Behavioral hyperventilation is a rarely recognized cause of central sleep apnea (CSA) among children. We report two pediatric patients who presented with prolonged central sleep apnea secondary to behavioral hyperventilation. One patient also had a prolonged corrected QT (QTC) interval resulting from hyperventilation.
Johnston TP, Tam-Williams J, Schmandt M, Patel AC, Cleveland C, Coste F, Kemp JS. Behavioral hyperventilation and central sleep apnea in two children. J Clin Sleep Med 2015;11(4):487–489.
Hyperventilation-induced apnea has been described in awake pediatric patients,1,2 but reports of the effects of hyperventilation on breathing during sleep are rare. Pevernagie et al. reported a 9-year-old girl with Cornelia de Lange syndrome with prolonged CSA events provoked by acute hyperventilation after witnessing a hospital roommate's sudden death.3 We present two cases of hyperventilation-induced CSA in children without obvious neurodevelopmental abnormalities.
REPORT OF CASES
A 9-year-old otherwise healthy girl came to an outside emergency department with chest pain, dyspnea, and dizziness. She had two observed apneas while sleeping with oxyhemoglobin saturation (SpO2%) as low as 46% and was transferred to our hospital. Her weight was 35.6 kg (85%tile) and her height was 137 cm (75%tile). Her BMI was 18.6 kg/m2, respiratory rate 20 breaths per minute, and she had small tonsils. The remainder of her examination was unremarkable. Polysomnography (PSG) showed 2 central apneas within the first 5 min of falling asleep, lasting 32 and 51 s, with SpO2% falling to 65% (Figure 1A). There were 8 brief central apneas, but no further hypoxemia during the remaining 6 h of sleep. At sleep onset, her end-tidal CO2 by infrared capnometry (PETCO2) was less than 20 mm Hg and rose within 13 min of sleep onset to 35 mm Hg. Laboratory studies indicated chronic respiratory alkalosis (venous blood gas pH = 7.48, PCO2 = 25 mm Hg, plasma HCO3 13 mg/dL).
(A) Patient 1 had 2 apneic events of 30 and 50 seconds within the first 90 seconds of sleep onset. Transcutaneous capnometry (TCM) range was 14–28 mm Hg, SpO2 nadir was 66%. (B) Patient 2 had a 79 second central apnea within 30 seconds of sleep onset. TCM range was 23–29 mm Hg. SpO2 nadir was 51%.
(A) Patient 1 had 2 apneic events of 30 and 50 seconds within the first 90 seconds of sleep onset. Transcutaneous capnometry (TCM) range was 14–28 mm Hg, SpO2 nadir was 66%. (B) Patient 2 had a 79 second central apnea within 30 seconds of sleep onset. TCM range was...
Her initial ECG showed a prolonged QTc interval of 627 ms.4 The QTC interval was normal on 2 subsequent ECGs. Brain and brainstem MRI with and without gadolinium-based contrast was normal. Urinalysis showed pH = 7.0, with calcium oxalate crystals. Urine calcium-to-creatinine ratio was normal, and renal ultrasound showed no nephrolithiasis. Urine toxicology screens were negative.
Coached “diaphragmatic” breathing helped slow her respiratory rate, and clonazepam was prescribed to manage her chronic anxiety. With these interventions, her awake PETCO2 rose to the 30s.
A 13-year-old boy had a past medical history of hypothyroidism, adequately treated with levothyroxine. He was referred by his pediatrician for an otolaryngologist's evaluation and PSG because of daytime sleepiness, snoring, apneic pauses in sleep, and sleepwalking. His weight was 49 kg (63%tile) and his height was 138 cm (< 3%tile); BMI was 24.0 kg/m2. He had enlarged tonsils. Before the PSG, he became acutely anxious, and his respiratory rate by PSG was 23/min, with PETCO2 = 22 mm Hg. PSG recorded 27 CSA events occurring primarily within the first 5 min after sleep onset (Figure 1B) The longest CSA lasted 79 s, with oxygen desaturation to 51%. Obstructive AHI was 5.8/h, and there was 6.4% of sleep time spent in periodic breathing. Brain MRI and echocardiogram were both normal. Urine toxicology screens were negative.
An adenotonsillectomy was performed 4 months after PSG because of the obstructive events, with improvement in his snoring and mouth breathing. Six months after the adenotonsillectomy, he had a repeat PSG during which he was less anxious. At sleep onset, his PETCO2 = 35.0 mm Hg and only 6 events, 5 central and 1 obstructive, were recorded; the longest central apnea was 15 seconds.
Behavioral hyperventilation is the apparent antecedent for CSA in both of our patients, without underlying central nervous system, metabolic, or cardiovascular abnormalities. While awake both patients had tachypnea with hypocarbia documented by capnography, and Patient 1 had respiratory alkalosis with metabolic compensation. Central apnea occurred during sleep initiation, with apnea as long as 51 and 79 seconds. This pattern is consistent with earlier cases of hyperventilation-associated CSA.3
In Case 1, in particular, the resolution of central apneas during the first several minutes of sleep, coincident with the gradual increase in PETCO2, emphasizes the pathophysiologic importance of the CO2 apneic threshold during sleep in CSA.
CSA is predictable among children with other reasons for hypocapnia, including those living at high altitude, in heart failure, or with renal failure on dialysis. Infants not infrequently have CSA longer than 20 seconds, but rarely as long as our subjects, and usually during REM sleep and not at sleep onset.5 Patients with Rett syndrome often hyperventilate and have central apnea when awake6 and occasionally, like our subjects, during wake-sleep transition.7 Behavioral hyperventilation-induced CSA has been rarely reported, however, and is sufficiently rare that it should be considered a diagnosis of exclusion after more threatening diagnoses have been ruled out. Nevertheless, failure to recognize behavioral hyperventilation-induced CSA could result in unnecessary diagnostic evaluations, such as cardiac electrophysiologic studies for prolonged QTC syndrome. Or, the presence of an incidental Chiari malformation might prompt surgical intervention without recognizing another straightforward cause of CSA.
This was not an industry supported study. The authors have indicated no financial conflicts of interest.