INTRODUCTION

TOP ABSTRACT INTRODUCTION CASE REPORTS DISCUSSION REFERENCES

Keywords: dysphagia; hiccups; continuous positive airway pressure; chlorpromazine; diaphragmatic flutter

Diaphragmatic flutter (DF) is a heterogeneous disorder that has occasionally been reported in the pediatric literature, including infants (1, 2), toddlers (3), and adolescents (4, 5). The clinical presentation, predominantly seen in adults, ranges from asymptomatic to palpitations, pain, and dyspnea (6). It is characterized by diaphragmatic contractions at a rate of 35 to 480/min (6). The reported respiratory patterns of DF are highly variable. Diaphragmatic contractions may be limited to inspiration (5), expiration (7), or occur throughout the respiratory cycle (2, 8). Involvement of additional respiratory muscles has also been observed (9). DF generally resolves during sleep (4, 5, 8), though it has rarely been reported to persist within all sleep stages (2). Gas exchange abnormalities are typically absent, but both hyperventilation (5, 8, 10) and hypoventilation (11, 12) have rarely been reported. The diaphragmatic contractions are generally asynchronous in relation to the heart beat (6) but may occasionally be synchronous (13). The duration of DF has varied from several days (2) to 18 yr (4). A range of therapies have been described including phrenic nerve crush (4, 5), diazepam (14), carbamazepine (15), chlorpromazine (11), and diphenylhydantoin (9).

Only four cases of DF in infants have been previously published (1, 2). Three of these were diagnosed retrospectively (2), and all resolved spontaneously without pharmacologic treatment. We report three neonates with respiratory flutter (RF) involving not only the diaphragm but also the intercostal muscles, and associated with dysphagia, laryngomalacia, and gastroesophageal reflux. We therefore prefer to use the more general term "respiratory flutter" rather than "diaphragmatic flutter." The infants presented with respiratory failure shortly after birth and had a favorable response to continuous positive airway pressure (CPAP). Chlorpromazine controlled the RF and was weaned off successfully by 4 to 7 mo of age. All infants have developed normally although the dysphagia has persisted.

	CASE REPORTS

TOP ABSTRACT INTRODUCTION CASE REPORTS DISCUSSION REFERENCES

Case 1

KM is an African-American male born at term without complications. In the delivery room, he had "irregular, noisy respirations." By 2 h of age, the infant developed grunting with a "fluttering" sound with each inspiration. His respiratory rate increased to 90/min as he developed progressively worsening stridor, flaring, and retractions. Examination of the thorax revealed jittery thoracic excursions associated with a regular underlying respiratory rhythm. Respiratory inductive plethysmography (RIP) (Non-invasive Monitoring Systems, Inc., Miami Beach, FL) revealed rapid fluctuations at a rate of 300/min during inspiration (Figure 1A). RF was confirmed by fluoroscopy. The RF was not synchronous with the heart rate. At times, "thoracic flutter" was observed, characterized by irregular, high-frequency thoracic tracings with regular abdominal excursions (Figure 1B). The RIP belts were positioned carefully and changed regularly. Fiberoptic laryngoscopy to evaluate stridor revealed normal vocal fold movement and laryngomalacia. Metabolic studies, a sepsis evaluation, and imaging studies of the chest, brain, and brainstem were normal.

View larger version (42K): [in this window] [in a new window]   Figure 1.   Case 1: 20-s epochs of RIP demonstrating RF evolving over a 5-min period. (A) DF limited to inspiration with paradoxical excursions in the thoracic tracings. (B) "Thoracic flutter" with regular abdominal excursions. (C) diaphragmatic "fibrillation" with marked increase in thoracic excursions. Heart rate 120 to 150/min, nonsynchronous with diaphragmatic contractions (not shown). Chest = rib cage excursion; Abd = abdominal excursion.

At 48 h, the infant was noted to have increasingly labored breathing. A capillary blood gas revealed a pH of 7.22, PCO₂ of 67 mm Hg, and PO₂ of 35 mm Hg. RIP tracings revealed a sudden shift to high-frequency, low-amplitude abdominal excursions with a marked increase in thoracic effort (Figure 1C). The child was placed on CPAP of 5 cm H₂O by means of nasal prongs with complete resolution of his RF. Attempts to wean the infant off CPAP during the next 2 d resulted in RF recurrence and respiratory failure. Though CPAP was effective in controlling the RF, it was difficult to maintain long-term and interfered with oral feeding. The child was started on chlorpromazine 1.5 mg/kg/d divided every 6 h enterally, which terminated the RF over several hours. A videofluoroscopic swallowing study demonstrated oropharyngeal dysphagia characterized by a delay in swallow onset, resulting in silent aspiration of thin liquids. The infant was managed successfully with thickened feeds for dysphagia and gastroesophageal reflux. Chlorpromazine was discontinued at 4 mo of age without RF recurrence. Now 26 mo of age, the child has developed normally although swallowing dysfunction persists as evidenced by coughing when drinking thin liquids.

Case 2

SK is a white male born at term after labor was induced for management of fetal bladder outlet obstruction. Hydronephrosis and hydroureter were confirmed by ultrasound and a catheter was placed in the bladder. The infant had mild respiratory distress immediately after delivery. On Day 2 of life, the infant was noted to develop fluttering inspiratory motions of his rib cage and abdomen. RIP monitoring confirmed the presence of RF (Figure 2). Bronchoscopy to evaluate stridor revealed fluttering of the epiglottis coincident with diaphragmatic contractions and laryngomalacia. Over the next 48 h, the infant developed worsening respiratory distress resulting in hypercarbia. CPAP improved his ventilation for 2 d but the infant subsequently required intubation. Chlorpromazine was started and the RF subsided over the next few hours permitting extubation. Occasional brief paroxysms of RF continued for several months that were tolerated without respiratory compromise. Attempts to reduce the dose of chlorpromazine during the first few months of life resulted in a resurgence of the RF with respiratory distress. Chlorpromazine was finally able to be discontinued at 7 mo of age without recurrence of the RF.

View larger version (24K): [in this window] [in a new window]   Figure 2.   Case 2: 15-s epoch of RIP demonstrating RF limited to inspiration in the abdominal channel unrelated to the heart rate. Abd = abdominal excursion; EKG = electrocardiogram.

The infant was subsequently diagnosed with pyloric stenosis and malrotation that were surgically repaired. Metabolic studies, a sepsis evaluation, and imaging of the brain and brainstem were normal. Marked dysphagia was evident clinically and severe gastroesophageal reflux was documented by a pH probe study. The infant was fed through a gastrostomy tube. Developmental milestones are normal at 8 mo of age, although oropharyngeal dysphagia has persisted.

Case 3

AS is an African-American female born at term without complications. At 4 h of age, the infant developed respiratory distress with feeding. On exam, the infant demonstrated ratchetlike inspiratory efforts that were evident on an impedance pneumographic monitor (Figure 3). RIP and fluoroscopy confirmed RF limited to inspiration. An arterial blood gas (ABG) test revealed a pH of 7.29, PCO₂ of 55 mm Hg, and PO₂ of 99 mm Hg on 30% supplemental oxygen. The infant was placed on CPAP, resulting in an improvement in oxygenation and a reduction in PCO₂ to 40 mm Hg. During the next 24 h, the infant demonstrated RF continuously and became more hypoxemic. Direct laryngoscopy to evaluate stridor revealed moderate laryngomalacia with edematous aryepiglottic cartilages. Metabolic studies, a sepsis evaluation, and imaging of the brain and brainstem were normal.

View larger version (23K): [in this window] [in a new window]   Figure 3.   Case 3: 15-s epoch of impedance pneumography during sleep demonstrating RF limited to inspiration at a rate of 300/min superimposed on an underlying respiratory rhythm of 40/min. Heart rate was 139/min, nonsynchronous with diaphragmatic contractions (not shown).

At 3 d of age, the infant deteriorated. An ABG showed a pH of 7.41, PCO₂ of 55 mm Hg, and PO₂ of 56 mm Hg on 65% supplemental oxygen. She was started on chlorpromazine that resolved the RF, and the infant was weaned to room air within 24 h. A pH probe documented gastroesophageal reflux during 45% of the recording time. A videofluoroscopic swallowing study demonstrated direct aspiration in the absence of a cough reflex. The infant was placed on antireflex medications and was fed through a gastrostomy tube. Chlorpromazine was discontinued at 8 mo of age without recurrence of the RF. Growth and developmental milestones were normal at 10 mo of age.

	DISCUSSION

TOP ABSTRACT INTRODUCTION CASE REPORTS DISCUSSION REFERENCES

We report three infants with RF that presented in the first several hours after delivery with respiratory failure. Dysphagia, laryngomalacia, and gastroesophageal reflux were subsequently evident in all infants. Patients with this association of findings have not been previously described. None of these neonates had evidence of sepsis, metabolic abnormalities, or central nervous system (CNS) lesions. The flutter was typically confined to the inspiratory phase of respiration. CPAP regularized the breathing pattern in one infant, and temporarily improved ventilation by reducing the amount of RF in the two others. Chlorpromazine effectively stopped the RF within 12 h in all three infants, though brief RF paroxysms persisted for several weeks to months. In addition to fluoroscopy, RIP was an accurate means of diagnosing RF. Growth and development were subsequently normal.

Rapid, rhythmic respiratory muscle contractions superimposed on a normal respiratory cycle have been observed in diverse conditions affecting the CNS, the phrenic nerve, and diaphragm (6). Primary disorders of the CNS, including encephalitis, tumors, and vascular accidents, are the most commonly reported causes in adults (6, 10). The pathophysiology is presumed to be the result of dysfunction of the respiratory rhythm generator. RF may also be caused by phrenic nerve irritation anywhere along its course in the neck or mediastinum. Finally, conditions resulting in direct diaphragmatic irritation such as peritonitis have been associated with RF (16).

Review of the literature revealed only four previous cases of RF in infants (1, 2). Adams and coworkers (2) reported RF in three infants with bronchopulmonary dysplasia hospitalized acutely for respiratory syncytial virus bronchiolitis. RF was an incidental RIP finding lasting no more than 5 min at a time, and resolved spontaneously within 4 d. The infants had DF rates of 225 to 270/min, intermittently superimposed on an apparently normal underlying rhythm. Riordan and coworkers (1), presented the case of an 8-day-old infant with RF who presented with "panting" and inspiratory stridor. Though respiratory tracings were not presented, it was stated that "inspirations were di- or triphasic, and abdominal palpation revealed two or three thrusts per inspiration." This pattern was clearly evident clinically and on the RIP tracings in our series (Figure 1A, Figure 2). Their patient was treated with supplemental oxygen during fluttering paroxysms, and symptoms ceased by 9 mo of age.

Diagnosis of RF has been reported using RIP (2, 17), fluoroscopy (1), and diaphragmatic electromyography (9). RIP, with a frequency response up to 32 Hz, accurately diagnoses high-frequency RF (2, 18). Though impedance pneumography is commonly employed in infants, its frequency response declines above 1.5 to 3 Hz, rendering it unsuitable for identifying most RF (2). Despite continuous impedance monitoring in all of our infants, only one of our three cases had an impedance tracing clearly showing RF (Figure 3). It is possible that the routine use of impedance pneumography in neonates rather than RIP results in the underrecognition of RF.

RF tends to occur most prominently during inspiration and involves chiefly the diaphragm. Our experience suggests that it is not confined to the diaphragm, however, as thoracic and other respiratory muscles may be similarly activated either with or without contraction of the diaphragm.

Hiccups are another pattern of abnormal respiratory muscle activation. However, the clonic contractions observed in RF, even at low frequencies, appear to be distinct from hiccups. A hiccup is characterized by glottic closure approximately 35 ms after a vigorous diaphragmatic contraction, resulting in the characteristic "hic" sound (19). By contrast, in RF sudden glottic closure sounds generally do not occur, nor were they present in our patients. Furthermore, with pathologic hiccups the frequency of diaphragmatic contractions is less than 60/min, whereas DF typically ranges from 35 to 480/min. Inspiratory airflow is minimal in unintubated infants during hiccups and minute ventilation may actually decline by 30% (20).

Hyperventilation (5, 8), hypoventilation (11, 12), hypoxemia (1), and respiratory failure (17) have only rarely been reported with DF. The present cases represent the first infants to present with hypercapnic ventilatory failure. The tidal volume in infants relies on a larger contribution from the diaphragm than in adults (21). Infants are noted to be poorly tolerant of diaphragmatic dysfunction, with unilateral or bilateral paralysis associated with a 20% and 50% mortality, respectively (22). Also, the mechanical advantage of the infant rib cage is limited by the relatively horizontal rib angles and the highly compliant chest wall. Thus, infants may be predisposed to respiratory failure with RF and therefore therapies involving interruption of the phrenic nerve are contraindicated.

Our infants with RF all had dysphagia, supporting the concept of a common brainstem dysfunction. The central pattern generator of swallowing is localized near the nucleus solitarius and nucleus ambiguous, just lateral to the medullary breathing center (23, 24). A complex interaction between the two brainstem areas normally coordinates breathing and swallowing. The combination of dysphagia and disordered control of breathing has been previously reported in infants with congenital central hypoventilation syndrome (CCHS) (25), respiratory dysrhythmia (26), and anoxic encephalopathy (27). Similar to the present series, infants with CCHS have normal brainstem magnetic resonance imaging (MRI) findings, suggesting a functional rather than a gross anatomic lesion. The normal brain MRI findings suggest a systems impairment characterized by shifting activation of inspiratory muscle groups. Alternatively, the dysphagia, gastroesophageal reflux, and laryngomalacia may have been secondary to the mechanical fluctuations in intrathoracic pressure induced by the RF.

The response of individual patients with RF to pharmacologic and nonpharmacologic therapy is highly variable. The resolution of the RF in Case 1 with CPAP suggests a reflex mechanism mediated by afferents in the upper airway, the chest wall, or lungs. In patients with pathologic hiccups, stimulation of the nasopharynx with a catheter has been reported to successfully interrupt the hiccups in most patients, presumably mediated by vagal pathways (28). Additionally, the ventilator flow rate has been shown to affect inspiratory time and respiratory rate in normal individuals (29). Resolution of RF during deep inspiration has been previously reported (30). However, previous adult patients with RF failed to respond to "mechanical hyperventilation" (17) and CPAP (8). Poor long-term tolerance of CPAP by Case 1 prompted a pharmacologic trial with chlorpromazine, a centrally acting dopamine antagonist. Chlorpromazine has been quite successful at treating recalcitrant hiccups, but adults with RF have had mixed results (11, 12). Chlorpromazine resulted in a marked reduction in the RF and permitted rapid weaning of ventilatory support. There were no side effects of chlorpromazine therapy.

In summary, we report the clinical and respiratory findings of three neonates with RF who presented with respiratory failure. The strikingly similar occurrence of RF, dysphagia, laryngomalacia, and gastroesophageal reflux suggests a distinct clinical entity that we have termed, the "respiratory flutter syndrome." The patients are now developmentally normal with the exception of swallowing dysfunction. RIP is recommended to confirm the diagnosis in infants with a compatible presentation. The frequent association of disordered control of breathing and dysphagia also suggests a common medullary brainstem dysfunction. Chlorpromazine effectively treated the RF and was well tolerated. The diagnosis of three infants with this presentation during an 18-mo period suggests that the "respiratory flutter syndrome" may be a more frequent cause of respiratory failure in newborns than previously recognized.