site and mechanics of spontaneous, sleep-associated obstructive apnea in infants

89:2453-2462, 2000. ;J Appl Physiol WatersGarrick W. Don, Turkka Kirjavainen, Catherine Broome, Chris Seton and Karen A.obstructive apnea in infantsSite and mechanics of spontaneous, sleep-associated

You might find this additional info useful...

43 articles, 23 of which you can access for free at: This article citeshttp://jap.physiology.org/content/89/6/2453.full#ref-list-1

4 other HighWire-hosted articles: This article has been cited by http://jap.physiology.org/content/89/6/2453#cited-by

including high resolution figures, can be found at: Updated information and serviceshttp://jap.physiology.org/content/89/6/2453.full

can be found at: Journal of Applied Physiology about Additional material and informationhttp://www.the-aps.org/publications/jappl

This information is current as of June 1, 2013.

http://www.the-aps.org/. © 2000 the American Physiological Society. ISSN: 8750-7587, ESSN: 1522-1601. Visit our website at year (monthly) by the American Physiological Society, 9650 Rockville Pike, Bethesda MD 20814-3991. Copyrightphysiology, especially those papers emphasizing adaptive and integrative mechanisms. It is published 12 times a

publishes original papers that deal with diverse area of research in appliedJournal of Applied Physiology

by guest on June 1, 2013http://jap.physiology.org/

Dow

nloaded from

http://jap.physiology.org/content/89/6/2453#cited-by

http://jap.physiology.org/

Site and mechanics of spontaneous, sleep-associatedobstructive apnea in infants

GARRICK W. DON,1 TURKKA KIRJAVAINEN,2 CATHERINE BROOME,3

CHRIS SETON,1 AND KAREN A. WATERS1,4

1David Read Sleep Unit, Department of Respiratory Medicine, The Royal Alexandra Hospitalfor Children, Westmead, New South Wales, Australia; 2Department of Pediatrics, University ofHelsinki, Helsinki 00029 HYKS, and University of Turku, Turku, Finland; 3Department of MedicalImaging, The Royal Alexandra Hospital for Children, Westmead, New South Wales 2124;and 4Departments of Medicine, and Paediatrics and Child Health,University of Sydney, New South Wales 2006, AustraliaReceived 17 February 2000; accepted in final form 28 June 2000

Don, Garrick W., Turkka Kirjavainen, CatherineBroome, Chris Seton, and Karen A. Waters. Site andmechanics of spontaneous, sleep-associated obstructive ap-nea in infants. J Appl Physiol 89: 2453–2462, 2000.—Toexamine the mechanics of infantile obstructive sleep apnea(OSA), airway pressures were measured using a triple-lumencatheter in 19 infants (age 1–36 wk), with concurrent over-night polysomnography. Catheter placement was guided bycorrelations between measurements of magnetic resonanceimages and body weight of 70 infants. The level of spontane-ous obstruction was palatal in 52% and retroglossal in 48% ofall events. Palatal obstruction predominated in infantstreated for OSA (80% of events), compared with 38.6% frominfants with infrequent events (P 5 0.02). During obstructiveevents, successive respiratory efforts increased in amplitude(mean intrathoracic pressures 211.4, 215.0, and 220.4cmH2O; ANOVA, P , 0.05), with arousal after only 29% ofthe obstructive and mixed apneas. The soft palate is com-monly involved in the upper airway obstruction of infantssuffering OSA. Postterm, infant responses to upper airwayobstruction are intermediate between those of preterm in-fants and older children, with infrequent termination byarousal but no persisting “upper airway resistance” andrespiratory efforts exceeding baseline during the event.

airway obstruction; soft palate; arousal

OBSTRUCTIVE SLEEP APNEA (OSA) occurs in infants as wellas older children, but for infants in their first year oflife OSA is not likely to be attributable to adenotonsil-lar hypertrophy. The exact prevalence of OSA in thisage group (infancy) is not well documented becauseepidemiological studies have concentrated on olderchildren (10). This may be due to the nonspecific natureof symptoms in infants or to difficulties in parentalassessment of symptoms that have been presentthroughout the life of the infant. These include feedingdifficulties, profuse sweating, and breath-holding spells(18). Known consequences of severe OSA in infancy

include failure to thrive, chronic respiratory failure,and irreversible developmental delay (13). Recentstudies suggesting that OSA may be linked with sud-den infant death syndrome (42, 27) emphasize thepotential importance of OSA during infancy and theneed to recognize and to understand the pathophysiol-ogy of the disease.

Previous studies of upper airway obstruction in in-fants have not considered the palate as a potential sitefor obstruction (26, 34). However, the retropalatal re-gion is a strong candidate as a site of airway collapse atany age, including infancy. It is a site of airway nar-rowing with little bony or cartilaginous support that isrendered prone to collapse when muscle tone decreasesduring sleep. The innervation of the palate is complex,involving at least five different neural centers (32).Palatal function in adults is affected by sleep, phase ofrespiration, and negative-pressure challenges. Thesefactors may also influence the activity of paltoglossusand levator palatini, which are the muscles primarilyresponsible for palatal position (16, 30, 36, 37, 40). Inadults with OSA, the frequency of palatal obstructionis approximately equal to that of retroglossal obstruc-tion (15). Equivalent, detailed information regardingpalatal information is not available in infants.

Early development is a period of rapid maturation insleep patterns and of many respiratory reflexes. In-fants have different reflex responses to many stimulicompared with older age groups, often with a predom-inance of inhibitory influences. Chemical responses aregenerally weak at birth and subsequently increase,whereas inhibitory reflexes are strong at birth andsubsequently decline (42). The mechanics of spontane-ous upper airway obstruction have been studied inadults (15), and many studies of airway collapsibilityhave been performed in older children (21). Few stud-ies in infancy have, however, included the palate as a

Address for reprint requests and other correspondence: K. A.Waters, David Read Laboratory, Blackburn Bldg., DO6, Univ. ofSydney, New South Wales 2006, Australia (E-mail: [email protected]).

The costs of publication of this article were defrayed in part by thepayment of page charges. The article must therefore be herebymarked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734solely to indicate this fact.

J Appl Physiol89: 2453–2462, 2000.

8750-7587/00 $5.00 Copyright © 2000 the American Physiological Societyhttp://www.jap.org 2453


Dow

nloaded from


potential site of upper airway obstruction, so currentliterature suggests that the retroglossal region is themost common site of obstruction in this age group (35).Many changes occur in chemical and reflex responsesduring infancy, so it is likely that infants would exhibitdifferences in the pathophysiological mechanisms of,and responses to, upper airway obstruction. Examplesof inhibitory influences on respiration include thestrongly inhibitory laryngeal chemoreflex (4), inhibi-tion of respiration during spontaneous sucking andswallowing (2, 22), and reduced activity of the genio-glossus at the onset of obstructive events (44).

We hypothesized that the airway obstruction in in-fants could predominate at the level of the soft palate.In addition, we postulated that inhibitory responseswould cause reduced respiratory effort in response tospontaneous airway occlusion and thus differ from theresponses exhibited in older children and adults. Fi-nally, we hypothesized that arousal would occur morecommonly after obstructive than central apneas, whencontinued respiratory efforts during obstructive apneawere defined using an esophageal pressure channel.

To evaluate these hypotheses, a triple-lumen pres-sure catheter was inserted during overnight sleepstudies of infants up to 8 mo of age. This permittedexamination of the relative contributions of palatal andthe retroglossal regions to obstructive apneas that oc-curred spontaneously during sleep in infants and per-mitted a detailed examination of mechanical responsesto spontaneous airway obstruction in infants.

METHODS

Study Design

To determine normal values for the distance from nares tooropharynx, magnetic resonance images (MRIs) of the upperairway were recalled from a reference group of 70 infants andcorrelated with growth parameters. These correlations wereused to assist placement of a triple-lumen catheter in afurther 19 infants, who were then studied to evaluate the siteand characteristics of upper airway obstruction in infants bymeasuring nasal, oropharyngeal, and esophageal pressureconcurrent with polysomnography (PSG). Apneic events andarousals were identified by PSG criteria. Manometric datawere analyzed for the period immediately before, during, andafter each apneic event.

Evaluation of Oropharyngeal Distance

There is no published method available to provide a meansfor estimating the distance from nares to oropharynx ininfants. Our study involved per nasal placement of a catheterin unsedated infants up to 8 mo of age. This required correctplacement of an oropharyngeal pressure port while minimiz-ing the amount of handling of the infant. To place the oro-pharyngeal lumen of the catheter, we evaluated MRIs todefine the distance from the nares to the oropharynx ininfants.

A regression equation was determined from the measure-ment of nose to oropharynx on MRIs of a reference group of70 infants. All digital records of cranial MRI available forinfants aged 6 mo or less were recalled from the 4-yr periodbetween 1994 and 1998, inclusive. Separate approval for thisaspect of the study was obtained from the Ethics Committee

of the Royal Alexandra Hospital for Children (RAHC), andthis did not include the infants investigated with manome-try.

The MRIs were acquired on a Phillips ACS-NT 1.5-T MRIunit (Powertrak 3000, software version 6.1.2). Images wereanalyzed using Siemens workstation and MagicView soft-ware (VA 31 C). This provided pixel-to-measurement corre-lations that were read in millimeters. Each MRI study wasexamined to determine whether a midline sagittal sectionshowing airway structures was available, and, if so, thedistances from nares to oropharynx were determined. Thedistance measured were from the nares to the posterior hardpalate, from posterior hard palate to the inferior tip of thesoft palate, and from the posterior hard palate to the superiortip of the epiglottis (Fig. 1). Adding these distances gave aminimum and maximum range for the oropharyngeal dis-tance, and the midpoint was taken to be the distance fromnares to oropharynx for each infant. The indications for MRIand physical characteristics of the infant at the time of thestudy were determined from the medical records.

The regression curve for age against the MRI measureddistance was used to approximate the distance from anteriornares to oropharynx in 19 infants undergoing PSG withconcurrent pressure measurements. Final analysis demon-strated that body weight was the best predictor for thepropharyngeal distance (Fig. 2). Observation of characteris-tic “swallowing spikes” on the oropharyngeal pressure chan-nel was taken as confirmation that the catheter was placed inthe oropharynx (38).

Analysis of Obstructive Apnea

Subjects. Infants presenting to the Sleep Unit at theRAHC, at least 36 wk postconceptional age and with a suffi-ciently normal upper airway to allow pressure recording,were invited to participate in this study. Of 45 infants eligi-ble for study, 30 were recruited and 15 parents declinedparticipation. There were 19 successful studies, 4 infants didnot tolerate the catheter, and 7 studies failed because oftechnical difficulties. The 19 infants with successful studieswere 15.9 6 2.0 (SE) wk of age (range 1–36 wk) and included10 male infants. Thirteen infants were born at term, and 6were premature with a mean gestation of 31.7 6 4.9 (SE) wk.Four infants were being investigated for suspected apneaafter an apparent life-threatening event, five because a sib-ling had died from the sudden infant death syndrome, andfive because of witnessed apneic events. The remaining fivehad clinically suspected OSA, four with failure to thrive(weight ,3rd percentile for age and gender) and one withmicrognathia. Approval for the study was obtained from theEthics Committee of the RAHC, and signed parental consentwas obtained before each study commenced.

PSG. All infants underwent multichannel overnight PSG,concurrent with airway pressure recordings. Sleep studieswere performed in the standard manner for the laboratory.A minimum of five channels were included for sleep staging[2 for electroencephalogram (EEG), 2 for electrooculogram(EOG), and 1 for submental electromyogram (EMG)], and fivechannels were monitored for respiratory analysis [airflow,respiratory effort from chest and abdomen, arterial O2 satu-ration (SaO2

), and transcutaneous CO2]. Surface electrodeswere used to record EEG, EOG, submental and diaphragmEMGs, and electrocardiogram. Nasal airflow was recordedvia nasal cannulas (intermediate infant, no. 1615, SalterLaboratories, Arvin, CA) connected to a low-level pressuretransducer referenced to atmospheric pressure (modelMP45–4 6 2 cmH2O, Validyne, Northridge, CA). Thoracic

2454 SITE AND MECHANICS OF OBSTRUCTIVE APNEA IN INFANTS


Dow

nloaded from


and abdominal respiratory effort were recorded using induc-tance plethysmography (Non-Invasive Monitoring Systems,Respitrace, Miami Beach, FL). SaO2 was recorded on thehand or foot (Ohmeda Biox 3700e Pulse Oximeter, Datex-Ohmeda, Homebush, Australia), and transcutaneous CO2levels were recorded from the upper chest or abdomen (modelTINA or TCM3, respectively, Radiometer, Copenhagen, Den-mark). All sleep study data were digitally filtered and ampli-

fied and were recorded onto a digital data acquisition system(Compumedics, Abbotsford, Victoria, Australia), for lateranalysis. Nasal pressure signal was recorded on the manom-etry and the PSG systems to assist in accurate time matchingof the respiratory events being analyzed.

Each 30-s epoch of the sleep studies was analyzed todetermine sleep state and the occurrence of respiratory andor arousal events. Sleep staging was determined using stan-dard infant criteria (1) for wakefulness, or quiet, active, andindeterminate sleep. Respiratory events were marked ac-cording to the standard criteria of the unit. That is, an apneais marked if there is a reduction in airflow $80% for a periodgreater than or equal to two respiratory cycles (based on therespiratory rate of the preceding minute) and associated witha $3% blood oxygen desaturation, an arousal, or both. Sig-nificant apneas were typed as obstructive, mixed, or centralon the PSG depending on the presence or absence of ongoingrespiratory efforts during the period of airflow cessation,respectively.

Two types of arousal were defined for the use in this study:“full EEG arousal” and “respiratory arousal.” A full EEGarousal was defined as a marked increase in EEG frequencyor amplitude for .1 s, associated with an increase in sub-mental EMG tone and a marked variability in respiratoryefforts and nasal airflow. A respiratory arousal was definedas a large change in two or more independent channels (e.g.,the EMGs and respiratory-effort channels), without EEGdisturbance, for .1 s. The incidence of full EEG and respi-ratory arousal after obstructive or mixed and central sleep

Fig. 1. Midline magnetic resonanceimage of an infant sagittal head sec-tion showing the distances measuredfor this study. A, tongue; B, hard pal-ate; C, soft palate; D, epiglottis; E, tra-chea.

Fig. 2. Raw data and regression line for body weight vs. distancefrom nares to oropharynx on magnetic resonance image. See text fordetails.

2455SITE AND MECHANICS OF OBSTRUCTIVE APNEA IN INFANTS


Dow

nloaded from


apnea was determined for each infant. The mean values weredetermined for the relevant group. Comparison between theincidence of arousals after central vs. obstructive apnea wasused to clarify whether such arousals depend on commonfactors (such as desaturation) or features unique to obstruc-tive events (such as stimulation of mechanoreceptors).

Airway manometry. A triple-lumen, saline-filled catheter(Critchley Electrical Products, Sydney, Australia) was usedto measure airway pressure fluctuations at the nasopharynx,the oropharynx, and the midthoracic esophagus. Standard-ized apertures were cut in the separate inner lumens at 1.5,12.5, and 16.5 cm from the sealed, distal end of the catheter.Each inner lumen was then connected to a separate pressuretransducer (Transpac IV, Abbot Critical Care, Dublin, Ire-land), and patency was maintained with a total saline flow of12.5 ml/h to prevent secretions from blocking the apertures.

The catheter was inserted via the nasal route into thethoracic esophagus, with the insertion distance for the oro-pharyngeal lumen of the catheter predicted from the MRIregression equation. Catheter placement was confirmed bythe observation of swallowing spikes on the oropharyngealchannel. Catheter apertures were thus placed in the naso-pharynx, oropharynx, and intrathoracic esophagus, and pres-sure measurements were recorded for a minimum of 3 h aftersleep onset. The catheter was left in place, without flow untilthe next full wakening, and then was removed for the re-mainder of the study. Recordings were made on a digital dataacquisition system (AMLAB v2.0, AMLAB International,Lane Cove, Sydney, Australia) that was time matched to thePSG. Sections of the PSG recording with the catheter in placevs. removed were analyzed separately to examine for theeffect of the catheter on the respiratory disturbance index(RDI).

Evaluation of apnea characteristics. Each apnea was iden-tified on the PSG and then identified and analyzed on themanometry recording. Each respiratory event was character-ized independently on the PSG and manometry recordings.For manometric analysis, pressure fluctuations were as-sumed to cease above the site of obstruction. Thus obstruc-tion at the level of the palate would be indicated by absence

of a nasal signal with continuing pressure fluctuations in theesophagus and the oropharynx (Fig. 3A). Retroglossal ob-struction was indicated by continued pressure fluctuationsonly at the level of the esophagus, with no fluctuations abovethis level, including the oropharyngeal and nasopharyngealsites (Fig. 3B). The use of a multichannel pressure catheterdoes not permit separation between obstruction isolated tothe retroglossal region and obstruction occurring simulta-neously at palatal and retroglossal regions, even when alaryngeal lumen is included (15).

The amplitude of respiratory efforts during obstructiveand mixed apnea was calculated from the pressure recordingin the esophagus. One or more obstructive events were seenin 16 (84%) of the 19 subjects, and the amplitude of thepressure swings was averaged for 5 breaths before and 5breaths after each event. During the event, individual respi-ratory efforts were analyzed. Mean values were calculatedfor each infant and then for relevant groups.

Statistical Analyses

All analyses were undertaken using SPSS for Windows(SPSS V9.0 for Windows, SPSS, Chicago, IL). For comparisonof event numbers between groups, a Student’s t-test wasused. For analysis of respiratory efforts during obstructiveapnea, changes in successive breaths were evaluated bytwo-way ANOVA without replication followed by Bonferro-ni’s test when the ANOVA showed significant difference. Theproportion of events followed by arousal was compared usingStudent’s t-tests for obstructive, mixed and central apneas.Results are presented as means 6 SE, unless otherwisestated. P values ,0.05 were considered statistically signifi-cant.

RESULTS

Evaluation of Oropharyngeal Distance

A total of 138 MRIs were recalled, and, of these, 70could be used for airway distance measurements (41 559% male infants, 29 5 41% female infants). The cor-

Fig. 3. Manometric recording of ob-structive apnea identified 2 ways. A:palatal obstruction terminated byarousal. Pressure fluctuations are ab-sent (or minimal) at nasopharyngeal(NPX) trace but continue at oropharyn-geal (OPX) and esophageal (ESO)traces. Uncalibrated nasal airflow (Na-sal) indicates the apnea indentified onpolysomnography. B: retroglossal ob-struction with continuing pressure fluc-tuations only in the ESO trace and aprogressive increase in respiratory ef-fort as the apnea progresses.



Dow

nloaded from


rected age of the infants was 5.3 wk premature to 27.0wk postterm [mean 11.7 6 8.4 (SD) wk]. Length wasrarely recorded, but weight was available for 55 sub-jects. The indication for the MRI included documented(e.g., seizures 5 16%), or suspected neurological condi-tions (e.g., hyper- or hypotonia 5 24%). Others condi-tions affected the cranial fossa but would not be ex-pected to affect facial growth (e.g., documented orsuspected hydrocephalus 5 22%, suspected space-occu-pying lesions 5 21%, or reduced cortical growth 512%). Lesions affecting the face tended to be outsidethe airway (orbits or cheek 5 4%). In two cases theMRI was undertaken to evaluate possible airway le-sions, one for a suspected tracheal lesion and anotherfor choanal atresia.

The mean distance from nares to oropharynx was6.2 6 0.35 cm. Regression analysis using weight in 55cases (37 5 67% male infants, and 18 5 33% femaleinfants) provided the best correlation with the oropha-ryngeal distance (R2 5 0.67, P , 0.001; Fig. 2). Cor-rected age also correlated significantly, but less well(R2 5 0.56, P , 0.001), and multiple-regression anal-ysis did not improve the correlation. Using this MRIdata, the best correlation for the distance to the oro-pharynx in infants up to the age of 27 wk postterm was

DORO 5 y

5 1.21 ln ~x! 1 4.14~R2 5 0.67, P , 0.001!

(Fig. 2).where x 5 body weight.

The catheter distance to the oropharynx (Doro) in the19 study infants fell within the 95% confidence inter-vals of the weight vs. MRI equation for 79% of cases.This result is consistent with our final analysis dem-onstrating that weight provided the best estimate ofnose-to-oropharyngeal distance. In the four outliers,the regression equation tended to underestimate thedistance from nose to oropharynx.

PSG

Infants with a minimum of four recorded obstructiveapneas were grouped into two categories, according tothe clinical recommendations following their study.Infants with few events (nontreatment) had an ob-structive RDI (ORDI) ,3 obstructions/h, and infants inwhom treatment was instituted (treatment) had anORDI .3 obstructions/h. Adequate manometry data

were available for nine subjects, five in the nontreat-ment and four in the treatment groups. The proportionof obstructive events occurring at palatal and retro-glossal regions was calculated for each subject, fol-lowed by group comparisons. Characteristics of infantsare presented in Table 1 and their sleep studies inTable 2.

The RDI was not affected by the presence of thecatheter [0.79 6 0.2 vs. 0.72 6 0.18 log10 RDI, catheterin place vs. catheter out, respectively; not significant(NS)], nor was the frequency of obstructive events(0.31 6 0.2 vs. 0.31 6 0.2 log10 ORDI catheter in placevs. catheter out, respectively; NS).

Site and Characteristics of Upper Airway Obstruction

The frequency of palatal and retroglossal obstruc-tions is presented for individual infants in Table 3.Overall, the soft palate was the primary site of obstruc-tion in 57 6 9.5% of obstructions. Accordingly, in 43 69.5% of obstructive apneas there was an obstruction atthe retroglossal level with or without simultaneouspalatal level obstruction. Infants in the nontreatmentgroup had palatal obstruction in 38.6 6 8.8% (SE) ofevents and retroglossal obstruction in 61.4 6 8.8% ofevents (NS). In contrast, the majority of obstructiveevents in the treatment group occurred at the palatallevel (80 6 9.8%) compared with 20 6 8.8% at theretroglossal level (NS). The differences in site of ob-struction were significant between the treatment andnontreatment groups (treatment vs. nontreatmentgroups, P 5 0.02 for palatal and for retroglossal ob-struction; Fig. 4).

Breathing Efforts During Obstructive Apneas

The mean breath amplitude over five breaths beforethe apnea was not different between groups (215.1 61.0 vs. 216.2 6 1.0%, nontreatment vs. treatment,respectively). The magnitude of respiratory efforts in-

Table 1. Subject characteristics, in treatment andnontreatment groups

Nontreatment Treatment P Value

n 14 5Age, wks 16.061.8 15.666.2 NSGender (M:F) 6:8 4:1Weight, kg 6.860.4 5.560.5 NSWeight SDS 0.360.5 21.461.5 NSWeight, % 55.5610.5 36.3621.2 NS

Values are means 6 SE; n, no. of subjects. M, male; F, female; SDS,standard deviation score; NS, not significant.

Table 2. Sleep characteristics in treatment andnontreatment groups

All Subjects Nontreatment Treatment P Value

TST, min 386.3611.0 399.1611.5 350.5621.3 NS%REM 36.161.5 37.161.3 33.464.6 NSRDI (total), events/hr 10.463.9 6.061.3 22.9614.0 0.04

REM 18.366.5 10.862.0 39.2623.3 0.03NREM 6.463.0 3.160.9 15.5611.1 0.05

ORDI (total), events/hr 4.762.7 1.260.2 14.569.7 0.02REM 8.664.8 2.360.5 26.3616.7 0.01NREM 2.861.9 0.560.1 9.367.1 0.03

CO2 range, mmHg 10.662.0 9.862.5 13.063.5 NSSaO2

averages, %Desaturation 3.760.3 3.460.3 4.660.7 NSREM (minimum) 95.160.7 95.860.6 93.061.5 NSNREM (minimum) 96.360.5 96.960.5 94.760.7 NSAwake (minimum) 95.560.5 96.060.5 94.061.0 NS

Values are means 6 SE. TST, total sleep time; %REM, percentTST spent in rapid-eye-movement (REM) sleep; RDI, respiratorydisturbance index; ORDI, obstructive respiratory disturbance index;NREM, non-REM sleep; SaO2

, oxygen saturation. CO2 range refers tothe range from stable wakefulness to the maximum value duringsleep (REM or non-REM).



Dow

nloaded from


creased with successive breaths during obstructive andmixed apnea (211.4 6 1.5, 215.0 6 1.6, and 220.4 62.7 cmH2O for the first, second, and third respiratoryefforts, respectively; Fig. 5A). After the apnea, respira-tory efforts were greater in the treatment group(216.2 6 1.0 vs. 222.4 6 2.9 cmH2O, nontreatment vs.treatment, respectively; P 5 0.02). The first effortmade during an obstructive event tended to have re-duced amplitude compared with the efforts precedingthe apnea (211.4 6 1.5 vs. 215.1 6 1.0 cmH2O, firstobstructed breath vs. baseline, respectively; P 5 0.002;Fig. 5B). We found no correlation between the ampli-tude of respiratory efforts at baseline with age, log10RDI, or log10 ORDI.

Arousal Frequency After Apneas

The occurrence of arousal at apnea termination wasdetermined from the PSG recording. From a total of341 obstructive and mixed events, full EEG or respira-tory arousal followed 34 and 66 events, respectively.From 388 central apneas, 31 and 56 events were fol-lowed by a full EEG or respiratory arousal, respec-tively. Full EEG arousal therefore followed 22.9 66.7% of obstructive and mixed vs. 7.3 6 1.8% of centralevents (P 5 0.04). Respiratory arousal followed 14.1 64.3% of obstructive and mixed and 20.5 6 6.0% ofcentral events (NS). When the two types of arousalwere combined, there was no difference between theproportion of events terminated by arousal accordingto apnea type (37.0 6 6.8 vs. 27.8 6 6.2%, obstructiveand mixed vs. central, respectively).

DISCUSSION

This study presents new data regarding measure-ments of the upper airway in infants on the basis ofMRI. Triple-lumen pressure catheters were placed ac-cording to these measurements, to measure the siteand characteristics of airway obstruction in infants.Our results demonstrate that the palate may be theprimary site of obstruction in young infants with clin-ically significant OSA, a site that has not been impli-cated previously in infants with OSA. Use of this newinvestigative method also provided information aboutthe responses of infants to airway obstruction, includ-ing initial inhibition of respiratory efforts and a ten-dency to EEG arousal after obstructive but not centralapnea.

Catheter Placement

The cephalocaudal growth gradient of cranial struc-tures is far advanced during infancy and early child-hood, with these structures nearer to final adult sizethan other parts of the body (5). There is a paucity ofdata regarding the distance to the oropharynx, withsome addressing childhood (12) but none assessing theearly period of exponential growth in infancy. Previousphysiological studies in infants do not provide predic-tive estimates of the distance a catheter should beinserted but rely entirely on the observation of swal-lowing spikes during the physiological recordings ofneonates (2, 3, 20). The infants in our study were older(up to age 8 mo) and were not sedated. In contrast toadults (15), it is not possible to directly visualize the

Fig. 4. Percentage of palatal and retroglossal obstruction betweennontreatment (open bars) and treatment (solid bars) groups. Onlysubjects with 4 or more apneas that could be analyzed for site ofobstruction were included in the comparison. Values are means 6SE.

Table 3. Location of upper airway obstruction in eachsubject

SubjectNo. ORDI,

Obstructions

Palatal,no.

Retroglossal,no.

Palatal,%

Retroglossal,%

1 0.2 2 2 50 502 0.2 1 1 50 503 0.3 1 1 50 504 0.5 0 3 0 1005 0.5 1 4 20 806 0.6 0 07 0.8 3 0 100 08 0.9 4 2 67 339 1.6 2 2 50 50

10 1.7 3 10 23 7711 1.9 3 0 100 012 2.1 0 2 0 10013 2.6 3 6 33 6714 2.6 0 3 0 10015 3.1 3 1 75 2516 4.5 12 10 55 4517 5.2 1 1 50 50*18 6.4 11 0 100 019 53.2 19 2 90 10

Subjects are presented in ascending order of the obstructive apneaindex (ORDI). Subjects 1–14 were not treated for obstructive sleepapnea and subjects 15–19 were. The ORDI was derived from poly-somnography, but proportions of palatal vs. retroglossal obstructionwere determined for events occurring during the period of the man-ometric recording. *This subject was not included in the groupcomparison because only 2 events could be analyzed on manometry.



Dow

nloaded from


oropharyngeal aperture of the catheter in infants be-cause of the anatomic characteristics of their oral air-way and the inability of infants to cooperate with theexamination.

To minimize the need for repositioning and refixingthe upper airway catheter, we evaluated a new methodof predicting the distance from the external nares tothe oropharynx. In our physiological studies, catheterplacement was predicted from the MRI data then con-firmed by observation of swallowing spikes on the oro-pharyngeal pressure trace. Thus noninvasive, predic-tive estimate derived from MRI data achieved the goals

of expediting catheter placement and minimizing thehandling required for each infant as well as avoidingthe need for imaging studies. Body weight is a readilyavailable growth parameter, which is particularly im-portant when proposing tools for clinical studies. If apressure catheter is being used, it remains advisable toconfirm catheter placement by the observation of char-acteristic swallowing pressure waves, at and below thelevel of the oropharynx (43).

Site and Response to OSA in Infants

The palatal region has not been implicated previ-ously as the site of spontaneous upper airway obstruc-tion in infants. This study has shown that the softpalate can be a primary site of upper airway obstruc-tion and that for a group of infants with clinicallysignificant OSA it is the most common site of obstruc-tion. Previous evidence suggests that inhibitory re-flexes are strong in preterm infants and may contrib-ute to the pathophysiology of the obstruction itself (10,11) and that arousals are infrequent in young infantsafter apnea (26). Our data show initial inhibition ofrespiratory efforts after the onset of obstruction butprogressively increasing amplitude of respiratory ef-forts during the event, with the efforts exceeding base-line before termination of the obstruction. We alsofound predominance of arousals with an EEG compo-nent after obstructive events compared with centralevents but no overall difference in the frequency ofarousals when EEG criteria were not included (27).

The demonstration that palatal obstruction occurs ininfant OSA is supported by adult studies using equiv-alent techniques showing that obstruction occurs atthe level of the palate and at the retroglossal region(15, 19). Most infants in this study had some obstruc-tion at palatal and at retroglossal levels. However,infants also tended to show a predominant site ofobstruction, and the incidence of palatal obstruction inthe treated group was higher than that reported foradult disease (17, 44). The predominance of palatalobstruction in infants with a higher obstructive RDIsuggests that palatal dysfunction plays a role in theincreased rate of obstruction experienced by those in-fants.

Obstructive events with retroglossal obstructionmay have included collapse (occlusion) of the palatalregion as well. Thus the incidence of palatal obstruc-tion may have been even greater than reported, andsuch events would involve a larger segment of theupper airway. If this were the case, infants with moresevere OSA would be expected to have more retroglos-sal than palatal events, and intrathoracic pressures atthe termination of the obstruction would be expected tobe greatest after retroglossal events. Given that nei-ther of these sequelae is supported by our data, weconclude that obstruction tended to occur at one or theother site. More extensive pressure catheters (morepressure sites) may still not help to clarify this issue(15), so that imaging (e.g., rapid MRI; Ref. 38) would be

Fig. 5. A: Amplitude of respiratory efforts during obstructive apnea.The efforts were increasing during successive breaths (P , 0.05). B:mean change in esophageal pressure amplitude compared with themean of the 5 breaths preceding the apnea (100% amplitude). Thepressure swing initially fell but exceeded baseline by the thirdobstructive effort.



Dow

nloaded from


required to distinguish isolated palatal from mixedpalatal and oropharyngeal collapse.

The pattern of respiratory efforts by these infantsduring obstruction suggests that they have less respi-ratory inhibition in response to spontaneous airwayobstruction compared with preterm infants. Preterminfants show reduced amplitude or respiratory effortsthroughout spontaneous obstructive episodes (9), butthe group in the present study exceeded baseline re-spiratory efforts by the third attempted breath. Exter-nally imposed upper airway occlusion stimulates adifferent pattern of respiratory effort as measured bydiaphragmatic EMG (9), so it is not clear whether theinhibition seen in older children in response to anexternal load reflects the same inhibitory reflexes (22).In this study, the first effort made during an obstruc-tive event tended to have reduced amplitude comparedwith the efforts preceding the apnea (Fig. 5B). How-ever, the infants went on to make substantial, andincreasingly greater, efforts against their spontaneousupper airway obstruction (Fig. 5, A and B). Thus post-term infants show active mechanical responses to air-way obstruction, whether the obstruction has occurredat the palate and/or the retroglossal region.

It is not clear whether adults and older childrenshow this pattern of initial inhibition in response toairway obstruction, although they do exhibit progres-sively greater respiratory efforts during externally in-duced airway occlusion (6). Possible mechanisms forthis increase in respiratory effort include hypoxia, hy-percarbia, and stimulation of mechanoreceptors by theairway occlusion itself (29, 33, 34). Infants with clini-cally significant OSA had greater amplitude of respi-ratory efforts, and infants with externally imposedairway occlusion also demonstrated an increased am-plitude of diaphragmatic EMG after the event (9). Itremains possible that mechanoreceptor-induced re-sponses were stronger in infants with more severeOSA.

Our study provides new data, quantifying the ampli-tude of efforts among postterm infants before, during,and after apnea. There was no evidence that the in-fants had persisting upper airway obstruction (“upperairway resistance”) between the discrete obstructiveevents. Older children clearly demonstrate increasedwork of breathing between the periods of discrete ob-struction (14). The lack of correlation between theamplitude of respiratory efforts at baseline with age,log10 RDI, or log10 ORDI is consistent with data in terminfants, suggesting that maturation of the response tolater obstructed breaths is due to maturation of che-moreflexes (8, 25) and not to pressure generation or theoccurrence of paradoxical respiration. We did not mea-sure acute changes in the SaO2

or CO2, but we found nodifferences in average SaO2

or CO2 values of the treat-ment vs. nontreatment groups to explain the increasedamplitude of postobstructive efforts (Table 2).

These infants showed few arousals at apnea termi-nation, consistent with other studies examining arousalresponses (26). Arousal is so common at apnea termi-

nation in adults that the two responses are often dealtwith interchangeably (17), and arousal from sleep isdeemed an important mechanism in averting poten-tially dangerous situations, such as hypoxia or hyper-capnia (28). We found that obstructive events weremore likely to precipitate arousals with EEG distur-bance than central apnea. We found a higher arousalindex than previous studies, but, importantly, the ap-neic events in the present study were selected on thebasis of the fact that they terminated with arousal ordesaturation (rather than duration). In addition, animportant new finding was that there was no overalldifference in the proportion of central or obstructiveapneas terminating in arousal when the definition ofarousals included disturbances without EEG shifts.The higher proportion of arousals after respiratoryevents in this study is almost certainly due to the samemethodological difference and inclusion of arousalswithout EEG shifts. The triggering mechanism forapnea termination may be hypoxia, hypercapnia orairway occlusion (3, 33, 34). Gleeson and associates(11) suggest that the point of apnea termination isdetermined by the amplitude of the respiratory effortsbeing made. Our results support the concept that ter-mination of apnea is independent of arousal in infantsand confirm that it is unlikely that arousal has afundamental role in apnea termination in infancy (26).The information regarding respiratory efforts in thisstudy suggests that, after term, apnea terminationmay be related to the amplitude of the respiratoryefforts or other mechanoreceptor pathways.

Small pressure fluctuations were sometimes seenabove the site of obstruction, which may have reflectedsome persisting airflow (see Fig. 1), but because theapnea was defined as a fall to #20% of the baseline thisdoes not preclude definition of the event as an apnea.These and other pressure changes observed during thestudy period were not due to transmission through thecatheter wall between the apertures, confirmed bymeasurement of catheter compliance and the demon-stration that this was negligible up to 100 cmH2O. Inaddition, the presence of the catheter did not affect therate of apnea (including obstructive events). Thesewere measured for the periods with the catheter inplace and without the catheter and did not change.Thus we confirmed that the presence of the catheter orof saline flow did not impact on the rate or site of upperairway obstruction observed in this study.

Summary

A regression equation was derived from MRI mea-surements of the upper airway and was used to assistthe placement of a nasal catheter into the oropharynxof infants. Pressure measurements from a triple-lumencatheter indicated that the soft palate is the mostcommon site of obstruction in infants with clinicallysignificant OSA, suggesting that palatal dysfunctioncontributes to the disease in infants. During spontane-



Dow

nloaded from


ous airway obstruction of infants ,1 yr of age, the firstrespiratory effort had reduced amplitude comparedwith baseline, but unlike preterm infants the ampli-tude of subsequent respiratory efforts returned to andthen exceeded baseline. Termination of apnea wasassociated with greater amplitude of respiratory ef-forts in the treatment group compared with infantswith few events. However, apnea termination was ac-companied by arousal in the minority of events, andthere was no evidence of increased amplitude of respi-ratory efforts between the discrete events. We specu-late that termination of apnea in infants is due toupper airway mechanoreceptor pathways or chemore-sponses, regardless of their interaction with arousalcenters.

The authors thank the parents of all infants who participated inthis study for their willingness to help our work and patience withour study. We also thank the staff of the David Read Sleep Unit fortheir assistance with the infant studies and for inserting the esoph-ageal catheter. Finally, we thank Dr. Jennifer Peat for advice re-garding statistical methodology and Kellie Tinworth for assistancewith preparation of the manuscript.

REFERENCES

1. Anders T, Emde R, and Parmelee A. A Manual of Standard-ized Terminology, Techniques and Criteria for Scoring of Statesof Sleep and Wakefulness in Newborn Infants. Los Angeles, CA:UCLA Brain Information Service, 1971.

2. Bamford O, Taciak V, and Gewolb IH. The relationshipbetween rhythmic swallowing and breathing during suckle feed-ing in term neonates. Pediatr Res 31: 619–624, 1992.

3. Berry RB and Gleeson K. Respiratory arousal from sleep:mechanisms and significance. Sleep 20: 654–675, 1997.

4. Davies AM, Koenig JS, and Thach BT. Upper airway che-moreflex responses to saline and water in preterm infants.J Appl Physiol 64: 1412–1420, 1988.

5. Fields HW. Craniofacial growth from infancy through adult-hood. Background and clinical implications. Pediatr Clin NorthAm 38: 1053–1088, 1991.

6. Frantz ID, Adler SM, Abroms IF, and Thach BT. Respira-tory response to airway occlusion in infants: sleep state andmaturation. J Appl Physiol 41: 634–638, 1976.

7. Frantz ID and Milic Emili J. The progressive response of thenewborn infant to added respiratory loads. Respir Physiol 24:233–239, 1975.

8. Gauda EB, Miller MJ, Carlo WA, Difiore JM, Johnsen DC,and Martin RJ. Genioglossus response to airway occlusion inapneic versus nonapneic infants. Pediatr Res 22: 683–687, 1987.

9. Gauda EB, Miller MJ, Carlo WA, DiFiore JM, and MartinRJ. Genioglossus and diaphragm activity during obstructiveapnea and airway occlusion in infants. Pediatr Res 26: 583–587,1989.

10. Gislason T and Benediktsdottir B. Snoring, apneic episodes,and nocturnal hypoxemia among children 6 months to 6 yearsold. An epidemiologic study of lower limit of prevalence. Chest107: 963–966, 1995.

11. Gleeson K, Zwillich CW, and White DP. The influence ofincreasing ventilatory effort on arousal from sleep. Am RevRespir Dis 142: 295–300, 1990.

12. Greenberg RS, Zahurak M, Belden C, and Tunkel DE.Assessment of oropharyngeal distance in children using mag-netic resonance imaging. Anesth Analg 87: 1048–1051, 1998.

13. Guilleminault C, Korobkin R, and Winkle R. A review of 50children with obstructive sleep apnea syndrome. Lung 159: 275–287, 1981.

14. Guilleminault C, Pelayo R, Leger D, Clerk A, and BocianRC. Recognition of sleep-disordered breathing in children. Pedi-atrics 98: 871–882, 1996.

15. Hudgel DW. Variable site of airway narrowing among obstruc-tive sleep apnea patients. J Appl Physiol 61: 1403–1409, 1986.

16. Hudgel DW and Hendricks C. Palate and hypopharynx—sitesof inspiratory narrowing of the upper airway during sleep. AmRev Respir Dis 138: 1542–1547, 1988.

17. Issa FG and Sullivan CE. Arousal and breathing responses toairway occlusion in healthy sleeping adults. J Appl Physiol 55:1113–1119, 1983.

18. Kahn A, Groswasser J, Sottiaux M, Rebuffat E, andFranco P. Mechanisms of obstructive sleep apneas in infants.Biol Neonate 65: 235–239, 1994.

19. Katsantonis GP, Moss K, Miyazaki S, and Walsh J. Deter-mining the site of airway collapse in obstructive sleep apneawith airway pressure monitoring. Laryngoscope 103: 1126–1131,1993.

20. Koenig JS, Davies AM, and Thach BT. Coordination ofbreathing, sucking, and swallowing during bottle feedings inhuman infants. J Appl Physiol 69: 1623–1629, 1990.

21. Marcus CL. Pathophysiology of childhood obstructive sleep ap-nea: current concepts. Respir Physiol 119: 143–154, 2000.

22. Marcus CL, Moreira GA, Bamford O, and Lutz J. Responseto inspiratory resistive loading during sleep in normal childrenand children with obstructive apnea. J Appl Physiol 87: 1448–1454, 1999.

23. Mathew OP. Upper airway negative-pressure effects on respi-ratory activity of upper airway muscles. J Appl Physiol 56:500–505, 1984.

24. Mathew OP, Roberts JL, and Thach BT. Pharyngeal airwayobstruction in preterm infants during mixed and obstructiveapnea. J Pediatr 100: 964–968, 1982.

25. Mathur R and Douglas NJ. Relation between sudden infantdeath syndrome and adult sleep apnoea/hypopnoea syndrome.Lancet 344: 819–820, 1994.

26. McNamara F, Issa FG, and Sullivan CE. Arousal patternfollowing central and obstructive breathing abnormalities ininfants and children. J Appl Physiol 81: 2651–2657, 1996.

27. Mograss MA, Ducharme FM, and Brouillette RT. Move-ment/arousals. Description, classification, and relationship tosleep apnea in children. Am J Respir Crit Care Med 150: 1690–1696, 1994.

28. Phillipson EA and Sullivan CE. Arousal: the forgotten re-sponse to respiratory stimuli. Am Rev Respir Dis 118: 807–809,1978.

29. Phillipson EA, Sullivan CE, Read DJ, Murphy E, andKozar LF. Ventilatory and waking responses to hypoxia insleeping dogs. J Appl Physiol 44: 512–520, 1978.

30. Rodenstein DO and Stanescu DC. The soft palate and breath-ing. Am Rev Respir Dis 134: 311–325, 1986.

31. Ruggins NR. Pathophysiology of apnoea in preterm infants.Arch Dis Child 66: 70–73, 1991.

32. Strutz J, Hammerich T, and Amedee R. The motor innerva-tion of the soft palate. An anatomical study in guinea pigs andmonkeys. Arch Otorhinolaryngol 245: 180–184, 1988.

33. Sullivan CE, Kozar LF, Murphy E, and Phillipson EA.Arousal, ventilatory, and airway responses to bronchopulmonarystimulation in sleeping dogs. J Appl Physiol 47: 17–25, 1979.

34. Sullivan CE, Murphy E, Kozar LF, and Phillipson EA.Waking and ventilatory responses to laryngeal stimulation insleeping dogs. J Appl Physiol 45: 681–689, 1978.

35. Suto Y, Matsuda E, and Inoue Y. MRI of the pharynx in youngpatients with sleep disordered breathing. Br J Radiol 69: 1000–1004, 1996.

36. Tangel DJ, Mezzanotte WS, and White DP. Influence of sleepon tensor palatini EMG and upper airway resistance in normalmen. J Appl Physiol 70: 2574–25781, 1991.

37. Tangel DJ, Mezzanotte WS, and White DP. Influences ofNREM sleep on activity of palatoglossus and levator palatinimuscles in normal men. J Appl Physiol 78: 689–695, 1995.



Dow

nloaded from


38. Thach BT, Davies AM, and Koenig JS. Pathophysiology ofsudden upper airway obstruction in sleeping infants and itsrelevance for SIDS. Ann NY Acad Sci 533: 314–328, 1988.

39. Tishler PV, Redline S, Ferrette V, Hans MG, and AltoseMD. The association of sudden unexpected infant death withobstructive sleep apnea. Am J Respir Crit Care Med 153: 1857–1863, 1996.

40. Wheatley JR, Tangel DJ, Mezzanotte WS, and White DP.Influence of sleep on response to negative airway pressure oftensor palatini muscle and retropalatal airway. J Appl Physiol75: 2117–2124, 1993.

41. White DP. Sleep-related breathing disorder. 2. Pathophysiologyof obstructive sleep apnoea. Thorax 50: 797–804, 1995.

42. Widdicombe J. Effects of hypoxia and hypercapnia on theHering-Breuer reflex of the conscious newborn rat. J ApplPhysiol 78: 3–4, 1995.

43. Wilson SL, Thach BT, Brouillette RT, and Abu Osba YK.Coordination of breathing and swallowing in human infants.J Appl Physiol 50: 851–858, 1981.

44. Wulbrand H, Von Zezschwitz G, and Bentele KH. Submen-tal and diaphragmatic muscle activity during and at resolutionof mixed and obstructive apneas and cardiorespiratory arousalin preterm infants. Pediatr Res 38: 298–305, 1995.



Dow

nloaded from


site and mechanics of spontaneous, sleep-associated obstructive apnea in infants

Documents