Archives of Disease in Childhood 1993; 69: 191-196

Breathing abnormalities in sleep inachondroplasia

K A Waters, F Everett, D Sillence, E Fagan, C E Sullivan

AbstractOvernight sleep studies were performedin 20 subjects with achondroplasia todocument further the respiratory abnor-malities present in this group. Somato-sensory evoked potentials (SEPs) wererecorded in 19 ofthe subjects to screen forthe presence of brainstem abnormalities,which are one of the potential aetiologicalmechanisms. Fifteen children aged 1 to14 years, and five young adults, aged20 to 31 years were included. All hadupper airway obstruction and 15 (75%)had a pathological apnoea index (greaterthan five per hour). Other sleep associ-ated respiratory abnormalities, includingpartial obstruction, central apnoea, andabnormal electromyographic activity ofaccessory muscles of respiration, alsoshowed a high prevalence. SEPs were

abnormal in eight (420/%), but there was no

correlation between abnormal SEPs andapnoea during sleep, either qualitativelyor quantitatively. A high prevalence ofboth sleep related respiratory abnor-malities and abnormal SEPs in youngsubjects with achondroplasia was demon-strated. However, the sleep related res-

piratory abnolrmalities do not alwaysresult in significant blood gas disturb-ances or correlate with abnormal SEPs inthis group.(Arch Dis Child 1993; 69: 191-196)

Camperdown,Australia, RoyalPrince Alfred HospitalSleep UnitK A WatersF EverettC E Sullivan

The Children'sHospitalKA WatersD SillenceE Fagan

Correspondence to:Professor Colin Sullivan,David Read Laboratory, TheUniversity of Sydney, NSW2006, Australia.

Accepted 11 March 1993

Achondroplasia is an autosomal dominantsyndrome of short limbed short stature and isthe most common type of chondrodysplasia.The bony abnormalities present in achondro-plasia result in a characteristic phenotype, butthe specific underlying defect is not yet known.The distinctive abnormalities of achondro-plasia that may predispose to upper airwayobstruction in sleep include a short cranialbase, and associated with this, a relativelyhypoplastic middle third of face. Respiratorycontrol centres can also be affected in this dis-order, when there is compression of the brain-stem at the level of the foramen magnum,I 2secondary to significant bony stenosis of theforamen magnum.

Recently a number of studies have suggestedthat sleep apnoea occurs commonly in thoseindividuals with achondroplasia.34 A range ofsleep disordered breathing is recognised, thecommonest being the upper airway obstructivevariety associated with snoring. The mech-anisms that cause apnoea are still uncertain,although structural abnormalities described

above, with subsequent narrowing of the upperairway, is considered to be an important oreven dominant cause of adult apnoea.The reports of apnoea in achondroplasia

include central apnoea (apnoea where there iscessation ofdiaphragmatic activity and airflow)and apneustic breaths. The occurrence of thistype of apnoea is suggestive of abnormal reflexcontrol of breathing in sleep. Becauseachondroplasia can be associated with stenosisof the foramen magnum and brainstem com-pression, it has been suggested that the apnoeain achondroplasia is the result of damage tobrainstem reflexes.4 In the infant, life threaten-ing apnoea is thought to be the result of criticalstenosis at the craniocervical junction, withcompressive damage to the respiratory controlcentres.The purpose of our study was to measure

and- characterise the type of apnoea inachondroplasia, and to measure somatosen-sory evoked potentials (SEPs) to see if theycorrelated with the type of apnoea. We providea detailed description of the types of res-piratory abnormalities present in sleep inachondroplasia. This helps to interpret furtherthe sleep disordered breathing in achondro-plasia, on the basis of either the limited upperairway dimensions or the neurological abnor-malities, secondary to bony malformations.

Patients and methodsIn the total study group there were 10 malesand 10 females. This included 15 children,eight boys and seven girls, aged 1 to 14 years,and five young adults, two male and threefemale, aged 19 to 31 years.The subjects were usually enrolled by

contact at the time of their regular review inour skeletal dysplasia clinic. Two patients of 22who have been approached to date havedeclined participation. The main reason givenin both cases was the parents' desire to limitthe number of investigations and procedurestheir children underwent. Three young adultswere recruited through the Little People'sAssociation, where the research study was dis-cussed, and one participant was contactedthrough a sibling.

Overnight sleep studies were performed in allpatients. A clinical history with specificquestioning regarding symptoms of obstructivesleep apnoea, apnoea of infancy, and othersleep disordered breathing was obtained.Clinical examinations were performed in each,with specific attention to the presence of eitherperipheral neurological or cranial nerveabnormalities. The studies subsequently took

191

on February 20, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.69.2.191 on 1 A

ugust 1993. Dow

nloaded from

Waters, Everett, Sillence, Fagan, Sullivan

place, overnight, in an established sleep unit,during the period of the subjects' normal nighttime sleep. Polysomnography included record-ings of sleep staging using electroencephalo-gram (EEG) (C3-A2), (02-Al), electro-oculogram (EOG), and chin electromyogram(EMGgg) or sternomastoid electromyogram(EMGsm) via surface electrodes. Respiratoryvariables included inductance plethysmography(Respitrace) of chest and abdomen (uncali-brated), nasal airflow (pressure wave), anddiaphragm EMG (EMGd). An electro-cardiogram (ECG) was recorded using chestleads and oxygen saturation (Sao2) using afinger probe (Ohmeda Biox 1000). Trans-cutaneous carbon dioxide (Tcco2) was recordedusing the TINA (Radiometer, Copenhagen)monitor. All studies were performed withoutsedation and recorded on a Grass Model multi-channel recorder.

Sleep studies were staged and scored by twoof the authors (KAW or FE) using the stagingcriteria of Rechtschaffen and Kales.5 Resultswere recorded for apnoea index, total apnoea,central apnoea, and obstructive apnoea times.The apnoea index is the total number ofapnoeas during sleep divided by the totalnumber of hours of sleep recorded. Anobstructive apnoea is one where there isabsence of airflow, but persistent respiratoryeffort, indicated by EMGd activity and chestand abdominal movements (Respitrace).Central apnoeas are those where the absentairflow is accompanied by absent diaphrag-matic activity and lack of thoracic andabdominal respiratory movement.The criteria used to define a respiratory

event were arousal or a change in Sao2 (>3%)on transcutaneous measures or a disruption ofthe regular pattern of respiration (for example,a pattern of repetitive central apnoea). As aresult of this, apnoea duration varied quitemarkedly across the age range of our patients.In a 12 month old infant an apnoea coveringthree normal breath intervals, but lasting onlythree seconds and resulting in a 4% oxygendesaturation would be included. We includedhypopnoeas (half average respiratory excur-sions) only if this occurred in association withan arousal and subsequent return to baselinerespiratory excursions. The hypopnoeas werenot usually associated with measured oxygenor carbon dioxide changes and as a result manyminor events were not included in the score.Other respiratory abnormalities were noted inthe results, but did not affect the scoringprocedure.The measurement of SEPs was used as an

index of brainstem dysfunction. SEPs wereusually recorded within 24 hours of the nighttime sleep study. A transcutaneous stimuluswas adjusted to motor threshold (indicated bya toe twitch). A stimulus range of 4-9 mAmpwas used, for 200 ,usec, at 2-1 Hz and appliedover the posterior tibial nerve at the ankle. Thepotential was recorded at the lumbar spine(referred to the iliac crest) and the corticalresponse was recorded at CZ 1 referred to FPZ(ground plate of leg). The data were analysedusing the Nicolet Compact Four evoked

potential system. Sensitivity set at 25 ,uV, andfilter 30-3000 Hz were used. Studies wereperformed without sedation; chloral hydrate(single dose 25 mg/kg) was used in two caseswhere the subjects were toddlers and unable tocooperate.The patient's height, leg length, and latency

values from the lumbar and cortical responseswere used to calculate conduction velocity. Forthe purpose of this study, the results weredivided into four groups. The criteria usedwere (i) presence of a definable response, (ii)clarity of the cortical response, and (iii) centralvelocity of conduction in relation to normalvalues.67 The four groups were defined as:grade 1 =normal; grade 2=present, notdelayed, but poorly defined; grade 3=presentbut delayed; and grade 4=absent.

These group classifications were used tocorrelate the SEPs with the respiratory scoresof apnoea index, total apnoea, central apnoea,and obstructive apnoea times. We alsoattempted to correlate the absolute time of theoccurrence of the evoked response with theapnoea indices.

ResultsCLINICAL HISTORY AND EXAMINATIONAll the subjects had a history of snoring insleep, and those with a history of witnessedapnoea had obstructive apnoea sufficient towarrant intervention. Not all of the subjectslater shown to have repetitive apnoea had ahistory of witnessed apnoea. Even in the mostsevere case, a 21 year old man, symptoms ofexcessive daytime somnolence and witnessedapnoea were not volunteered as they werechronic and considered a part of havingachondroplasia. In the group with respiratoryabnormalities not thought to require treat-ment, the history did not indicate the severityof the abnormality present, and auxiliarysymptoms such as restlessness in sleep,sweating, and loud snoring were not discrimi-natory. None of these individuals had soughtmedical attention for these symptoms, whichon specific questioning included excessivedaytime somnolence and witnessed apnoea.

Clinical examination did not discriminatewhich subjects would have obstruction.None of the group had any evidence of bulbardysfunction to clinical examination, and all ofthe group had brisk tendon reflexes of thelower limbs. Only one of the adults hadsymptoms of headache and upper limb painthat had been diagnosed as secondary tocervical spinal canal stenosis; there were noassociated upper limb neuromuscular abnor-malities. Two of the subjects had a history oflumbar claudication, and one of these hadobstructive sleep apnoea, the other hadrepetitive cycles of mixed apnoea.

SLEEP STUDIESResults of all of the studies are presented inthe table. A total of 20 individuals withachondroplasia were included. An apnoeaindex of greater than five per hour8 was

192

on February 20, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.69.2.191 on 1 A

ugust 1993. Dow

nloaded from

Breathing abnormalities in sleep in achondroplasia

recorded in 15 studies. The majority hacrespiratory abnormalities during sleep iing several disorders ranging from srabnormal EMG activity of the accmuscles of respiration during sleepreference), and periodic cycles of apnoe;

All of the subjects snored during theof their study, and therefore had some o

of partial upper airway obstruction.'evidence for partial upper airway obstron the recorded sleep study was incEMGgg activity (and EMGsm where tialso measured) indicating use of accrespiratory muscles during inspiration.indicators included acute carbon dretention associated with onset of eitheror a particular sleep stage, a characi'flow limited' pattern on the airflow tra1A), and active expiration on EMGd.findings were most pronounced in slovsleep, and were severe in six (40%) ofchildren. Respiration during slow wavein these six children was characterisedpresence of all of these components toiin each period of slow wave sleep, an(

50

_ iJ di, I J Ii,

SW

i other not affected by posture. This pattern ofinclud- obstruction in slow wave sleep was associatedioring, in all six cases with carbon dioxide retention:essory of up to 17 mm Hg (2-27 kPa) recorded on(EMG Tcco2.a. Twelve of the children (80%) demonstratedperiod a pattern of phasic EMGgg in rapid eye move-degree ment sleep (fig 2A). This sleep stage was9 The associated with a notable diminution of theuction tonic EMGgg activity (fig 2B, at change of:reased sleep state) but the children demonstrated par-uis was tial obstruction, without apnoeas, in the-essory presence of this increased hyoid andOther sternomastoid activity, a phenomena not seenlioxide in adults. All other features were unequivocalrsleep, of rapid eye movement sleep during theseteristic intervals with low amplitude high frequencyce (fig EEG, phasic eye movements, and reducedThese EMG activity. In five children no rapid eyev wave movement period was free of this phasicthe 15 EMGgg activity.e sleep A pattern of repetitive central/mixed apnoeaby the was seen in nine children (60%). Thesegether, periods were characteristic of the occurrence ofd were regular central apnoeic events, interrupted by

brief periods of apparently normal breathing.The number of epochs of repetitive apnoeaand the total duration of the combined epochsis included in the table. These cycles were ofvariable duration, extending up to 25 minutesin a 7 year old boy. In five children this patternoccurred during rapid eye movement sleep.Isolated central apnoeas may be normal, butnot repetitive.1I

Complete obstructive apnoea, with com-plete cessation of airflow despite continuing

,| respiratory effort, was observed in a total ofseven (35%) of the group. This led to activeintervention in those where there was either anapnoea index of greater than 20 per hour,repetitive blood gas changes (as measuredusing transcutaneous monitoring) of greaterthan 10% desaturation and greater than 10

Eimm Hg (1.33 kPa) carbon dioxide retention,or an arousal index of greater than 20 per hour.Four children were diagnosed with severeobstructive apnoea, and all demonstratedrepetitive obstruction associated with oxygendesaturation and carbon dioxide retention.This occurred predominantly in rapid eye

Summary of sleep study abnormalities

Apnoea index (per hour)

Figure 1 (A) Upper airway obstruction in slow wave sleep. Sleep study tracings showingslow wave sleep with EEG waveforms of high amplitude and low frequency. Surface EMGactivity ofgenioglossus and diaphragm show phasic inspiratory (EMGgg) activity.Expiratory EMG activity on the trace EMGd shows expiratory activityfrom abdominalmuscles. Airflow signal is measured as an airflow wave and shows a flattened inspiratory(flow limited) waveform (1 cm=7-6 secs). (B) Carbon dioxide retention in slow wave sleep(SWS). Sao2 and Tcco2 levels are recorded simultaneously but on a slow recorder output.There is a phase delay between the output of these signals caused by a combination ofpenposition and signal response time of the measuring device. There is acute carbon dioxideretention during the period ofslow wave sleep, with return to baseline during light sleep. Thenormal range of carbon dioxide variation is 7 mm Hg (0-93 kPa) throughout the night.

Sleep respiratory abnormalitiesTotal(A)Accessory EMG*REMNREM

(B) Cyclic apnoeatDuration (min)Mean % total sleep timeNo of episodes (total)

(C) Apneustic breathsSEP grade

1234

Mean (SD) 18-2 (14 8)(range 0-57)

2012

1011

1232-3 (15 7)7-4 (3-9)10-5 (4 9)

6

11314

*Accessory EMG=phasic activity of EMG of accessorymuscles, sternomastoid, chin, +/- abdominal.tCyclic apnoea=periodic apnoea (not obstructive); valuesshown are mean duration (SD). REM=rapid eye movement;NREM=non-rapid eye movement sleep.

A

Sao2 (%)

Respitrace(abdomen)

Nasal airflow(inspiratory

EEG

EMGgg

EMGd

B

N0c,,

100

80

60-

40-

0)IEEI-9

1 hour

193

on February 20, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.69.2.191 on 1 A

ugust 1993. Dow

nloaded from

Waters, Everett, Sillence, Fagan, Sullivan

movement sleep. There was concurrentevidence of severe obstruction in slow wavesleep in only one of these cases.Three young adults had repetitive apnoea

and hypopnoea. This pattern was pre-dominantly in the wake to sleep transitionstate, and recurred throughout the night inlight sleep stages. This resulted in markedsleep disruption, without significant oxygendesaturation or carbon dioxide retention.One 30 year old woman had no apnoea, buthypopnoeas associated with very higharousability, resulting in high sleep disturbance.

SOMATOSENSORY EVOKED POTENTIALSSEPs were performed in 19 subjects. Theywere not measured in the 12 month old infant,

V A, ' ,iN

EOG 5OpV[AVE if< *J

EEGi A OfEEG 50 gv E

EMGgg 50 AV[Al- -s1

EMGd 50gVC

.I ! 8(

- - vW. 1*.IbAW 1 .lJ-Jo.o.-

1... . , - W . . .. . ._S...............Co " I,, Wil ,l, ai II ii II 'L!Xl w ! 1''''',';

t

Figure 2 (A) Phasic EMGgg in rapid eye movement sleep. Sleep record showing rapideye movement sleep, as indicated by phasic eye movements on the EOG (divergent signals),high frequency low amplitude EEG, and reduced EMGgg activity. Phasic inspiratoryactivity ofEMGgg is seen, and mixed hypopnoeas showing reduced EMGd, reducedRespitrace (sum of chest and abdominal movement), and markedly reduced aitflow.(B) Phasic EMGgg in rapid eye movement sleep, persisting but with augmentation in lightsleep; arrow indicates change ofsleep state. This figure includes a sleep state transition,from rapid eye movement sleep to light sleep at the arrow pointer. There is loss ofphasic eyemovements and appearance of sleep spindles in the EEG signal. EMGgg activity is phasicthroughout, but is augmented after the transition to light sleep. This clearly shows thatgenioglossal activity is reduced, but not lost, and remains phasic during rapid eye movementsleep. (In (A) and (B) 1 cm=10 secs.)

as we had no normal values to correlate withthe results, making such results uninter-pretable. Results are shown in the table. Eightsubjects had abnormal SEPs with four havingabsent responses. Notably, all of these four hadsubstantial apnoea. However, the subjects withnormal responses had a wide range of apnoeafrom none to clinically significant. There wasno clear correlation between either theincidence, type (central or obstructive), orduration of apnoea and abnormalities found onthe studies of SEPs. There was also no correla-tion between the absolute delay of the evokedresponse and the severity of the apnoea(apnoea index or apnoea duration). The childwith the most central apnoea had normalSEPs. Only one of the young adults treated forobstructive apnoea and hypopnoea also hadabnormal SEPs. All the subjects with abnormalSEPs had abnormal sleep study results, but thepresence of an abnormal sleep study (by any ofour criteria) was not associated with abnormalresults of the SEPs.

DiscussionOur study of 20 subjects with achondroplasia,who were not selected on clinical criteria, hasshown that sleep apnoea and sleep disorderedbreathing are extremely common in this con-dition. All of our subjects demonstrated vari-able degrees of upper airway obstruction, andin seven subjects the extent of the apnoea wassufficiently severe to require treatment witheither adenotonsillectomy or nasal continuouspositive airway pressure (CPAP). Previousstudies have reported a high prevalence in clinicpatients with achondroplasia. The probablereason that our study showed a higher level ofobstructive sleep apnoea is that we used fullovernight sleep studies and more recentlyaccepted scoring criteria for use in children.

Stokes et al found a 10% incidence of severerespiratory abnormalities in their clinicpatients.3 In the study by Nelson et al 34%/o ofpatients had abnormal polysomnographicstudies.4 In this study, upper airway obstruc-tion was demonstrated and all of the respira-tory abnormalities demonstrated can beaccounted for by upper airway obstruction.Possible reasons for the different incidence ofsignificant sleep obstruction include eitherdifferent patient selection or that our studyutilised full overnight sleep studies. Further-more, the adult scoring criteria used in otherstudies24 to define significant apnoea wouldexclude some of our clearly illustratedexamples of sleep disordered breathing inchildren and young adults. If we were touse the requirements of desaturation >10%,bradycardia to 50% of baseline, or apnoeagreater than 10 seconds, we would notconsider three of our young adults to havesleep apnoea. These patients had multipleapnoeas/hypopnoeas without oxygen desatura-tions, resulting in frequent arousal, significantsleep disturbance, and daytime somnolence.The presence of such sleep disturbance inthese cases has led us to commence nasalCPAP, with marked clinical improvement.

A

Sao2 (%)100 F

50

Respitrace(sum)

Nasal airflow(inspiratory 1-)

B

Sao2 (%)100

Respitrace(sum)

Nasal airflow ) ry-

inspiratoy 1

EEG 50 gVE

EMGgg

EMGd

1-10 i 'I - -'-' ---- -- ---- - -- - -A-- -Ah-- jL -1- - -, - ., -

-i "i. L.hwljjwj,uliwjil ii

rTTFRIylflnqllfmi ll'"Iff"lff iTTTII ITTI ITI IrTl if Fvq ITTVI "IMIN

F-06 50 9VE....rl,.j

-.d ., .. --m-04 optip-MAIMI'll 0- -4

-,

194

II.,#"'. /-1 .

v

IIII II II lil III II Ilillitril III 1"I III iql II,,. III'mu m ilm

IIlr., A

---,i --) ""\-i \,

'i

,1. -111, -1 J,

-'fl -illILL.

on February 20, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.69.2.191 on 1 A

ugust 1993. Dow

nloaded from

Breathing abnormalities in sleep in achondroplasia

There are no adequate data on theprevalence of apnoea in the various age rangescovered by our patients, so it is uncertain ifthe abnormalities are coincidental. In recentprevalence estimates in the age and sex groupclinically thought to dominate sleep apnoea(men aged 40 to 65), the prevalence has beenestimated to be up to 10%.11 All of our patientswere younger than this known high risk group,and half were female. Clearly, on any of thesecriteria it is highly likely that sleep disorderedbreathing occurs much more commonly insubjects with achondroplasia.

Obesity is a known contributing factor inadult sleep apnoea. Our subjects varied inphysiognomy from thin to solid, but none

appeared obese in the familiar context of adultobstructive apnoea. The body mass indexmeasurement was of limited value as there wasno basis for comparison in dwarves.

This particular group provided a uniqueopportunity to compare the responses ofchildren to that of (young) adults. The charac-teristic physiological picture of adult sleepapnoea is repetitive upper airway obstructionassociated with arousal at the cessation of theapnoea. The typical patient profile is an over-

weight middle aged male drinker, with womenbeing affected only in the postmenopausalperiod. Children, in contrast, 'sleep through'upper airway obstruction, and their responsesto upper airway obstruction differ qualitativelyfrom our experience of adult apnoea. These

children progressed through all sleep stages(including rapid eye movement) withoutarousing, despite significant upper airwayobstruction.

Unlike adults, the infants maintainedEMGgg activity during obstructed breathing inrapid eye movement; possibly this EMGggresponse allowed the normal sleep stage toprogress without arousal. Complete obstructionand apnoea appear to be prevented by maintain-ing upper airway tone, so forestalling arousaland consequent disruption of rapid eye move-ment sleep (fig 2). Clearly, the partial upperairway obstruction caused a reduction inalveolar ventilation and the consequent bloodgas changes activated chemoreflexes leading toactivation of expiratory muscles and EMGgg.Other potential mechanisms include upperairway sensory reflexes resulting in increasedupper airway tone'2 13; reflexes that this studysuggests are more active in children or are main-tained in rapid eye movement sleep in thepaediatric population where they activate thisphasic muscle activity. This phasic EMGresponse to obstruction is not seen in adults.The difference may simply lie in the ability ofchildren to respond to these reflexes, withspecific phasic increase of muscle activity,despite general loss of muscle tone in rapid eyemovement sleep. While adults may compensate

for their upper airway obstruction in slow wave

sleep with phasic upper airway muscle(EMGgg) activity, this activity did not extend torapid eye movement sleep. Rapid eye movementsleep was markedly disrupted in the adults in thepresence of obstruction, consistent with theresults of other studies of adult sleep apnoea."I

These young adults also demonstrated anapparently heightened arousal response toobstruction in light sleep, resulting in signifi-cant sleep disruption. Two premenopausalwomen aged 19 and 30 years respectively, hadrepetitive hypopnoea leading to arousal, with-out any demonstrable blood gas changes ontranscutaneous measures. Presumably, reflexesother than chemoresponses, for exampleoriginating in the upper airway'3 causedarousal in these subjects. The combined effectof these responses to obstruction resulted insleep disturbance sufficient to cause markedabnormalities of both daytime function anddocumented sleep records. There was a clearresponse to treatment with nasal CPAP in boththese parameters. The abnormalities describedare not detectable without a full overnightsleep recording as used in this study, and pre-menopausal women would usually be con-sidered at very low risk of developing sleepapnoea.These results may simply represent the

evolution of sleep apnoea, by cross sectionalsurvey of a population at high risk, and there-fore be applicable to populations withoutachondroplasia. Three of our patients havedemonstrated the typical picture of sleepapnoea and the associated blood gas dis-turbances. In affected children, we describeapnoeas being most marked in rapid eye move-ment sleep, but associated with new findings ofsleep not disrupted, despite significant bloodgas fluctuations with obstructive episodes. A21 year old man demonstrated classic obstruc-tive sleep apnoea on the polygraphic recording,but with atypical features of no oxygen desatu-ration to less than 90%/o and yet carbon dioxideretention (confirmed on a morning arterialblood gas).

Such a rise in Tcco2 without a correspond-ing fall in Sao2 will occur when the initialarterial oxygen tension has been high, becausethe relatively small change in partial pressureof oxygen does not alter the saturation at thishigh level. In fig 1B, the saturation has fallenapproximately 3% during the period corres-ponding to that of slow wave sleep and carbondioxide retention. A time lag occurs betweenthe changes in Sao2 and Tcco2 on our record-ing because of technical differences in themethods of measuring and recording the twovariables. Acute desaturations do occur inassociation with more rapid carbon dioxideretention. We were careful to obtain blood gascross checks of our Tcco2, in order to confirmtrends indicating significant carbon dioxideretention. We believe that the apparent dis-crepancy sometimes seen, of carbon dioxideretention without a corresponding fall in Sao2,is in the inaccuracy of the oximeter in its highranges. For example, it does not truly andreliably record saturation in the 90% range andtherefore we would not rely on this at thesehigh values. Our blood gases have confirmedthe key point that some of these subjectsdevelop carbon dioxide retention.Both Reid et al2 and Hecht and Butler'4

have postulated that the brainstem compres-sion in achondroplasia is the principle cause of

195

on February 20, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.69.2.191 on 1 A

ugust 1993. Dow

nloaded from

Waters, Everett, Silence, Fagan, Sullivan

the sleep disordered breathing. Nelson reportsa fall in apnoea index after decompressivesurgery, on follow up sleep studies in four ofeight patients. The average apnoea index forthat particular group was 11 6 per hour, andfell to 7-2 per hour.4 Our results showing ahigh incidence of sleep breathing disorderssupport the possibility that brainstem com-pression is an important contributing factor.The high frequency of central apnoea in ourstudy group could argue in support of brain-stem dysfunction disrupting respiratorycontrol. Another feature supporting thehypothesis is apneustic breaths'5; this was seenonly in six of our subjects. However, it is alsopossible that airway/midfacial dimensionabnormality is an important factor.

In the available literature, the most commonpresentation for SEP values is of absolute delayin msec. Because of the wide range of ages inthe patient population studied here, we feltthat this measurement was too vulnerable tothe effects of development and growth inheight, to be valid across the entire group.Normal values for SEP velocities have beenprovided by studies in normal Japanese school-children, and so we felt that the classificationofnormal or abnormal would be more accurateon these criteria. This choice of classificationhas affected the decision to call the valuesnormal or abnormal but seems to be the mostaccurate. Using the criteria of absolute delay inmsec, all of the subjects in this group wouldhave been classified as having abnormal SEPresults. While this fits with our finding of allsubjects have respiratory abnormalities, therewas still no qualitative relationship of severitybetween the two abnormalities.Our subjects as a group had clear evidence

of abnormal SEPs but we showed no simplecorrelation between these and the respiratoryabnormalities. The presence of absent SEPsdid correlate with the presence of disorderedbreathing in sleep. But, the presence, type, orseverity of the respiratory abnormalities foreither individuals or the group as a whole didnot relate to abnormal SEP responses. Clearlythe abnormal SEPs reflect dysfunction at thecervicomedullary junction, and it is probablethat more subtle damage is involved in otherbrainstem reflex systems such as respiratorycentres and upper airway muscle control.Decompression of the foramen magnum(posterior cranial fossa) has been proposed andused as a treatment for sleep apnoea on thebasis that brainstem dysfunction underlies theupper airway obstruction. We do not under-take this surgical form of treatment inachondroplasia because we remain uncertainof whether the aetiology of the apnoea isneurological or in the structure of the upperairway (or a combination of both). Rather, webelieve that the presence of these brainstem

and neurological problems potentiates theupper airway obstruction that is common inthis disorder, but an abnormal upper airway isstill the primary cause.

CONCLUSIONUpper airway obstruction occurs in a significantproportion of individuals with achondroplasia,and this proportion is only determined accu-rately when overnight studies are undertaken.

While cases of typical obstructive sleepapnoea occur with oxygen desaturation, andcarbon dioxide retention do occur, there maybe other respiratory disturbances present thatwill result in significant sleep disruption with-out alteration of blood gases (on transcuta-neous measurement).With the high frequency of airway obstruc-

tion seen in this and other groups of peoplewith achondroplasia, airway obstruction mustbe considered an integral part of the syndromerather than an occasional complication, andfull sleep studies are required to demonstratethe abnormalities present. SEPs are not anappropriate screening test for the presence ofrespiratory abnormalities.We would like to thank Mrs Cheryl Cochineas for performingthe SEP studies in these subjects.

1 Cohen MM-Jr, Walker GF, Phillips C. A morphometricanalysis of the craniofacial configuration in achondro-plasia. J Craniofac Genet Dev Biol Suppl 1985; 1: 139-65.

2 Reid CS, Reed EP, Kopitz SE, et al. Cervicomedullarycompression in young patients with achondroplasia: valueof comprehensive neurological and respiratory evaluation.JPediatr 1987; 110: 522-30.

3 Stokes DC, Phillips JA, Leonard CO, et al. Respiratorycomplications of achondroplasia. J Pediatr 1983; 102:534-41.

4 Nelson FW, Hecht J, Horton WA, Butler IJ, Goldie WD,Miner M. Neurological basis of respiratory complicationsin achondroplasia. Ann Neurol 1988; 24: 89-93.

5 Rechtschaffen A, Kales A. A manual of standardised termin-ology, techniques and scoring systems for sleep stages ofhumansubjects. UCLA, Los Angeles: Brain InformationService/Brain Research Institute, 1968.

6 Nelson F, Goldie WD, Hecht JT, et al. Short-latencysomatosensory evoked potentials in the management ofpatients with achondroplasia. Neurology 1984; 34:1053-8.

7 Mutoh K, Okuno T, Mikawa H, Hojo H. Maturation ofsomatosensory evoked potentials upon posterior tibialnerve stimulation. Pediatr Neurol 1988; 4: 342-9.

8 Guilleminault C, Tilkian A, Dement WC. The sleep apneasyndromes. Annu Rev Med 1976; 27: 465-84.

9 Lugaresi E, Crignotta F, Coccagna G, Montagna P. Clinicalsignificance of snoring. In: Saunders NA, Sullivan CE,eds. Sleep and breathing. New York: Marcel Dekker, 1984:283-98.

10 Sheldon SH, Spire JP, Levy HB. Sleep disordered respira-tion in childhood. In: Sheldon SH, Spire JP, Levy HB,eds. Pediatric sleep medicine. Philadelphia: WE Saunders,1992: 136-50.

11 Krieger J. Obstructive sleep apnea: clinical manifestationsand pathophysiology. In: Thorpy MJ, ed. Handbook ofsleep disorders. New York: Marcel Dekker, 1990: 259-84.

12 Plowman L, Lauff DC, Berthon-Jones M, Sullivan CE.Waking and genioglossus muscle responses to upper air-way pressure oscillation in sleeping dogs. J Appl Physiol1990; 68: 2564-73.

13 Henke KG, Sullivan CE. Effects ofhigh-frequency pressurewaves applied to upper airway on respiration in centralapnea.JAppl Physiol 1992; 73: 1141-5.

14 Hecht JT, Butler U. Neurologic morbidity associated withachondroplasia. J Child Neurol 1990; 5: 84-97.

15 Mador MJ, Tobin MJ. Apneustic breathing. A characteris-tic feature of brainstem compression in achondroplasia.Chest 1990; 97: 877-83.

196

on February 20, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.69.2.191 on 1 A

ugust 1993. Dow

nloaded from