Obstructive sleep apneas in relation to severity of cervical spinal cord injury

Obstructive sleep apneas in relation to severity of cervical spinal cordinjury

B Klefbeck1, M Sternhag2, J Weinberg3, R Levi4, C Hultling4 and J Borg2

Departments of 1Physical Therapy, Karolinska Institute, Stockholm, 2Clinical Neuroscience, Karolinska Hospital,Stockholm, 3Neurology, Huddinge Hospital, Huddinge, 4Spinalis SCI Research Unit, Karolinska Institute, Stockholm,Sweden

Thirty-three subjects (28 men, ®ve women) with complete or incomplete cervical cord injuryrepresenting a wide range of neurological impairment were investigated with regard to theprevalence of Obstructive Sleep Apnea (OSA). The relation between OSA and neurologicalfunction, respiratory capacity, body mass index and symptoms associated with OSA werestudied. Overnight sleep recordings employed combined oximetry and respiratory movementmonitoring. Pulmonary function tests included static and dynamic spirometry, maximal staticinspiratory and expiratory pressures at the mouth. The subjects answered a questionnaireconcerning sleep quality and tiredness. The prevalence of OSA was 15% (5/33) in this non-obese cervical cord injury study population. Nine percent of the subjects (3/33) ful®lled thecriteria for obstructive sleep apnea syndrome, but daytime sleepiness or fatigue were alsocommon in subjects without OSA. There was an inverse correlation between oxygendesaturation index and American Spinal Injury Association (ASIA) motor score in thesubjects with complete injury, while there was no such correlation in the whole study group.There were signi®cant correlations between maximal inspiratory and expiratory pressures andvital capacity and between ASIA motor score and vital capacity.

Keywords: ASIA motor score; oxygen desaturation; sleep apnea syndrome; sleep; tetraplegia

Introduction

Cervical spinal Cord Injury (CCI) might causerespiratory insu�ciency due to respiratory muscleparesis. Lesions above the C5 level are most severedue to paresis of the diaphragm and might cause areversible or persistent need for mechanical ventila-tion.1,2 In CCI patients with respiratory muscle paresis,but without need for assisted ventilation, an increasedincidence of Obstructive Sleep Apneas (OSA), whencompared to a non-injured population, has beenreported.3 ± 6

OSA is characterised by sleep-related intermittentupper airway obstruction which may be associated withepisodes of oxygen desaturations and sleep fragmenta-tion.7 In obstructive sleep apnea syndrome this iscombined with symptoms such as snoring, excessivedaytime sleepiness and cardiovascular sequelae.8 Thede®nition of respiratory events di�ers slightly indi�erent studies.9 ± 12 Young et al10 and Gislason etal11 de®ne the apnea-hypopnea score as either acomplete cessation of air¯ow lasting 10 s or more(apnea) or a reduction in respiratory air¯ow accom-panied by a decrease of 4% or more in oxygensaturation (hypopnea). The apnea-hypopnea score isthe average of apneas and hypopneas per hour of sleep.

Using this score, Young et al found that 9% of womenand 24% of men had an apnea-hypopnea score of 55.When they combined this with daytime hypersomno-lence, they estimated the prevalence of obstructive sleepapnea syndrome to 2% among women and 4% amongmen. In a community-based study by Gislason et al11

the prevalence of obstructive sleep apnea syndrome wasestimated at 1.3% among Swedish men.Patients with cervical cord injuries often complain

of sleeping problems and daytime sleepiness.13

McEvoy et al3 reported that nine of 40 patients withspinal cord injuries at C8 level and above hadobstructive apneas and six of the 40 patientscomplained of daytime sleepiness which was directlyrelated to the frequency of sleep arousals. Cahan et al5

noted that ®ve of six quadriplegic patients with night-time hypoxia had an increased daytime sleepinesscompared to six of ten quadriplegic patients withnonhypoxia.The pathophysiology of OSA is not fully under-

stood.7 Obesity, ethanol, male gender, age andconditions that narrow the upper airway areconsidered strong risk factors in the non-injuredpopulation.14,15 In the CCI population, data concern-ing relations between sleep-disordered breathing andcharacteristics of the CCI is limited. An associationbetween time spent below 90% SaO2 and body mass

Correspondence: B Klefbeck, Department of Physical Therapy,Karolinska Institute, Novum, S-141 57 Huddinge, Sweden

Spinal Cord (1998) 36, 621 ± 628 1998 International Medical Society of Paraplegia All rights reserved 1362 ± 4393/98 $12.00

http://www.stockton-press.co.uk/sc

index, level of injury and maximum expiratorypressure6 as well as between sleep-disordered breath-ing and neck circumference and supine sleep posture3

have been reported.The aims of this study were to further elucidate the

prevalence of OSA and obstructive sleep apneasyndrome in subjects with CCI by use of a screeningmethod for OSA and to study the relation toneurological function, respiratory capacity, body massindex and sleeping position.

Subjects

The criteria for participation were traumatic tetraplegia,one year or more post injury, motor injury level C4-T1with complete or incomplete tetraplegia (A-D accordingto the American Spinal Injury Association (ASIA)impairment scale)16,17 20 ± 60 years of age, no pulmon-ary or cardiovascular disorders. Subjects with ventila-tory treatment due to hypoventilation were excluded.The subjects were invited to participate in the study

as a complement to an annual routine follow up (May1995 to November 1995) o�ered to all spinal cordinjured subjects living in the Stockholm region.13 Ofthirty-seven consecutive subjects, 33 subjects gaveinformed consent. There were 28 men and ®vewomen. Mean age was 37.9 (range 22 ± 59) years andmean duration of injury was 13.9 (range 1 ± 33) years.Individual data in groups of complete and incompleteinjury are given in Table 1.

Methods

Neurological examination for determination of com-pleteness according to the ASIA impairment, sensoryand motor scoring scales16,17 was performed.Overnight sleep recording was performed in a single

room in hospital. The sleeping position was repeatedlyobserved. However, the majority of the subjects sleptall night in the same position because they were unableto turn. A simpli®ed sleep apnea investigationconsisting of combined ear oximetry (Ohmeda Biox3740, Health Care, Louisville, USA) and respiratoryand body movement monitoring (Static ChargeSensitive Bed) as described by Svanborg et al8 wasperformed. The body movement monitoring andoximetry data were collected by use of a pen-recorder(Graftec Thermal Array-corder WR 770, GraphiticCorporation, Yokohama, Japan) using slow paperspeed (10 mm/min) to clearly demonstrate changes inoxygen saturation (SaO2) and respiration movementpatterns. Oxygen Desaturation Index (ODI) wasde®ned as the average number of oxygen desatura-tions of 4% or more per sleeping hour.8 Periods ofobstructive apneas form a typical diamond-shapedpattern in the body movement monitoring (periodicrespiration pattern), usually accompanied by repetitiveoxygen desaturations.8 The recordings were analyzedmanually. The number of signi®cant desaturations andtotal time with periodic respiration in minutes were

calculated. The nadir SaO2 value (the lowest valueduring the night) was also noted. Sleeping time wasesimated from the patterns in the movement recording.Total sleeping time had to be 4 h or more, otherwisethe examination was repeated. Criterion for pathologicrecording, ie indicating OSA, was the combination ofODI56 and 445% periodic respiration time out ofthe total estimated sleeping time.8 If the subjects wereunable to sleep in hospital, they were o�ered to sleepat home with the ApnoLog system (C-A Tegner Inc,Stockholm, Sweden) including pulse oximetry (Biox3760), respiration and body movements (PVDF

Table 1 Characteristics of the subjects, nos 1 ± 17 completeinjuries, nos 18 ± 33 incomplete injuries

Age Time Level of injurySubject Gender (years) (years) (motor) (sens)

12345678910

MaleMaleMaleFemaleMaleFemaleMaleMaleMaleMale

26465925362625294628

12830920461327

C4C7

C5/C6C5C5C6C6

C6/C5C7C6

C3C7C3C4C6C6C5C5C6C5

11121314151617

MaleMaleMaleMaleMaleFemaleMale

30322423225128

2715342188

Th1Th1C6C6C5C5C4

C7C7C5C4C4C3C4

MeanSD

32.711.0

12.611.0

1819202122232425

MaleMaleMaleFemaleMaleMaleMaleMale

3024433934455453

112254153334

C7C6C5C5C6C8*C5

C7C6C4C4C4C7C3C3

2627282930313233

MaleMaleMaleMaleMaleMaleFemaleMale

3943535047583053

16241033233227

C6C5C7C7C6C4C5C5

C7C5C7C7C4C4Th5Th7

MeanSD

43.410.1

15.312.2

Total MeanSD

37.911.8

13.911.5

Time (years)=time elapsed since injury, motor=motor levelof injury, sens=sensory level of injury, *=not possible toassess with ASIA, SD=standard deviation

Obstructive sleep apnea in tetraplegiaB Klefbeck et al

622

pressure sensor mattress).18 To make certain bothmethods were correlated, four subjects performedrecordings both at home and in hospital. The resultsof the two methods were in good accordance even inestimating sleeping time.A pulmonary function test with static and dynamic

spirometry was performed and Maximal InspiratoryPressure (MIP) and Maximal Expiratory Pressure(MEP) at the mouth (Precision Medical Ltd, UK)19 ± 21

were measured. Body mass index (kg/m2) wascalculated from the subject's weight in kg and lengthin meters.22 For the Swedish population, body massindex between 20.1 and 25.0 in men and 18.7 and 23.8in women is considered normal.22

A detailed drug report including present use ofantispastic and sedative drugs was obtained.At the time of the sleep recordings, a questionnaire

concerning sleep and tiredness during the last threemonths was distributed. The Basic Nordic SleepQuestionnaire23,24 was used together with threequestions from a Sleep and Health Questionnairedeveloped by Kump et al (Likert scale 1 ± 6).25 InKump's Questionnaire, the subject was asked toquantify the extent of her/his sleepiness, energy level,or performance impairment due to sleepiness.Regression analyses were used to describe the

relation between ODI, periodic respiration, ASIAmotor score and pulmonary function data. Pearson'sor Spearman's correlation coe�cient was used toidentify the relationship between variables. To analysethe answers from the questionnaire, they weresubjectively divided into two groups (normal andpathologic) and a non-parametric test, Mann ±Whitney, was used. P50.05 was considered sign-®cant.

Results

Sleep respiratory recordingsThirty-one of the subjects performed the sleeprecordings during one night. One patient had to dothe sleep recording an extra night due to too shortsleeping time during the ®rst night. One further subjecthad sleeping di�culties in hospital and performed asleep recording at home.Mean sleeping time was 6.4 (SD 1.0) h and ranged

4.0 ± 8.2 h. Further individual sleep recording data arepresented in Table 2. Mean awake SaO2 was 96.2%(SD 1.2) and ranged 94 ± 98. Mean nadir SaO2 was88.4% (SD 6.4) and ranged 69 ± 95.Median ODI was 1 and ODI ranged 0 ± 29. Median

periodic respiration was 4% and ranged 0 ± 89%.When only subjects with ODI40 (n=17) wereconsidered, there was a statistically signi®cantcorrelation between ODI and nadir SaO2 (rs=0.82,P50.05). When all subjects with ODI40 andperiodic respiration445% were considered therewere statistically signi®cant correlations betweenODI and periodic respiration (rs=0.50, P50.05), as

well as in the two subgroups with complete andincomplete injuries (rs=0.67, P50.05 and rs=0.68,P50.05). Five of the 33 subjects (15%), all men, hadODI46 and periodic breathing445% (values thusconsidered as pathological).8 Thus ®ve of the 28 men(18%) ful®lled the diagnostic criteria for OSA. Fiveadditional subjects had pathological amounts ofperiodic breathing of clearly obstructive type (21 ±71%). A sixth subject had ODI=5 and periodicrespiration of 19%. These six subjects are regarded asborderline cases.8

ASIA motor score and sensory score and ODI/sleeprespiratory recordingsMean ASIA motor score was 36.6 (SD 26.9) and ranged0 ± 90. Mean ASIA sensory score was for light touch 33.7(SD 26.7) and ranged 8 ± 110 and for pin prick 40.0 (SD29.3) and ranged 8 ± 110. Individual data are given inTable 2. There was no signi®cant correlation betweensleep respiratory pathology and motor or sensory scores.Of the ®ve subjects with OSA, two had ASIA motorscores of 0 and 9, while the other three subjects hadASIA motor scores of 43, 68 and 72. Two had completeinjury and three had incomplete injury.In the 17 subjects with complete injury (cf. Table 1)

there was a signi®cant (r=0.48, P50.05) correlationbetween ODI and ASIA motor score.

Vital capacity, MIP, MEP and ODI/sleep respiratoryrecordingsSpirometric pulmonary function tests were performedin 27 patients, and MIP and MEP were performed in30 patients.Mean Vital Capacity (VC) was 3.39 (SD 1.0) and

ranged 1.21 ± 4.85 litre. Mean forced expiratoryvolume during 1 s was 2.95 (SD 0.8) and ranged1.15 ± 3.90 litre. Mean MIP was 76.7 (SD 28.2) andranged 17 ± 131 cm H2O. Mean MEP was 70.4 (SD31.7) and ranged 19 ± 149 cm H2O. Individual data aregiven in Table 3. There were no signi®cant correlationsbetween ODI or periodic respiration and any of thelung function tests. There were signi®cant correlationsbetween MIP and VC (r=0.59, P50.01), MEP andVC (r=0.66, P50.001) and between ASIA motorscore and VC (r=0.60, P50.01).

Body mass index, age and ODI/sleep respiratoryrecordingsMean Body Mass Index (BMI) was 21.2 (SD 3.8) andranged 13.6 ± 30.0 kg/m2. It should be noted that noneof the subjects was morbidly obese (BMI430).Individual data are given in Table 4. Three men hadBMI425 and one female had BMI=30.0, none ofthese had ODI46. The age of the subjects ranged from22 ± 59 years. There were no signi®cant correlationsbetween ODI or periodic respiration and BMI, age ortime elapsed since injury.

Obstructive sleep apnea in tetraplegiaB Klefbeck et al

623

Sleeping position and ODI/sleep respiratory recordingsTwenty-seven of 33 subjects slept in the supineposition and only ®ve of them alternated betweenthe supine and lateral positions or between the proneand supine positions. Six subjects spent all night inthe lateral position. Individual data are given in Table4. Of ®ve subjects with OSA, four spent all night in

the supine position and the ®fth subject spent part ofthe night in the supine position. Of the six subjectswith borderline results, four slept all night in thesupine position, one altered between the lateral andsupine position and the sixth subject slept in thelateral position. None of the 22 subjects slept in theelevated position.

Table 2 Individual sleep breathing data and motor and sensory scores, subjects nos 1 ± 17 complete injuries, subject nos 18 ± 33incomplete injures

PeriodicAwake Nadir respiration Motor Sensory score

Subject SaO2 (%) SaO2 (%) ODI (%) score light t pin p

12345678910

95979798969597979697

77888192869288959391

292110101001

89077444303011

03898191816144019

12301814201814162922

12301815221214162922

11121314151617

95969595949894

93929093879575

0000105

003020019

4648233415114

2633172217810

2533202221810

MeanMedianSD

96

1.3

88.7

6.00 4

21.3

14.7

19.2

7.1

19.4

7.2

1819202122232425

9696969897949896

9292909291807969

0100123722

0416309744879

2544122417437268

2825361647831260

285952626282±60

2627*282930313233

9696959897959798

8093938990919493

50021001

71002124504

6876903087±8865

74110713486137614

811101083574438014

MeanMedianSD

96.4

0.2

88

7.11 12.5

53.9

27.4

49.1

31.2

63.3

27.1

Total MeanMedianSD

96.2

1.2

88.4

6.41 4

36.6

26.9

33.7

26.7

40.0

29.3

Awake SaO2=oxygen saturation at start of the registration, Nadir SaO2=lowest registered saturation, ODI=total number ofdesaturations divided by the sleeping time in hours, light t=light touch, pin p=pin prick, *=home sleep recording, ±=missingvalue, SD=standard deviation

Obstructive sleep apnea in tetraplegiaB Klefbeck et al

624

Drugs and ODI/sleep respiratory recordingsEleven subjects used baclofen (70 ± 125 mg) anddiazepam (1.25 ± 5 mg) for reducing spacticity. Onesubject with OSA and three borderline cases usedbaclofen. A routine clinical evaluation revealed nohistory of alcohol or psychotropic drug abuse.Individual data are presented in Table 4.

Questionnaire, snoring and ODI/sleep respiratoryrecordingsThree of the ®ve subjects with ODI46 and seven ofthe 28 subjects with ODI56 reported excessivetiredness and sleep disturbances. Three of 33 (9%)subjects therefore ful®lled the criteria for obstructivesleep apnea syndrome. Three subjects with ODI46

Table 3 ODI and data from pulmonary function tests, subjects nos 1 ± 17 complete injuries, subjects nos 18 ± 33 incompleteinjuries

VC FEV1,0 FIV1,0 FRC MIP MEPSubject ODI (litre) pred (litre/s) (litre/s) (litre) (cm H20) (cm H20)

12345678910

292110101001

1.213.78±

1.223.002.923.932.92±

3.81

2075

2953746451

66

1.203.75±

1.152.462.673.642.70±

3.23

±3.52±

1.022.622.683.312.42±

3.20

3.965.24±

1.863.102.704.06±±

4.18

4868704251805874±71

1963392538513263±71

11121314151617

0000105

±4.214.42±

2.661.282.20

7572

473637

±±

3.75±

2.431.19±

±±±±

2.42±±

±2.523.88±

3.58±±

1311291075662±41

1091031055832±48

MeanMedianSD

02.89

1.1

53.8

18.9

2.56

1.0

2.64

0.8

3.5

1.0

72.5

28.6

57.1

29.2

1819202122232425

0100123722

±4.133.743.113.314.853.964.57

67689266968193

±3.683.232.863.103.803.253.90

±3.913.012.862.994.473.663.83

±3.853.483.293.943.943.414.07

66123114478885±

101

679774446580±

1042627282930313233

40021001

4.063.833.343.32±

3.053.954.69

84727462

7992107

3.323.402.693.11±

2.043.323.85

3.693.182.942.97±

1.543.753.65

2.523.762.664.29±

2.872.364.31

100557392109178163

1087814975118307888

MeanMedianSD

13.85

0.6

80.9

13.5

3.3

0.5

3.32

0.7

3.48

0.7

80.9

28.2

83.7

29.2

Total MeanMedianSD

13.39

1.0

67.9

21.2

2.95

0.8

3.07

0.8

3.49

0.8

76.7

28.2

70.4

31.7

ODI=total number of desaturations divided by the sleeping time in hours, VC=Vital capacity, pred=percent of predictednormal, FEV1,0=Forced expiratory volume during 1 s, FIV1,0=Forced inspiratory volume during 1 s, FRC=Functionalresidual capacity, MIP=Maximal inspiratory pressure, MEP=Maximal expiratory pressure, ±=missing values, SD=standarddeviation

Obstructive sleep apnea in tetraplegiaB Klefbeck et al

625

and four subjects with ODI56 reported of snoring.Individual data of symptoms and snoring are presentedin Table 4. There were no statistically signi®cantcorrelations between ODI and data from the ques-tionnaire.

Discussion

The study group can be considered as unselected andrepresentative for the cervical cord injured subjects inthe Stockholm area as the subjects were recruited

Table 4 ODI in relation to BMI, sleeping position, symptoms, snoring and used muscle relaxant, subjects nos 1 ± 17 completeinjuries, subjects nos 18 ± 33 incomplete injuries

Subject ODIBMI

(kg/m2)Sleepingposition Symptom Snoring

Musclerelaxant Dosage

12345678910

292110101001

14.920.122.515.523.423.716.219.122.014.8

supprone/sup

suplatsuplatsupsuplatsup

+

+

+

+

+

+

bac

bacbacbac

125 mg

75 mg30 mg75 mg

11121314151617

0000105

21.617.619.918.620.116.925.2

latlatsupsupsupsup

lat/sup

+

+bac

bac

75 mg

100 mg

MeanMedianSD

019.5

3.2

1819202122232425

0100123722

23.020.524.619.621.622.523.121.1

supsupsupsupsupsup

lat/supsup

++ +

dia

bac+dia

bac

5 mg

70+1.25 mg

75 mg2627282930313233

40021001

23.423.527.713.625.623.030.024.9

supsup

lat/suplatsup

lat/supsupsup

+++

+

+

+bacdia

125 mg5 mg

MeanMedianSD

123.0

3.6

Total MeanMedianSD

121.2

3.8

ODI=total number of desaturations divided by the sleeping time in hours, BMI=Body Mass Index, sup=supine, lat=lateral,+=symptom present as excessive tiredness and sleep disturbances or snoring, bac=baclofen, dia=diazepam, Dosage=dosageof muscle relaxant, SD=standard deviation

Obstructive sleep apnea in tetraplegiaB Klefbeck et al

626

consecutively from the yearly routine follow-up. Therewere 28 men (85%) in this study, as compared to 81%men in an inventory of the spinal cord injuredpopulation in the same region in Sweden.13 Subjectsincluded had complete or incomplete injuries atdi�erent cervical levels representing a wide spectrumof neurological and respiratory dysfunction.In the present study the prevalence of cases who

ful®lled the laboratory criteria for OSA was 15%(5/33), that is within the same order as in a communitybased study by Young et al10 of non-injured subjectswhere the prevalence of OSA was 9% in women and24% in men. However, when comparing this data twopoints must be considered. First, Young's as well asother studies of non injured populations include obesesubjects known to have an increased risk forOSA.12,14,26,27 In our study there were no morbidlyobese subjects. Four subjects had BMI above normalrange, corresponding to overweight for three men andobesity for one woman,22 and those four all hadnormal nocturnal respiration. In a population-basedsurvey of non-injured subjects by Kripke et al27 theprevalence of ODI520 in subjects with BMI529 was3%, while the corresponding prevalence in our studywas 9%. Thus, the prevalence of OSA seems to behigher in non-obese CCI-subjects than in non-obese,non-injured subjects. Further, in our study there weresix additional border-line cases with high amounts ofobstructive breathing but virtually normal oximetry.Some of these might have more than 10 s cessation ofair¯ow and thus ful®l the criteria for OSA accordingto Young et al.10

There was a signi®cant negative correlation betweenthe degree of obstructive respiration and ASIA motorscore in the subpopulation of CCI subjects with moresevere injuries. Previous studies also indicate that theprevalence of OSA increases with more severe injuries(cf. further discussion below).There was no correlation between ODI and data

from the questionnaire. Ten subjects reported symp-toms in accordance with obstructive sleep apneasyndrome. Three of them had OSA and reportedsleep disturbances and excessive daytime tiredness/sleepiness and snoring and thus ful®lled the criteria forobstructive sleep apnea syndrome, thus the estimatedprevalence of obstructive sleep apnea syndrome in thestudy population was 9% (3/33), that is higher thanthe estimated 1.3% in a Swedish prevalence study11 ofnon injured subjects. Seven subjects reported sleepdisturbances and excessive tiredness, but had normalsleep recordings, indicating that other causes ofdisturbed sleep are common in this group.The observed prevalence of OSA in the present

study was lower than reported in previous studies ofCCI subjects.3 ± 6 This might be due to di�erentmethods and patient selection. The method used forscreening of OSA and the criteria used for classifica-tion of OSA in the present study are well establishedand only slightly di�erent from those in some of theprevious studies3 ± 5 and should have only marginal, if

any, signi®cance for the di�ering results. However,previous studies have included subjects with a highermean age and/or more severe injury. The observedprevalence of OSA was 22% in a study by McEvoy etal3 in CCI-subjects with Frankel grades A ±C and amean vital capacity of 54.3% of predicted normal,45% in a study by Short et al4 in CCI-subjects whohad a mean age of 58.5 years, a median vital capacityof 2.47 litres and Frankel grades A ±C and 38% in astudy by Cahan et al5 in subjects with a mean age of48 years. Flavell6 reported severe nocturnal desatura-tions in 30% in subjects with injury at C4 ±C6 levelwith Frankel grades A ±B and a mean vital capacity of45.7% of predicted normal. Thus, we suggest thathigher age and more severe injury might explain thehigher prevalence of OSA observed in these previousstudies. This is also in agreement with our observationof an inverse correlation between ODI and ASIAmotor score in the subjects with complete injury in thepresent study and thus in accordance with previous®ndings where the level of injury or vital capacity wascorrelated to OSA.5,6

It has been shown that normal subjects may exhibitOSA after topical oropharyngeal anesthesia28 and thatthe temperature sensitivity is selectively impaired inthe oropharynx of patients with obstructive sleepapnea syndrome.29 The CCI subjects had no sensoryinvolvement of the oropharynx and there was nosigni®cant correlation between ODI or periodicbreathing and degree of sensory loss, as de®ned inthe ASIA sensory score. This indicates that the sensorydisturbance due to spinal cord injury has nosigni®cance for OSA.There was a signi®cant correlation between ASIA

motor score and VC, as expected, as well as betweenVC and MEP and between VC and MIP in the presentstudy. The signi®cant correlation seen between VC andMIP is in contrast to the study of Roth et al,30 whodid not ®nd any correlation between these parametersin 52 patients with complete C4-Th6 spinal cordinjury. The discrepancy might be explained by agreater range of motor functions in the presentstudy, which included subjects with both completeand incomplete injuries.Of the subjects with OSA four of ®ve spent all night

supine, one subject was supine part of the night, andhad ODI=7. However, the majority22,33 of thetetraplegic subjects in the present study had difficul-ties in changing sleeping position during the night andthus spent all night in the supine position. Previousstudies of non-injured subjects indicate that the supineposition increases the risk for OSA31 while the lateralposition reduces the risk.32 Even if our data does notpermit any conclusion, it is in accordance with these®ndings.Eleven of 33 subjects in the present study were using

muscle relaxants but only one of them was classi®ed asOSA and three as borderline cases. In the study ofShort et al4 six of 22 patients with more than ®vedesaturations per hour were using baclofen. It has

Obstructive sleep apnea in tetraplegiaB Klefbeck et al

627

been shown that a single dose of 25 mg baclofen altersthe sleep architecture and produces a small reductionin mean sleep oxygen saturation although it does notsigni®cantly increase sleep-disordered breathing.33

Thus, baclofen might be a risk factor for OSA butour data indicates that moderate doses are not a majorrisk factor.

Conclusion

The prevalence of OSA in CCI subjects with a widerange of motor loss (ASIA motor score 0 ± 90) andrespiratory impairment (VC 1.21 ± 4.85 litre) was 15%and higher than in a corresponding, non-obese, non-injured population. There was an inverse relationbetween ODI and ASIA motor score in the subgroupwith complete injury. The prevalence of obstructivesleep apnea syndrome was 9%, which is higher than inthe estimated prevalence in Swedish non-injuredsubjects. Symptoms associated with OSA were notrelated to the results of the respiratory recordings andare thus not appropriate for predicting OSA in thispatient group.

Acknowledgements

This study was supported by grants from The SpinalisFoundation, Karolinska Institute and the Swedish Asso-ciation for Tra�c and Polio Disabled. Eva Svanborg MD,PhD is acknowledged for her valuable comments.

References

1 Carter RE. Respiratory aspects of spinal cord injury manage-ment. Paraplegia 1987; 25: 262 ± 266.

2 Fuhrer MJ et al. Postdischarge outcomes for ventilator-dependent quadriplegics. Arch Phys Med Rehabil 1987; 68:353 ± 356.

3 McEvoy DR et al. Sleep apnoea in patients with quadriplegia.Thorax 1995; 50: 613 ± 619.

4 Short DJ, Stradling JR, Williams SJ. Prevalence of sleep apnoeain patients over 40 years of age with spinal cord lesions. J NeurolNeurosurg Psychiatry 1992; 55: 1032 ± 1036.

5 Cahan C et al. Arterial oxygen saturation over time and sleepstudies in quadriplegic patients. Paraplegia 1993; 31: 172 ± 179.

6 Flavell H et al. Hypoxia episodes during sleep in high tetraplegia.Arch Phys Med Rehabil 1992; 73: 623 ± 627.

7 Strohl KP, Redline S. Recognition of obstructive sleep apnea.Am J Respir Crit Care Med 1996; 154: 279 ± 289.

8 Svanborg E, Larsson H, Carlsson-Nordlander B, Pirskanen R. Alimited diagnostic investigation for obstructive sleep apneasyndrome. Oximetry and static charge sensitive bed. Chest1990; 98: 1341 ± 1345.

9 Lavie P. Incidence of sleep apnea in a presumably healthyworking population: A signi®cant relationship with excessivedaytime sleepiness. Sleep 1983; 6: 312 ± 318.

10 Young T et al. The occurance of sleep-disordered breathingamong middle-aged adults. N Engl J Med 1993; 328: 1230 ± 1235.

11 Gislason T et al. Prevalence of sleep apnea syndrome amongSwedish men: an epidemiological study. J Clin Epidemiol 1998;41: 571 ± 576.

12 Strandling JR, Crosby JH. Predictors and prevalence ofobstructive sleep apnoea and snoring in 1001 middle aged men.Thorax 1991; 46: 85 ± 90.

13 Levi R, Hultling C, Nash MS, Seiger AÊ . The Stockholm spinalinjury study: 1. Medical problems in regional SCI population.Paraplegia 1995; 33: 308 ± 315.

14 Block A, Boysen PG, Wynne JW, Hunt LA. Sleep apnea,hypopnea and oxygen desaturation in normal subjects. N Engl JMed 1979; 300: 513 ± 517.

15 Deegan PC, McNicholas WT. Pathophysiology of obstructivesleep apnoea. Eur Respir J 1995; 8: 1161 ± 1178.

16 Ditunno JF, Young W, Donovan WH, Creasey G. Theinternational standards booklet for neurological and functionalclassi®cation of spinal cord injury. Paraplegia 1994; 32: 70 ± 80.

17 El Masry WS et al. Validation of the American Spinal InjuryAssociation (ASIA) Motor Score and the National Acute SpinalCord Injury Study (NASCIS) Motor Score. Spine 1996; 21: 614 ±619.

18 Svanborg E, Pa al T, MoÈ ller J. Apno-Log - a portable monitoringsystem for home recordings of obstructive sleep apneas. SleepRes 1992; 1: 225.

19 Black LF, Hyatt RE. Maximal static respiratory pressures ingeneralized neuromuscular disease. Am Rev Respir Dis 1971; 103:641 ± 650.

20 Black LF, Hyatt RE. Maximal respiratory pressures: Normalvalues and relationship to age and sex. Am Rev Respir Dis 1969;99: 696 ± 702.

21 HamnegaÊ rd C-H et al. Portable measurement of maximummouth pressures. Eur Respir J 1994; 7: 398 ± 401.

22 Kuskowska WA, RoÈ ssner S. Body mass distribution of arepresentative adult population in Sweden. Diab ResearchClinical Practice 1990; 10: S37 ± S41.

23 Gisalson T et al. Basic Nordic Sleep Questionnaire. ScandinavianSociety for Sleep Research. 1988.

24 Biering-SoÈ rensen F, Biering-SoÈ rensen M, Hilden J. Reproduci-bility of Nordic Sleep Questionnaire in spinal cord injured.Paraplegia 1994; 32: 780 ± 786.

25 Kump K et al. Assessment of the validity and utility of a sleep-symptom questionnaire. Am J Respir Crit Care Med 1994; 150:735 ± 741.

26 Vgontzas AN et al. Sleep apnea and sleep disruption in obesepatients. Arch Intern Med 1994; 154: 1705 ± 1711.

27 Kripke DF et al. Prevalence of sleep-disordered breathing in ages40 ± 64 years: A population-based survey. Sleep 1997; 20: 65 ± 76.

28 McNicholas W et al. Upper airway obstruction during sleep innormal subjects after selective topical oropharyngeal anesthesia.Am Rev Respir Dis 1987; 135: 1316 ± 1319.

29 Larsson H et al. Temperature thresholds in the oropharynx ofpatients with obstructive sleep apnea syndrome. Am Rev RespirDis 1992; 146: 1246 ± 1249.

30 Roth EJ et al. Pulmonary function testing in spinal cord injury:correlation with vital capacity. Paraplegia 1995; 33: 451 ± 457.

31 Svanborg E et al. Hur vanlig aÈ r positionskaÈ nslighet vidobstruktivt soÈ mnapne syndrom? Abstract LaÈkarstaÈmman, Stock-holm 1995.

32 McKenzie Neill A, Angus SM, Sajkov D, McEvoy RD. E�ects ofsleep posture on upper airway stability in patients withobstructive sleep apnea. Am J Respir Crit Care Med 1997; 155:199 ± 204.

33 Finnimore AJ, Roebuck M, Sajkov D, McEvoy RD. The e�ectsof the GABA agonist, baclofen, on sleep and breathing. EurRespir J 1995; 8: 230 ± 234.

Obstructive sleep apnea in tetraplegiaB Klefbeck et al

628