95 22.7 6 8.4 5.9 6 4.7 61

Yumino, Kasai, Kimmerly, et al.: Stroke Volume and Sleep Apnea

437

Figure 2. Change in stroke volume (SV; upper panel) and cardiac output (CO; lower panel) from baseline during obstructive apneas (OAs), obstructive hypopneas (OHs), central apneas (CAs), and central hypopneas (CHs). Whereas SV and CO decreased during OHs (25.6 6 0.6% and 23.7 6 0.5%), and decreased further during OAs (29.2 6 1.2% and 27.5 6 1.3%), they increased during CHs (12.1 6 0.6% and 10.9 6 0.5%), and increased further during CAs (12.9 6 0.05% and 13.4 6 0.9%). *P , 0.05 from baseline to end of events; **P , 0.001 from baseline to end of events.

greater the fall in SV independent of other factors. This indicates that the greater the LV systolic compromise, and the greater the duration of respiratory events and degree of hypoxia to which the compromised LV is exposed during obstructive events, the greater the deterioration in SV and CO. The adverse effect of hypoxia on SV and CO during obstructive events was most likely due to a local, direct myocardial depressant effect, rather than to peripheral chemoreceptor stimulation, which

would increase cardiac sympathetic activity and augment rather than suppress SV and CO. Our study is subject to some limitations. First, although we validated the accuracy of the DPP for detecting the change in SV during respiratory maneuvers by echocardiography in healthy subjects, we did not determine this in patients with HF. However, there is no obvious reason why DPP would not track changes in SV as precisely as in patients with HF as it does in healthy subjects. Second, we did not measure intrathoracic pressure by esophageal manometry during sleep. This is because there is a limit to the number of factors that can be measured in patients with HF without disrupting their sleep. Being instrumented for full polysomnography and wearing DPP with the arm splinted all night can be quite uncomfortable, and we did not wish to add further discomfort by insertion of an esophageal balloon that might have disrupted patients’ sleep, and compromised our ability to collect the essential data. Nevertheless, it is well known from previous studies that much greater degrees of negative intrathoracic pressure are generated during obstructive than during central events in patients with HF (3, 5, 9, 28, 32). Third, it is possible that differing levels of sympathetic activity before respiratory events might have influenced the SV and CO responses to them. However, the only means of testing acute alteration in sympathetic activity in humans is via invasive measurement of muscle sympathetic activity via the peroneal nerve. As with esophageal manometry, because of the multiple measurements to which our patients were subjected, it was not feasible to measure muscle sympathetic activity during our protocol. In conclusion, our findings indicate that, in patients with HF, obstructive and central respiratory events are associated with opposite effects on SV and CO. During obstructive events, SV and CO fall, whereas, during central events, they rise relative to the baseline ventilatory period. These findings suggest that OSA has substantial adverse hemodynamic effects in patients with HF that may be amenable to CPAP therapy (5, 9, 24, 28, 29). On the other hand, our findings also raise the possibility that CSA may not have adverse hemodynamic effects, and indeed may have no or slightly positive hemodynamic effects in patients with HF, which may have some clinical implications. For example, among patients with HF who convert from OA to CA either overnight (23) or over longer periods (33), elimination of negative intrathoracic pressure swings when converting from obstructive to central events could reduce right ventricular preload and LV afterload, negating the adverse effects of OAs and hypopneas on SV and CO. Further studies will be required to test this intriguing possibility. Finally, our findings also suggest greater potential for improvement in cardiovascular function, and possibly morbidity and mortality, through treatment of OSA than of CSA in patients with HF, as elimination of OSA

TABLE 4. FACTORS ASSOCIATED WITH REDUCTION IN STROKE VOLUME DURING OBSTRUCTIVE EVENTS Univariable Analysis

Multivariable Analysis

Factors

b-Coefficient (95% CI)

P Value

b-Coefficient (95% CI)

P Value

Age, per 1-yr increase Male Ischemic etiology LV ejection fraction, per 1% decrease Apnea (compared to hypopnea) Presence of arousal Duration of respiratory events, per 1-s increase Oxygen desaturation, per 1% decrease

0.04 3.01 20.43 0.36 4.08 2.25 0.32 0.79

0.503 0.032 0.755 ,0.001 0.001 0.063 ,0.001 ,0.001

0.02 1.81 22.01 0.31 1.90 0.68 0.15 0.47

0.742 0.183 0.177 ,0.001 0.113 0.486 0.036 ,0.001

(20.07 to 0.15) (0.27 to 5.76) (23.13 to 2.17) (0.24 to 0.49) (1.60 to 6.56) (20.12 to 4.63) (0.18 to 0.45) (0.57 to 1.01)

Definition of abbreviations: CI ¼ confidence interval; LV ¼ left ventricular.

(20.10 to 0.14) (20.86 to 4.47) (24.93 to 0.91) (0.18 to 0.43) (20.46 to 4.26) (22.38 to 3.68) (0.01 to 0.29) (0.22 to 0.73)

438

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

by CPAP would have a greater cardiovascular unloading effect than elimination of CSA. Author disclosures are available with the text of this article at www.atsjournals.org.

References 1. Yumino D, Wang H, Floras JS, Newton GE, Mak S, Ruttanaumpawan P, Parker JD, Bradley TD. Prevalence and physiological predictors of sleep apnea in patients with heart failure and systolic dysfunction. J Card Fail 2009;15:279–285. 2. Yumino D, Wang H, Floras JS, Newton GE, Mak S, Ruttanaumpawan P, Parker JD, Bradley TD. Relationship between sleep apnoea and mortality in patients with ischaemic heart failure. Heart 2009;95:819– 824. 3. Yumino D, Bradley TD. Central sleep apnea and Cheyne-Stokes respiration. Proc Am Thorac Soc 2008;5:226–236. 4. Bradley TD, Floras JS. Obstructive sleep apnoea and its cardiovascular consequences. Lancet 2009;373:82–93. 5. Bradley TD, Hall MJ, Ando S, Floras JS. Hemodynamic effects of simulated obstructive apneas in humans with and without heart failure. Chest 2001;119:1827–1835. 6. Ross J Jr. Afterload mismatch and preload reserve: a conceptual framework for the analysis of ventricular function. Prog Cardiovasc Dis 1976; 18:255–264. 7. Stoohs R, Guilleminault C. Cardiovascular changes associated with obstructive sleep apnea syndrome. J Appl Physiol 1992;72:583–589. 8. Brinker JA, Weiss JL, Lappe DL, Rabson JL, Summer WR, Permutt S, Weisfeldt ML. Leftward septal displacement during right ventricular loading in man. Circulation 1980;61:626–633. 9. Hall MJ, Ando S, Floras JS, Bradley TD. Magnitude and time course of hemodynamic responses to Mueller maneuvers in patients with congestive heart failure. J Appl Physiol 1998;85:1476–1484. 10. Yumino D, Kimmerly D, Redolfi S, Su MC, Ruttanaumpawan P, Amirthalingam V, Maybury B, Smith S, Floras JS, Bradley TD. Acute effects of obstructive and central apneas on stroke volume in heart failure. Am J Respir Crit Care Med 2009;179:A3607. 11. Imholz BP, Langewouters GJ, van Montfrans GA, Parati G, van Goudoever J, Wesseling KH, Wieling W, Mancia G. Feasibility of ambulatory, continuous 24-hour finger arterial pressure recording. Hypertension 1993;21: 65–73. 12. EEG arousals: scoring rules and examples: a preliminary report from the sleep Disorders Atlas Task Force of the American Sleep Disorders Association. Sleep 1992;15:173–184. 13. Rechtschaffen A, Kales A. A manual of standardized terminology, techniques and scoring for sleep stages of human subjects. Los Angeles: UCLA Brain Information Service/Brain Research Institute; 1968. 14. Yumino D, Redolfi S, Ruttanaumpawan P, Su MC, Smith S, Newton GE, Mak S, Bradley TD. Nocturnal rostral fluid shift: a unifying concept for the pathogenesis of obstructive and central sleep apnea in men with heart failure. Circulation 2010;121:1598–1605. 15. Tolle FA, Judy WV, Yu PL, Markand ON. Reduced stroke volume related to pleural pressure in obstructive sleep apnea. J Appl Physiol 1983;55:1718–1724.

VOL 187

2013

16. Peters J, Fraser C, Stuart RS, Baumgartner W, Robotham JL. Negative intrathoracic pressure decreases independently left ventricular filling and emptying. Am J Physiol 1989;257:H120–H131. 17. Buda AJ, Pinsky MR, Ingels NB Jr, Daughters GT II, Stinson EB, Alderman EL. Effect of intrathoracic pressure on left ventricular performance. N Engl J Med 1979;301:453–459. 18. Virolainen J, Ventila M, Turto H, Kupari M. Effect of negative intrathoracic pressure on left ventricular pressure dynamics and relaxation. J Appl Physiol 1995;79:455–460. 19. Franklin KA, Nilsson JB, Sahlin C, Naslund U. Sleep apnoea and nocturnal angina. Lancet 1995;345:1085–1087. 20. Cargill RI, Kiely DG, Lipworth BJ. Adverse effects of hypoxaemia on diastolic filling in humans. Clin Sci (Lond) 1995;89:165–169. 21. Chaudhary BA, Ferguson DS, Speir WA Jr. Pulmonary edema as a presenting feature of sleep apnea syndrome. Chest 1982;82:122–124. 22. Wang H, Parker JD, Newton GE, Floras JS, Mak S, Chiu KL, Ruttanaumpawan P, Tomlinson G, Bradley TD. Influence of obstructive sleep apnea on mortality in patients with heart failure. J Am Coll Cardiol 2007;49:1625–1631. 23. Tkacova R, Niroumand M, Lorenzi-Filho G, Bradley TD. Overnight shift from obstructive to central apneas in patients with heart failure: role of PCO2 and circulatory delay. Circulation 2001;103:238–243. 24. Kaneko Y, Floras JS, Usui K, Plante J, Tkacova R, Kubo T, Ando S, Bradley TD. Cardiovascular effects of continuous positive airway pressure in patients with heart failure and obstructive sleep apnea. N Engl J Med 2003;348:1233–1241. 25. Maze SS, Kotler MN, Parry WR. Doppler evaluation of changing cardiac dynamics during Cheyne-Stokes respiration. Chest 1989;95:525–529. 26. Chen L, Scharf SM. Comparative hemodynamic effects of periodic obstructive and simulated central apneas in sedated pigs. J Appl Physiol 1997;83:485–494. 27. Tarasiuk A, Scharf SM. Cardiovascular effects of periodic obstructive and central apneas in dogs. Am J Respir Crit Care Med 1994;150: 83–89. 28. Malone S, Liu PP, Holloway R, Rutherford R, Xie A, Bradley TD. Obstructive sleep apnoea in patients with dilated cardiomyopathy: effects of continuous positive airway pressure. Lancet 1991;338:1480– 1484. 29. Mansfield DR, Gollogly NC, Kaye DM, Richardson M, Bergin P, Naughton MT. Controlled trial of continuous positive airway pressure in obstructive sleep apnea and heart failure. Am J Respir Crit Care Med 2004;169:361–366. 30. Bradley TD, Logan AG, Kimoff RJ, Series F, Morrison D, Ferguson K, Belenkie I, Pfeifer M, Fleetham J, Hanly P, et al. Continuous positive airway pressure for central sleep apnea and heart failure. N Engl J Med 2005;353:2025–2033. 31. Naughton MT, Liu PP, Bernard DC, Goldstein RS, Bradley TD. Treatment of congestive heart failure and Cheyne-Stokes respiration during sleep by continuous positive airway pressure. Am J Respir Crit Care Med 1995;151:92–97. 32. Takasaki Y, Orr D, Popkin J, Rutherford R, Liu P, Bradley TD. Effect of nasal continuous positive airway pressure on sleep apnea in congestive heart failure. Am Rev Respir Dis 1989;140:1578–1584. 33. Tkacova R, Wang H, Bradley TD. Night-to-night alterations in sleep apnea type in patients with heart failure. J Sleep Res 2006;15:321–328.