Changes in ventilation and its components in normal subjects … · sleep usually occurred within 90minutes andlasted about 30minutes. If this wasimmediately followed bySWsleep stages

Thorax 1985;40:364-370

Changes in ventilation and its components in normalsubjects during sleepJR STRADLING, GA CHADWICK, AJ FREW

From the Osler Chest Unit, Churchill Hospital, Oxford

ABSTRACT Non-invasive measurements were made of ventilation, its derivatives, the contribu-tions of abdomen and ribcage and arterial oxygen saturation in six healthy normal men whilstawake and during sleep. Minute ventilation fell significantly during slow wave (SW) sleep andrapid eye movement (REM) sleep (awake = 6.3 1 min-', SW sleep = 5 7 1 min-', REM sleep =

5.4 1 min- 1; p < 0.04). Mean inspiratory flow also fell significantly but timing was unchanged. Theabdominal (diaphragmatic) contribution to ventilation fell very significantly during SW sleep butreturned to awake levels during REM sleep (awake 54%, SW sleep 38%, REM sleep 56%; p <0.007). There were also significant falls in arterial oxygen saturation during SW and REM sleep(awake 97*3%, SW sleep 96*5%, REM sleep 96-2%; p < 0.002). These falls represent reduc-tions in arterial oxygen tension similar to those seen in patients with chronic airways obstructionand can be accounted for entirely by the associated reduction in ventilation.

Interest is growing in the relationship betweenabnormalities of nocturnal breathing and thedevelopment of daytime symptoms. This relation-ship is most clearly seen in the obstructive sleepapnoea syndrome.' It is, however, known thatpatients with hypoxic lung disease become evenmore hypoxaemic during sleep.23 By analogy withobstructive sleep apnoea, it has been suggested thatthis extra hypoxaemia may hasten the developmentof cor pulmonale.45 Although the extra hypoxaemiawas initially thought to occur in those with abnormalnocturnal breathing, later studies suggested that thedrop in oxygen tension was similar throughout alarge and varied group of patients with chronic air-ways obstruction and that the severity of the dropsin oxygen saturation simply reflected their positionon the oxyhaemoglobin dissociation curve.6 Thusthe question arose whether patients with nocturnalhypoxaemia and normal subjects show similar fallsin oxygen tension and in alveolar ventilation.

Patients with bilateral diaphragm paralysis suffersevere nocturnal hypoxaemia during rapid eyemovement (REM) sleep.' The mechanism proposedis based on limited evidence that during REM sleep,breathing in normal subjects is mainly diaphragma-

Address for reprint requests: Dr JR Stradling, Churchill Hospital,Oxford.

Accepted 3 December 1984

tic owing to physiological inhibition of the inter-costal muscles.89 Consequently a patient dependingentirely on his intercostal muscles would hypoventi-late considerably during REM sleep. Patients withchronic airways obstruction and hyperinflation havediaphragms that are flat and at a mechanical disad-vantage, leading to a smaller contribution by thediaphragm to their tidal volumes than in normal sub-jects.'" These patients may behave in a qualitativelysimilar way to patients with bilateral diaphragmparalysis, in that they may be more vulnerable to theinhibition of intercostal muscles during REM sleepand thus hypoventilate more than normal subjects.It is therefore important to know the extent to whichthe compartmental contribution to ventilationchanges normally with sleep.

Previous studies of ventilation during sleep havehad various limitations, the most important beingthe use of mouthpiece or masks. These are known toproduce changes in the pattern of breathing, -'13 buttheir different effects on ventilation during wakeful-ness and sleep are not known. Recent evidence sug-gests that a mask enclosing both nose and mouth haslittle effect on respiratory rate during sleep, but thatthe sleep is more disturbed.'4 The best availablestudy in normal subjects used a tight fitting maskand gave no information on the contributions fromribcage and diaphragm.'5 That study was performedat altitude of 1600 m, so a degree of hypoxaemiawas present that may have increased ventilation

364

on February 22, 2021 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.40.5.364 on 1 M

ay 1985. Dow

nloaded from

Changes in ventilation and its components in normal subjects during sleepmore during the wakefulness than during asleep,since hypoxic sensitivity is diminished duringsleep.'6 1 The only study using body surface motion 5to derive ventilation in normal subjects,'8 withoutdisturbing the face, was performed before it hadbecome clear how seriously changes in posture -affect the accuracy of the calibration with this tech-nique.'9 20 Thus the values for changes in ventilationand the relative ribcage and abdominal contribu-tions found in this study may have been incorrectowing to errors in calibration.We have studied non-invasively the changes in

ventilation and its derivatives, in the relative con-tributions of the ribcage and abdomen, and in arter-ial oxygen saturation that occur during sleep in nor-mal subjects. We believe that the techniques we Figl Modified"Buxton"tippingchairformaintenanceofhave used to measure ventilation and the relative a fied posture throughout the studies.contributions of ribcage and abdomen to tidal vol-ume are more appropriate than those used in previ-ous studies. motion (V-M) coefficients) the electrical signals are

calibrated to represent the volume of air displaced inMethods the lungs. If the respiratory pump behaves with two

degrees of freedom then these two belts will cor-SUBJECTS rectly represent all the air going in and out of theSix healthy men aged 18-32 years were studied. lungs. The RIP was calibrated by multiple linearNone were smokers and none had had a recent regression,amicrocomputerbeingused.22 Theaccu-respiratory tract infection. One had experienced racy of the V-M coefficients was tested by compar-mild asthma when younger but not in the previous ing RIP tidal volume measurements obtained byyear and all had normal spirometric indices. None respiratory inductance plethysmography with thosehad taken any medication or receivedcs±imulant obtained by spirometry. Some of the comparisonsdrinks for four hours before the study. Obesity indi- were done during normal breathing and some withces (weight (kg)/height (m)2) ranged from 23 to 29, deliberate changes in relative ribcage and abdominalthe normal range being 23-26. None had any com- contributions.23 This procedure was repeated at theplaints related to sleeping. The subjects were fully end of the experiment with a range of ribcage andaware of the nature and purpose of the study. Four abdominal contributions covering the range seenof the six subjects had only five hours sleep the night during the study and any implied change in V-Mbefore the study; the other two had had normal coefficients was calculated. If the drift was morenights. than 5% then the data were recalculated with the

new V-M coefficients since the data actually ana-TECHNIQUES lysed were always nearer the end of the study thanThe subjects were studied recumbent in a tipping the beginning.chair (Buxton type) with the angle between the seat Arterial oxygen saturation (Sao2) was measuredand the back opened out to 1200 (fig 1). This effec- with an ear lobe oximeter (Biox). Although at hightively prevented any change in posture that could saturations this instrument has an error of +2%,24have altered the calibration of the respiratory induc- the changes in Sao2 in an individual are more accu-tance plethysmograph: in particular, the spinal angle rately represented.25was fixed. This chair, although very comfortable, Continuous monitoring of electroencephalogramdisturbed sleep to some extent because it limited (EEG), submental electromyogram (EMG), andmovement. electro-oculogram (EOG) was performed to stage

Respiratory inductance plethysmography was the sleep and define wakefulness according to stan-used to measure ventilation on the basis of,2' dard criteria.26 This staging was done at the time ofchanges in self inductance of wire "zigzags" on belts the study from the signals displayed on a largearound the chest and abdomen. The belts were indi- monitor or from a polygraph recorder (Beckmann)vidually prepared from stretchy sticking plaster and at 25 mm/s. Changes from one state to another werefirmly stuck to the skin to prevent slippage. By -the - not sudden and only data from periods of at least 10application of appropriate constants (volumez---minutes' -clear wakefulness, clear SW sleep (stage 3

365

on February 22, 2021 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.40.5.364 on 1 M

ay 1985. Dow

nloaded from

366

or 4), or clear REM sleep were analysed. Continu-ous slow recordings (0.1 mm/s) were made of rib-cage and abdominal movement, Sao2, EOG, EMG,and EEG, with faster periods for documentationpurposes at 25 mm/s.

PROTOCOLEach study was supervised by two of the authors.Each subject reported at about 23.00 hours. Heemptied his bladder and the plethysmograph beltswere applied above the nipples and over theumbilicus. The subject sat in the chair while thesilver disc electrodes and ear oximeter were applied.The chair was tilted backwards to near horizontaland the subject made comfortable with pillowsunder the head and legs. Calibration and testing ofthe respiratory inductance plethysmograph was thenperformed. The lights were turned off at about23.45 hours. A sensitive intercom allowed any

periods of snoring to be detected and these were notanalysed. Ventilation was monitored breath bybreath with the aid of the microcomputer23 and themean values were printed out each minutes togetherwith the Sao2 at that time; a signal was also passedto the polygraph for synchronisation. Sleep stages 1or 2 with periodic arousals were always seen ini-tially, followed by SW sleep stages 3 or 4. REMsleep usually occurred within 90 minutes and lastedabout 30 minutes. If this was immediately followedby SW sleep stages 3 or 4 then a further 15 minuteswas recorded, the subjects was woken, and 15minutes of "awake" data were recorded, followedby calibration tests. If the subject awoke spontane-ously after the REM sleep period then the 15minute awake period was immediately recorded, fol-lowed by calibration tests. In this way consecutiveperiods of REM and SW sleep and wakefulness (orSW and REM sleep and wakefulness) were availablefor analysis within a period of less than 80 minutes.This further minimised any possible effects ofchanges in calibration of the respiratory inductanceplethysmograph.

DATAAverages of minute ventilation, frequency, tidal

Stradling, Chadwick, Frew

volume, mean inspiratory flow (VT/TI), fractionalinspiratory time (T/TProT), abdominal contribution(as a volumetric fraction determined after applica-tion of V-M coefficients) were calculated eachminute by the computer. A minimum of 10 minutes'data for each stage were further averaged by handand the standard deviation was calculated. Thisallowed the minute to minute coefficient of variation(SD/mean) for each variable during each state to becalculated. Changes were tested for statisticalsignificance by means of paired Student's t tests:awake versus SW sleep, awake versus REM sleep,and SW versus REM sleep.

Alveolar ventilation was estimated approximatelyby assuming an anatomical dead space of 150 ml. Amulticompartment computer simulation of gasexchange27 was used to assess whether the changesobserved in Sao2 could be explained on the basis ofchanges in calculated alveolar ventilation. Thiscomputer model-particularly in conjunction with a

fixed estimate of anatomical dead space-will, ofcourse, be able to predict changes in arterial bloodgases with only limited accuracy.

Results

The table and figure 2 show the changes observedduring SWS and REMS. Significant falls in ventila-tion and Sao2 occurred. The calculated fall in alveo-lar ventilation during REM sleep relative to awakeperiods (-20%) is exactly the amount required toproduce a fall in Sao2 from 97-3% to 96-2% accord-ing to the computer model (an approximate fall inPao2 from 96 5 mm Hg (12.9 kPa) to 82'5 mm Hg(11.0 kPa)). These falls in ventilation resultedalmost entirely from changes in VT and VT/TI andnot from changes in timing (f or TI/TTOT). The frac-tional abdominal contribution to ventilation fell dur-ing SW sleep but returned to awake levels duringREM sleep. This implies that the actual ribcage andabdominal contributions to ventilation were equallyreduced during REM sleep compared and awakeperiods. The ribcage contribution to ventilation,however, was actually higher during SW sleep thanduring awake periods even though overall ventila-

Ventilation and oxygen saturation during sleep

Awake SWS (% change) REM (% change)

Minute ventilation (1 min-') 6-28 5 67 (49.7) 5 44 (413-0)Alveolar ventilation (I min-') 4-02 3-38 (416) 3-21 (,120)Frequency/min 15-1 15-2 (11) 14-9 (11-4)Tidal volume (VT) (ml) 420 373 (411) 367 (413)VT/TI (1 min-') 17-2 16-7 (43) 14-7 (415)TI/TTOT 0-367 0-350 (45) 0-374 (T2)Abdominal contribution 0-54 0-38 (430) 0 56 (14)Arterial oxygen saturation (%) 973 96 5 (41) 96-2 (41)

TI-inspiratory time; TTOT-total respiratory cycle; SWS-slow wave sleep (stages 3 and 4); REM-rapid eye movement sleep.

on February 22, 2021 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.40.5.364 on 1 M

ay 1985. Dow

nloaded from

Changes in ventilaton and its components in normal subjects during sleep

7

6

5

MinuteVentilation

TidalVolume

0,)7E

20

18

16

14

VT/TI* 0.7

0.6

0.5

0.4

367

AbdominalContribution

A SWS REM

Mean ± SEM

NS = Not significant*p<0.05, ** p<0.01

*** p<O.001

Frequency17 -

NS

15 -

NS NS13 _

I I RIEA SWS REM

Fig 2 Effect ofsleep state on ventilation and its subdivisions and on arterial oxygen saturation. A-awake periods;SWS-slow wave sleep (stages 3 and 4); REM-rapid eye movement sleep.

tion fell.Snoring had a profound effect the proportions of

ribcage and abdominal contributions. In the onesubject who snored briefly during an early SW sleepperiod (not included in the analysis) the abdominalfractional contribution rapidly rose from 035 to0-70, with a slight increase in minute ventilation thatmight have been an artefact due to rarefaction ofintrathoracic gas with each breath. The changesfrom one sleep state to another were usually gradualand accompanied by progressive changes in theabdominal contribution (fig 3). These findings showthe importance of analysing periods that clearly rep-resent one particular state and are not complicatedby snoring.

Figure 4 shows the changes in the coefficient ofvariation (minute to minute) for all the variablesawake-periods and SW and REM sleep. The var-

iability of most of the measurements was less duringSW than during REM sleep and awake periods. The

REM

variability of Sao, was highest during REM sleep,perhaps confirming that ventilation during this stateis relatively independent of chemical drives.'61728

Discussion

Respiratory inductance plethysmography is more

difficult to use accurately than was originallythought.'92023 Changes in the relationship betweenbody surface motion and the volume of air movingin and out of the lung may be considerable, particu-larly with the usual changes in posture that occur

during sleep.20 Simple comparisons made with a

spirometer in the same posture before and aftersleep do not ensure accuracy in the interim, whenboth posture and the ribcage-abdominal contribu-tions may alter. Small changes in an initially smallVT may produce relatively large changes in alveolarventilation and thus accurate measurements of ven-

tilation are required to analyse the causes of blood

Arousal

0.4

cOc = ~ sws sws

.2 0.2-

.0Cinutes

Subject G

Fig 3 Minute to minute changes in abdominal contribution to ventilationduring different sleep states in one subject. Definitions as in figure 2.

TI/TTOT0.4

0.37

0.34

SaO298

97

96

95A SWS REMA SWS REM

on February 22, 2021 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.40.5.364 on 1 M

ay 1985. Dow

nloaded from

368

Minute

16Ventilation

,-NS\12 -

4 - E

A SWS REM

TidalVolume

NS

VT/TI

\ NS

0.A*\

Stradling, Chadwick, Frew

AbdominalContribution

NSL*0

NS

TI/fTOT16

12

% 8

NS = Not significant*p<0.05, ** p<O.01* p<O.001 0 l l

A SWSREM

^^ -e- NSlSSE

A SWSREM

0.8 SaO2

0.6 *

% 0.4 -

0.2 NS

O I I I

A SWS REM

Fig 4 Effect ofsleep state on the minute to minute coefficient ofvariation ofventilation and its subdivisions and onarterial oxygen saturation. Definitions as in figure 2.

gas disturbances. We have shown that multiplelinear regression calibration is at least as good asisovolume techniques in deriving V-M coefficients23and that deliberate changes in the ribcage andabdominal contributions are necessary to test vol-ume motion coefficients adequately. We believethat, in conjunction with posture restraint andclosely adjacent recording periods, these techniquesprovide the best accuracy in respiratory inductanceplethysmography during sleep.

In human subjects abdominal and ribcage move-ments probably reflect diaphragm and intercostalmuscle activities respectively.29 The lower ribcagecan move either with the bulk of the chest or withthe abdomen but the relationship is probably fixedin the one posture and there is evidence to suggestthat it is not changed by sleep.30 Multiple linearregression calibration weights the two degrees offreedom according to the displaced volume rep-resented by each. If the movement of the lower rib-cage results from the action of the diaphragm thenthe volume of lung gas displaced by this movementwill be included with the calculated "abdominal"contribution to ventilation. Isovolume techniques tocalibrate the respiratory inductance plethysmographmay well make this relationship different from therelationship found in normal breathing.23We have found significant falls in ventilation dur-

ing SW and REM sleep and these were adequate toexplain the observed reductions in Sao2. The calcu-lated mean fall in Pao2 from awake periods to SWsleep was 1-4 kPa (10.5 mm Hg) and from awake

periods to REM sleep 1 9 kPa (14.0 mm Hg). Thesefalls are quantitatively similar to those estimated byoximetry in a previous study on patients withchronic airways obstruction.6 Although the latterwere measured by respiratory inductance plethys-mography, with a calibration which would now beregarded as incorrect, the falls in ventilation seen inpatients with chronic airways disease during REMsleep were considerably greater and more thanenough to account for the degree of extra hypox-aemia.631 Reasons for this are being sought, butprobably these patients never reach a steady stateduring REM sleep and periods of increased ventila-tion save them from even worse hypoxaemia, pro-ducing characteristic swings in Sao2.36

Results of studies on ventilation during sleep innormal subjects are not entirely similar to ours. Thestudies of Douglas et al'5 were performed at analtitude of 1600 m with a facemask and the conse-quent hypoxaemia (end tidal Po2 11.1 kPa (84 mmHg)) and 80 ml extra dead space probably accountfor the higher resting ventilation (7.7 v 6*3 1 min-')and the larger fall during REM sleep (16% v 13%,or 39% v 20% if a calculated alveolar ventilation isconsidered). In some respects these subjects mayresemble patients with hypoxic lung disease whoseventilation is being driven by hypoxaemia, and whenthis drive is reduced by REM sleep'61 an exagger-ated fall in ventilation might be expected. Theincrease in respiratory rate during sleep found inDouglas's study was perhaps due to the use of maskswhich are known to decrease frequency in awake

on February 22, 2021 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.40.5.364 on 1 M

ay 1985. Dow

nloaded from

Changes in ventilation and its components in normal subjects during sleep

subjects, an effect that might disappear with loss ofconsciousness.

Tabachnik et al,'8 using respiratory inductanceplethysmography, obtained results closer to ours.Their measurements of minute ventilation (L/min- )were 6-2 (awake), 5-6 (SW sleep) and 5 8 (REMsleep) and of the abdominal contribution 060(awake), 034 (SW sleep) and 0-64 (REM sleep).Where differences exist, particularly the larger per-centage change in the abdominal contribution andthe higher ventilation rate during REM than SWsleep, they would be explicable if the abdominalV-M coefficient had been consistently overesti-mated. We have calculated from available data'9that this would be so if the incorrect two posturecalibration technique (erect and supine) is used for asingle position supine study. Although it is notstated, this calibration technique is assumed to havebeen used in Tabachnik's study. This study alsofound similar changes in Sao, and in the variabilityof ventilation with its subdivisions according to sleepstate.

While awake, a patient with poor diaphragmaticfunction can recruit intercostal muscles and acces-sory muscles of respiration. The 32% reduction inthe ribcage contribution that we have observed dur-ing REM sleep could be critical to such patients inmaking these adaptations with primarily posturalmuscles no longer available. This change in patternclearly needs to be confirmed in such patients beforebeing put forward as the explanation for the largerfalls in ventilation seen in patients with chronic air-ways obstruction.6 It is worth noting that thechanges in derived Pao2 we have seen in the normalsubjects are themselves sufficient to produce drama-tic falls in Sao2 if the patient's daytime Sao2 is on thesteep part of the oxyhaemoglobin dissociation curve(<85%).

In conclusion, we have found falls in ventilationand Sao2 during sleep in normal subjects that aresufficient to account for the larger falls in Sao2 seenin patients with hypoxic lung disease. The falls inventilation were, however, not as great as those seenin some patients with chronic airways obstruction.We have shown changes with sleep state in the rib-cage and abdominal contribution to ventilation thatcould have important consequences when dia-phragmatic action is embarrassed.

References

1 Guilleminault C, Tilkian A, Dement WC. The sleepapnea syndromes. Annu Rev Med 1976;27:465-84.

2 Trask CH, Cree EM. Oximeter studies on patients withchronic obstructive emphysema, awake and during

sleep. N Engl J Med 1962;266:639-42.3 Douglas NJ, Calverley PMA, Leggett RJE, et al. Trans-

ient hypoxaemia during sleep in chronic bronchitis andemphysema. Lancet 1979;i: 1-4.

4 Flenley DC. Clinical hypoxia: causes, consequences,and correction. Lancet 1978;i:542-6.

5 Block AJ, Boysen PG, Wynne JW. The origins of corpulmonale. A hypothesis. Chest 1979;75:109.

6 Stradling JR, Lane DJ. Nocturnal hypoxaemia inchronic obstructive pulmonary disease. Clin Sci1983;64:213-22.

7 Skatrud J, Iber C, McHugh W, Rasmussen H, NicholasD. Determinants of hypoventilation during wakeful-ness and sleep in diaphragm paralysis. Am Rev RespirDis 1979;121:587-93.

8 Tusiewicz K, Moldofsky H, Bryan AC, Bryan MH.Mechanics of the ribcage and diaphragm during sleep.J Appl Physiol 1977;43:600-2.

9 Parmeggiani PL, Sabattini L. Electromyographicaspect of postural respiratory and thermoregulatorymechanisms in sleeping cats. Electroencephalogr ClinNeurophysiol 1972;3:1-3.

10 Tobin MJ, Chadha TS, Jenouri G, Birch SJ, GazerogluBS, Sackner MA. Breathing patterns. 2. Diseased sub-jects. Chest 1983;84:286-94.

11 Askanazi J, Silverberg PA, Foster RJ, Hyman Al,Milic Emili J, Kinney JM. Effects of respiratoryapparatus on breathing pattern. J Appl Physiol1980;48:577-80.

12 Gilbert R, Auchincloss JH, Brodsky J, Boden W.Changes in tidal volume, frequency and ventilationinduced by their measurement. J Appl Physiol1972;33:252-4.

13 Sackner JD, Nixon AJ, Davis B, Atkins N, SacknerMA. Effects of breathing through external dead spaceon ventilation at rest and during exercise. II. Am RevRespir Dis 1980; 122:933-40.

14 Krieger J, Kurtz D. Effects of pneumotachygraphicrecording of breathing on sleep and respiration duringsleep. Bull Eur Physiopathol Respir 1983; 19:641-4.

15 Douglas NJ, White DP, Pickett CK, Weil J, Zwillich C.Respiration during sleep in normal man. Thorax1982;37:840-4.

16 Douglas NJ, White DP, Weil JV, et al. Hypoxic ven-tilatory response decreases during sleep in normalmen. Am Rev Respire Dis 1982;125:286-9.

17 Berthon-Jones M, Sullivan CE. Ventilatory andarousal responses to hypoxia in sleeping humans. AmRev Respir Dis 1982; 125:632-9.

18 Tabachnik E, Muller NL, Bryan AC, Levison H.Changes in ventilation and chest wall mechanics duringsleep in normal adolescents. J Appl Physiol 1981;51:557-64.

19 Zimmerman PV, Connellan SJ, Middleton HC,Tabona MV, Goldman MD, Pride NB. Posturalchanges in ribcage and abdominal volume-motioncoefficients and their effect on the calibration of arespiratory inductance plethysmograph. Am RevRespir Dis 1983;127:209-14.

20 Stradling JR, Chadwick GA. Inaccuracies in the use ofbody surface movements to measure ventilation duringsleep. Bull Eur Physiopathol Respir 1984;20 (5): 12A(abstract).MA. A transducer for non-invasive monitoring of therespiration. In: Stott F, Raftery EB, Sleight P, Gould-ing L, eds. ISAM Proceedings of the Second Interna-

369

on February 22, 2021 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.40.5.364 on 1 M

ay 1985. Dow

nloaded from

370

tional Symposium on Ambulatory Monitoring. Lon-don: Academic Press 1978;-119-28.

22 Stradling JR, Nichol CG, Cover D, Davies EE, HughesJMB, Pride NB. The effects of oral almitrine on pat-tern of breathing and gas exchange in patients withchronic obstructive pulmonary disease. Clin Sci1984;66:435-42.

23 Stradling JR, Chadwick GA. Simplifying the use ofrespiratory inductance plethysmography. Thorax1984;39:233-4.

24 Rebuck AS, Chapman KR, D'Urzo A. The accuracyand response characteristics of a simplified ear oxime-ter. Chest 1983;:860-4.

25 Ries AL, Farrow JT, Clausen JL. Accuracy of twooximeters at rest and during exercise in pulmonarypatients.Am Rev Respir Dis 1984; 129:A79 (abstract).

26 Rechtshaffen A, Kales A. A manual of standardisedterminology, techniques and scoring system for sleep

Stradling, Chadwick, Frew

stages in human subjects. Washington DC: US Gov-ernment Printing Office, 1968. (National Institutes ofHealth Publication 204).

27 West JB. Ventilation-perfusion inequality and overallgas exchange in computer models of the lung. RespirPhysiol 1969;7:88-110.

28 Phillipson EA. Control of breathing during sleep. AmRev Respir Dis 1978; 118:909-39.

29 Gilbert R, Auchincloss JH, Peppi D. Relationship ofribcage and abdominal motion to diaphragm functionduring quiet breathing. Chest 1981;80:607-12.

30 Mortola JP, Anch AM. Chest wall configuration insupine man: wakefulness and asleep. Respir Physiol1978;35:201-13.

31 Hudgel DW, Martin RJ, Capehart M, Johnson B, HillP. Contribution of hypoventilation to sleep oxygendesaturation in chronic obstructive pulmonary disease.J Appl Physiol 1983; 55:669-77.

on February 22, 2021 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.40.5.364 on 1 M

ay 1985. Dow

nloaded from