Autonomic dysfunction (AD) has been associated with both obstructive and central sleep apneas. Several mechanisms have been proposed for the emergence of sleep apnea in AD, which include impaired sensory input, compromised local reflexes, and altered central processing. We present a case of a 70-year-old woman who had experienced cardiac arrest four times related to hypoxic events due to the apparent sudden onset of obstructive sleep apnea (OSA) in the setting of AD. The episodes of OSA were profoundly prolonged and a tracheostomy was ultimately needed due to the inability of positive airway pressure therapy with supplemental oxygen to control events. We think that this case is unique because of the extreme duration of the obstructive apneas (up to 233.8 seconds), which almost certainly reflects lack of protective autonomic control in terminating these events.
Aulakh PK, Westerman DE, Dedhia RC. The longest obstructive apnea you have ever seen: a patient with new-onset autonomic dysfunction. J Clin Sleep Med. 2018;14(5):893–895.
Treatment of severe obstructive sleep apnea (OSA) can be challenging. Abrupt onset of OSA is rare but has been encountered in conditions with primary autonomic disorder such as multiple system atrophy (MSA).1
Persons with autonomic dysfunction (AD) appear to have susceptibility to airway obstruction during sleep due to several factors that may include aberrant afferent and central pathways leading to impaired brainstem reflexes. Patients with AD may have vocal cord paralysis necessitating tracheostomy. In the patient discussed here, tracheostomy was necessary for failure of bilevel positive airway pressure (BPAP) and oxygen to eliminate the sleep-disordered breathing (SDB), which included prolonged obstructive apneas of previously unreported duration.
REPORT OF CASE
A 70-year-old nonobese woman was transferred to Emory University Hospital in March 2016 after experiencing cardiac arrest four times related to hypoxic events of unclear etiology. Comprehensive work-up, which included echocardiogram with contrast; electroencephalogram; magnetic resonance imaging/ magnetic resonance angiogram of the head/neck; computed tomography of the head, chest, abdomen, and pelvis; Doppler ultrasound of the carotids; and nuclear medicine myocardial perfusion study failed to reveal the etiology of the recurrent hypoxic events leading to cardiac arrest. A permanent pacemaker had been placed following the third cardiac arrest.
The patient also complained of paresthesia of the left periauricular region and generalized hyperesthesia of the upper trunk and neck. Marked swelling of the tongue was present, accompanied by dysarthria and difficulty swallowing. Nerve conduction studies revealed evidence of bilateral trigeminal neuropathies, but no evidence of a generalized large fiber polyneuropathy. Additionally, during her hospitalization, the patient experienced a syncopal episode. Orthostatic vital signs revealed profound orthostatic hypotension. Due to poor response to intravenous fluid resuscitation, the patient was started on fludrocortisone.
In the hospital, the patient was noted to have intermittent desaturations while awake and more frequently during sleep. Pulmonary function tests showed a decreased diffusion capacity. The patient was discharged with 3 L/min supplemental oxygen.
The patient underwent a split-night sleep study in April 2016 at Northside Hospital Sleep Disorders Center. The apneahypopnea index (AHI) was 27 events/h. The obstructive apnea index was 22 events/h, hypopnea index 5 events/h, and central apnea index zero events/h. The longest obstructive apnea duration was 233.8 seconds (Figure 1). Mean oxygen saturation during the study was 88% and the lowest oxygen saturation was 51%. Percentage total sleep time with oxygen saturation ≤ 89% during diagnostic portion was 21.5%. Supplemental oxygen was added about 10 minutes into the study and titrated up to 3 L/min. Continuous positive airway pressure (CPAP) titration was initiated at 5 cm H2O, transitioned to BPAP that was titrated up to 22/14 cm H2O, then switched to BPAP in spontaneous timed mode. No supplemental oxygen was used during the titration period. Using BPAP in spontaneous timed mode with pressure of 24/14 cm H2O the residual AHI was 5.2. (3 central, 2 mixed, and 2 hypopneas) The patient spent 81.5 minutes at this pressure setting. Because of OSA severity, the patient was loaned an auto-titrating BPAP (inspiratory positive airway pressure max 24, expiratory positive airway pressure min 12, pressure support 8) after the study, until the patient could obtain a device through her insurance. The patient was already on 3 L/min home oxygen, and this regimen was continued because of isolated desaturations below 80% during the titration study.
A 10-minute tracing depicting an obstructive apnea 233.8 seconds in duration.
Note the oxygen desaturation to less than 50% from the preceding apnea before supplemental oxygen was added.
A 10-minute tracing depicting an obstructive apnea 233.8 seconds in duration.
There were no further incidents of respiratory arrest but compliance data now indicated the emergence of central events on BPAP of 25/17 cm H2O. The BPAP pressure was reduced to 22/14 cm H2O. The patient was subsequently referred to the Mayo Clinic in Rochester, Minnesota for further evaluation, where a CPAP titration study confirmed the presence of severe SDB most profound during rapid eye movement (REM) sleep and associated with central apneas lasting up to 140 seconds. Oxygen levels that were reasonably well maintained in non-REM sleep plummeted to 43% during REM sleep. BPAP was ineffective in eliminating events during REM sleep and the patient was transitioned to an adaptive servoventilation device with 1 L/min supplemental oxygen, with evidence of hypoventilation. The patient underwent a repeat titration study at Mayo Clinic 3 days later. During the repeat titration study, initially BPAP in spontaneous mode was utilized, with marked desatu-rations due to obstructive events and few central apneas. The patient was subsequently transitioned to BPAP in spontaneous timed mode and the optimal settings were determined to be BPAP 21/9 cm H2O, backup rate 12 breaths per minute along with 3 L/min supplemental oxygen.
Evaluation of the tongue swelling failed to establish any specific pathology. Spirometry results were within normal limits. The flow volume loop was normal, indicating no significant extrathoracic airway obstruction. The only abnormality on the pulmonary function tests was reduced diffusion capacity at 61% of predicted value. The maximum inspiratory pressure was 60 (predicted 87%) and maximum expiratory pressure was 108 (predicted 81%), indicative of adequate diaphragm and abdominal muscle strength. The patient was evaluated by otolaryngology at Mayo Clinic and Emory University. There was no evidence of vocal cord dysfunction. Pulse oximetry at rest showed an oxygen saturation of 99% with a pulse rate of 84 beats per minute. Exercise challenge failed to result in oxygen desaturation or an increase in heart rate.
Severe autonomic abnormalities were detected on the patient's autonomic reflex screen, including an abnormal thermoregulatory sweat test and elevated protein in cerebrospinal fluid but a negative oligo bands and immunoglobulin G index. A trial of immunotherapy for an autoimmune autonomic neuropathy failed to elicit any improvement.
Severe hypoxemia persisted despite nocturnal BPAP therapy with supplemental oxygen at 3 L/min (oxygen nadir 59%) and a tracheostomy was therefore performed at Emory Hospital in August 2016.
Repeat polysomnography (PSG) was performed on August 30, 2016 with the tracheostomy tube open. The study was started with breathing room air but oxygen desaturations in the 60% range occurred. Twenty-eight percent oxygen was added and saturations remained ≥ 90% for almost the entire study duration. A few central apneas were present and the overall AHI was only 2 events/h. An arterial blood gas analysis on room air on arising for evaluation of hypoventilation included a pH 7.46, PO2 71 mmHg, and PCO2 44 mmHg. Repeat PSG in November 2016 still showed mild SDB with an AHI of 5.6 events/h. The longest event was a hypopnea in REM sleep lasting 134 seconds. The baseline oxygen saturation breathing room air while awake was 100% and the lowest saturation was 63%. Two L/min oxygen was added and the saturations remained above 90%. The patient has experienced no further incidents of respiratory arrest. Her tracheostomy is occluded during the day and opened at night, and she remains on supplemental nocturnal oxygen. Her tongue is no longer swollen and the etiology of the swelling was never established. However, the patient still has numbness of the chest and face as well as some difficulty with mastication.
This case report illustrates abrupt onset of OSA associated with autonomic dysfunction. Our patient was noted to have severe changes on the autonomic reflex screen and thermoregulatory sweat test along with orthostatic hypotension. Etiology of airway obstruction in patients with OSA and autonomic dysfunction may be related to impaired input from sensory fibers in the upper airway which play a vital role in preserving the patency of the airway.2 A recent study by Hilz et al.3 demonstrated a higher prevalence of OSA in individuals with familial dysautonomia as compared to the control group Their study noted that the obstructive apnea index was six times higher than the central apnea index in the familial dysautonomia patient group. Our patient was likewise noted to have predominantly obstructive events on diagnostic PSG. The respiratory events were extraordinarily prolonged and associated with severe oxygen desaturation. The longest obstructive apnea was 233.8 seconds in duration (Figure 1), which we believe to be the longest obstructive apnea reported in the literature.
Sudden nocturnal death was the most common cause of fatality in a cohort of patients with MSA, a primary autonomic disorder, studied by Nishizawa and Shimohata.4 They also noted that CPAP therapy failed to prevent sudden death in patients with MSA. Their finding highlights the critical role of an intact autonomic nervous system in preventing sudden life-threatening events during sleep in the appropriate setting. This reinforces our belief that the extremely long obstructive apneas experienced by our patient represented a failure of autonomic responses in terminating these events, resulting in repeated bouts of acute near-fatal respiratory failure episodes.
Interestingly, a study by Bowes et al.5 demonstrated failure of arousal from sleep in dogs, with carotid body denervation, during profound hypoxia induced by rebreathing technique, with oxygen saturation of 60% in slow wave sleep and 50% in REM sleep. A few dogs required resuscitation when rebreathing was continued after reaching the aforementioned oxygen saturation. One could hypothesize that carotid body dysfunction may have led to failure of arousal from sleep despite profound hypoxia in this patient.
Tracheostomy has been used in patients with MSA to overcome obstruction at a glottic level,6 and was one of the earliest interventions in the treatment of OSA prior to the introduction of noninvasive approaches. It nevertheless remains an important alternative in selected patients with life-threatening OSA.
All authors have seen and approved the manuscript. The authors report no conflicts of interest.