Issue Navigator

Volume 14 No. 02
Earn CME
Accepted Papers

Scientific Investigations

Obstructive Sleep Apnea in Patients With Congenital Central Hypoventilation Syndrome Ventilated by Diaphragm Pacing Without Tracheostomy

Annie Wang, MD1; Sheila Kun, RN, MS2; Bonnie Diep, MD1; Sally L. Davidson Ward, MD1,2; Thomas G. Keens, MD1,2; Iris A. Perez, MD1,2
1Keck School of Medicine of USC, University of Southern California, Los Angeles, California; 2Division of Pulmonology and Sleep Medicine, Department of Pediatrics, Children's Hospital Los Angeles, Los Angeles, California


Study Objectives:

To determine presence of obstructive sleep apnea (OSA) in patients with congenital central hypoventilation syndrome (CCHS) ventilated by diaphragm pacing (DP) without tracheostomy, and to determine if OSA can be improved by DP setting changes.


We reviewed polysomnography (PSG) results of 15 patients with CCHS from October 2001 to April 2014, age 15.4 ± 7.8 years, body mass index 22.0 ± 6.0 kg/m2, and 60% female.


Of the 22 PSG results obtained for the 15 patients with CCHS, 9 were performed with tracheostomy capped, and 13 were performed after patients underwent decannulation. OSA was present on 6 of 9 tests in patients with tracheostomy capped, including 3 patients with immediate, severe OSA necessitating that the studies be completed with tracheostomy uncapped. OSA was present on 2 of 13 tests in patients in whom decannulation had been performed. Hypoventilation was seen on only one test without OSA. On 2 of 5 tests showing OSA, OSA improved by decreasing DP amplitude settings; apnea-hypopnea index decreased from 11.1 ± 2.5 to 1.8 ± 2.5 events/h; PETCO2 decreased from 57.5 ± 3.5 to 38.5 ± 0.7 torr; SpO2 increased from 76.5 ± 0.7% to 93.0 ± 7.1%. OSA improved in one patient with slight increase in respiratory rate. Settings were manipulated in 4 tests showing OSA; no changes were attempted in the remaining study. One patient was placed on bilevel positive airway pressure with temporary suspension of DP. Age (P < .119), previous adenotonsillectomy (P < .211), and body mass index (P < .112) did not significantly contribute to OSA.


OSA occurs in patients with CCHS ventilated by DP. However, decreasing DP amplitude settings can lessen upper airway obstruction without compromising gas exchange.


Wang A, Kun S, Diep B, Davidson Ward SL, Keens TG, Perez IA. Obstructive sleep apnea in patients with congenital central hypoventilation syndrome ventilated by diaphragm pacing without tracheostomy. J Clin Sleep Med. 2018;14(2):261–264.


Current Knowledge/Study Rationale: There are limited data on the presence of obstructive sleep apnea in patients receiving diaphragm pacing without tracheostomy. Therefore, we sought to assess whether obstructive sleep apnea occurs in patients with congenital central hypoventilation syndrome who are ventilated by diaphragm pacing, and whether obstructive apneas can be resolved with diaphragm pacer setting changes.

Study Impact: In some centers decannulation has not been performed in patients ventilated by diaphragm pacing due to risk of obstructive sleep apnea; however, this study shows that diaphragm pacing without tracheostomy can be achieved and obstructive sleep apnea can be alleviated.


Congenital central hypoventilation syndrome (CCHS) is a rare genetic disorder of the autonomic nervous system due to a mutation in the PHOX2B gene.1 CCHS is characterized by alveolar hypoventilation during sleep or during both sleep and wakefulness.1,2 All patients with CCHS require lifelong ventilatory assistance during sleep; a subset of patients also require daytime assistance.1,2 In some patients with CCHS who only need ventilatory support during sleep, diaphragm pacing (DP) can be used to manage respiratory insufficiency.3 One advantage of this approach is that DP can obviate the need for tracheostomy.

However, several reports have found that patients with DP have concurrent upper airway obstruction at the pharyngolaryngeal site, such as vocal cord adduction and glottic obstruction.46 During normal inspiration, there is synchronous contraction of upper airway skeletal muscles to maintain an open upper airway. However, DP sends electrical current directly to the diaphragm via the phrenic nerve and bypasses the brainstem respiratory center without stimulating synchronous neuronal activity to the upper airway skeletal muscles.2,5 Thus, pacing can induce upper airway obstruction in patients with CCHS because there is no stiffening of the upper airway in anticipation of the negative intraluminal pressure generated by the diaphragm contraction.

There are 4 case reports of upper airway obstruction in patients undergoing DP.47 We hypothesize that although obstructive sleep apnea (OSA) can occur in patients with CCHS who are ventilated by DP without tracheostomy, the OSA can be resolved by modifying diaphragm pacer settings and thereby reducing the negative intraluminal pressure imposed on the upper airway.


We performed a retrospective study of 15 patients with CCHS, who were ventilated by DP (Model: Mark IV external transmitter, Avery Biomedical Devices, Commack, New York, United States) during sleep and from whom polysomnography (PSG) results were obtainable from October 2001 to April 2014. The diagnosis of CCHS was confirmed by genetic testing in 14 patients. Polysomnography was carried out in patients without tracheostomy as part of periodic objective assessment of ventilation during sleep or with uncuffed tracheostomy capped to address readiness for decannulation. The following data were collected: (1) demographic, (2) PHOX2B genotype, (3) body mass index (BMI), (4) nasal steroid use, (5) ventilatory status at time of sleep study, and (6) history of adenotonsillectomy. The following data from PSG tests were analyzed: (1) total sleep time, (2) baseline and lowest oxygen saturation (SpO2), (3) baseline and highest end-tidal carbon dioxide tension (PETCO2), (4) obstructive and mixed apneas, (5) obstructive hypopneas, (6) apnea-hypopnea index (AHI), and (7) diaphragm pacer settings at the time of study (rate and right and left amplitudes). This study was approved by the Institutional Review Board of the Children's Hospital Los Angeles.

Overnight Polysomnography

Surface electrodes were connected to the patient to monitor electroencephalogram, electrocardiogram, chin electromyogram, and electro-oculogram to stage sleep. In addition, calibrated respiratory inductance plethysmography of the chest and abdomen and pulse oximetry were applied and recorded. A nasal air pressure transducer, oronasal thermal sensor, and snore transducer were placed at the nose and mouth to evaluate combined airflow. Expired carbon dioxide tension was measured continuously with a sampling catheter taped to the nose or mouth. During the study, pacer settings were adjusted based on our goals of keeping PETCO2 between 30 to < 40 torr and oxygen saturation of 95% or higher.

Just prior to bedtime, patients were connected to the monitors while still awake, as noted. In those with tracheostomy, the tracheostomy tube was capped. The antennae that are connected to the transmitter were placed on the receivers on each side. The pacing was initiated by turning the pacers on both sides with the patient still awake. The study was started on the diaphragm pacer settings that the patient had at home. If SpO2 fell < 95% and/or PETCO2 rose > 40 torr, pacers were adjusted by increasing the amplitudes by 0.2 to 0.4 unit on each side and/or by increasing the rate. If the patient were to experience shoulder pain with the pacer breath, this indicated that the amplitude was set too high and that the setting needed to be turned back down by 0.5 to 1.0 unit. If obstructive events were present, the amplitudes were decreased by 0.2 to 0.4 unit on each side. In this instance, the rate might need to be increased by 1–2/min to meet PETCO2 goals. In some patients, the amplitude changes were in smaller increments. If snoring alone was present, but with adequate gas exchange, SpO2 ≥ 95% and/or PETCO2 ≤ 40 torr, the pacer settings were not changed. During the sleep study, the tidal volumes were not measured, but indirectly assessed by adequacy of gas exchange, chest wall motion, diaphragmatic excursion, and improvement in paradoxical breathing.

Data Analysis

We used paired t tests to compare SpO2, PETCO2, and AHI before and after adjustments in amplitude. Chi-square analysis was used to compare proportions of patients with OSA. Calculations were performed using Excel's Descriptive Statistical Tool, 2010, version 14.4.6 (Microsoft Corp, Redmond, Washington, United States). Statistical significance was set at P < .05.


Patient Demographics

Fifteen patients with CCHS, who were ventilated by DP, were included in the study. Of the 15 patients, 6 were being paced without tracheostomy, and 9 had sleep studies performed with capped tracheostomy as part of the evaluation for readiness for decannulation. The mean age of our cohort at the time of the sleep study was 15.4 ± 7.8 years (range, 5 to 29 years), BMI was 22.0 ± 6.0 kg/m2 (range 14.2–34.0), and 60% were female (Table 1).

Patient demographics and polysomnogram findings.


table icon
Table 1

Patient demographics and polysomnogram findings.

(more ...)

Fourteen patients with CCHS had confirmed PHOX2B gene mutations with a heterozygous polyalanine repeat mutation (PARM). The genotypes are as follows: 9 patients had 20/25 PARM, 2 patients had 20/26 PARM, 2 patients had 20/27 PARM, and 1 patient had 20/33 PARM. One patient's PHOX2B genotype was not available (Table 1).

Of the 15 patients with CCHS, 8 were using nasal steroids, 1 had history of adenotonsillectomy, 2 had both history of adenotonsillectomy and used nasal steroids, and the remaining 4 patients did not have surgery nor required nasal steroids (Table 1).

Polysomnography Findings

Twenty-two PSG tests were obtained from the 15 patients (12 overnight and 10 daytime sleep studies). Five patients had more than one test.

Of the 22 tests, 9 (41%) were obtained with tracheostomy capped, and 13 (59%) were obtained after patients had undergone decannulation. OSA—defined as AHI > 1.5 events/h in the patients younger than 18 years—was found in 6 of 9 tests (67%) from patients in whom the tracheostomy had been capped. Three of these patients experienced immediate, severe OSA necessitating uncapping of the tracheostomy during the sleep study; thus, AHI values were not available for these tests but the other characteristics of these patients were included in the other analyses (Table 1). OSA was present in 2 of 13 tests (15%) for patients in whom decannulation had been performed.

Age (P < .119), previous adenotonsillectomy (P < .211), and BMI (P < .112) did not significantly contribute to OSA.

OSA was noted in 8 tests. AHI was measured in 5 of the 8 tests and the mean was 14.1 ± 16.3 events/h (range 2.6–42.3 events/h). In 6 tests (75%), hypoxemia was present (SpO2 < 92%) with a mean lowest SpO2 of 70.2 ± 11.9% (range, 49% to 79%). Hypercapnia was present in 4 of 8 tests with OSA (50%), with mean maximal PETCO2 of 58.0 ± 2.4 torr (range 55–60 torr) and 1 of 14 tests without OSA (7%).

During PSG in 4 of 5 patients with OSA, adjustments in diaphragm pacer settings (decreasing amplitude) were made during the study, with improvement in OSA in 2 of the 5 tests (40%) (Figure 1). After diaphragm pacer setting changes were made, AHI decreased from 11.1 ± 2.5 events/h to 1.8 ± 2.5 events/h; PETCO2 decreased from 57.5 ± 3.5 torr to 38.5 ± 0.7 torr; and SpO2 nadir increased from 76.5 ± 0.7% to 93.0 ± 7.1%. OSA improved in one patient with only a slight increase in respiratory rate. No changes were attempted in one of the five patients. One patient had to be placed on bilevel positive airway pressure because of severe OSA despite diaphragm pacer setting changes during the study.

Improvements in AHI, SpO2, and PETCO2 following a decrease in amplitude (voltage) in diaphragm pacer settings for two patients with OSA.

AHI = apnea-hypopnea index, PETCO2 = end-tidal carbon dioxide, OSA = obstructive sleep apnea, SpO2 = oxygen saturation.


Figure 1

Improvements in AHI, SpO2, and PETCO2 following a decrease in amplitude (voltage) in diaphragm pacer settings for two patients with OSA.

(more ...)

We also reviewed individual patient results because some accounted for more than one test in the results provided in the previous paragraphs. Five patients (3 who had undergone decannulation, 2 with capped tracheostomy) experienced OSA during polysomnography. Three of the five patients with OSA (60%) improved by decreasing the diaphragm pacer amplitude settings or rate. As mentioned previously, one patient was transitioned to bilevel positive airway pressure. Subsequently, her diaphragm pacer settings were readjusted, and her follow-up PSG showed improved OSA with new diaphragm pacer settings.

The mean BMI of the 8 patients with CCHS at the time of PSG with OSA was 18.7 ± 3.5 kg/m2 (range 14.2–25.1 kg/m2), and was not significant compared to those without OSA at the time of PSG: 23.9 ± 5.9 kg/m2 (range 15.5–32.1 kg/m2) (P < .112). Of the patients with OSA, one was overweight with a BMI of 25.1 kg/m2 and had an AHI of 42.3 events/h. Her lowest SpO2 was 49%, and highest PETCO2 was 60 torr. Her obesity could have contributed to the severe hypoxemia, hypoventilation, and OSA while on DP as well as limited the effectiveness of pacing as a mode of ventilation. This patient was placed on bilevel positive airway pressure.


Our study shows that OSA can accompany DP for patients with CCHS without tracheostomy or with capped tracheostomy during sleep. During normal inspiration, the brainstem's respiratory center coordinates upper airway skeletal muscle contraction with diaphragm contraction to maintain upper airway patency.35 During DP, this synchronous contraction is bypassed, predisposing the upper airway to collapse due to the negative intrathoracic pressure created by the diaphragm contraction and absent upper airway skeletal muscle contraction. Despite smaller airway dimensions and midposition of their vocal cords at rest,3 younger children and infants are not necessarily more susceptible to upper airway collapse. Being overweight, which is a contraindication to DP, can also contribute to obstructive apnea, as seen in one of our patients.

Some centers have not performed decannulation in patients ventilated by DP due to risk of OSA; however, at our center, we recently showed that DP without tracheostomy can be successfully achieved in patients with CCHS.8 Home care is simplified and morbidity is decreased when DP eliminates the need for tracheostomy. Our study indicates that changing the diaphragm pacer settings can alleviate obstructive apneas if present. By decreasing the amplitude setting, the force of inspiration is decreased with each diaphragm contraction, enabling adequate ventilation and oxygenation. By improving airway obstruction along with the OSA, hypoventilation improved as well.

In conclusion, upper airway obstruction occurs during DP without tracheostomy. However, this can often be improved by decreasing amplitude settings.


Work for this study was performed at Children's Hospital Los Angeles (CHLA), Los Angeles, California. The authors report no conflicts of interest.



apnea-hypopnea index


body mass index


congenital central hypoventilation syndrome


diaphragm pacing


obstructive sleep apnea


polyalanine repeat mutation


end-tidal carbon dioxide




oxygen saturation



Weese-Mayer DE, Berry-Kravis EM, Ceccherini I, Keens TG, Loghmanee DA, Trang H. An official ATS clinical policy statement: congenital central hypoventilation syndrome: genetic basis, diagnosis, and management. Am J Respir Crit Care Med. 2010;181(6):626–644. [PubMed]


Chen ML, Keens TG. Congenital central hypoventilation syndrome: not just another rare disorder. Paediatr Respir Rev. 2004;5(3):182–189. [PubMed]


Chen ML, Tablizo MA, Kun S, Keens TG. Diaphragm pacers as a treatment for congenital central hypoventilation syndrome. Expert Rev Med Devices. 2005;2(5):577–585. [PubMed]


Glenn WW, Gee JB, Cole DR, Farmer WC, Shaw RK, Beckman CB. Combined central alveolar hypoventilation and upper airway obstruction: treatment by tracheostomy and diaphragm pacing. Am J Med. 1978;64(1):50–60. [PubMed]


Hyland RH, Hutcheon MA, Perl A, et al. Upper airway occlusion induced by diaphragm pacing for primary alveolar hypoventilation: implications for the pathogenesis of obstructive sleep apnea. Am Rev Respir Dis. 1981;124(2):180–185. [PubMed]


Domanski MC, Preciado DA. Vocal cord collapse during phrenic nerve-paced respiration in congenital central hypoventilation syndrome. F1000Res. 2012;1:42. [PubMed Central][PubMed]


Reverdin AK, Mosquera R, Colasurdo GN, Jon CK, Clements RM. Airway obstruction in congenital central hypoventilation syndrome. BMG Case Rep. 2014;2014.


Diep B, Wang A, Kun S, et al. Diaphragm pacing without tracheostomy in congenital central hypoventilation syndrome patients. Respiration. 2015;89(6):534–538. [PubMed]