Impaired Pulmonary Diffusing Capacity and Hypoxia in Heart Failure Correlates With Central Sleep Apnea Severity*

Original Research SLEEP MEDICINE

Impaired Pulmonary Diffusing Capacity and Hypoxia in Heart Failure Correlates With Central Sleep Apnea Severity* Irene Szollosi, BSc; Bruce R. Thompson, PhD; Henry Krum, MBBS, PhD; David M. Kaye, MBBS, PhD; and Matthew T. Naughton, MD

Background: Heart failure (HF) is often associated with interstitial pulmonary edema and structural changes, resulting in thickening of the alveolar-capillary membrane and reductions in diffusing capacity of the lung for carbon monoxide (DLCO). Reduced DLCO reflects an impaired efficiency of gas exchange, which may increase plant gain, influence ventilatory control stability, and result in central sleep apnea (CSA). In this study, we test the hypothesis that reductions in DLCO would be associated with increased apnea-hypopnea index (AHI) in patients with CSA. Methods: Overnight polysomnography, pulmonary function tests, and arterial blood gas analyses were performed in 45 patients with chronic, stable HF. Univariate and multivariate regression analyses were performed in those patients with predominant CSA to test which variables were associated with AHI. Results: Patients had a mean (ⴞ SD) age of 52.7 ⴞ 8.9 years, a mean left ventricular ejection fraction of 26.5 ⴞ 9.9%, and a mean AHI of 22.0 ⴞ 17.4 events per hour. In CSA patients, DLCO and PaO2 both correlated with total AHI (r ⴝ ⴚ 0.43, p ⴝ 0.046 and r ⴝ ⴚ 0.53, p ⴝ 0.011, respectively) and with supine AHI (r ⴝ ⴚ 0.56, p ⴝ 0.009 and r ⴝ ⴚ 0.60, p ⴝ 0.004, respectively). In a forward stepwise estimation model, DLCO, PaO2, and body mass index were independent predictors of total AHI, explaining 51% of variability, as was supine AHI, explaining 64% of variability. DLCO and PaO2 accounted for 37% of the variability in total AHI and 49% of the variability in supine AHI. Conclusions: In patients with HF and CSA, reductions in DLCO and PaO2 are independently associated with respiratory disturbance during sleep. The increase in ventilatory instability may be due to plant gain effects. (CHEST 2008; 134:67–72) Key words: central sleep apnea; heart failure; plant gain Abbreviations: AHI ⫽ apnea-hypopnea index; BMI ⫽ body mass index; CSA ⫽ central sleep apnea; Dlco ⫽ diffusing capacity of the lung for carbon monoxide; FRC ⫽ functional residual capacity; HF ⫽ heart failure; Kco ⫽ diffusing capacity of the lung for carbon monoxide corrected for alveolar volume; LVEF ⫽ left ventricular ejection fraction; OSA ⫽ obstructive sleep apnea; Spo2 ⫽ pulse oximetric saturation; VA ⫽ alveolar volume

failure (HF) is associated with a high H eart prevalence of breathing abnormalities during sleep. In a pattern dependent on the severity of HF, between 37% and 75% of patients have central sleep apnea (CSA) with periodic breathing or CheyneStokes respiration, with a further 10 to 25% of patients also having obstructive sleep apnea (OSA).1–3 CSA is attributed to ventilatory instability arising from increased circulation time and increased loop gain in the metabolic control of ventilation,4,5 while OSA is attributed to upper airway collapse, with some studies6,7 showing that increased loop gain may contribute www.chestjournal.org

to its pathogenesis in some patients. In control theory, the loop gain is the sum of the controller gain and the plant gain of the control system. In the ventilatory control system, controller gain is analogous to chemoFor editorial comment see page 7 receptor sensitivity, which is found to be heightened in patients with CSA,8 and results in brisk ventilatory responses to small changes in blood gas levels. Plant gain refers to the influence of the efficiency of the gas exchange system, particularly in relation to the CHEST / 134 / 1 / JULY, 2008

67

way in which changes in ventilation impact on blood gas levels. Theoretically, an increase in plant gain or a reduction in the efficiency of gas exchange may contribute to respiratory instability; however, this relationship has not been previously investigated. HF is typically accompanied by a degree of elevation of the pulmonary capillary wedge pressures,9 depending on the severity of the underlying heart disease and the efficacy of therapy. In association with this, interstitial pulmonary edema and structural changes result in progressive thickening of the alveolar-capillary membrane.10 The function of the alveolar-capillary membrane is assessed by the diffusing capacity of the lung for carbon monoxide (Dlco), which is shown to be impaired in HF patients11 and is associated with decrements in exercise capacity.12–15 Dlco is a measure of the ease of transfer for pulmonary gas exchange; therefore, alterations in Dlco are likely to affect plant gain and influence the stability of the ventilatory control system. Thus, we hypothesized that in HF patients with predominant CSA during sleep, where ventilatory control instability rather than upper airway collapsibility plays a key pathophysiologic role, reductions in Dlco would be associated with more severe respiratory disturbance during sleep. Materials and Methods

Polysomnography Overnight polysomnography was performed using a data acquisition and analysis system (Profusion-2, Compumedics Eseries; Compumedics; Abbottsford, NSW, Australia). The physiologic signals recorded were as follows: EEG (leads C4-A1, C3-A2, O2-A1, and O1-A2); electrooculogram; ECG; submental electromyogram; oronasal flow measured with thermistor (model 971; Nellcor Puritan Bennett; Minneapolis, MN) and a nasal pressure cannula (Salter Laboratories; Arvin, CA); ribcage and abdominal effort using piezoelectric sensors; snoring sounds measured with a microphone; pulse oximetric saturation (Spo2) [Oxypleth; Novametrix; Wallingford, CT] with the average time set at 3 s; calibrated body position; and continuous digital video monitoring during sleep. Sleep stages and arousals were scored manually according to standard criteria.16,17 Respiratory events were scored according to modified Chicago criteria18 with hypopneas defined as a reduction in oronasal flow for a duration of at least 10 s accompanied by desaturation of ⱖ2% and/or arousal. Hypopneas were classified as obstructive if paradoxical motion or notching occurred in the respiratory bands and there was evidence of upper airway resistance, such as snoring or inspiratory flow limitation in the nasal pressure signal, and central if respiratory movements remained in phase without evidence of upper airway resistance. Subjects were classified as having stable breathing if they had an apnea-hypopnea index (AHI) of ⬍ 10 events per hour, CSA if ⬎ 75% of events were central, and OSA if ⬎ 25% of events were obstructive. Mixed events were grouped with obstructive events to yield a purely central group for analysis. An arterial blood gas sample was obtained from a radial artery while the subject was supine, awake, and had rested for 10 min on the night of the polysomnography. Pulmonary Function

Subjects We evaluated 45 consecutive HF patients who had been referred for polysomnography for investigation of sleep-disordered breathing, meeting the following inclusion criteria: age, 18 to 75 years; and documented ischemic or idiopathic dilated cardiomyopathy of at least 6 months in duration, with left ventricular ejection fraction (LVEF) ⬍ 50%, New York Heart Association class II–IV, and stable medical condition at the time of assessment (ie, no hospital admissions or changes in medication for at least 2 weeks). The exclusion criteria were primary valvular, congenital, or restrictive cardiomyopathy, recent (within 3 months) myocardial infarction, significant coexisting pulmonary disease (forced expiratory ratio, ⬍ 70% predicted), and neurologic impairment. The study was approved by the Alfred Hospital’s Ethics Committee and patients provided written informed consent. *From the Departments of Allergy Immunology and Respiratory Medicine (Ms. Szollosi, and Drs. Thompson and Naughton), and Cardiology (Drs. Krum and Kaye), Alfred Hospital, Melbourne, VIC, Australia. This research was supported by an Australian Postgraduate Award (to Ms. Szollosi). The authors have reported to the ACCP that no significant conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article. Manuscript received June 13, 2007; revision accepted December 2, 2007. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml). Correspondence to: Matthew T. Naughton, MD, Alfred Hospital, Respiratory Medicine, Commercial Rd, Prahran, VIC 3181, Australia; e-mail: [email protected] DOI: 10.1378/chest.07-1487 68

Spirometry and single-breath Dlco was measured (Profiler; Medgraphics; St. Paul, MN) according to American Thoracic Society/European Respiratory Society guidelines,19 with an inspired gas mixture containing 0.3% carbon monoxide, 0.5% neon, and 20.6% oxygen, with the balance being nitrogen. At least two measurements were performed that had to agree within 3 mL/ min/mm Hg or 10%, whichever was greater, with at least a 4-min interval between repeat tests. All Dlco results were corrected for hemoglobin using the method described by Cotes20 using hemoglobin values obtained from arterial blood gas sampled from the radial artery. Dlco corrected for alveolar volume (Kco) was calculated as the Dlco divided by the alveolar volume (VA). Statistical Analysis All statistical analyses were performed using statistical software package (Stata9; StataCorp LP; College Station, TX). Differences between groups were assessed using the ␹2 test for categoric variables, and one-way analysis of variance for continuous variables with post hoc analyses were performed using the Bonferroni correction. Pearson correlation coefficients were calculated between AHI and age, body mass index (BMI), LVEF, pulmonary function, and arterial blood gas data in patients with predominant CSA. A forward stepwise estimation was performed in which significant predictors as well as age, BMI, LVEF, and VA were included to identify which parameters were independently associated with AHI.

Results Subject characteristics are presented in Table 1, with all results presented as the mean ⫾ SD. There Original Research

Table 1—Subject Characteristics*

Characteristics Age, yr Gender Male Female Dilated cardiomyopathy Ischemic Idiopathic Heart rhythm SR Paced AF NYHA class II III IV BMI, kg/m2 LVEF, % ESS score Total sleep time, min Sleep efficiency, % AHI, events/h Total Supine position Nonsupine position Sleep status, % Supine position Stage 1 Stage 2 SWS REM Spo2, % Mean Minimum Sleep heart rate, beats/min FEV1, % FVC, % FER, % TLC,§ % FRC,㛳 % VA, % Dlco, % Kco, % pH Paco2 Pao2

StableBreathing Group (n ⫽ 13) 50 ⫾ 12

OSA Group (n ⫽ 16) 54 ⫾ 6

CSA Group (n ⫽ 16) 53 ⫾ 8

9 4

14 2

14 2

6 7

4 12

8 8

6 5 2

9 6 1

6 7 3

4 6 3 23.4 ⫾ 3.2 25.9 ⫾ 7.5 7.8 ⫾ 3.3 273 ⫾ 72 66 ⫾ 17 5.8 ⫾ 2.7

6 6 4 29.0 ⫾ 6.3† 28.7 ⫾ 8.3 10.2 ⫾ 3.6 304 ⫾ 69 74 ⫾ 18 23.2 ⫾ 15.1†

1 12 3 24.9 ⫾ 4.1 24.8 ⫾ 12.6 6.1 ⫾ 3.4‡ 270 ⫾ 98 62 ⫾ 22 33.8 ⫾ 16.7†

15.4 ⫾ 16.7 3.6 ⫾ 2.6

35.8 ⫾ 23.6† 14.1 ⫾ 19.1

46.3 ⫾ 16.7† 30.2 ⫾ 20.2†‡

28 ⫾ 26 10 ⫾ 7 52 ⫾ 11 24 ⫾ 12 14 ⫾ 8

55 ⫾ 27† 14 ⫾ 9 58 ⫾ 13 14 ⫾ 11 15 ⫾ 7

53 ⫾ 31 23 ⫾ 20† 49 ⫾ 13 17 ⫾ 11 11 ⫾ 9

97.7 ⫾ 0.8 92.0 ⫾ 5.8 61 ⫾ 13

96.3 ⫾ 1.6 84.3 ⫾ 5.3† 68 ⫾ 10

95.4 ⫾ 2.0† 83.3 ⫾ 8.0† 67 ⫾ 10

79 ⫾ 12 86 ⫾ 17 92 ⫾ 9 96 ⫾ 13 98 ⫾ 22 85 ⫾ 15 62 ⫾ 18 74 ⫾ 13 7.43 ⫾ 0.03 41.8 ⫾ 5.6 99 ⫾ 13

92 ⫾ 18 94 ⫾ 15 96 ⫾ 8 94 ⫾ 11 78 ⫾ 11† 92 ⫾ 13 64 ⫾ 17 68 ⫾ 17 7.43 ⫾ 0.03 40.5 ⫾ 4.5 88 ⫾ 15

76 ⫾ 14‡ 84 ⫾ 16 90 ⫾ 8 89 ⫾ 11 87 ⫾ 13 82 ⫾ 11 54 ⫾ 17 65 ⫾ 19 7.46 ⫾ 0.04‡ 35.7 ⫾ 4.4†‡ 79 ⫾ 13†

*Values are given as mean ⫾ SD or No., unless otherwise indicated. SR ⫽ sinus rhythm; AF ⫽ atrial fibrillation; NYHA ⫽ New York Heart Association; ESS ⫽ Epworth sleepiness scale; SWS ⫽ slowwave sleep; REM ⫽ rapid eye movement; FER ⫽ forced expiratory ratio; TLC ⫽ total lung capacity. †p ⬍ 0.05 compared to the stable-breathing group. ‡p ⬍ 0.05 compared to the OSA group. §The number if patients in each group are 10, 13, and 12, respectively. 㛳A total of 16 of 22 patients were included in the regression analysis.

were 13 patients in the stable-breathing group, 16 patients in the OSA group, and 16 patients in the CSA group. Patients with OSA had higher BMI and reduced functional residual capacity (FRC) percent predicted than patients in the stable-breathing group. The CSA and OSA groups had higher AHI and lower minimum Spo2 during sleep than the stable-breathing group. The CSA group also showed reduced Paco2 compared to the OSA and stablebreathing groups, and reduced Pao2 compared to the stable-breathing group. The groups were similar in most other respects, including medication use, which was as follows for all groups: angiotensinconverting enzyme inhibitors, 76%; diuretics, 96%; ␤-blockers, 78%; amiodarone, 33%; angiotensin II receptor antagonist, 22%; and digoxin, 58%. There was no difference in smoking history between groups with 42% never-smokers, 53% ex-smokers, and 4% current smokers. Six of 13 patients who were classified as having stable breathing had mainly central events with an AHI of ⬍ 10; thus, a total of 22 patients were included in the regression analyses. Using Pearson correlation analyses, there were significant correlations between Dlco and Pao2, and total AHI (Dlco percentage: r ⫽ ⫺ 0.43; p ⫽ 0.046; Pao2: r ⫽ ⫺ 0.53; p ⫽ 0.011); and an additional significant correlation between Paco2 and Kco with supine AHI (Dlco percentage: r ⫽ ⫺ 0.56; p ⫽ 0.009; Pao2: r ⫽ ⫺ 0.60; p ⫽ 0.004; Paco2: r ⫽ ⫺ 0.48; p ⫽ 0.038; Kco percent predicted: r ⫽ ⫺ 0.63; p ⫽ 0.003) was found. In the forward stepwise estimation model, Pao2, Dlco, and BMI were all independent predictors of total AHI, explaining 51% of the variability, as well as independent predictors of supine AHI, explaining 64% of the variability (Table 2). Dlco and Pao2 accounted for 37% of the variability in total AHI and 49% of the variability in supine AHI. Figure 1 shows the relationships among supine AHI, Dlco, and Pao2 in patients with predominant CSA. Table 2—Multiple Regression Analyses* Variables AHI total Pao2 Dlco % BMI Constant AHI supine Pao2 Dlco % BMI Constant

Correlation Coefficient

SE

p Value

⫺ 0.51 ⫺ 0.39 1.93 43.3

0.22 0.18 0.87 28.6

0.031 0.043 0.040 0.147

⫺ 0.69 ⫺ 0.50 2.40 63.6

0.21 0.18 0.86 28.1

0.005 0.010 0.012 0.036

R2 Value 0.508

0.645

*Variables use in the regression analysis were Dlco percentage, Paco2, Pao2, age, BMI, LVEF, and VA percentage. www.chestjournal.org

CHEST / 134 / 1 / JULY, 2008

69

Figure 1. Relationship between Dlco and Pao2 and total AHI, illustrating the independent effect of Dlco and Pao2 on supine AHI.

Discussion This is the first study to find a significant relationship between Dlco and CSA in patients with HF, thus providing clinical evidence that impaired gas exchange is associated with, and may contribute to, ventilatory instability during sleep. We found that impaired gas exchange across the pulmonary capillary membrane, as measured by Dlco, was independently associated with increased respiratory disturbance during sleep in CSA. Furthermore, although patients with CSA had reduced Paco2, an independent negative correlation in multivariate analyses was found between Pao2, and not Paco2, and the AHI. For more than half a century, models of periodic breathing have been developed to explain CheyneStokes respiration and other forms of ventilatory instability.4,5,21–23 Mathematical modeling and computer simulations have identified enhanced loop gain as a key determinant of instability in ventilatory control, with increased circulatory delay contributing to inappropriate ventilatory responses that result in periodic breathing. Loop gain has two elements (controller gain and the plant gain), of which it is the sum. Controller gain, or chemoreceptor sensitivity, determines the magnitude of ventilatory response to a given chemical stimulus, with Paco2 and Pao2 both being important variables in the metabolic control of ventilation. Increased controller gain has been well documented in the clinical setting, with increased hypercapnic ventilatory responses demonstrated in patients with idiopathic CSA,24 patients with CSA secondary to HF,8,25 and patients with HF exhibiting 70

periodic breathing during wakefulness.23 In addition to predisposing the patient to ventilatory instability, increased chemosensitivity may also contribute to the excessive ventilatory responses during exercise that were observed in patients with HF,26,27 and may explain the correlation between AHI and the ventilatory response to exercise28 that has been documented in patients with CSA secondary to HF. The concept of plant gain is less well understood and has been less commonly studied, and refers to the efficiency of the gas exchange system, or how changes in ventilation translate into changes in blood gas levels. While increased plant gain contributes to increased loop gain in all models of ventilatory instability, evidence for this phenomenon has been lacking in the clinical setting. Nevertheless, the importance of plant gain has been described in the instability model of periodic breathing, with acknowledgment of the relationship between lung volumes and plant gain.4,23 Reductions in lung volume are thought to increase plant gain since smaller lung volumes are less effective at damping changes in Paco2 and Pao2, thus favoring instability. For example, the rate of change in arterial gas tensions is less rapid during breathholding at vital capacity than at smaller lung volumes.29 Similarly, reduced Pao2 will also increase plant gain, as a greater drop in Pao2 will occur for a given change in ventilation when Pao2 is reduced, due to the shape of the oxygen dissociation curve. Indirect clinical evidence for the importance of lung volume changes has come from observations30 that the expression of CSA has a large postural component and that central apneas experienced while the patient is in the supine position produce more marked Spo2 desaturation than those experienced during sleep in lateral positions. Given that VA and FRC did not have significant effects on respiratory disturbance, Pao2 and Dlco effects on plant gain may be more important than lung volume effects in the pathogenesis of CSA. Interestingly, in the present study of HF patients, those with OSA were found to have significantly lower FRC than those with stable breathing, which may be explained by the higher BMI observed in this group. Other physiologic factors that are thought to alter plant gain are metabolic rate, which alters O2 utilization and CO2 production, as well as cardiac output, which affects circulation time and thus the feedback response. It is possible that reduced Pao2 and Dlco may be the byproducts of more severe HF; thus, circulation time, and not plant gain effects, may be responsible for explaining the observed results. However, we believe that this is unlikely, given that we did not find a relationship between Pao2 or Dlco, either alone or in combination, with either LVEF or cardiac output. Furthermore, unlike the Original Research

potential postural effects on plant gain, presently there is little evidence that postural effects on cardiac output may be responsible for the increased ventilatory instability that is observed with patients in the supine position. In healthy subjects, it has been reported that, compared to the supine position, the left lateral position increased left atrial diameter, facilitated venous return, and decreased mean systemic arterial pressure.31 Thus, although it is unlikely, it cannot be ruled out that increased venous return in HF patients in the lateral position may contribute to maintaining increased cardiac output and decreasing circulation time, thus favoring ventilatory stability. In the current study, we provide evidence that Dlco is independently associated with CSA severity in patients with HF. The association between AHI and Dlco, as well as between AHI and Pao2, was strengthened when supine AHI instead of total AHI was considered in the relationship. Furthermore, Paco2 was associated with supine AHI and not total AHI. We believe that this is explained by the influence of sleeping position on the variability of total AHI; therefore, when supine AHI is considered the association between independent predictors is strengthened. Previous studies1,2,32–34 have found that HF patients with CSA had a significantly lower Paco2 than those without CSA and that Pao2 was not different between groups. Some interventional studies35–37 that raised Paco2 during sleep showed that increasing Paco2 stabilized breathing, which reinforced the importance of hypocapnia in the pathogenesis of CSA. In the current study, Paco2 was also found to be lower in patients with CSA; however, patients with CSA were also found to have lower Pao2 than those with stable breathing; low Dlco, although not statistically significant, may in part explain this observation. Interestingly, in multivariate analyses, Pao2 not Paco2 was correlated with AHI. The current findings support evidence from mathematical modeling that hypocapnia per se does not destabilize respiratory control unless it is a byproduct of increased chemoreflex gain.22,23 Clinical evidence38,39 from sleeping dogs also supports data from mathematical modeling studies that hyperventilation and hypocapnia favor respiratory control system stability; however, hypoxia sensitizes the ventilatory response to CO2. Study Limitations As with all multivariate regression analyses, the results of this study only show that an independent association exists between Pao2 and AHI, as well as between Dlco and AHI, in patients with predominant CSA and HF. The presence of such an associwww.chestjournal.org

ation does not prove a cause-and-effect relationship. As this is a hypothesis-generating observational study, it is important in that it provides the foundations for further interventional studies to address the potential causal relationships. Furthermore, larger studies with measures also assessing controller gain are required to confirm the relative contributions of plant gain and controller gain to CSA pathogenesis. In conclusion, the present study is the first to provide direct clinical evidence that plant gain effects may be more important in CSA pathogenesis than previously thought. The independent association between reduced Pao2 and reduced Dlco with the severity of CSA supports the concept that increased plant gain can contribute to the development of respiratory control instability in patients with HF. However, further studies are required to assess the relative contribution of plant gain and controller gain to CSA pathogenesis.

References 1 Solin P, Bergin P, Richardson M, et al. Influence of pulmonary capillary wedge pressure on central apnea in heart failure. Circulation 1999; 99:1574 –1579 2 Sin DD, Fitzgerald F, Parker JD, 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; 160:1101–1106 3 Javaheri S. Sleep disorders in systolic heart failure: a prospective study of 100 male patients; the final report. Int J Cardiol 2006; 106:21–28 4 Khoo MC. Determinants of ventilatory instability and variability. Respir Physiol 2000; 122:167–182 5 Khoo MC, Gottschalk A, Pack AI. Sleep-induced periodic breathing and apnea: a theoretical study. J Appl Physiol 1991; 70:2014 –2024 6 Hudgel W, Gordon E, Thanakitcharu S, et al. Instability of ventilatory control in patients with obstructive sleep apnea. Am J Respir Crit Care Med 1998; 158:1142–1149 7 Wellman A, Jordan AS, Malhotra A, et al. Ventilatory control and airway anatomy in obstructive sleep apnea. Am J Respir Crit Care Med 2004; 170:1225–1232 8 Solin P, Roebuck T, Johns DP, et al. Peripheral and central ventilatory responses in central sleep apnea with and without congestive heart failure. Am J Respir Crit Care Med 2000; 162:2194 –2200 9 Gehlbach BK, Geppert E. The pulmonary manifestations of left heart failure. Chest 2004; 125:669 – 682 10 West JB, Mathieu-Costello O. Stress failure of pulmonary capillaries: role in lung and heart disease. Lancet 1992; 340:762–767 11 Guazzi M. Alveolar-capillary membrane dysfunction in heart failure: evidence of a pathophysiologic role. Chest 2003; 124:1090 –1102 12 Puri S, Baker BL, Dutka DP, et al. Reduced alveolar-capillary membrane diffusing capacity in chronic heart failure: its pathophysiological relevance and relationship to exercise performance. Circulation 1995; 91:2769 –2774 13 Siegel JL, Miller A, Brown LK, et al. Pulmonary diffusing capacity in left ventricular dysfunction. Chest 1990; 98: 550 –553 CHEST / 134 / 1 / JULY, 2008

71

14 Wright RS, Levine MS, Bellamy PE, et al. Ventilatory and diffusion abnormalities in potential heart transplant recipients. Chest 1990; 98:816 – 820 15 Agostoni P, Bussotti M, Cattadori G, et al. Gas diffusion and alveolar-capillary unit in chronic heart failure. Eur Heart J 2006; 27:2538 –2543 16 American Sleep Disorders Association. EEG arousals: scoring and examples: a preliminary report from the Sleep Disorders Atlas Task Force of the American Sleep Disorders Association. Sleep 1992; 15:173–184 17 Rechtschaffen A, Kales A. A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. Washington DC: US Government Printing Office, 1968 18 American Academy of Sleep Medicine. Sleep-related breathing disorders in adults: recommendations for syndrome definition and measurement techniques in clinical research; the report of an American Academy of Sleep Medicine task force. Sleep 1999; 22:667– 689 19 MacIntyre N, Crapo RO, Viegi G, et al. Standardisation of the single-breath determination of carbon monoxide uptake in the lung. Eur Respir J 2005; 26:720 –735 20 Cotes J. Effect of variability in gas analysis on the reproducibility of the pulmonary diffusing capacity by the single breath method. Thorax 1963; 18:151–154 21 Longobardo GS, Cherniack NS, Fishman AP. Cheyne-Stokes breathing produced by a model of the human respiratory system. J Appl Physiol 1966; 21:1839 –1846 22 Manisty CH, Willson K, Wensel R, et al. Development of respiratory control instability in heart failure: a novel approach to dissect the pathophysiological mechanisms. J Physiol 2006; 577:387– 401 23 Francis DP, Willson K, Davies LC, et al. Quantitative general theory for periodic breathing in chronic heart failure and its clinical implications. Circulation 2000; 102:2214 –2221 24 Xie A, Rutherford R, Rankin F, et al. Hypocapnia and increased ventilatory responsiveness in patients with idiopathic central sleep apnea. Am J Respir Crit Care Med 1995; 152:1950 –1955 25 Javaheri SA. Mechanism of central sleep apnea in patients with heart failure. N Engl J Med 1999; 341:949 –954 26 Sullivan MJ, Higginbotham MB, Cobb FR. Increased exercise ventilation in patients with chronic heart failure: intact

72

27

28

29 30

31

32

33

34 35 36 37 38 39

ventilatory control despite hemodynamic and pulmonary abnormalities. Circulation 1988; 77:552–559 Weber KT, Kinasewitz GT, Janicki JS, et al. Oxygen utilization and ventilation during exercise in patients with chronic cardiac failure. Circulation 1982; 65:1213–1223 Arzt M, Harth M, Luchner A, et al. Enhanced ventilatory response to exercise in patients with chronic heart failure and central sleep apnea. Circulation 2003; 107:1998 –2003 Mithoefer JC. Lung volume restriction as a ventilatory stimulus during breath holding. J Appl Physiol 1959; 14:701–705 Szollosi I, Roebuck T, Thompson B, et al. Lateral sleeping position reduces severity of central sleep apnea/CheyneStokes respiration. Sleep 2006; 29:1045–1051 Pump B, Talleruphuus U, Christensen NJ, et al. Effects of supine, prone, and lateral positions on cardiovascular and renal variables in humans. Am J Physiol Regul Integr Comp Physiol 2002; 283:R174 –R180 Lorenzi-Filho G, Azevedo ER, Parker JD, et al. Relationship of carbon dioxide tension in arterial blood to pulmonary wedge pressure in heart failure. Eur Respir J 2002; 19:37– 40 Naughton M, Benard D, Tam A, et al. Role of hyperventilation in the pathogenesis of central sleep apneas in patients with congestive heart failure. Am Rev Respir Dis 1993; 148:330 –338 Hanly P, Zuberi N, Gray R, Pathogenesis of Cheyne-Stokes respiration in patients with congestive heart failure: relationship to arterial Pco2. Chest 1993; 104:1079 –1085 Khayat RN, Xie A, Patel AK, et al. Cardiorespiratory effects of added dead space in patients with heart failure and central sleep apnea. Chest 2003; 123:1551–1560 Szollosi I, Jones M, Morrell M, et al. Effect of CO2 inhalation on central sleep apnoea and arousals on sleep. Respiration 2004; 71:493– 498 Szollosi I, O’Driscoll DM, Dayer MJ, et al. Adaptive servoventilation and deadspace: effects on central sleep apnoea. J Sleep Res 2006; 15:199 –205 Dempsey JA, Smith CA, Przybylowski T, et al. The ventilatory responsiveness to CO2 below eupnoea as a determinant of ventilatory stability in sleep. J Physiol (Lond) 2004; 560:1–11 Nakayama H, Smith CA, Rodman JR, et al. Effect of ventilatory drive on carbon dioxide sensitivity below eupnea during sleep. Am J Respir Crit Care Med 2002; 165:1251–1260

Original Research