High prevalence and persistence of sleep apnoea in patients referred for acute left ventricular failure and medically treated over 2 months

Transcription

1 European Heart Journal (1999) 20, Article No. euhj , available online at on High prevalence and persistence of sleep apnoea in patients referred for acute left ventricular failure and medically treated over 2 months F. Tremel*, J.-L. Pépin, D. Veale, B. Wuyam, J.-P. Siché*, J. M. Mallion* and P. Lévy *Department of Cardiology, Department of Respiratory Medicine and Sleep Laboratory, University Hospital, Grenoble, France Aims Cardiac failure patients were studied systematically using polysomnography 1 month after recovering from acute pulmonary oedema, and again after 2 months of optimal medical treatment for cardiac failure. Methods and Results This prospective study of consecutive patients was conducted in a cardiac care unit of a university hospital. VO 2 measurements and left ventricular ejection fraction were recorded. Thirty-four patients, initially recruited with pulmonary oedema, improved after 1 month of medical treatment to NYHA II or III. They were aged less than 75 years and had a left ventricular ejection fraction less than 45% at the time of inclusion. Age was 62 (9) years, body mass index=27 (5) kg. m 2 and an ejection fraction=30 (10)%. Eighteen of the 34 patients (53%) had coronary artery disease. Twenty-eight of the 34 had sleep apnoea syndrome with an apnoea+hypopnoea index >15. h 1 of sleep. Thus, the prevalence of sleep apnoea in this population was 82%. Twenty-one of 28 (75%) patients had central sleep apnoea and seven of 28 (25%) had obstructive sleep apnoea. Patients with central sleep apnoea had a lower PaCO 2 than those with obstructive sleep apnoea (33 (5) vs 37 (5) mmhg, P<0 005). Significant correlations were found between apnoea+hypopnoea index and peak exercise oxygen consumption (r= 0 73, P<0 01), and apnoea+hypopnoea index and PaCO 2 (r= 0 42, P=0 03). When only central sleep apnoea patients were considered, a correlation between apnoea+hypopnoea index and left ventricular ejection fraction was also demonstrated (r= 0 46, P<0 04). After 2 months of optimal medical treatment only two patients (both with central sleep apnoea) showed improvement (apnoea+hypopnoea index <15. h 1 ). Conclusions We have demonstrated a high prevalence of sleep apnoea, which persisted after 2 months of medical treatment, in patients referred for acute left ventricular failure. Central sleep apnoea can be considered a marker of the severity of congestive heart failure. (Eur Heart J 1999; 20: ) Key Words: Left ventricular failure, sleep apnoea, Cheyne Stokes respiration. See page 1140 for the Editorial comments on this article Introduction Congestive heart failure is a common condition [1] and is associated with high morbidity and mortality [2,3]. Sleepdisordered breathing has been reported as one of the many factors which may contribute to the declining course of congestive heart failure. Two different patterns Revision submitted 20 January 1999, and accepted 3 February This study was funded by PHRC and DRRC CHU Grenoble and Région Rhone Alpes (Hypoxie). Correspondence: DrFrédéric Tremel, Département de Cardiologie, CHU de Grenoble, BP 217 X, 38043, Grenoble, France X/99/ $18.00/0 of sleep respiratory disturbances may be found in association with cardiac failure. Obstructive sleep apnoea is characterized by recurrent collapse of the pharyngeal airway during sleep resulting in snoring and daytime sleepiness. It affects 4% of men and 2% of women between the ages of 30 and 60 years [4]. Obstructive sleep apnoea has important pathological consequences, such as an increased incidence of hypertension [5,6], arrhythmias [7,8], myocardial infarction [9] and stroke [10,11]. Mortality is raised [12] and thought to be largely secondary to cardiovascular causes [13]. Thus coronary heart diseases associated with obstructive sleep apnoea could explain, in part, the occurrence of chronic left ventricular failure in these disorders The European Society of Cardiology

2 1202 F. Tremel et al. In contrast Cheyne Stokes respiration with central sleep apnoea is commonly seen in patients with congestive heart failure [14 18]. Proposed mechanisms include: (i) increased central controller gain (i.e. raised responses to CO 2 and hypoxia favouring respiratory instability), (ii) increased circulation time and (iii) hypocapnia. The published prevalence rates of Cheyne Stokes respiration with central sleep apnoea in patients with left ventricular failure varies from 30 to 100% [19]. However the difference in prevalence rate between acute and chronic conditions and the stability of sleep respiratory disturbances in patients with congestive heart failure remain to be established. Thus, we have designed a study to (i) systematically screen by complete polysomnograhy consecutive patients referred to a cardiovascular unit for acute left ventricular failure (ii) reassess these patients after 2 months of optimal medical treatment for cardiac failure. The aim was to define the prevalence of sleep respiratory disturbance in patients after an episode of acute left ventricular failure and the subsequent change after heart failure therapy. Methods Patients This was a prospective study in consecutive patients (age <75 years, left ventricular ejection fraction <45%) referred to a cardiovascular unit of a university teaching hospital with an acute left ventricular failure episode (New York Heart Association: Class IV). Patients underwent polysomnography, spirometry and an assessment of blood gases at 1 month after the acute event (month T1). In patients in whom the first polysomnogram demonstrated significant sleep apnoea syndrome (apnoea+hypopnoea index >15. h 1 of sleep), a follow-up polysomnogram was undertaken a further month later (month T2) after optimization of therapy. Polysomnography Continuous recordings were taken by electroencephalogram of electrode positions C 3 /A 2 -C 4 /A 1 -C z /O 1 according to the International Electrode Placement System, and of eye movements. A chin electromyogram, and an electrocardiogram with a modified V 2 lead were also recorded. Respiration was monitored with uncalibrated inductance respiratory plethysmography. Airflow was measured by the sum of buccal and nasal thermistor signals, and oxygen saturation was measured with a Biox-Ohmeda 3700 oximeter. The polysomnogram was scored manually according to standard criteria [20]. Episodes of apnoea were defined as complete cessation of airflow for 10 s or more and hypopnoea as a greater than 50% decrease in oronasal airflow lasting for at least 10 s. Apnoea events were classified as central, obstructive or mixed according to the absence or presence of breathing efforts. The variation in oxygen saturation overnight was quantified as the Delta index, calculated by the method validated by Lévy et al. [21]. Periodic breathing was defined as a gradual waxing and waning of respiration followed by central apnoea or hypopnoea [19]. The number of minutes of Cheyne Stokes respiration with central sleep apnoea while asleep and while awake was manually scored. The number of minutes while asleep was divided by the total sleep time to calculate the percentage of sleep spent in Cheyne Stokes respiration with central sleep apnoea. Patients were labelled as Cheyne Stokes respiration with central sleep apnoea when more than 75% of the apnoeas or hypopnoeas were central in nature. In contrast, patients were defined as having obstructive sleep apnoea when more of 25% of the respiratory events were obstructive. Left ventricular ejection fraction As an objective measure of left ventricular function, left ventricular ejection fraction was assessed at rest by radionuclide angiography (mean delay (SD): 6 (6) days after admission). With this technique, left ventricular ejection fraction was determined by in vivo labelling of red blood cells with 99mTc, using the classical equilibrium method [22]. Normal left ventricular ejection fraction values for our laboratory have already been established [23]. Exercise testing Exercise studies were performed, as per our routine, at a mean of 43 days after the episode of acute left ventricular failure (n=26/34). Seated on an electronic-braked cycle ergometer, the patient started with a 2-min period of unloaded cycling (20 Watts), followed by load increments of 10 Watts. min 1. Exercise was stopped when subjects reached their exercise fatigue limit and maximal exercise was determined using standard criteria [24].VO 2 was determined during the stress test by a cycle to cycle method (Medical Graphics). VO 2 measurements were expressed as a percentage of the predicted normal values for sedentary adults (VO 2 max). Statistics All data are presented as mean (SD). Results at T1 month and T2 months are shown as a box plot, which depicts three main features of the variable: its centre, its spread, and its outliers. The top and the bottom of the box are the 25th and the 75th percentiles. The length of the box is thus the interquartile range (IQR), that is, the box represents the middle 50% of the data. The line drawn through the middle of the box corresponds to the median. The upper adjacent value is the largest observation that is less than or equal to the 75th percentile

3 Sleep apnoea in LV failure 1203 Table 1 Demographic, cardiac and respiratory functional data plus medications All patients (n=34) No SDB (n=6) CSR-CSA (n=21) OSAS (n=7) P Age (years) 62 (9) 60 (10) 63 (11) 63 (6) 0 74 BMI (kg. m 2 ) 27 2 (4 5) 32 4 (4 6) 25 5 (3 8) 27 7 (3 4) Gender (men/women) (nb) 28/6 2/4 19/2 7/ Medications Diuretics 100% 100% 100% 100% Nitrates 79% 83% 76% 86% 0 84 Digoxin 21% 33% 19% 14% 0 67 Amiodarone 29% 17% 24% 57% 0 19 ACE inhibitors 68% 83% 67% 57% 0 59 Blood gases PaCO 2 35 (5) 40 (4) 33 (5) 37 (5) PaO 2 80 (10) 73 (9) 81 (11) 82 (7) 0 31 ph 7 44 (0 05) 7 41 (0 01) 7 45 (0 06) 7 44 (0 03) 0 11 SaO 2 96 (2) 95 (2) 96 (1) 96 (1) 0 25 Functional data Vital capacity (ml) 3214 (764) 3054 (614) 3217 (782) 3314 (900) 0 88 FEV 1 (ml) 2337 (710) 2362 (392) 2336 (771) 2322 (782) 0 99 Single-breath diffusing capacity 79 (27) 89 (10) 72 (27) 93 (34) 0 28 for CO/Alveolar ventilation (%) Cardiac data LVEF (%) 30 (10) 32 (7) 30 (11) 30 (11) 0 90 VO 2 max (% of predicted value) (n=26) 70 (5) 92 (11) 66 (13) 72 (11) Comparison of clinical variables between the groups of patients who demonstrated no sleepdisordered breathing, Cheyne Stokes respiration with central sleep apnoea or obstructive sleep apnoea were carried out using the Kruskal Wallis test. Data are presented as mean (SD). BMI=body mass index; AHI=apnoea+hypopnoea index; LVEF=left ventricular ejection fraction; CSR-CSA=Cheyne Stokes respiration with central sleep apnoea; OSAS=obstructive sleep apnoea syndrome; SDB=sleep disordered breathing; VO 2 max=% of the predicted normal value for peak oxygen consumption during exercise testing; FEV 1 =forced ventilatory capacity during one second. plus 1 5 times IQR. The lower adjacent value is the smallest observation that is greater than or equal to the 25th percentiles minus 1 5 times IQR. The adjacent values are displayed as T-shaped lines that extend from each end of the box. Comparisons of variables from the first polysomnography recording (T1 month) with those of the second (T2 months) were made using the paired t-test or the Wilcoxon signed-rank test. To examine the possibility of a relationship between the apnoea+ hypopnoea index and variables of cardiac or respiratory function, a correlation analysis was made. A comparison of clinical variables between the groups of patients who demonstrated Cheyne Stokes respiration with central sleep apnoea vs obstructive sleep apnoea was carried out using the Mann Whitney or the Kruskal Wallis test. A P value of <0 05 was considered significant. The study was approved by the local ethical committee. Informed written consent was obtained from all the patients. Results Prevalence of sleep apnoea 1 month after acute left ventricular failure The mean interval between admission with an acute episode of left ventricular failure and the first polysomnogram was 25 (16) days. Optimal medical, therapy including maximally tolerated doses of angiotensinconverting enzyme inhibitors, nitrates, digoxin and diuretics were prescribed as clinically appropriate (Table 1). The demographic, cardiac and respiratory functional data plus medications are shown in Table 1. Patients were initially in NYHA class IV and improved after 1 month of medical treatment to NYHA III or II. Eighteen (53%) of the patients had coronary artery disease and 16 (47%) had cardiomyopathy. The results of the 1 month sleep study are shown in Table 2. For the group as a whole the mean apnoea+hypopnoea index was 45 (29). h 1 of sleep. Cheyne Stokes respiration occurred during a mean of 20% of the total sleep time. Twenty-eight of the patients had sleep apnoea syndrome with an apnoea+hypopnoea index of >15. h 1 of sleep (Table 2). Thus, the prevalence of significant sleep apnoea in this population 1 month after an episode of acute left ventricular ejection fraction was 28/34(82%). Twenty-one of these 28 (75%) patients had Cheyne Stokes respiration with central sleep apnoea vs 7/28 (25%) obstructive sleep apnoea (i.e. a prevalence of 7/34 (20 6%)). The Cheyne Stokes respiration with central sleep apnoea tended to have a lower PaCO 2 : 33 (5) mmhg, than obstructive sleep apnoea: 37 (5) mmhg, P<0 005) (Table 1). The majority of the patients exhibiting Cheyne Stokes respiration

4 1204 F. Tremel et al. Table 2 Results of the 1 month after left ventricular failure sleep study Polysomnography All patients (n=34) No SDB (n=6) CSR-CSA (n=21) OSAS (n=7) P TST (minutes) 303 (111) 379 (63) 270 (117) 335 (90) Stage 1+2 (% of TST) 82 (12) 81 (10) 82 (13) 87 (13) 0 66 Stage 3+4 (% of TST) 3 (5) 5 (5) 3 (5) 3 (3) 0 49 REM sleep (% of TST) 14 (10) 14 (10) 14 (11) 11 (9) 0 68 Micro arousals (nb/night) 161 (74) 91 (56) 159 (69) 215 (63) 0 02 Sleep stage changes (nb/night) 392 (169) 257 (156) 371 (150) 551 (153) Mean nocturnal SaO 2 (%) 94 (2) 93 (3) 94 0 (1 6) 94 4 (3 0) 0 36 Minimal nocturnal SaO 2 (%) 83 (9) 81 (7) 85 (5) 77 (16) 0 30 AHI 45 (29) 5 (4) 54 (25) 51 (24) Apnoea index 15 (16) 1 (1) 18 (18) 18 (6) Hypopnoea index 30 (25) 4 (3) 36 (24) 33 (25) Comparison of clinical variables between the groups of patients who demonstrated no sleepdisordered breathing, Cheyne Stokes respiration with central sleep apnoea or obstructive sleep apnoea syndrome were carried out using the Kruskal Wallis test. TST=total sleep time; Micro-arousals=scored according to the rules of the American sleep disorders association ASDA (Sleep 1992; 15: ); AHI=apnoea+hypopnoea index; CSR- CSA=Cheyne Stokes respiration with central sleep apnoea; OSAS=obstructive sleep apnoea syndrome; SDB=sleep disordered breathing; REM=rapid eye movement. 120 r = 0 73 P = VO 2 max AHI Figure 1 Correlation between apnoea+hypopnoea index and VO 2 max in 26 patients: 17 Cheyne Stokes respiration with central sleep apnoea ( ), six obstructive sleep apnoea ( ), three negative ( ). Significant correlation was found between apnoea+hypopnoea index and VO 2 max (r= 0 73; P<0 01; CI ( 0 87; 0 48)). The correlation was mainly explained by the patients exhibiting Cheyne Stokes respiration with central sleep apnoea (r= 0 73; P<0 01; CI ( 0 90; 0 38)). It is clear that obstructive sleep apnoea would not produce any negative correlation with VO 2 max (r=0 186; CI: 0 73; 0 87); P=0 72). AHI=apnoea+hypopnoea index; VO 2 max=percentage of predicted maximal VO 2 during exercise testing; CSR-CSA=Cheyne Stokes respiration with central sleep apnoea; OSA= obstructive sleep apnoea. 120 with central sleep apnoea were newly diagnosed with congestive heart failure (74%). Significant correlations were found for the whole group between the apnoea+hypopnoea index and peak exercise VO 2 (r= 0 73, P<0 01) (Fig. 1), and the apnoea+hypopnoea index and PaCO 2 (r= 0 42, P=0 03). A trend towards a correlation between the apnoea+hypopnoea index and left ventricular ejection fraction was also demonstrated (r= 0 29, P=0 09) for the whole group. When only Cheyne Stokes respiration with central sleep apnoea patients were considered this correlation became significant (r= 0 46, P<0 04). Age,

5 Sleep apnoea in LV failure 1205 Table 3 Evolution of sleep apnoea 2 months after acute left ventricular failure CSR-CSA (n=15) OSAS (n=4) PSG 1 PSG 2 P PSG 1 PSG 2 P TST (min) 299 (108) 307 (76) ns 347 (122) 317 (52) ns Stage 1+2 (% of TST) 79 (11) 79 (12) ns 87 (17) 79 (13) ns Stage 3+4 (% of TST) 4 (5) 4 (5) ns 3 (5) 1 (1) ns REM sleep (% of TST) 16 (9) 16 (9) ns 10 (13) 21 (13) ns Mean nocturnal sleep SaO 2 (%) 93 9 (1 8) 94 0 (1 8) ns 93 9 (4 0) 91 8 (3 5) 0 12 Minimal nocturnal SaO 2 (%) 84 (6) 81 (8) ns 71 (20) 75 (12) ns AHI 49 5 (25 1) 48 7 (32 7) ns 53 1 (28 9) 47 4 (20 9) ns Apnoea index 17 5 (13 5) 15 5 (19 2) ns 16 3 (7 1) 17 9 (18 5) ns Hypopnoea index 32 2 (21 0) 33 2 (23 6) ns 36 8 (30 4) 29 5 (24 1) ns Comparison of variables from the first polysomnography recording (month T1) with those of the second (month T2) were made using the Wilcoxon signed-rank test. The mean apnoea+hypopnoea index was unchanged (50 (25). h 1 vs 48 (30). h 1 : ns). The sleep structure and the nocturnal SaO 2 were similar between the two sleep studies for both the Cheyne Stokes respiration with central sleep apnoea and the obstructive sleep apnoea syndrome patients. TST=total sleep time; Micro-arousals=scored according to the rules of the American sleep disorders association ASDA (Sleep 1992; 15: ); AHI=apnoea+hypopnoea index; CSR- CSA=Cheyne Stokes respiration with central sleep apnoea; OSAS=obstructive sleep apnoea syndrome; SDB=sleep disordered breathing; REM=rapid eye movement. body mass index, spirometry values were not correlated with the apnoea+hypopnoea index in the whole group. Prevalence of sleep apnoea two months after acute left ventricular failure Of the 34 patients studied at month T1, six had an apnoea+hypopnoea index less than 15. h 1, one patient died, one remained clinically unstable, one was lost for follow-up and six refused to undergo a second study. Hence 15 patients did not undergo polysomnography at months T2. The two subgroups of patients reassessed or not by polysomnography at 2 months were similar in terms of left ventricular ejection fraction, pulmonary and cardiac functional data, and anthropometric values. Finally polysomnography at T2 was thus recorded for 19 patients (15 Cheyne Stokes respiration with central sleep apnoea, four obstructive sleep apnoea), at a mean delay of 37 (16) days from month T1 (Table 3). The clinical status and medications of the patients studied remained unchanged between the two polysomnography studies, although left ventricular ejection fraction demonstrated a trend towards a moderate increase (31 (12) vs 35 (12)%, ns, n=7). In only two patients did Cheyne Stokes respiration with central sleep apnoea disappear (apnoea+ hypopnoea index <15. h 1 ) at the time of polysomnography. T2 (Fig. 2), and the mean apnoea+hypopnoea index did not change (50 (25). h 1 vs 48 (30). h 1 : ns). The sleep structure and the mean and minimal nocturnal SaO 2 were similar between the two sleep studies for both Cheyne Stokes respiration with central sleep apnoea and the obstructive sleep apnoea patients (Table 3). Discussion We have demonstrated a high prevalence of sleep apnoea in a group of 34 consecutive patients studied after recovering from acute left ventricular failure. The sleep respiratory disturbances were mainly central sleep apnoea (75%). The severity of sleep apnoea was inversely correlated with VO 2 max, PaCO 2 levels and left ventricular ejection fraction. These nocturnal abnormalities, which are primarily a marker of the severity of congestive heart failure, persisted after 2 months despite optimal medical heart failure treatment. Prevalence of sleep apnoea 1 month after acute left ventricular failure The prevalence rates reported in the literature vary from 30% to 100% [19]. These differences may be explained by the number, the patients selected and the criteria used to score hypopnoea. Lofaso et al. [25] found that 45% of patients on a heart transplant waiting list had central sleep apnoea. Their patients were particularly severe in terms of left ventricular ejection fraction (mean left ventricular ejection fraction: 13%). In two prospective longitudinal studies Jahaveri et al. [18,26] also reported a prevalence of 45% and 51%, respectively, in stable patients medically treated without recent left ventricular failure. Our study is the first prospective study reporting such a high prevalence of sleep apnoea in consecutive patients 1 month after acute left ventricular failure. The proximity to a recent episode of left ventricular failure could explain this high rate of sleep apnoea in comparison to other studies. The majority of the patients

6 1206 F. Tremel et al. AHI OSA CSR-CSA AHI = 15 0 Month T1 Month T2 Figure 2 Prevalence of sleep apnoea 2 months after acute left ventricular failure (n=19): 15 Cheyne Stokes respiration with central sleep apnoea, four obstructive sleep apnoea. Left: T1 month ( ) and T2 month ( ) Cheyne Stokes respiration with central sleep apnoea disappeared in only two patients (apnoea+hypopnoea index <15. h 1 ) at the time of polysomnography T2 and the mean apnoea+hypopnoea index was unchanged (50 (25). h 1 vs 48 (30). h 1 : ns). Right: Box plots of apnoea+hypopnoea index at months T1 and T2. The line drawn through the middle of the boxes corresponds to the median. The top and the bottom of the boxes are the 25th and the 75th percentiles. AHI=apnoea+hypopnoea index; CSR-CSA=Cheyne Stokes respiration with central sleep apnoea; OSA=obstructive sleep apnoea. exhibiting central sleep apnoea were newly diagnosed patients and the benefits of medical treatment may not be fully attained by 1 month. Cheyne Stokes respiration with central sleep apnoea vs obstructive sleep apnoea syndrome As in previous studies [14 18], we found a majority of patients exhibiting central sleep apnoea (75%). Our study is the first since the recent study of Jahaveri et al. [26] to prospectively examine the prevalence of both obstructive sleep apnoea and central sleep apnoea in congestive heart failure. Jahaveri and colleagues [26] have found that 11% of male patients with stable heart failure suffer from obstructive sleep apnoea [26]. Our study has demonstrated that seven out of 34 (20 6%) patients referred with acute left ventricular failure had undiagnosed obstructive sleep apnoea. This rate is high in comparison to the 9% expected in the general population of the same age [4] but seems logical when one considers the established association between hypertension, coronary heart disease and obstructive sleep apnoea. Another recent study on diastolic heart failure [27] found that 55% of the patients had significant sleep-disordered breathing which was mainly obstructive apnoea. It has been suggested that synergy between sleep-disordered breathing and daytime hypertension may precipitate the development of diastolic heart failure [27]. These patients with obstructive sleep apnoea are probably those in whom the benefits of nasal continuous positive airway pressure could be greater, at least in terms of improvement in symptoms and quality of life. Conversion of Cheyne Stokes ventilation during sleep to obstructive sleep apnoea after heart transplantation has been previously described [28]. Upper airway instability may also play a role in central sleep apnoea patients. It has been shown that upper airway collapse occurs in some patients during Cheyne Stokes respiration [29]. After heart failure treatment the pre-existent upper airway instability may lead to obstructive sleep apnoea in some patients. Persistence of sleep apnoea after a further month of medical treatment A significant improvement in cardiac function is expected to lead to the disappearance of central sleep apnoea. Harrison et al. [30] showed that central sleep apnoea decreased gradually and finally disappeared with clinical improvement in the heart failure. There are reports that central sleep apnoea improved after heart transplantation [31]. Dark et al. [32] found a resolution of breathing pattern abnormalities in 50% of a small group of six patients following medical treatment. Their initial sleep study was conducted in the first 48 h after admission for acute left ventricular failure. Thus central sleep apnoea intensity probably changes with marked variations in cardiac function. In this context, it would have been of interest to re-study not only the 21 patients in our study with an initial apnoea+hypopnoea index >15. h 1 but also the six survivors with a first apnoea+hypopnoea index of less than 15. h 1 of sleep. According to the principle of regression to the mean, the results of all the potentially suitable people (21+6) would probably have demonstrated that sleep respiratory disturbances were likely, in reality, to be more stable over time. Conversely, our study demonstrated that patients with stable chronic heart failure exhibited stable

7 Sleep apnoea in LV failure 1207 sleep respiratory disturbances. The demonstration of persistent central sleep apnoea episodes after optimization of cardiac function represents a first step in justifying a specific treatment for this condition. However, further studies are needed to show how much this treatment may affect the prognosis of cardiac heart failure. Explanatory factors for Cheyne Stokes respiration with central sleep apnoea Proposed mechanisms for central sleep apnoea include: (i) increased gain of central control, (ii) increased circulation time, (iii) hypocapnia, and (iv) reduced buffering capacity of arterial blood gases. Our data are in agreement with these reported predictive factors. A low PaCO 2 promotes ventilatory instability and the occurrence of central apnoeas is consistent with the physiological notion of the apnoeic threshold. Thus, similar to Javaheri and Corbett [33] we have demonstrated a significant awake hypocapnia in our central sleep apnoea patients and a correlation between PaCO 2 levels and the apnoea+hypopnoea index. The positive predictive value ofapaco 2 <35 mmhg for central apnoea is 78% [33]. Moreover the prevalence of ventricular tachycardia has been demonstrated to be 20 times greater in hypocapnic patients than in eucapnic patients [33]. We found a significant negative correlation between left ventricular ejection fraction and the apnoea+ hypopnoea index only in the central sleep apnoea subgroup. Thus central sleep apnoea seems to be related to the severity of left ventricular failure. Conversely, the existence of obstructive sleep apnoea in these patients, as in the general population, is probably related to obesity and upper airway anatomical or functional abnormalities. VO 2 max is an independent prognostic factor in congestive heart failure patients [34] and our study has demonstrated a close correlation between the apnoea+ hypopnoea index and VO 2 max, mostly explained by the central sleep apnoea patient subgroup. Andreas et al. [35] also found a negative correlation between the importance of Cheyne Stokes respiration and peak oxygen consumption. Moreover, when the patients were treated with nocturnal oxygen they found a correlation between the improvement in Cheyne Stokes respiration and the increase in peak oxygen consumption during exercise. Oxygen has many effects on muscle metabolism owing to an increase in tissue oxygen content. However, Andreas et al. suggested that treatment of central sleep apnoea by oxygen with concomitant reduction in arousals, desaturations and a reduction in sympathetic overactivity, is the explanation for the exercise capacity improvement found in congestive heart failure [35]. Sympathetic activation is an important pathophysiological and prognostic factor in heart failure [36]. Central sleep apnoea per se is an additional factor for sympathetic activation. It can be speculated that treatment of Cheyne Stokes respiration with central sleep apnoea may reduce sympathetic tone and thus lead to improved survival in such patients. Sleep respiratory disturbances as a prognostic factor in left ventricular failure Increased mortality has been reported in congestive heart failure patients exhibiting central sleep apnoea during sleep. Central sleep apnoea itself may be sufficient to impair cardiac function by different mechanisms. Changes in pleural pressure with negative swings at the peak of hyperventilation affect both preload and afterload. Sympathoadrenal activation also occurs as a result of arousals, hypoxaemia and hypercapnia. These phenomena may result in an imbalance between oxygen delivery, particularly in the presence of coronary artery disease, and myocardial oxygen demand [37]. Chen and Scharf [38] have demonstrated that central apnoeas produced in an animal model revealed greater depression in cardiac output and greater changes of afterload-related left ventricular dysfunction than did obstructive apnoeas. Thus it seems logical that central sleep apnoea during sleep should appear as an independent prognostic factor. Findley et al. [14] found that all six patients with central sleep apnoea died within 6 months compared to only three of nine patients without Cheyne Stokes respiration. These two subgroups of patients had similar levels of left ventricular ejection fraction. Ancoli-Israel et al. [39] have shown that 108 men with severe Cheyne Stokes respiration had reduced survival and enlarged cardiac size. Hanly and co-workers [40] found a significantly reduced cumulative survival when Cheyne Stokes respiration patients were followed up for 4 5 years. Conversely, a recent study by Andreas and colleagues [41] found no significant prognostic impact of nocturnal Cheyne Stokes respiration with central sleep apnoea on survival. For these authors only Cheyne Stokes respiration present during the daytime suggested a high likelihood of dying within a few months. However, owing to the correlations which we have found between Cheyne Stokes respiration with central sleep apnoea and left ventricular ejection fraction and VO 2 max, the sleep respiratory disturbances may at least be a marker of the severity of congestive heart failure. Conclusion We have demonstrated a high prevalence and persistence of sleep respiratory disturbances in patients referred for acute left ventricular failure and medically treated over 2 months. Large-scale controlled studies are needed to demonstrate more clearly the role of central sleep apnoea as an independent prognostic factor and to identify the effects of specific treatment modalities on the outcome of heart failure. No such data are likely to be

8 1208 F. Tremel et al. produced in the immediate future. In the meantime, a case for treating patients might be made on the grounds of alleviating disruption of sleep physiology, or of attenuating non-pharmacologically stimuli to sympathetic overactivity, a known maladaptive phenomenon in congestive heart failure. Grants: Clinical Research Funding: PHRC and DRRC CHU Grenoble and Région Rhone Alpes (Hypoxie). We would like to thank Mrs Christelle Deschaux for her statistical assistance. References [1] Ho KKL, Pinsky JL, Kannel WB et al. The Epidemiology of Heart Failure: The Framingham Study. J Am Coll Cardiol 1993; 22: 6A 13A. [2] The Digitalis Investigation Group. The Effects of Digoxin on Mortality and Morbidity in Patients with Heart Failure. N Engl J Med 1997; 336: [3] The SOLVD investigators. Effects of Enalapril on survival in patients with reduced left ventricular ejection fractions and congestive heart failure. N Engl J Med 1991; 325: [4] Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The occurrence of sleep-disordered breathing among middleaged adults. N Engl J Med 1993; 328: [5] Fletcher EC, DeBehneke RD, Lovoi MS, Gorin AB. Undiagnosed sleep apnoea in patients with essential hypertension. Ann Intern Med 1985; 103: [6] Hla KM, Young TB, Bidwell T, Palta M, Skatrud GB, Dempsey J. Sleep apnoea and hypertension. A populationbased study. Ann Intern Med 1994; 120: [7] Guilleminault C, Connolly SJ, Winkle RA. Cardiac arrhythmia and conduction disturbances during sleep in 400 patients with sleep apnoea syndrome. Am J Cardiol 1983; 52: [8] Flemons WW, Remmers JE, Gillis AM. Sleep apnoea and cardiac arrhythmias. Is there a relationship? Am Rev Respir Dis 1993; 148: [9] Hung J, Whitford EG, Parsons RW, Hillman DR. Association of sleep apnoea with myocardial infarction in men. Lancet 1990; 336: [10] Konsenvuo M, Kaprio J, Telakivi T, Partinen M, Heikkil K, Sarna S. Snoring as a risk factor for ischaemic heart disease and stroke in men. Br Med J 1987; 294: [11] Palomaki H, Partinen M, Juvela S, Kaste M. Snoring as a risk factor for sleep-related brain infarction. Stroke 1989; 20: [12] He J, Kryger MH, Zorick FJ, Conway W, Roth T. Mortality and apnoea index in obstuctive sleep apnoea-experience in 385 male patients. Chest 1988; 94: [13] Partinen M, Jamieson A, Guilleminault C. Long term outcome for obstructive sleep apnoea patients: mortality. Chest 1988; 94: [14] Findley LJ, Zwillich CW, Ancoli-Israel S, Kripke D, Tisi G, Moser KM. Cheyne Stokes breathing during sleep in patients with left ventricular heart failure. South Med J 1985; 78: [15] Yasuma F, Nomura H, Hayashi H, Okada T, Tsuzuki M. Breathing abnormalities during sleep in patients with chronic heart failure. Jpn Circ J 1989; 53: [16] Hanly PJ, Millar TW, Steljes DG, Baert R, Frais A, Kryger MH. Respiration and abnormal sleep in patients with congestive heart failure. Chest 1989; 96: [17] Yamashiro Y, Kryger MH. Review: Sleep in heart failure. Sleep 1993; 16: [18] Javaheri S, Parker TJ, Wexler L et al. Occult sleep-disordered breathing in stable congestive heart failure. Ann Intern Med 1995; 122: [19] Quaranta AJ, D Alonzo GE, Krachman SL. Cheyne Stokes respiration during sleep in congestive heart failure. Chest 1997; 111: [20] Rechtshaffen A, Kales, eds. A manual of standardized terminology. Technique and scoring system for sleep stages of normal subjects. Bethesda. National Insitutes of Health, Publication N204, [21] Lévy P, Pépin J-L, Deschaux-Blanc C, Paramelle B, Brambilla C. Accuracy of oximetry for detection of respiratory disturbances in sleep apnoea syndrome. Chest 1996; 109: [22] Green MV, Brody WR, Douglas MA, Borer JS, Ostrow HG, Line BR. Ejection fraction by count rate from gated images. J Nucl Med 1978; 19: 880. [23] Machecourt J, Comet M, Bourlard P et al. Place de l angioscintigraphie cavitaire gauche à l effort dans le diagnostic de la maladie coronarienne. Comparaison de l effort statique et dynamique. Arch Mal Coeur 1984; 77: [24] Wasserman K, Hansen JA, Sue DY, Whipp BJ. Principles of exercise testing and interpretation. Philadelphia: Lea and Febiger, 1987: [25] Lofaso F, Verschueren P, Dubois Rande, Harf A, Goldenberg F. Prevalence of sleep-disordered breathing in patients on a heart Transplant waiting list. Chest 1994; 106: [26] Jahaveri S, Parker TJ, Liming JD et al. Sleep apnea in 81 ambulatory male patients with stable heart failure. Types and their prevalences, consequences, and presentations. Circulation 1998; 97: [27] Chan J, Sanderson J, Chan W et al. Prevalence of sleepdisordered breathing in diastolic heart failure. Chest 1997; 111: [28] Callop NA. Cheyne Stokes ventilation converting to obstructive sleep apnoea following heart transplantation. Chest 1993; 104: [29] Alex C, Onal E, Lopata M. Upper airway occlusion during sleep in patients with Cheyne Stokes respiration. Am Rev Respir Dis 1986; 133: [30] Harrison TR, King CE, Calhoun JA et al. Congestive heart failure as the cause of paroxysmal dyspnea at the onset of sleep. Arch Intern Med 1934; 53: [31] Murdock DK, Lawless CE, Loeb HS, Scanlon PJ, Pifarre R. The effect of heart transplantation on Cheyne Stokes respiration associated with congestive heart failure. J Heart Transplant 1986; 5: [32] Dark DS, Pingleton SK, Kerby GR, Crabb JE, Gollub SB, Glatter TR. Breathing pattern abnormalities and arterial oxygen desaturation during sleep in the congestive heart failure syndrome. Chest 1987; 91: [33] Javaheri S, Corbett WS. Association of low PaCO2 with central sleep apnoea and ventricular arrhythmias in ambulatory patients with stable heart failure. Ann Intern Med 1998; 128: [34] Mancini DM, Eisen H, Kussmaul W, Mull R, Edmunds LH, Wilson JR. Value of peak exercise oxygen consumption for optimal timing of cardiac transplantation in ambulatory patients with heart failure. Circulation 1991; 83: [35] Andreas S, Clemens C, Sandholzer Figulla HR, Kreuzer H. Improvement of exercise capacity with treatment of Cheyne Stokes respiration in patients with congestive heart failure. J Am Coll Cardiol 1996; 27: [36] Cohn JN, Johnson GR, Shabetai et al. Ejection fraction, peak exercise oxygen consumption, cardiothoracic ratio, ventricular arrhythmias, and plasma norepinephrine as determinants of prognosis in heart failure. Circulation 1993; 87: VI-5 VI-16. [37] Javaheri S. Central sleep-apnoea-hypopnoea syndrome in heart failure: Prevalence, impact, and treatment. Sleep 1996; 19: S229 S231. [38] Chen L, Scharf SM. Comparative hemodynamic effects of periodic obstructive and simulated central apnoeas in sedated pigs. J Appl Physiol 1997; 83: [39] Ancoli-Israel S, Engler RL, Friedman PJ, Klauber MR, Ross PA, Kripke DF. Comparison of patients with central sleep apnoea with and without Cheyne Stokes respiration. Chest 1994; 106:

9 Sleep apnoea in LV failure 1209 [40] Hanly PJ, Zuberi-Khokhar NS. Increased mortality associated with Cheyne Stokes respiration in patients with congestive heart failure. Am J Respir Crit Care Med 1996; 153: [41] Andreas S, Hagenah G, Müller C, Werner GS, Kreuzer H. Cheyne Stokes respiration and prognosis in congestive heart failure. Am J Cardiol 1996; 78: