Acta Paediatr 81: 609-12. 1992

Obstructive sleep apnea in Arnold-Chiari malformation treated with acetazolamide J Milerad, H Lagercrantz and P Johnson’ Department of Pediatrics, Karolinska Hospital, Stockholm, Sweden and the Nufield Institute for Obstetrics and Gynecology. Oxford, U K ’

Milerad J, Lagercrantz H, Johnson P.Obstructive sleep apnea in Arnold-Chiari malformation treated with acetazolamide. Acta Pzdiatr 1992;81:609-12. Stockholm. ISSN 0803-5253 We studied respiratory patterns and transcutaneous gas pressures in two infants with Arnold-Chiari type I1 malformation referred to us due to repeated episodes of stridor and cyanosis. During both active and quiet sleep, respiration was irfegular and absent or inverse thoracic breathing movements and frequent decreases in oxygen saturation to below 80% were observed. When breathing air with 2% COZor when given acetazolamide 10 mg/kg, chest wall movements normalized and oxygenation increased to near normal levels. After three months of treatment with acetazolamide20 mg/kg/24 h no further episodes of hypoventilation or hypoxemia were observed and further treatment could be discontinued. We conclude that stimulation of respiration by COz or by acetazolamide appears to recruit chest wall muscles and promote upper airway patency in Arnold-Chiari malformation. A treatment trial with acetazolamide seems justifiable in these infants when respiratory problems are present. 0 Acetazolamide. apnea. control of breathing, hydrocephalus, infants, myelomeningocele J Milerad, Department of Pediatrics, Karolinska Hospital, S-104 01 Stockholm, Sweden

Arnold-Chiari malformation of the brain and brainstem is present in most cases of myelomeningocele and hydrocephalus (1). Type IT Chiari anomaly is characterized by a kinking and caudal displacement of the cerebellum and brainstem into the cervical canal. Stridor and airway obstructions, frequent clinical findings in these infants (2-6) have been attributed to compression of the brainstem and the cranial nerve efferents. However, the respiratory problems are not limited to malfunction of the upper airway. In many cases respiratory symptoms may persist even after tracheostomy and relief of brainstem compression by cervical laminectomy (7). Due to the lack of efficient treatment, the long-term outcome of Arnold-Chiari infants with respiratory problems has remained poor and according to some authors almost one-third may die in respiratory arrests during the first year of life (2). In this study we investigated the possibility that stimulation of respiratory drive with CO2 or the carbonic anhydrase inhibitor acetazolamide, may improve respiratory function in these infants. This assumption was based on the fact that a low respiratory drive, as reported previously in these infants (6, 8) is known to compromise both respiratory mechanics and upper airway patency.

Methods Patients

Case I. This female infant was born at term with low lumbar myelomeningocele and hydrocephalus. Arnold-

Chiari type IT malformation was diagnosed by CT scan. The lumbar defect was successfully closed at the second postnatal day and she received a ventriculoperitoneal shunt at two weeks of age. The shunt was revised after another two weeks, and her ventricular index stayed thereafter within normal limits. Stridor and apneas with color changes appeared within the first days and increased in severity over the next weeks. During wakefulness stridor and cyanotic spells could be elicited by crying or feeding, and bagging was often required to terminate these events. A respiratory investigation at six weeks of age showed partial upper airway obstruction, hypoventilation and recurrent severe hypoxemia. Respiratory stimulants such as theophylline or progesterone were not successful and cervical laminectomy performed at eight weeks of age resulted in only a minor improvement. The airway obstructions could be abolished entirely when breathing CO2-enriched gas mixtures and similar effects were seen after a single dose of acetazolamide 10 mg/kg. After two weeks of treatment with acetazolamide 20 mg/kg/24 h she was discharged. Her plasma pH stayed within normal limits (7.35-7.45) during treatment, while capillary PCOZdecreased to 4.6-4.8 kPa. The treatment was continued until the age of 5.5 months. Slight somnolence and lethargy were the main side effects. The girl is now five years old and is doing very well. She still has frequent mixed and obstructive apneas during sleep but she is no longer hypoxemic during these episodes. Case 11. This female infant was born with a lumbosacral myelomeningocele and hydrocephalus. The diag-

6 10

J Milerad el at.

ACTA PRDIATR 81 (1992)

nosis of Arnold-Chiari type I1 malformation was made by MRT imaging. She received a ventriculoperitoneal shunt at one week of age. Heavy and irregular breathing during sleep was noted at around seven weeks of age. At nine weeks of age she was referred for a respiratory investigation after an apneic attack which required intubation and assisted ventilation. Her breathing during sleep was irregular with stridor and frequent mixed and obstructive apneic episodes. These respiratory arrests were accompanied by decreases in transcutaneous PO2 (tcPO2) to 7-8 kPa and in oxygen saturation to below 80%. Her tcPCO2 was within the normal range (5.6-6 kPa). When breathing COz-enriched air or when given an oral dose of acetazolamide 10 mg/kg, respiratory patterns normalized and no hypoxic episodes were observed. No side effects of the drug were observed in this infant. Acetazolamide therapy 20 mg/kg/day was continued until five months of age and no clinically significant obstructions occurred after discontinuation of treatment. The girl is now four years of age, she is walking without support and is doing well. 5.02

Procedures

The techniques used for the respiratory recordings have been described in detail previously (9) and therefore will only be outlined briefly here. All studies were performed during spontaneous sleep with the infants lying prone in the cot. A trained nurse scored the sleep states every minute by visual observation. Breathing movements were monitored with inductive plethysmography (Respitrace), the chest coil placed over the mamillae and the abdominal coil around the umbilicus. The Respitrace unit was electronically calibrated in order to give an equal signal amplitude for an equal increase in coil circumference. Oxygen saturation was measured with a pulse oximeter (Studley Data Systems, Oxford, England). Arterial oxygen and carbon dioxide partial pressures were estimated with transcutaneous PO2 (tcP02) and transcutaneous PCOz (tcPC02) electrodes (Radiometer TCM2, Copenhagen, Denmark) heated to 45 and 43"C, respectively. CO2-enriched humidified gas mixtures were obtained by mixing pure C02 with air. This gas mixture was delivered through a 10-1 plastic hood. The C02 content of inspired air was analyzed with a capnograph (Datex, Helsinki, Finland).

Results Both infants spent most of their sleep time in active or undetermined sleep but the abnormalities observed were not confined to that particular state. Typically, the thoracic signal was absent or inverse while the Respitrace sum signal (corresponding to tidal volume) was smaller than the abdominal signal. The small tidal volumes and the rapid decreases in oxygen saturation, often to below 8O%, suggested that the inward suction

tCPO2

F

2%

GO2

9; Fig. 1. A representative recording of active sleep during: (A) air breathing; and (B) 2% CO2 in air. Chest and abdominal movements are out phase in (A) and in phase in (B) as indicated by arrows. The second recording was obtained 10 min after the first. The reversal of paradoxical breathing is indicated by arrows.

of the chest was due to inspiratory efforts against a partially closed airway resulting in a displacement of air from the thorax to the abdomen (Figs 1A, 2A). Also, the increased tcPCOz and the low tcPO2 were suggestive of alveolar hypoventilation. When I .5% COZor more was added to the inspired air chest and abdominal movements synchronized, tidal volumes increased. The tcP02 level increased and the tcPCO2 level decreased in spite of C02 being added. The number of central apneas decreased and the decreases in oxygen saturation were less marked (Fig. 1B). Similar effects on breathing could be observed 30 min after a single oral dose of acetazolamide (10 mg/kg, Fig. 2B). Lower doses had less effect on respiratory patterns. The effect of various treatments in infant no. 1 is summarized in Fig. 3. The mean number of hypoxic events, defined as a decrease in oxygen saturation to less than 80%, per hour of sleep is shown. During CO2 breathing the number of hypoxic events decreased from 59 to 19 per hour. After spinal decompression and cervical laminectomy 42 events per hour were observed and during acetazolamide treatment the number decreased to 22 per hour.

Sleep apnea treated wirh acerazolamide

ACTA PEDIATR 81 (1992)

(A)

SBC

I I

I I I

I

I

8

I I I ,

1 , 1 , , , , , 1 , 1 , , 1 , 1 1 1 1 1 1 1

, , , , I

EKG

6I 1

Mean number of hypoxic eventa per hour of deep

chest

60 ~~

abd

40

8U

so02 +0-

100

20 \

no treatment

0 T

-----.;

,4 I L

tCPCO2

ocdozolornida

larnlnactorny

Fig. 3. Mean number of hypoxic events per hour of sleep during two consecutive sleep epochs of active and quiet sleep. I = N o treatment, air breathing; 2 =no treatment, CO2 breathing; 3 =after laminectomy, air breathing; 4 = laminectomy and acetazolamide 20 mg/kg/day air breathing.

acetazolamide tCP02

CO2 breathing

no tnatmmt

tcPCO2

-,

Fig. 2. A representative recording of active sleep: (A) before; and (B) 30 min after a single oral dose of acetazolamide. There is a marked increase in tidal volume (sum of chest and abdominal excursions),

tcPO2 and oxygen saturation after treatment.

Discussion While a large number of abnormal ventilatory patterns during sleep have been described in infants with Arnold-Chiari malformation (6, 8, 10) the mechanisms of the respiratory dysfunction are not clear. Furthermore, for reasons outlined below they seem to be different from those of patients with obstructive sleep apnea from other causes. While COz may aggravate the obstructions in patients with, for example, tonsillar hypertrophy, we found that breathing C02 relieved the symptoms in both infants studied. This improvement was related to the reversal of paradoxical breathing. Although paradoxical respiration is a normal finding in infants during active sleep, we have several reasons to believe that the inward chest movements recorded in our patients were caused by partial airway obstruction. Sleep state-related paradoxi-

cal breathing is usually present only during part of active sleep and does not compromise oxygenation or ventilation (1 I). In our patients the paradoxical breathing was present during both sleep states and ventilation was severely compromised by this breathing pattern. Finally, the rapid switch to synchronous breathing when C02 was given is not consistent with a sleep-related phenomenon. Our recording technique did not enable us to determine the anatomical site of the airway obstruction. However, the clinical symptoms suggested that laryngeal dysfunction played a major role. The improved airway patency during C02 challenge can be attributed to several mechanisms. CO2 inhalation is known to increase the activity of the posterior cricoarytenoid muscles which are the main dilators of the glottic aperture ( I 2). Furthermore, C02 breathing in neonates has been shown to increase the contribution of the chest wall to the breathing effort (13, 14) which in turn increases upper airway patency through vagal feedback mechanisms (1 5). Since effects on breathing were seen within a few breaths it seems probable that these CO2 effects were mediated via peripheral chemoreceptors or airway COz receptors (16); a central mechanism would require a longer latency. The carbonic anhydrase inhibitor acetazolamide, has been used therapeutically in the treatment of hypoventilation syndrome ( 1 7). More recently, beneficial effects in obstructive sleep apnea (18) have been described. Acetazolamide increases respiratory drive by stimulating both the peripheral and central chemoreceptors, the latter effect being due to the lowering of pH in the CSF (19). Its effects on breathing are thus similar to CO2 inhalation although the longer latency between administration and

6 12

J Milerad el al.

effect would suggest a central rather than a peripheral chemoreceptor effect. However, acetazolamide also dilates the cerebral vessels (20) and thus affects both the synthesis and composition of CSF fluid (21). Therefore, that some of the respiratory effects of acetazolamide in patients with hydrocephalus may be attributed to changes in intracranial pressure and brain blood flow cannot be excluded. In contrast to untreated cases, who often have a prolonged course of respiratory problems, we observed no relapse in our patients after discontinuation of acetazolamide treatment. We would therefore speculate that this long-term stimulation had led to an “entrainment” of the upper airway muscles analogous to other types of repeated stimulation. Since no efficient treatment is available at present we feel that a trial with acetazolamide is justifiable in all infants with respiratory problems. Respiratory investigation should be considered even in the absence of clinical stridor since signs of abnormal respiratory control can be recorded even in the absence of clinically significant apneas (7). AcknowIed~emen/s.-This work was supported by grants from the National Foundation against Heart and Lung Diseases and the First of May Flower campaign for Children’s Health. The help of Dr P Kogner in collecting the data is gratefully acknowledged.

References I . Samuelsson L, Bergstrom K, Thuomas K, Hemmingsson A, Wallensten R. MR imaging of syringohydromyelia and chiari malformations in myelomeningocele patients with scoliosis. AJNR 1987;8:539-46 2. Wealthall SR, Whittaker GE, Greenwood N. The relationship of apnoea and stridor in spina bifida occulta to the unexplained infant’s deaths. Dev Med Child Neurol 1974;16(Suppl):107-16 3. Hesz N, Wolraich M. Vocal cord paralysis and brainstem dysfunction in children with spina bifida. Dev Med Child Neurol 1985;27:522-3 1 4. Holinger PC, Holinger LD, Reichert JJ, Holinger JJ. Respiration obstruction and apnea in infants wih bilateral vocal cord paraly-

ACTA PEDIATR 81 (1992)

sis. myelomeningocele, hydrocephalus and Arnold-Chiari malformation. J Pediatr 1978;92:368-73 5. Morley AR. Laryngeal stridor. Arnold-Chiari malformation and medullary hemorrhages. Dev Med Child Neurol 1985;11:4714 6. Davidson Ward SL, Jocobs RA, Gates EP, Hart LD, Keens TG. Abnormal ventilatory patterns during sleep in infants with myelomeningocele. J Pediatr 1986;109:631-4 7. Charney EB, Rorke LB, Sutton LN, S h u t L. Management of Chiari I1 complications in infants with myelomeningocele. J Pediatr 1987;111:36471 8. Ruff ME, Oakes WJ, Fisher SR, Spock A. Sleep apnea and vocal cord paralysis secondary to type I chiari malformation. Pediatrics 1987;80:231-4 9. Milerad J, Wennergren G, Hertzberg T, Lagercrantz H. Respiratory and arousal responses in apneic infants reinvestigated. Eur J Pediatr 1989;148565-70 10. Davidson Ward SL, Nickerson BG, Hal van der A, Rodriguez AM, Jacobs RA, Keens TG. Absent hypoxic and hypercapneic arousal responses in children with myelomeningocele and apnea. Pediatrics 1986;78:4450 1 1 . Gaultier C, Praud J, Canet E, Delaperche M, DAllest A. Paradoxical inward rib cage movement during rapid eye movement sleep in infants and young children. J Dev Physiol 1987;9:391-7 12. Bartlett D. Effects of hypercapnia and hypoxia on laryngeal resistance to airflow. Respir Physiol 1979;37:293-302 13. Davi M, Sankaran K, MacCallum M, Cates D, Rigatto H. The effect of sleep state on chest distortion and on the ventilatory response to COz in neonates. Pediatr Res 1979;13:982-6 14. Anderson D, Gennser G, Johnson P. The effect ofcarbon dioxide inhalation on phase characteristics of breathing movements in healthy newborn infants. J Dev Physiol 1986;8: 147-57 15. Bartlett D. Effects of vagal afferents on laryngeal responses to hypercapnia and hypoxia. Respir Physiol 1980;42:189-98 16. Johnson P. Evidence for lower airwaychemoreceptors in newborn lambs. Pediatr Res 1976;10:462 17. White DP, Zwillich CW, Pickett ChK, Douglas NJ, Findley LJ, Weil JV. Central sleep apnea, improvement with acetazolamide therapy. Arch Intern Med 1982;142:1816-19 18. Inoue Y. Hazama H. Sasaki T. Effects of acetazolamide on S ~ apnea syndrome. Sleep 1987;16:362 19. Teppema LJ, Rochette F, Demedts M. Ventilatory response to carbonic anhydrase inhibition in cats: effects of acetazolamide in intact vs peripherally chemodenervated animals. Respir Physiol 1988;74:373-82 20. Jaavaheri S,Wejne J, Demeester G, Leusen I. Effects ofacetazolamide on ionic composition of cisternal fluid during acute respiratory acidosis. J Appl Physiol 1984;57:85-91 21. Severinghaus JW, Cotev S. Carbonic acidosis and cerebral vasodilation after Diamox. %and J Clin Lab Invest 1968; (Suppl)lO2 Received April 2, 1991. Accepted Sep. 18, 1992

D