- Email: [email protected]
Predicting the severity of congenital high airway obstruction syndrome
Journal of Pediatric Surgery (2010) 45, 1633–1639
www.elsevier.com/locate/jpedsurg
Predicting the severity of congenital high airway obstruction syndrome Jessica L. Roybal 1 , Kenneth W. Liechty, Holly L. Hedrick, Michael W. Bebbington, Mark P. Johnson, Beverly G. Coleman, N. Scott Adzick, Alan W. Flake ⁎ Center for Fetal Diagnosis and Treatment, Children's Hospital of Philadelphia, PA 19104-4318, USA Received 4 November 2009; revised 17 January 2010; accepted 18 January 2010
Key words: CHAOS; Hydrops fetalis; EXIT; Ex utero intrapartum therapy
Abstract Background: Congenital high airway obstruction syndrome (CHAOS) is caused by complete or nearcomplete obstruction of the fetal airway. Obstruction sets in motion a sequence of events that can ultimately lead to fetal demise. However, on rare occasions in utero airway decompression occurs, reversing syndromic findings and improving the prognosis. In our relatively large series of CHAOS patients, we have observed a spectrum of clinical severity. The aim of this study was to identify the prenatal characteristics of CHAOS predictive of a milder postnatal course. Methods: The medical charts of all fetuses observed at our institution with the diagnosis of CHAOS were reviewed for radiologic findings, delivery information, perinatal course, autopsy or discharge report, and long-term follow-up. Results: Between 1996 and 2008, 12 fetuses with CHAOS were identified. Four fetuses had no evidence of hydrops on initial imaging. Of the 8 fetuses displaying hydrops, 3 were terminated, 2 died in utero, and 1 with multiple anomalies died at birth. Six fetuses were delivered via the ex utero intrapartum therapy procedure for attempted salvage, and 5 of the 6 survived the neonatal period including all 4 fetuses without hydrops. Serial prenatal imaging demonstrated less severe signs of CHAOS in 3 fetuses, and in 2 of them, direct laryngoscopy revealed a tiny opening in the airway. All 3 fetuses that showed improvement on serial imaging survived the neonatal period and were discharged home by 2-10 weeks of age. Conclusions: Although the natural history of CHAOS is variable, trends in prenatal ultrasound findings are highly predictive of postnatal outcome and are a valuable guide to prenatal counseling. © 2010 Elsevier Inc. All rights reserved.
Congenital high airway obstruction syndrome (CHAOS) is caused by complete or near-complete obstruction of the fetal airway. The term was first introduced by Hedrick and ⁎ Corresponding author. Department of Surgery, Philadelphia, PA 19104-4318, USA. Tel.: +1 215 590 3671; fax: +1 215 590 3324. E-mail addresses: [email protected] (J.L. Roybal), [email protected] (A.W. Flake). 1 Department of Surgery, Abramson Research Bldg, Rm 1114, 3615 Civic Center Blvd, Philadelphia, PA 19104-4318, USA. 0022-3468/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.jpedsurg.2010.01.022
colleagues [1] in 1994 to describe 4 fetuses with upper airway obstruction and a constellation of ultrasound findings considered incompatible with survival. Fifteen years later, CHAOS remains a rare diagnosis but is no longer uniformly fatal. Better prenatal and perinatal management based on a greater understanding of the natural history of the syndrome have led to improved outcomes. In our series of 12 patients, the largest reported to date, we have observed that CHAOS involves a spectrum of clinical severity.
1634 Congenital high airway obstruction syndrome is thought to occur when the upper airway fails to recanalize around the 10th week of gestation thereby causing obstruction [1]. The fetal lung normally secretes fluid that is cleared through the larynx. Obstruction of the upper airway blocks the flow of fetal lung fluid and raises the intratracheal pressure, which in turn, leads to distension of the tracheobronchial tree and secondary proliferative lung growth [2]. As the lungs expand, the diaphragms flatten and may even evert, and the heart becomes compressed in the midline. Elevated intrathoracic pressure causes decreased venous return and fetal cardiac failure leading to ascites, placentomegaly, and eventually hydrops fetalis. A subtype of CHAOS with a less severe natural history has recently been described [2,3]. A total of 6 patients, reported by 5 different groups, have had a minor pharyngotracheal or laryngotracheal fistula that serves as a pop-off valve for partial decompression of the respiratory tract leading to resolution of sonographic findings and a milder postnatal course. However, most CHAOS cases lack a decompressive tract, and the complete airway occlusion is maintained throughout gestation. The purpose of this study was to determine the spectrum of disease among a relatively large series of fetuses diagnosed with CHAOS and observed at our institution. Given the recent reports of a less severe subtype, we hoped to identify prenatal characteristics predictive of a milder postnatal course.
1. Materials and methods A retrospective search of the fetal database of the Center for Fetal Diagnosis and Treatment of the Children's Hospital of Philadelphia (Pa) was performed for all patients with a prenatal diagnosis of CHAOS after approval was obtained from the institutional review board (IRB 08-006152). Twelve patients were identified. Charts were reviewed for initial diagnosis, prenatal radiologic findings, operative reports, perinatal inpatient course, and autopsy or discharge reports. Serial prenatal ultrasound reports were compared, and resolution of ultrasound findings was defined as an improvement in one or more of the sonographic features of CHAOS including resolution of ascites, normalization of the amniotic fluid volume, restoration of a normal diaphragm contour, or return of the heart to the left chest. Long-term follow-up of survivors through July 2008 was also reviewed.
2. Results Between January 1996 and July 2008, 12 patients were diagnosed with CHAOS. Ten of these fetuses were previously reported by our institution [2-4]. Three fetuses were female and 9 were male. All of the fetuses were referred from outside institutions based on ultrasound findings. Most
J.L. Roybal et al. fetuses were referred with a diagnosis of bilateral congenital cystic adenomatoid malformation or echogenic lungs, but 4 (Table 1; C, F, H, I) had a diagnosis of CHAOS made at their referring institution. Gestational age at initial diagnosis of CHAOS (either at our institution or the referring institution) ranged from 19 to 29 weeks. Evaluation included ultrasound, magnetic resonance imaging (MRI), and echocardiogram. All fetuses on presentation had large, echogenic lungs, 10 had ascites, 9 had signs of airway obstruction (dilated airways), 8 had flattened or everted diaphragms, and 8 had hydrops (Table 1). Based on MRI, obstruction was identified at the level of the midneck in most fetuses. Three fetuses (J, K, L) were terminated after counseling for the diagnosis of CHAOS with associated hydrops. Two fetuses (H, I) died in utero at 21 and 22 weeks of gestation, and both were hydropic. One fetus (G) had multiple other congenital anomalies and was not offered the ex utero intrapartum therapy (EXIT) procedure but carried to term and failed resuscitation at delivery. The remaining 6 fetuses underwent elective or emergent EXIT procedures for attempted salvage. Of the 6 fetuses (A-F) delivered by the EXIT procedure at our institution, 5 survived the neonatal period. One fetus with hydrops (F) was delivered by emergent EXIT for maternal mirror syndrome at 32 weeks of gestation. The fetus was thought to have mid-neck obstruction based on MRI but died 17 minutes after delivery because of nearly complete tracheal agenesis with consequent inability to obtain an airway. At autopsy, the infant was found to have complete agenesis of the trachea from just below the larynx to approximately 1.5 cm above the carina. Retrospective review of the MRI confirmed complete tracheal agenesis extending into the thoracic cavity. Among the successful EXIT procedures, 3 were scheduled and 2 were performed emergently for preterm labor at 34 and 29 weeks (C, D). No fetuses survived the neonatal period without the EXIT procedure. The diagnosis of hydrops in association with CHAOS was a poor prognostic sign. Hydrops is defined by fluid accumulation in a body cavity in addition to the peritoneum (ie, pleural or pericardial fluid) and edema of the skin and/or scalp. Of the 8 fetuses in our series with hydrops at presentation, 3 underwent termination. Of the remaining 5 fetuses, 2 died in utero, 1 had multiple anomalies and failed resuscitation at term, and 1 had tracheal agenesis and died during an unsuccessful EXIT procedure. Only one fetus with coexisting hydrops survived after an EXIT procedure at 31 weeks (E). In contrast, the absence of hydrops in 4 fetuses (A-D) was a favorable prognostic sign, with all 4 surviving the neonatal period after delivery by the EXIT procedure. The presence of isolated ascites without hydrops in 3 of these fetuses (B-D) did not preclude survival. In 3 fetuses (H, I, K), the etiology of ultrasound findings consistent with CHAOS could not be determined because of fetal demise or termination without consent for autopsy. In 9 fetuses, the cause of CHAOS was determined by laryngoscopy, autopsy, or both. The most frequent etiology was
Predicting the severity of congenital high airway obstruction syndrome Table 1
1635
CHAOS patients (grouped according to the presence or absence of hydrops on initial ultrasound)
Other Resolution of Cause of Pinhole Fetus Gestational Hydrops Ascites obstruction fistula on on initial on initial anomalies ultrasound age at laryngoscopy findings diagnosis ultrasound ultrasound A
22
No
No
No
Yes
B
22
No
Yes
No
Yes
C
22
No
Yes
No
Yes
Laryngeal atresia
D
19
No
Yes
Yes
No, but stable
Laryngeal atresia
E
20
Yes
Yes
No
No, worse
Tracheal atresia
F
20
Yes
Yes
No
No, worse
Tracheal agenesis
G
23
Yes
Yes
Multiple
No, worse
Laryngeal atresia
H I J
22 21 20
Yes Yes Yes
Yes Yes Yes
No Yes No
N/A N/A N/A
K L
20 19
Yes Yes
Yes No
No Yes
N/A N/A
Unknown Unknown Laryngeal atresia Unknown Laryngeal cyst
laryngeal atresia found in 4 fetuses (C, D, G, J). Other diagnoses found in one fetus each included tracheal atresia (E), tracheal agenesis (F), a laryngeal web (B), an obstructing laryngeal cyst (L), and complete subglottic stenosis (A). A laryngotracheal fistula with a pinhole visualized in an atretic larynx could only be confirmed in 3 fetuses (A, B, D). Associated anomalies were present in 4 fetuses (D, G, I, L). One fetus (I), found to have left renal agenesis but no other findings of Fraser syndrome, had an early in utero demise at 22 weeks of gestation. A second fetus (L), with ultrasound findings suggestive of caudal regression syndrome, was terminated and autopsy was declined. A third fetus (G) had musculoskeletal, genitourinary, and gastrointestinal anomalies and could not be resuscitated at term. Only one fetus with associated anomalies (D), mandibular hypoplasia and omphalocele, survived the neonatal period (see below). Four fetuses demonstrated improvement or stabilization of the diagnostic features of CHAOS on serial imaging during gestation (A-D) and are described in more detail. Fetus A was referred for echogenic lungs at 21 weeks of gestation and was found at 22 weeks to have polyhydram-
Complete subglottic stenosis Laryngeal web
Yes
Delivery
Scheduled EXIT at 37 wk Yes Scheduled EXIT at 37 wk No Emergent EXIT at 34 wk Yes Emergent EXIT at 29 wk No Scheduled EXIT at 31 wk No Failed emergent EXIT at 31 wk, autopsy Failed No, laryngeal resuscitation atresia with tracheoesophageal at birth, autopsy fistula N/A Demise N/A Stillborn N/A Termination, autopsy N/A Termination No Termination, autopsy
N1 mo of Survival ventilatory to N1 y support No
Yes
No
Yes
No
N/A
Yes
Yes
Yes
Yes
N/A
No
N/A
No
N/A N/A N/A
No No No
N/A N/A
No No
nios, enlarged lungs, a fluid-filled dilated airway, compression of the heart with slight midline shift, and flattened diaphragms but no ascites or hydrops. For the next 14 weeks, the amniotic fluid volume normalized, compression of the heart disappeared, and the diaphragms took a more normal contour. At 37 2/7 weeks, the fetus was delivered by a scheduled EXIT procedure, and tracheostomy was performed. Direct laryngoscopy identified fused vocal cords with a tiny slitlike opening in the glottis leading to nearcomplete subglottic stenosis. The patient remained on the ventilator with minimal support for a few weeks, followed by continuous positive airway pressure (CPAP) for 2 months. She was discharged from the hospital with a tracheostomy on room air by 10 weeks of age. At 13 months of age, direct laryngoscopy/bronchoscopy revealed grade IV subglottic stenosis due to a complete subglottic web, congenital midline fusion of the vocal cords with almost no motion, and a tiny slitlike opening in the glottis. The decision was made to wait until an older age for airway reconstruction. Preoperatively, a gastrostomy tube was placed to decrease the aspiration risk after reconstruction.
1636 At 2.5 years of age, a double-stage laryngotracheoplasty was performed with anterior and posterior costal cartilage grafts over a Montgomery t-tube. Two months after reconstruction, the t-tube was removed; the airway was balloon dilated, and the tracheostomy tube was replaced. She was transitioned back to oral feeds slowly. As of last follow-up, she has required airway dilatation with an angioplasty balloon 4 times because of restenosis. She currently tolerates a regular diet and uses the gastrostomy tube for medications only. Because of congenital vocal cord fusion, the patient is nonverbal and uses sign language to communicate. Fetus B was referred for CHAOS at 21 weeks of gestation and had large, echogenic lungs, a fluid-filled trachea, everted diaphragms, and gross ascites, with no hydrops (Fig. 1). By 34 weeks of gestation, the ascites resolved, lung volumes remained stable, the diaphragms were no longer everted but remained flattened, and the trachea remained dilated and filled with fluid with a region of tracheal narrowing. The fetus was delivered and underwent tracheostomy by EXIT procedure at 37 2/7 weeks of gestation. Direct laryngoscopy revealed a severe laryngeal stenosis with a small posterior opening in the larynx and tracheomalacia at the level of the carina. The patient required ventilatory support for 2 days, followed by less than 2 weeks of CPAP. She was discharged on room air 20 days after EXIT. At home, she tolerated oral feeds. At 2 years of age, direct laryngoscopy/bronchoscopy revealed mobile vocal cords and a thick core of cartilage occupying almost the entire cricoid lumen with a pinhole-
J.L. Roybal et al. sized posterior airway, so she underwent partial cricotracheal resection with primary anastomosis and suprastomal stent placement. Postoperatively, the patient was fed via nasogastric tube but aspirated oral feeds, so a gastrostomy tube was placed to decrease her aspiration risk. The airway stent was removed 3 weeks after reconstruction, and the patient was slowly transitioned back to oral feeds. Five months after reconstruction, bronchoscopy revealed a large suprastomal polyp that was endoscopically excised with tracheostomy revision. At last follow-up, the patient was almost 3 years of age, tracheostomy dependent, and remained nonverbal. Fetus C was diagnosed with CHAOS at 22 weeks with findings of hyperechoic lungs, a dilated trachea, flattened diaphragms, mild to moderate ascites, and mild displacement of the heart to the midline. The patient was first evaluated at our institution at 31 weeks of gestation and was confirmed to have polyhydramnios, enlarged echogenic lungs, flattened hemidiaphragms, and ascites. Repeat ultrasound 2 weeks later showed persistently enlarged lungs and flattened diaphragms but stable amniotic fluid volumes and resolution of ascites, suggesting possible decompression of upper airway obstruction given the rapid resolution of ascites. Because of premature rupture of membranes, the fetus was delivered by emergency EXIT at 34 weeks of gestation and underwent tracheostomy. Direct laryngoscopy revealed complete atresia of the vocal cords with no definite laryngeal opening visualized. Chest xray revealed demonstrated hyperinflated lungs with flat diaphragms. The patient remained on a mechanical ventilator
Fig. 1 Ultrasound images of fetus B at 22 weeks of gestation (A-C) and 33 weeks of gestation (D-E). (A and B) The 22-week sagittal scan demonstrates enlarged echogenic lungs (cursors), eversion of the diaphragm (d), the liver (l) surrounded by gross ascites (a), and a dilated fluidfilled trachea (arrowheads) measuring 2.3 mm in AP diameter. (C) Transverse scan shows enlarged echogenic lungs expanding the thorax (cursors) and the dilated fluid-filled trachea (arrowhead). (D) Follow-up scan at 32 weeks of gestation demonstrates complete resolution of ascites, stable volumes of enlarged lungs (cursors), and flattening of the previously everted diaphragm (d). (E) Sagittal scan of the thorax demonstrates a persistently dilated fluid-filled trachea (arrow).
Predicting the severity of congenital high airway obstruction syndrome for 5 days, followed by CPAP for 9 days, and then room air. He was transitioned slowly to oral feeds and was discharged from the hospital at 9 weeks of age. At 6 months of age, laryngoscopy revealed no vocal fold motion, a complete web and atresia involving the larynx, and fusion of the arytenoids. As of 12 months of age, the patient remains tracheostomy dependent with airway reconstruction tentatively scheduled for 2 years of age. He is on full oral feeds. Interestingly, the presence of a fistula did not always correlate with a favorable postnatal course. Fetus D was referred after diagnosis of CHAOS at 19 weeks with echogenic lungs, bilateral pleural effusions, ascites, nuchal edema, and hydrops. Our evaluation at 24 weeks showed markedly enlarged lungs, a fluid-filled trachea extending into the neck, everted diaphragms, ascites, a small amount of nuchal edema, and compression of the heart but no hydrops. Associated anomalies of mandibular hypoplasia and an omphalocele were also noted. Normal male karyotype was confirmed. Approximately 6 weeks later, repeat ultrasound and MRI revealed no worsening but no significant improvement in any previous findings. An emergent EXIT with tracheostomy was performed at 29 4/7 weeks for premature rupture of membranes. Initial laryngoscopy at 1 week of age revealed an atretic larynx with a posterior pinhole. Although he did survive the neonatal period, the patient had significant respiratory compromise requiring continuous ventilatory support for 7 months and intermittent support after that. He died at home at 13 months of age as a result of a tracheostomy complication. Of the 5 fetuses that survived the neonatal period, the one that showed no signs of improvement on prenatal imaging, fetus E, had tracheal atresia but no evidence of an airway fistula. By comparison to the other surviving fetuses, his postnatal course was long and burdened with many complications that were previously reported [3]. At last publication, he was 32 months of age and tracheostomy dependent but awaiting decannulation. He was briefly decannulated but subsequently developed left lower lobe bronchiectasis necessitating replacement of the tracheostomy, a period of ventilatory support, and long-term antibiotics. At last follow-up, he was 10 years old, still tracheostomy dependent, tolerating soft oral feeds, and using a Passy-Muir valve for speech.
3. Discussion The true incidence of CHAOS is unknown; many of the affected fetuses die in utero or are stillborn. From 1826 to 1994, only 48 cases of laryngeal atresia and CHAOS were reported and only one fetus survived the immediate postnatal period [1]. Survival is dependent on early prenatal diagnosis, close prenatal follow-up, and fetal or immediate postnatal intervention. Understanding the natural history of the disease is essential. Although a few fetuses with CHAOS have been
1637
saved by emergency tracheostomy immediately after delivery, use of the EXIT procedure has most dramatically improved survival with CHAOS [3,5-14]. The EXIT approach preserves the uteroplacental circulation to the fetus while an airway is established. Excluding cases of Fraser syndrome, since 1994, 35 cases of laryngeal atresia and/or CHAOS have been reported in the English language literature [1-23]. Thirteen of the reported cases were delivered by the EXIT procedure, and 4 additional cases were salvaged by early tracheostomy after non-EXIT delivery. Our institution has performed 6 EXIT procedures for CHAOS with tracheostomy at the time of EXIT. After tracheostomy and postnatal recovery, tracheal reconstruction is generally planned as an elective procedure at a later age. Serial fetal imaging with ultrasound and MRI has been the mainstay of our prenatal management of CHAOS. Sonographic findings, which can be present as early as 16 weeks of gestation, include dilated airways distal to the obstruction, large echogenic lungs, flattened or everted diaphragms, fetal ascites, and hydrops [1]. Congenital high airway obstruction syndrome is most often misdiagnosed as bilateral congenital cystic adenomatoid malformation [2]. Although prenatal imaging can identify the location of upper airway obstruction, the cause of the obstruction is often unclear. The most common underlying cause of upper airway obstruction is laryngeal atresia. Other structural etiologies include subglottic stenosis or atresia, laryngeal or tracheal webs, laryngeal or tracheal agenesis, and rarely, an occluding laryngeal cyst. Occasionally, CHAOS occurs as part of Fraser syndrome, an autosomal recessive disorder characterized by laryngeal atresia, cryptophthalmos, syndactyly, and urogenital defects [24]. This series emphasizes the spectrum of severity of CHAOS. Obstruction of the fetal airway sets in motion a sequence of events that can ultimately lead to fetal demise. The sequence, however, is occasionally interrupted. Early presentation with hydrops or associated anomalies confers a poor prognosis, with a high rate of fetal demise and poor survival even with the EXIT procedure. However, the prognosis for survival of fetuses that present with isolated airway obstruction without hydrops is relatively favorable. In this setting, the observation of prenatal resolution of the features of CHAOS portends a relatively favorable postnatal course, often with a fairly brief period of postnatal ventilator support. In the absence of hydrops and other anomalies, fetuses with ultrasound findings that do not improve but do not progress are likely to survive but require extensive and prolonged postnatal support. The presence of hydrops remains an ominous sign, but under appropriate circumstances, early EXIT with tracheostomy may salvage selected patients. Two hydropic fetuses that we delivered with early EXIT (one emergently) were both found to have complete tracheal atresia. Fetus F had tracheal agenesis with intrathoracic obstruction that is an unsalvageable condition. Fetus E had subglottic obstruction that could be treated with tracheostomy. Early EXIT at 31
1638 weeks of gestation with tracheostomy rescued the fetus. The infant had a prolonged course in the intensive care unit with multiple complications but survived. In our series of patients, this was the only survivor with fetal hydrops. He is now the oldest survivor of our series of CHAOS patients, but his prolonged postoperative course, recurrent airway infections, and 10-year tracheostomy requirement demonstrate the potential long-term consequences of severe CHAOS. In such cases, extensive and accurate nondirective counseling is mandatory because of the high likelihood of death or quality of life impacting morbidity. At the other end of the spectrum, there is clearly a subtype of CHAOS patients that have resolution of prenatal ultrasound findings and a mild clinical course. Although the presence of a pinhole laryngotracheal fistula explained 2 cases, we were unable to identify a fistula in a third fetus that showed improvement in prenatal imaging. It seems intuitive that in order for CHAOS to resolve, decompression of the obstructed fetal airway must occur. We speculate that the failure to identify a fistula in such patients most likely represents the limitations of postnatal laryngoscopy in visualizing small openings in the often distorted anatomy. The presence of other anomalies generally predicts more severe CHAOS [2]. We observed this in 4 fetuses. As previously described, fetus D had several congenital anomalies, and despite the presence of a pinhole laryngotracheal fistula and some reversal of the prenatal features of CHAOS, several months of mechanical ventilation was required. Because fetus G had several anomalies in the presence of hydrops and severe oligohydramnios, the diagnosis was unclear and EXIT was not offered. Autopsy confirmed severe CHAOS. Other than fetus D, no fetuses that survived had any other congenital anomalies. In every patient with CHAOS, preoperative identification of the level of obstruction by MRI or ultrasound is extremely important but sometimes imperfect. Fetus F failed the EXIT procedure because prenatal imaging falsely identified the point of obstruction in the midneck. Retrospective review of the MRI identified agenesis of the trachea extending into the thorax. Had the anatomy been correctly identified prenatally, an EXIT procedure would not have been attempted. It is interesting to speculate regarding the relevance of these findings to experimentally or clinically induced prenatal tracheal occlusion for treatment of congenital diaphragmatic hernia. Just as CHAOS has been referred to as an “experiment of nature” replicating prenatal tracheal occlusion to induce lung growth, CHAOS with spontaneous resolution could be considered analogous to prenatal tracheal occlusion and release. There has been speculation regarding the induction of polyalveolar or nonfunctional lungs by prenatal tracheal occlusion [25,26]. If data from our CHAOS patients can be extrapolated to induced tracheal occlusion, our data would suggest that there is minimal if any adverse effect on long-term lung function caused by prenatal tracheal occlusion and release, as our patients with prenatal spontaneous reversal of CHAOS in the absence of associated
J.L. Roybal et al. anomalies had minimal lung disease despite seemingly prolonged periods of tracheal occlusion. For those fetuses that show no improvement on serial imaging, there may be a role for the creation of a laryngotracheal fistula or tracheostomy prenatally. Two groups have described fetoscopic decompression of the trachea using balloon dilatation, laser decompression [27,28], or tracheostomy [29]. It must be noted that the favorable neonatal outcome described for fetuses with an observed or presumed fistula is relative. In our experience, the infants who were quickly weaned from the ventilator still faced a number of hurdles: tracheostomy dependence, the need for complicated, staged airway reconstruction, a delay in oral feeding, and speech limitations. All of our surviving patients were tracheostomy dependent at last follow-up. Permanent decannulation must be postponed until airway reconstruction has been completed. The reconstruction process is generally deferred until the child is at least 2 years of age and then usually requires a staged procedure [30]. Postoperative complications including airway restenosis requiring dilatation and the development of airway polyps or granulation tissue postpone potential decannulation. Oral feeding in these patients is possible but is often delayed. In the newborn period and in the period after airway reconstruction, all of our survivors were initially fed enterally to limit the risk of aspiration with oral feeds. Over time, in consultation with a speech therapist, they have all been transitioned to oral intake. In 2 cases, a gastrostomy tube was placed before airway reconstruction, but in time, both patients returned to oral feeds. Speech impairment is universal. Reconstructive surgery can restore the anatomy of the airway, but in the absence of vocal cords or vocal cord mobility, speech cannot be restored. Patients will be nonverbal and must be taught sign language. For those patients with subglottic stenosis or tracheal atresia, recovery of speech may be a possibility with capping of the tracheostomy, but speech therapy is essential. Therefore, even in the best case scenarios, CHAOS patients will have definite functional limitations affecting their quality of life. These issues should be addressed during prenatal counseling. Although the natural history of CHAOS is variable, patterns of disease exist. Presence of a pinhole laryngotracheal fistula in some fetuses may allow airway decompression with resolution of ascites and diaphragmatic eversion. In other fetuses, airway decompression can occur without a visible fistula. As long as optimal perinatal management is performed, trends in the ultrasound findings of CHAOS are highly predictive of postnatal outcome and are a valuable guide to prenatal counseling.
References [1] Hedrick MH, Ferro MM, Filly RA, et al. Congenital high airway obstruction syndrome (CHAOS): a potential for perinatal intervention. J Pediatr Surg 1994;29:271-4.
Predicting the severity of congenital high airway obstruction syndrome [2] Lim FY, Crombleholme TM, Hedrick HL, et al. Congenital high airway obstruction syndrome: natural history and management. J Pediatr Surg 2003;38:940-5. [3] Crumbleholme TM, Sylvester K, Flake AW, et al. Salvage of a fetus with congenital high airway obstruction syndrome by ex utero intrapartum treatment (EXIT) procedure. Fetal Diagn Ther 2000;15 (5):280-2. [4] Mong A, Johnson AM, Kramer SS, et al. Congenital high airway obstruction syndrome: MR/US findings, effect on management, and outcome. Pediatr Radiol 2008;38(11):1171-9. [5] Vidaeff AC, Szmuk P, Mastrobattista JM, et al. More or less CHAOS: case report and literature review suggesting the existence of a distinct subtype of congenital high airway obstruction syndrome. Ultrasound Obstet Gynecol 2007;30(1):114-7. [6] DeCou JM, Jones DC, Jacobs HD, et al. Successful ex utero intrapartum treatment (EXIT) procedure for congenital high airway obstruction syndrome (CHAOS) owing to laryngeal atresia. J Pediatr Surg 1998;33:1563-5. [7] Bui TH, Grunewald C, Frenckner B, et al. Successful EXIT (ex utero intrapartum treatment) procedure in a fetus diagnosed prenatally with congenital high-airway obstruction syndrome due to laryngeal atresia. Eur J Pediatr Surg 2000;10:328-33. [8] Paek BW, Callen PW, Kitterman J, et al. Successful fetal intervention for congenital high airway obstruction syndrome. Fetal Diagn Ther 2002;17:272-6. [9] Kanamori Y, Kitano Y, Hashizume K, et al. A case of laryngeal atresia (congenital high airway obstruction syndrome) with chromosome 5p deletion syndrome rescued by ex utero intrapartum treatment. J Pediatr Surg 2004;39:25-8. [10] Vanhaesebrouck P, De Coen K, Defoort P, et al. Evidence for autosomal dominant inheritance in prenatally diagnosed CHAOS. Eur J Pediatr 2006;165(10):706-8. [11] Glynn F, Sheahan P, Hughes J, et al. Successful ex utero intrapartum treatment (EXIT) procedure for congenital high airway obstruction syndrome (CHAOS) owing to a large oropharyngeal teratoma. Ir Med J 2006;99(8):242-3. [12] Shimabukuro F, Sakumoto K, Masamoto H, et al. A case of congenital high airway obstruction syndrome managed by ex utero intrapartum treatment: case report and review of the literature. Am J Perinatol 2007;24(3):197-201. [13] Colnaghi M, Condo V, Gagliardi L, et al. Prenatal diagnosis and postnatal management of congenital laryngeal atresia in a preterm infant. Ultrasound Obstet Gynecol 2007;29(5):583-5. [14] Kuwashima S, Kitajima K, Kaji Y, et al. MR imaging appearance of laryngeal atresia (congenital high airway obstruction syndrome): unique course in a fetus. Pediatr Radiol 2008;38(3):344-7. [15] Tang PTM, Meagher SE, Khan AA, et al. Laryngeal atresia: antenatal diagnosis in a twin pregnancy. Ultrasound Obstet Gynecol 1996;7: 371-3.
1639
[16] Kalache KD, Chaoui R, Tennstedt C, et al. Prenatal diagnosis of laryngeal atresia in two cases of congenital high airway obstruction syndrome (CHAOS). Prenat Diagn 1997;17:577-81. [17] Morrison PJ, Macphail S, Williams D, et al. Laryngeal atresia or stenosis presenting as second-trimester fetal ascites—diagnosis and pathology in three independent cases. Prenat Diagn 1998;18:963-7. [18] Meizner I, Sherizly I, Mashiach R, et al. Prenatal sonographic diagnosis of laryngeal atresia is association with single umbilical artery. J Clin Ultrasound 2000;28:435-8. [19] Witters I, Moerman PH, Fryns JP. Prenatal echographic diagnosis of laryngeal atresia as part of a multiple congenital anomalies (MCA) syndrome. Genetic Counseling 2000;11:215-9. [20] Önderoglu L, Saygan Karamürsel B, Bulun A, et al. Prenatal diagnosis of laryngeal atresia. Prenat Diagn 2003;23:277-80. [21] Minior VK, Gagner JP, Landi K, et al. Congenital laryngeal atresia associated with partial diaphragmatic obliteration. J Ultrasound Med 2004;23:291-6. [22] Zhang P, Herring D, Cook L, et al. Fetal laryngeal stenosis/atresia and congenital high airway obstructive syndrome (CHAOS): a case report. J Perinatol 2005;25:426-8. [23] Erickson V, Messner A. Radiology quiz case 2. Congenital high airway obstruction syndrome (CHAOS). Arch Otolaryngol Head Neck Surg. 2007 Mar;133(3):299, 300-1. [24] Berg C, Geipel A, Germer U, et al. Prenatal detection of Fraser syndrome without cryptophthalmos: case report and review of the literature. Ultrasound Obstet Gynecol 2001;18:76-80. [25] DiFiore JW, Fauza DO, Slavin R, et al. Experimental fetal tracheal ligation reverses the structural and physiological effects of pulmonary hypoplasia in congenital diaphragmatic hernia. J Pediatr Surg 1994;29 (2):248-56 [discussion 256-7]. [26] Davey MG, Hedrick HL, Bouchard S, et al. Temporary tracheal occlusion in fetal sheep with lung hypoplasia does not improve postnatal lung function. J Appl Physiol 2003;94(3):1054-62. [27] Kohl T, Hering R, Bauriedel G, et al. Fetoscopic and ultrasoundguided decompression of the fetal trachea in a human fetus with Fraser syndrome and congenital high airway obstruction syndrome (CHAOS) from laryngeal atresia. Ultrasound Obstet Gynecol 2006;27(1):84-8 [discussion 88]. [28] Kohl T, Van de Vondel P, et al. Percutaneous fetoscopic laser decompression of congenital high airway obstruction syndrome (CHAOS) from laryngeal atresia via a single trocar—current technical constraints and potential solutions for future interventions. Fetal Diagn Ther 2009;25(1):67-71. [29] Fowler SF, Sydorak RM, Albanese CT, et al. Fetal endoscopic surgery: lessons learned and trends reviewed. J Pediatr Surg 2002;37(12): 1700-2. [30] Hartnick CJ, Rutter M, Lang F, et al. Congenital high airway obstruction syndrome and airway reconstruction: an evolving paradigm. Arch Otolaryngol Head Neck Surg 2002;128(5):567-70.
www.elsevier.com/locate/jpedsurg
Predicting the severity of congenital high airway obstruction syndrome Jessica L. Roybal 1 , Kenneth W. Liechty, Holly L. Hedrick, Michael W. Bebbington, Mark P. Johnson, Beverly G. Coleman, N. Scott Adzick, Alan W. Flake ⁎ Center for Fetal Diagnosis and Treatment, Children's Hospital of Philadelphia, PA 19104-4318, USA Received 4 November 2009; revised 17 January 2010; accepted 18 January 2010
Key words: CHAOS; Hydrops fetalis; EXIT; Ex utero intrapartum therapy
Abstract Background: Congenital high airway obstruction syndrome (CHAOS) is caused by complete or nearcomplete obstruction of the fetal airway. Obstruction sets in motion a sequence of events that can ultimately lead to fetal demise. However, on rare occasions in utero airway decompression occurs, reversing syndromic findings and improving the prognosis. In our relatively large series of CHAOS patients, we have observed a spectrum of clinical severity. The aim of this study was to identify the prenatal characteristics of CHAOS predictive of a milder postnatal course. Methods: The medical charts of all fetuses observed at our institution with the diagnosis of CHAOS were reviewed for radiologic findings, delivery information, perinatal course, autopsy or discharge report, and long-term follow-up. Results: Between 1996 and 2008, 12 fetuses with CHAOS were identified. Four fetuses had no evidence of hydrops on initial imaging. Of the 8 fetuses displaying hydrops, 3 were terminated, 2 died in utero, and 1 with multiple anomalies died at birth. Six fetuses were delivered via the ex utero intrapartum therapy procedure for attempted salvage, and 5 of the 6 survived the neonatal period including all 4 fetuses without hydrops. Serial prenatal imaging demonstrated less severe signs of CHAOS in 3 fetuses, and in 2 of them, direct laryngoscopy revealed a tiny opening in the airway. All 3 fetuses that showed improvement on serial imaging survived the neonatal period and were discharged home by 2-10 weeks of age. Conclusions: Although the natural history of CHAOS is variable, trends in prenatal ultrasound findings are highly predictive of postnatal outcome and are a valuable guide to prenatal counseling. © 2010 Elsevier Inc. All rights reserved.
Congenital high airway obstruction syndrome (CHAOS) is caused by complete or near-complete obstruction of the fetal airway. The term was first introduced by Hedrick and ⁎ Corresponding author. Department of Surgery, Philadelphia, PA 19104-4318, USA. Tel.: +1 215 590 3671; fax: +1 215 590 3324. E-mail addresses: [email protected] (J.L. Roybal), [email protected] (A.W. Flake). 1 Department of Surgery, Abramson Research Bldg, Rm 1114, 3615 Civic Center Blvd, Philadelphia, PA 19104-4318, USA. 0022-3468/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.jpedsurg.2010.01.022
colleagues [1] in 1994 to describe 4 fetuses with upper airway obstruction and a constellation of ultrasound findings considered incompatible with survival. Fifteen years later, CHAOS remains a rare diagnosis but is no longer uniformly fatal. Better prenatal and perinatal management based on a greater understanding of the natural history of the syndrome have led to improved outcomes. In our series of 12 patients, the largest reported to date, we have observed that CHAOS involves a spectrum of clinical severity.
1634 Congenital high airway obstruction syndrome is thought to occur when the upper airway fails to recanalize around the 10th week of gestation thereby causing obstruction [1]. The fetal lung normally secretes fluid that is cleared through the larynx. Obstruction of the upper airway blocks the flow of fetal lung fluid and raises the intratracheal pressure, which in turn, leads to distension of the tracheobronchial tree and secondary proliferative lung growth [2]. As the lungs expand, the diaphragms flatten and may even evert, and the heart becomes compressed in the midline. Elevated intrathoracic pressure causes decreased venous return and fetal cardiac failure leading to ascites, placentomegaly, and eventually hydrops fetalis. A subtype of CHAOS with a less severe natural history has recently been described [2,3]. A total of 6 patients, reported by 5 different groups, have had a minor pharyngotracheal or laryngotracheal fistula that serves as a pop-off valve for partial decompression of the respiratory tract leading to resolution of sonographic findings and a milder postnatal course. However, most CHAOS cases lack a decompressive tract, and the complete airway occlusion is maintained throughout gestation. The purpose of this study was to determine the spectrum of disease among a relatively large series of fetuses diagnosed with CHAOS and observed at our institution. Given the recent reports of a less severe subtype, we hoped to identify prenatal characteristics predictive of a milder postnatal course.
1. Materials and methods A retrospective search of the fetal database of the Center for Fetal Diagnosis and Treatment of the Children's Hospital of Philadelphia (Pa) was performed for all patients with a prenatal diagnosis of CHAOS after approval was obtained from the institutional review board (IRB 08-006152). Twelve patients were identified. Charts were reviewed for initial diagnosis, prenatal radiologic findings, operative reports, perinatal inpatient course, and autopsy or discharge reports. Serial prenatal ultrasound reports were compared, and resolution of ultrasound findings was defined as an improvement in one or more of the sonographic features of CHAOS including resolution of ascites, normalization of the amniotic fluid volume, restoration of a normal diaphragm contour, or return of the heart to the left chest. Long-term follow-up of survivors through July 2008 was also reviewed.
2. Results Between January 1996 and July 2008, 12 patients were diagnosed with CHAOS. Ten of these fetuses were previously reported by our institution [2-4]. Three fetuses were female and 9 were male. All of the fetuses were referred from outside institutions based on ultrasound findings. Most
J.L. Roybal et al. fetuses were referred with a diagnosis of bilateral congenital cystic adenomatoid malformation or echogenic lungs, but 4 (Table 1; C, F, H, I) had a diagnosis of CHAOS made at their referring institution. Gestational age at initial diagnosis of CHAOS (either at our institution or the referring institution) ranged from 19 to 29 weeks. Evaluation included ultrasound, magnetic resonance imaging (MRI), and echocardiogram. All fetuses on presentation had large, echogenic lungs, 10 had ascites, 9 had signs of airway obstruction (dilated airways), 8 had flattened or everted diaphragms, and 8 had hydrops (Table 1). Based on MRI, obstruction was identified at the level of the midneck in most fetuses. Three fetuses (J, K, L) were terminated after counseling for the diagnosis of CHAOS with associated hydrops. Two fetuses (H, I) died in utero at 21 and 22 weeks of gestation, and both were hydropic. One fetus (G) had multiple other congenital anomalies and was not offered the ex utero intrapartum therapy (EXIT) procedure but carried to term and failed resuscitation at delivery. The remaining 6 fetuses underwent elective or emergent EXIT procedures for attempted salvage. Of the 6 fetuses (A-F) delivered by the EXIT procedure at our institution, 5 survived the neonatal period. One fetus with hydrops (F) was delivered by emergent EXIT for maternal mirror syndrome at 32 weeks of gestation. The fetus was thought to have mid-neck obstruction based on MRI but died 17 minutes after delivery because of nearly complete tracheal agenesis with consequent inability to obtain an airway. At autopsy, the infant was found to have complete agenesis of the trachea from just below the larynx to approximately 1.5 cm above the carina. Retrospective review of the MRI confirmed complete tracheal agenesis extending into the thoracic cavity. Among the successful EXIT procedures, 3 were scheduled and 2 were performed emergently for preterm labor at 34 and 29 weeks (C, D). No fetuses survived the neonatal period without the EXIT procedure. The diagnosis of hydrops in association with CHAOS was a poor prognostic sign. Hydrops is defined by fluid accumulation in a body cavity in addition to the peritoneum (ie, pleural or pericardial fluid) and edema of the skin and/or scalp. Of the 8 fetuses in our series with hydrops at presentation, 3 underwent termination. Of the remaining 5 fetuses, 2 died in utero, 1 had multiple anomalies and failed resuscitation at term, and 1 had tracheal agenesis and died during an unsuccessful EXIT procedure. Only one fetus with coexisting hydrops survived after an EXIT procedure at 31 weeks (E). In contrast, the absence of hydrops in 4 fetuses (A-D) was a favorable prognostic sign, with all 4 surviving the neonatal period after delivery by the EXIT procedure. The presence of isolated ascites without hydrops in 3 of these fetuses (B-D) did not preclude survival. In 3 fetuses (H, I, K), the etiology of ultrasound findings consistent with CHAOS could not be determined because of fetal demise or termination without consent for autopsy. In 9 fetuses, the cause of CHAOS was determined by laryngoscopy, autopsy, or both. The most frequent etiology was
Predicting the severity of congenital high airway obstruction syndrome Table 1
1635
CHAOS patients (grouped according to the presence or absence of hydrops on initial ultrasound)
Other Resolution of Cause of Pinhole Fetus Gestational Hydrops Ascites obstruction fistula on on initial on initial anomalies ultrasound age at laryngoscopy findings diagnosis ultrasound ultrasound A
22
No
No
No
Yes
B
22
No
Yes
No
Yes
C
22
No
Yes
No
Yes
Laryngeal atresia
D
19
No
Yes
Yes
No, but stable
Laryngeal atresia
E
20
Yes
Yes
No
No, worse
Tracheal atresia
F
20
Yes
Yes
No
No, worse
Tracheal agenesis
G
23
Yes
Yes
Multiple
No, worse
Laryngeal atresia
H I J
22 21 20
Yes Yes Yes
Yes Yes Yes
No Yes No
N/A N/A N/A
K L
20 19
Yes Yes
Yes No
No Yes
N/A N/A
Unknown Unknown Laryngeal atresia Unknown Laryngeal cyst
laryngeal atresia found in 4 fetuses (C, D, G, J). Other diagnoses found in one fetus each included tracheal atresia (E), tracheal agenesis (F), a laryngeal web (B), an obstructing laryngeal cyst (L), and complete subglottic stenosis (A). A laryngotracheal fistula with a pinhole visualized in an atretic larynx could only be confirmed in 3 fetuses (A, B, D). Associated anomalies were present in 4 fetuses (D, G, I, L). One fetus (I), found to have left renal agenesis but no other findings of Fraser syndrome, had an early in utero demise at 22 weeks of gestation. A second fetus (L), with ultrasound findings suggestive of caudal regression syndrome, was terminated and autopsy was declined. A third fetus (G) had musculoskeletal, genitourinary, and gastrointestinal anomalies and could not be resuscitated at term. Only one fetus with associated anomalies (D), mandibular hypoplasia and omphalocele, survived the neonatal period (see below). Four fetuses demonstrated improvement or stabilization of the diagnostic features of CHAOS on serial imaging during gestation (A-D) and are described in more detail. Fetus A was referred for echogenic lungs at 21 weeks of gestation and was found at 22 weeks to have polyhydram-
Complete subglottic stenosis Laryngeal web
Yes
Delivery
Scheduled EXIT at 37 wk Yes Scheduled EXIT at 37 wk No Emergent EXIT at 34 wk Yes Emergent EXIT at 29 wk No Scheduled EXIT at 31 wk No Failed emergent EXIT at 31 wk, autopsy Failed No, laryngeal resuscitation atresia with tracheoesophageal at birth, autopsy fistula N/A Demise N/A Stillborn N/A Termination, autopsy N/A Termination No Termination, autopsy
N1 mo of Survival ventilatory to N1 y support No
Yes
No
Yes
No
N/A
Yes
Yes
Yes
Yes
N/A
No
N/A
No
N/A N/A N/A
No No No
N/A N/A
No No
nios, enlarged lungs, a fluid-filled dilated airway, compression of the heart with slight midline shift, and flattened diaphragms but no ascites or hydrops. For the next 14 weeks, the amniotic fluid volume normalized, compression of the heart disappeared, and the diaphragms took a more normal contour. At 37 2/7 weeks, the fetus was delivered by a scheduled EXIT procedure, and tracheostomy was performed. Direct laryngoscopy identified fused vocal cords with a tiny slitlike opening in the glottis leading to nearcomplete subglottic stenosis. The patient remained on the ventilator with minimal support for a few weeks, followed by continuous positive airway pressure (CPAP) for 2 months. She was discharged from the hospital with a tracheostomy on room air by 10 weeks of age. At 13 months of age, direct laryngoscopy/bronchoscopy revealed grade IV subglottic stenosis due to a complete subglottic web, congenital midline fusion of the vocal cords with almost no motion, and a tiny slitlike opening in the glottis. The decision was made to wait until an older age for airway reconstruction. Preoperatively, a gastrostomy tube was placed to decrease the aspiration risk after reconstruction.
1636 At 2.5 years of age, a double-stage laryngotracheoplasty was performed with anterior and posterior costal cartilage grafts over a Montgomery t-tube. Two months after reconstruction, the t-tube was removed; the airway was balloon dilated, and the tracheostomy tube was replaced. She was transitioned back to oral feeds slowly. As of last follow-up, she has required airway dilatation with an angioplasty balloon 4 times because of restenosis. She currently tolerates a regular diet and uses the gastrostomy tube for medications only. Because of congenital vocal cord fusion, the patient is nonverbal and uses sign language to communicate. Fetus B was referred for CHAOS at 21 weeks of gestation and had large, echogenic lungs, a fluid-filled trachea, everted diaphragms, and gross ascites, with no hydrops (Fig. 1). By 34 weeks of gestation, the ascites resolved, lung volumes remained stable, the diaphragms were no longer everted but remained flattened, and the trachea remained dilated and filled with fluid with a region of tracheal narrowing. The fetus was delivered and underwent tracheostomy by EXIT procedure at 37 2/7 weeks of gestation. Direct laryngoscopy revealed a severe laryngeal stenosis with a small posterior opening in the larynx and tracheomalacia at the level of the carina. The patient required ventilatory support for 2 days, followed by less than 2 weeks of CPAP. She was discharged on room air 20 days after EXIT. At home, she tolerated oral feeds. At 2 years of age, direct laryngoscopy/bronchoscopy revealed mobile vocal cords and a thick core of cartilage occupying almost the entire cricoid lumen with a pinhole-
J.L. Roybal et al. sized posterior airway, so she underwent partial cricotracheal resection with primary anastomosis and suprastomal stent placement. Postoperatively, the patient was fed via nasogastric tube but aspirated oral feeds, so a gastrostomy tube was placed to decrease her aspiration risk. The airway stent was removed 3 weeks after reconstruction, and the patient was slowly transitioned back to oral feeds. Five months after reconstruction, bronchoscopy revealed a large suprastomal polyp that was endoscopically excised with tracheostomy revision. At last follow-up, the patient was almost 3 years of age, tracheostomy dependent, and remained nonverbal. Fetus C was diagnosed with CHAOS at 22 weeks with findings of hyperechoic lungs, a dilated trachea, flattened diaphragms, mild to moderate ascites, and mild displacement of the heart to the midline. The patient was first evaluated at our institution at 31 weeks of gestation and was confirmed to have polyhydramnios, enlarged echogenic lungs, flattened hemidiaphragms, and ascites. Repeat ultrasound 2 weeks later showed persistently enlarged lungs and flattened diaphragms but stable amniotic fluid volumes and resolution of ascites, suggesting possible decompression of upper airway obstruction given the rapid resolution of ascites. Because of premature rupture of membranes, the fetus was delivered by emergency EXIT at 34 weeks of gestation and underwent tracheostomy. Direct laryngoscopy revealed complete atresia of the vocal cords with no definite laryngeal opening visualized. Chest xray revealed demonstrated hyperinflated lungs with flat diaphragms. The patient remained on a mechanical ventilator
Fig. 1 Ultrasound images of fetus B at 22 weeks of gestation (A-C) and 33 weeks of gestation (D-E). (A and B) The 22-week sagittal scan demonstrates enlarged echogenic lungs (cursors), eversion of the diaphragm (d), the liver (l) surrounded by gross ascites (a), and a dilated fluidfilled trachea (arrowheads) measuring 2.3 mm in AP diameter. (C) Transverse scan shows enlarged echogenic lungs expanding the thorax (cursors) and the dilated fluid-filled trachea (arrowhead). (D) Follow-up scan at 32 weeks of gestation demonstrates complete resolution of ascites, stable volumes of enlarged lungs (cursors), and flattening of the previously everted diaphragm (d). (E) Sagittal scan of the thorax demonstrates a persistently dilated fluid-filled trachea (arrow).
Predicting the severity of congenital high airway obstruction syndrome for 5 days, followed by CPAP for 9 days, and then room air. He was transitioned slowly to oral feeds and was discharged from the hospital at 9 weeks of age. At 6 months of age, laryngoscopy revealed no vocal fold motion, a complete web and atresia involving the larynx, and fusion of the arytenoids. As of 12 months of age, the patient remains tracheostomy dependent with airway reconstruction tentatively scheduled for 2 years of age. He is on full oral feeds. Interestingly, the presence of a fistula did not always correlate with a favorable postnatal course. Fetus D was referred after diagnosis of CHAOS at 19 weeks with echogenic lungs, bilateral pleural effusions, ascites, nuchal edema, and hydrops. Our evaluation at 24 weeks showed markedly enlarged lungs, a fluid-filled trachea extending into the neck, everted diaphragms, ascites, a small amount of nuchal edema, and compression of the heart but no hydrops. Associated anomalies of mandibular hypoplasia and an omphalocele were also noted. Normal male karyotype was confirmed. Approximately 6 weeks later, repeat ultrasound and MRI revealed no worsening but no significant improvement in any previous findings. An emergent EXIT with tracheostomy was performed at 29 4/7 weeks for premature rupture of membranes. Initial laryngoscopy at 1 week of age revealed an atretic larynx with a posterior pinhole. Although he did survive the neonatal period, the patient had significant respiratory compromise requiring continuous ventilatory support for 7 months and intermittent support after that. He died at home at 13 months of age as a result of a tracheostomy complication. Of the 5 fetuses that survived the neonatal period, the one that showed no signs of improvement on prenatal imaging, fetus E, had tracheal atresia but no evidence of an airway fistula. By comparison to the other surviving fetuses, his postnatal course was long and burdened with many complications that were previously reported [3]. At last publication, he was 32 months of age and tracheostomy dependent but awaiting decannulation. He was briefly decannulated but subsequently developed left lower lobe bronchiectasis necessitating replacement of the tracheostomy, a period of ventilatory support, and long-term antibiotics. At last follow-up, he was 10 years old, still tracheostomy dependent, tolerating soft oral feeds, and using a Passy-Muir valve for speech.
3. Discussion The true incidence of CHAOS is unknown; many of the affected fetuses die in utero or are stillborn. From 1826 to 1994, only 48 cases of laryngeal atresia and CHAOS were reported and only one fetus survived the immediate postnatal period [1]. Survival is dependent on early prenatal diagnosis, close prenatal follow-up, and fetal or immediate postnatal intervention. Understanding the natural history of the disease is essential. Although a few fetuses with CHAOS have been
1637
saved by emergency tracheostomy immediately after delivery, use of the EXIT procedure has most dramatically improved survival with CHAOS [3,5-14]. The EXIT approach preserves the uteroplacental circulation to the fetus while an airway is established. Excluding cases of Fraser syndrome, since 1994, 35 cases of laryngeal atresia and/or CHAOS have been reported in the English language literature [1-23]. Thirteen of the reported cases were delivered by the EXIT procedure, and 4 additional cases were salvaged by early tracheostomy after non-EXIT delivery. Our institution has performed 6 EXIT procedures for CHAOS with tracheostomy at the time of EXIT. After tracheostomy and postnatal recovery, tracheal reconstruction is generally planned as an elective procedure at a later age. Serial fetal imaging with ultrasound and MRI has been the mainstay of our prenatal management of CHAOS. Sonographic findings, which can be present as early as 16 weeks of gestation, include dilated airways distal to the obstruction, large echogenic lungs, flattened or everted diaphragms, fetal ascites, and hydrops [1]. Congenital high airway obstruction syndrome is most often misdiagnosed as bilateral congenital cystic adenomatoid malformation [2]. Although prenatal imaging can identify the location of upper airway obstruction, the cause of the obstruction is often unclear. The most common underlying cause of upper airway obstruction is laryngeal atresia. Other structural etiologies include subglottic stenosis or atresia, laryngeal or tracheal webs, laryngeal or tracheal agenesis, and rarely, an occluding laryngeal cyst. Occasionally, CHAOS occurs as part of Fraser syndrome, an autosomal recessive disorder characterized by laryngeal atresia, cryptophthalmos, syndactyly, and urogenital defects [24]. This series emphasizes the spectrum of severity of CHAOS. Obstruction of the fetal airway sets in motion a sequence of events that can ultimately lead to fetal demise. The sequence, however, is occasionally interrupted. Early presentation with hydrops or associated anomalies confers a poor prognosis, with a high rate of fetal demise and poor survival even with the EXIT procedure. However, the prognosis for survival of fetuses that present with isolated airway obstruction without hydrops is relatively favorable. In this setting, the observation of prenatal resolution of the features of CHAOS portends a relatively favorable postnatal course, often with a fairly brief period of postnatal ventilator support. In the absence of hydrops and other anomalies, fetuses with ultrasound findings that do not improve but do not progress are likely to survive but require extensive and prolonged postnatal support. The presence of hydrops remains an ominous sign, but under appropriate circumstances, early EXIT with tracheostomy may salvage selected patients. Two hydropic fetuses that we delivered with early EXIT (one emergently) were both found to have complete tracheal atresia. Fetus F had tracheal agenesis with intrathoracic obstruction that is an unsalvageable condition. Fetus E had subglottic obstruction that could be treated with tracheostomy. Early EXIT at 31
1638 weeks of gestation with tracheostomy rescued the fetus. The infant had a prolonged course in the intensive care unit with multiple complications but survived. In our series of patients, this was the only survivor with fetal hydrops. He is now the oldest survivor of our series of CHAOS patients, but his prolonged postoperative course, recurrent airway infections, and 10-year tracheostomy requirement demonstrate the potential long-term consequences of severe CHAOS. In such cases, extensive and accurate nondirective counseling is mandatory because of the high likelihood of death or quality of life impacting morbidity. At the other end of the spectrum, there is clearly a subtype of CHAOS patients that have resolution of prenatal ultrasound findings and a mild clinical course. Although the presence of a pinhole laryngotracheal fistula explained 2 cases, we were unable to identify a fistula in a third fetus that showed improvement in prenatal imaging. It seems intuitive that in order for CHAOS to resolve, decompression of the obstructed fetal airway must occur. We speculate that the failure to identify a fistula in such patients most likely represents the limitations of postnatal laryngoscopy in visualizing small openings in the often distorted anatomy. The presence of other anomalies generally predicts more severe CHAOS [2]. We observed this in 4 fetuses. As previously described, fetus D had several congenital anomalies, and despite the presence of a pinhole laryngotracheal fistula and some reversal of the prenatal features of CHAOS, several months of mechanical ventilation was required. Because fetus G had several anomalies in the presence of hydrops and severe oligohydramnios, the diagnosis was unclear and EXIT was not offered. Autopsy confirmed severe CHAOS. Other than fetus D, no fetuses that survived had any other congenital anomalies. In every patient with CHAOS, preoperative identification of the level of obstruction by MRI or ultrasound is extremely important but sometimes imperfect. Fetus F failed the EXIT procedure because prenatal imaging falsely identified the point of obstruction in the midneck. Retrospective review of the MRI identified agenesis of the trachea extending into the thorax. Had the anatomy been correctly identified prenatally, an EXIT procedure would not have been attempted. It is interesting to speculate regarding the relevance of these findings to experimentally or clinically induced prenatal tracheal occlusion for treatment of congenital diaphragmatic hernia. Just as CHAOS has been referred to as an “experiment of nature” replicating prenatal tracheal occlusion to induce lung growth, CHAOS with spontaneous resolution could be considered analogous to prenatal tracheal occlusion and release. There has been speculation regarding the induction of polyalveolar or nonfunctional lungs by prenatal tracheal occlusion [25,26]. If data from our CHAOS patients can be extrapolated to induced tracheal occlusion, our data would suggest that there is minimal if any adverse effect on long-term lung function caused by prenatal tracheal occlusion and release, as our patients with prenatal spontaneous reversal of CHAOS in the absence of associated
J.L. Roybal et al. anomalies had minimal lung disease despite seemingly prolonged periods of tracheal occlusion. For those fetuses that show no improvement on serial imaging, there may be a role for the creation of a laryngotracheal fistula or tracheostomy prenatally. Two groups have described fetoscopic decompression of the trachea using balloon dilatation, laser decompression [27,28], or tracheostomy [29]. It must be noted that the favorable neonatal outcome described for fetuses with an observed or presumed fistula is relative. In our experience, the infants who were quickly weaned from the ventilator still faced a number of hurdles: tracheostomy dependence, the need for complicated, staged airway reconstruction, a delay in oral feeding, and speech limitations. All of our surviving patients were tracheostomy dependent at last follow-up. Permanent decannulation must be postponed until airway reconstruction has been completed. The reconstruction process is generally deferred until the child is at least 2 years of age and then usually requires a staged procedure [30]. Postoperative complications including airway restenosis requiring dilatation and the development of airway polyps or granulation tissue postpone potential decannulation. Oral feeding in these patients is possible but is often delayed. In the newborn period and in the period after airway reconstruction, all of our survivors were initially fed enterally to limit the risk of aspiration with oral feeds. Over time, in consultation with a speech therapist, they have all been transitioned to oral intake. In 2 cases, a gastrostomy tube was placed before airway reconstruction, but in time, both patients returned to oral feeds. Speech impairment is universal. Reconstructive surgery can restore the anatomy of the airway, but in the absence of vocal cords or vocal cord mobility, speech cannot be restored. Patients will be nonverbal and must be taught sign language. For those patients with subglottic stenosis or tracheal atresia, recovery of speech may be a possibility with capping of the tracheostomy, but speech therapy is essential. Therefore, even in the best case scenarios, CHAOS patients will have definite functional limitations affecting their quality of life. These issues should be addressed during prenatal counseling. Although the natural history of CHAOS is variable, patterns of disease exist. Presence of a pinhole laryngotracheal fistula in some fetuses may allow airway decompression with resolution of ascites and diaphragmatic eversion. In other fetuses, airway decompression can occur without a visible fistula. As long as optimal perinatal management is performed, trends in the ultrasound findings of CHAOS are highly predictive of postnatal outcome and are a valuable guide to prenatal counseling.
References [1] Hedrick MH, Ferro MM, Filly RA, et al. Congenital high airway obstruction syndrome (CHAOS): a potential for perinatal intervention. J Pediatr Surg 1994;29:271-4.
Predicting the severity of congenital high airway obstruction syndrome [2] Lim FY, Crombleholme TM, Hedrick HL, et al. Congenital high airway obstruction syndrome: natural history and management. J Pediatr Surg 2003;38:940-5. [3] Crumbleholme TM, Sylvester K, Flake AW, et al. Salvage of a fetus with congenital high airway obstruction syndrome by ex utero intrapartum treatment (EXIT) procedure. Fetal Diagn Ther 2000;15 (5):280-2. [4] Mong A, Johnson AM, Kramer SS, et al. Congenital high airway obstruction syndrome: MR/US findings, effect on management, and outcome. Pediatr Radiol 2008;38(11):1171-9. [5] Vidaeff AC, Szmuk P, Mastrobattista JM, et al. More or less CHAOS: case report and literature review suggesting the existence of a distinct subtype of congenital high airway obstruction syndrome. Ultrasound Obstet Gynecol 2007;30(1):114-7. [6] DeCou JM, Jones DC, Jacobs HD, et al. Successful ex utero intrapartum treatment (EXIT) procedure for congenital high airway obstruction syndrome (CHAOS) owing to laryngeal atresia. J Pediatr Surg 1998;33:1563-5. [7] Bui TH, Grunewald C, Frenckner B, et al. Successful EXIT (ex utero intrapartum treatment) procedure in a fetus diagnosed prenatally with congenital high-airway obstruction syndrome due to laryngeal atresia. Eur J Pediatr Surg 2000;10:328-33. [8] Paek BW, Callen PW, Kitterman J, et al. Successful fetal intervention for congenital high airway obstruction syndrome. Fetal Diagn Ther 2002;17:272-6. [9] Kanamori Y, Kitano Y, Hashizume K, et al. A case of laryngeal atresia (congenital high airway obstruction syndrome) with chromosome 5p deletion syndrome rescued by ex utero intrapartum treatment. J Pediatr Surg 2004;39:25-8. [10] Vanhaesebrouck P, De Coen K, Defoort P, et al. Evidence for autosomal dominant inheritance in prenatally diagnosed CHAOS. Eur J Pediatr 2006;165(10):706-8. [11] Glynn F, Sheahan P, Hughes J, et al. Successful ex utero intrapartum treatment (EXIT) procedure for congenital high airway obstruction syndrome (CHAOS) owing to a large oropharyngeal teratoma. Ir Med J 2006;99(8):242-3. [12] Shimabukuro F, Sakumoto K, Masamoto H, et al. A case of congenital high airway obstruction syndrome managed by ex utero intrapartum treatment: case report and review of the literature. Am J Perinatol 2007;24(3):197-201. [13] Colnaghi M, Condo V, Gagliardi L, et al. Prenatal diagnosis and postnatal management of congenital laryngeal atresia in a preterm infant. Ultrasound Obstet Gynecol 2007;29(5):583-5. [14] Kuwashima S, Kitajima K, Kaji Y, et al. MR imaging appearance of laryngeal atresia (congenital high airway obstruction syndrome): unique course in a fetus. Pediatr Radiol 2008;38(3):344-7. [15] Tang PTM, Meagher SE, Khan AA, et al. Laryngeal atresia: antenatal diagnosis in a twin pregnancy. Ultrasound Obstet Gynecol 1996;7: 371-3.
1639
[16] Kalache KD, Chaoui R, Tennstedt C, et al. Prenatal diagnosis of laryngeal atresia in two cases of congenital high airway obstruction syndrome (CHAOS). Prenat Diagn 1997;17:577-81. [17] Morrison PJ, Macphail S, Williams D, et al. Laryngeal atresia or stenosis presenting as second-trimester fetal ascites—diagnosis and pathology in three independent cases. Prenat Diagn 1998;18:963-7. [18] Meizner I, Sherizly I, Mashiach R, et al. Prenatal sonographic diagnosis of laryngeal atresia is association with single umbilical artery. J Clin Ultrasound 2000;28:435-8. [19] Witters I, Moerman PH, Fryns JP. Prenatal echographic diagnosis of laryngeal atresia as part of a multiple congenital anomalies (MCA) syndrome. Genetic Counseling 2000;11:215-9. [20] Önderoglu L, Saygan Karamürsel B, Bulun A, et al. Prenatal diagnosis of laryngeal atresia. Prenat Diagn 2003;23:277-80. [21] Minior VK, Gagner JP, Landi K, et al. Congenital laryngeal atresia associated with partial diaphragmatic obliteration. J Ultrasound Med 2004;23:291-6. [22] Zhang P, Herring D, Cook L, et al. Fetal laryngeal stenosis/atresia and congenital high airway obstructive syndrome (CHAOS): a case report. J Perinatol 2005;25:426-8. [23] Erickson V, Messner A. Radiology quiz case 2. Congenital high airway obstruction syndrome (CHAOS). Arch Otolaryngol Head Neck Surg. 2007 Mar;133(3):299, 300-1. [24] Berg C, Geipel A, Germer U, et al. Prenatal detection of Fraser syndrome without cryptophthalmos: case report and review of the literature. Ultrasound Obstet Gynecol 2001;18:76-80. [25] DiFiore JW, Fauza DO, Slavin R, et al. Experimental fetal tracheal ligation reverses the structural and physiological effects of pulmonary hypoplasia in congenital diaphragmatic hernia. J Pediatr Surg 1994;29 (2):248-56 [discussion 256-7]. [26] Davey MG, Hedrick HL, Bouchard S, et al. Temporary tracheal occlusion in fetal sheep with lung hypoplasia does not improve postnatal lung function. J Appl Physiol 2003;94(3):1054-62. [27] Kohl T, Hering R, Bauriedel G, et al. Fetoscopic and ultrasoundguided decompression of the fetal trachea in a human fetus with Fraser syndrome and congenital high airway obstruction syndrome (CHAOS) from laryngeal atresia. Ultrasound Obstet Gynecol 2006;27(1):84-8 [discussion 88]. [28] Kohl T, Van de Vondel P, et al. Percutaneous fetoscopic laser decompression of congenital high airway obstruction syndrome (CHAOS) from laryngeal atresia via a single trocar—current technical constraints and potential solutions for future interventions. Fetal Diagn Ther 2009;25(1):67-71. [29] Fowler SF, Sydorak RM, Albanese CT, et al. Fetal endoscopic surgery: lessons learned and trends reviewed. J Pediatr Surg 2002;37(12): 1700-2. [30] Hartnick CJ, Rutter M, Lang F, et al. Congenital high airway obstruction syndrome and airway reconstruction: an evolving paradigm. Arch Otolaryngol Head Neck Surg 2002;128(5):567-70.