November 28, 2004

Death During Polysomnography of a Patient With Cheyne-Stokes Respiration, Respiratory Acidosis, and Chronic Heart Failure*

A patient with chronic heart failure and chronic respiratory failure (CRF) underwent ambulatory polysomnography at home. She was found dead on the morning after the recording. The tracings confirmed severe sleep apnea syndrome. After 8 h of incessant Cheyne- Stokes respiration during sleep, respiratory arrest occurred, followed 7 min later by asystole. This report illustrates a case of respiratory drive failure during sleep as the mode of death in a patient with heart failure, sleep apnea syndrome, and CRF. (CHEST 2004; 126:1698-1700)

Key words: Cheyne-Stokes respiration; heart failure; sleep apnea syndrome

Abbreviations: CHF = chronic congestive heart failure; CRF = chronic respiratory failure; CSR = Cheyne-Stokes respiration

Sleep apnea syndrome in patients with chronic congestive heart failure (CHF) is predominantly central, manifest as Cheyne-Stokes respiration (CSR), and associated with a shorter average survival.1- 3 However, despite suffering from prolonged and frequent episodes of apnea and oxygen desaturation, patients with sleep apnea syndrome in the presence or absence of CHF are not known to be at risk of death immediately attributable to apnea and dips in oxygen. Furthermore, we are aware of no previous report of death occurring during polysomnography in patients suffering from Pickwickian syndrome or CHF.


A 60-year-old woman was hospitalized for an episode of severe left-sided and right-sided cardiac decompensation. A review of her records revealed that she had such episodes previously, and had been placed on a long-term regimen of medications for the management of CHF. Her medical history included diabetes and systemic hypertension and, although whether her presenting episode was primarily due to left-sided or right-sided CHF was not formally clarified, it was attributed to her known ischemic heart disease and documented myocardial infarction followed by coronary artery by pass grafting 10 years earlier. At that time, the angiographically measured left ventricular ejection fraction was 40%. Because she weighed 166 kg and her body mass index was 60.24, follow-up echocardiography was limited by poor cardiac imaging, no additional investigations were undertaken, and she was treated symptomatically when she began presenting with CHF. Before the last episode, she was stable in New York Heart Association functional class III CHF.

On presentation, while receiving 0.5 L/min of nasal oxygen, the patient's PaO^sub 2^ was 40 mm Hg; PCO^sub 2^, 82 mm Hg; pH 7.31; HCO^sub 3^^sup -^ 40 mnol/L; base excess, + 10.2; oxygen saturation, 67%; and hemoglobin, 16.5 g/dL, findings consistent with chronic respiratory failure (CRF) and partly compensated respiratory acidosis. After treatment of the acute episode of heart failure, the room air oxygen saturation, measured by pulse finger oximetry, stabilized at 60%. This patient had never undergone pulmonary function testing, and her blood gas abnormalities were attributed mostly to combined left-heart failure and obesity (Pickwickian syndrome). After clinical stabilization, the patient complained of severe daytime sleepiness and nocturnal restlessness, and was advised to first undergo polysomnography. Because of subsequent events, no other pulmonary investigations were undertaken. After relief of her dyspnea, the patient refused further oxygen supplementation, and polysomnography was recorded at home (Hypnotrace Tyco; Healthcare Puritan Bennett; Industry, CA) after she had requested to be discharged from the hospital. She was found dead on the morning alter polysomnography.

The overall duration of polysomnography was 570 min (the recording began at 9:16 PM and ended at 6:46 AM), 505 min of which, up to the time of respiratory arrest, were analyzed. The last 65 min of the recording were those after final breathing, a time not studied in terms of hypnogram and apnea-hypopnea index to obtain polysomnography analysis and scoring when the patient was alive. Sleep time, according to the hypnogram, was 434 min. The tracings, obtained with measurements of nasal pressure airflow by nasal cannula and no thermocouple, confirmed the diagnosis of severe CSR with an apnea-hypopnea index of 93.3. This breathing pattern was evident from the beginning of our recording during wakefulness with airflow waveforms consistent with CSR, with a waxing-waning pattern and without respiratory pause. Fixed CSR with frank apnea developed throughout the recording after the patient fell asleep (Fig 1, left, A). No obvious obstructive exent was observed in this obese and hypercapnic patient. Oxygen saturation on room air was 61% in the beginning of the recording and averaged 58.7% during the night. At 5:22 AM, after a last complete episode of CSR, absence of the waxing phase of the next cycle and of the negative component of the airflow waveform was observed, as active expiratory movements ceased (Fig 1, right, B). In addition, the amplitude of the waveform decreased, with lengthening of the pauses between respiratory cycles. The progressive and regular decrease in the amplitude of the waveform ended in complete absence of airflow movement 19 min later (Fig 2, left, A). Cardiac arrest was documented 7 min alter the respiratory arrest (Fig 2, right, B).

FIGURE 1. Polysomnographic recordings. Left, A: The airflow waveform is consistent with CSR. Right, B: At 5:22 AM, the waxing phase and negative component of the CSR airflow waveform are no longer recorded.

FIGURE 2. Final polysomnographic recordings. Left, A: Cessation of airflow movements. Right, B: Cardiac arrest is recorded 7 min later.


It is not known whether patients with CHF and CSR are at higher risk of death from pure respiratory arrest. To date, the risk of death in this population has been mostly attributed to ventricular tachyarrhythmias and ischemic or hemodynamic events.4,5 Systematic polysomnography recordings in patients with CHF may provide information regarding their risk of respiratory death, in the presence or absence of CRF.

Patients with CHF often present with chronic hypocapnia, partly due to hyperventilation. This leads to CSR with central apneas caused by falls in PCO^sub 2^ below the apneic threshold. Therefore, a compensatory respiratory acidosis is unusual in these patients. Our patient presented with severe hypoventilation manifested by profound hypoxeinia and respiratory acidosis. This uncommon clinical presentation associated with CSR may be due to obesity hypoventilation syndrome or associated CRF from undocumented intrinsic pulmonary disease.

Severe hypoxemia may cause generalized and profound enough cerebral depression to cause death. Therefore, though no autopsy was performed, we believe that hypoxemia was this patient s likely cause of death. The coincidence of death and polysomnography was undoubtedly fortuitous, and no other event preceding the respiratory arrest was identified, including acute cardiac decompensation or cardiac arrhythmias, to explain the interruption in rhythmic respiratory drive. Brain dysfunction was likely related to chronic hypoxemia and recurrent dips in oxygen during apneas, facilitated by a heightened sensitivity to stress from CHF and secondary organ hypoperfusion. Nevertheless, questions remain unanswered, since the low oxygen concentration measured during the recording was similar to that observed during daytime, it is highly likely that CSR observed had been present regularly before this fatal episode. The airflow waveform consistent with CSR recorded in the beginning of polysomnography, when the patient was awake, was likely a sign of persistent and severe brain hypoxemia, and may be one of the signs that the patient was at high risk of respiratory death. In conclusion, besides the classic fatal cardiac events, failure of respiratory drive during sleep may be a cause of death in patients presenting with CHF and CSR, associated with Pickwickian syndrome in the present case.

* From Hpital Cardiologique du Haut-Lvque, University of Bordeaux, Pessac, France.


1 Bradley TD, Floras JS. Sleep apnea and heart failure: part I. Obstructive sleep apnea. Circulation 2003; 107:1671-1678

2 Bradley TD, Floras JS. Sleep apnea and heart failure: part II. Central sleep apnea. Circulation 2003: 107: 1822-1826

3 Sin DD, Logan AG, Fitzgerald FS, et al. Effects of continuous positive airway pressure on cardiovascular outcomes in heart failure patients with and without Cheyne-Stokes respiration. Circulation 2000; 102:61-66

4 Stevenson WG, Epstein LM. Predicting sudden death risk for heart failure patients in the implantable cardioverter- defibrillator age. Circulation 2003; 107:514-516

5 Boehmer JP. Device therapy for heart failure. Am J Cardiol 2003; 91:53D-59D

Philippe Bordier, MD; Stephane Garrigue, MD; Sylvain Reuter, MD; Pierre Bordachar, MD; and Jacques Clementy, MD

Manuscript received August 26, 2003: revision accepted June 1, 2004.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (e-mail: [email protected]).

Correspondence to: Philippe Bordier, MD, Hpital Cardiologique du Haut-Lvque, avenue de Magellan, 33604 Pessac cedex, France; e-mail: [email protected]

Copyright American College of Chest Physicians Nov 2004