British Journal of Anaesthesia 1993; 70: 259-266

EFFECT ON LUNG VOLUMES OF OXYGEN CONCENTRATIONWHEN BREATHING IS RESTRICTED

A. B. BAKER, A. McGINN AND C. JOYCE

SUMMARY

We have examined the effect of the fractionalconcentration of nitrogen (FlN2) on the decrease inlung volumes which occurs during restrictedbreathing with oxygen-nitrogen mixtures. Con-scious human subjects breathed one of five gasmixtures of oxygen and nitrogen for 15 min on eachof five occasions. For the final 5 min of each 15-minperiod, functional residual capacity (FRC) wasreduced by one tidal volume, by external pressur-ization. After return to normal breathing, thesubjects showed a statistically significant decreasein FRC and residual volume (RV), which becamelarger as F/o increased from 30% to 100%. Thisreduction in lung volumes was resistant to early re-expansion. The maximum decrease in both FRCand RV was found with 100% oxygen, and was10% of the control lung volumes. The use of amixture of 75 % nitrogen and 25 % oxygen preventedthis decrease in lung volumes. Nitrogen 50%reduced the decrease in FRC, but did not affect thedecrease in RV. The difference in effect on lungvolumes between F/02 25% and 30% was stat-istically significant, indicating a watershed area forF/^ between 70% and 75%. (Br. J. Anaesth. 1993;70: 259-266)

KEY WORDSLung: functional residual capacity, nitrogen splinting. Oxygenconcentration. Ventilation: restricted.

For more than a century there has been interest inthe effect of breathing different gases on the rate ofalveolar gas absorption leading to alveolar atelectasis[1,2]. In 1879, Lichtheim showed that alveolarcollapse occurred in 45 min when open-chested dogsbreathed oxygen, whilst with air breathing it oc-curred after 24 h [3]. In 1932, Coryllos and Birn-baum carried out more extensive studies in bothopen and closed chest dogs which agreed sub-stantially with the findings of Lichtheim [4]. Daleand Rahn, in 1952, investigated the rate of gasabsorption and confirmed that gas was removed 60times more quickly from the lung during oxygenbreathing than during air breathing [5]. Morerecently, evidence of alveolar collapse has been notedafter oxygen breathing at low lung volumes [6—10],with increased gravitational forces [11—14]; and withhead-out immersion [15—17], although one study[18] showed no difference between air and oxygen.

Decreases in functional residual capacity (FRC)have been reported after 100% oxygen breathing[19-22]. In 1983, Baker and Restall [19] described adecrease in residual volume (RV) in subjects whohad breathed 100 % oxygen at reduced lung volume.

Studies have investigated the effect of differentconcentrations of inert gases in the alveolar mixtureon atelectasis. Green [13] showed that 20% nitrogenreduced the rate of absorption of gas compared withoxygen, and that 40 % nitrogen reduced it evenfurther. At reduced ambient pressure ("5psi")3DuBois and colleagues [7] showed that 5 % nitrogenin oxygen prevented alveolar collapse in one subject,whereas 2.5 % nitrogen did not. This finding wasconfirmed in the same subject by Turaids, Nobregaand Gallagher [23]. In a preliminary study in ourlaboratory, Smith [24] showed, for subjects who hadbreathed either 100% oxygen or 50% oxygen-nitrogen mixtures at reduced lung volumes, thatthere was more reduction in FRC and RV with100% oxygen. Even so, there were statisticallysignificant reductions in FRC and RV with the 50 %mixture.

We have therefore studied in greater detail theeffect of various oxygen-nitrogen mixtures, to de-lineate the effect of increasing nitrogen concentrationon stabilization of the absorption process (nitrogensplinting of the lung), by measurement of FRC andRV when subjects had breathed the different gasmixtures at reduced lung volumes. We used a studydesign for reduced lung breathing developed byNunn [10] and used previously by us [19]. Thisdesign is relevant to the post-anaesthesia state, toaviation and diving, because it temporarily reducesFRC whilst oxygen is breathed in a fashion similar toimmersion forces in diving or acceleration forces inaviation when the subject is wearing an anti-gravitysuit [12, 17]. The hypothesis tested was that theextent to which nitrogen reduces or prevents thedecrease in lung volumes which occurs duringoxygen breathing at reduced FRC was proportionalto the nitrogen concentration in the inspired gasmixture (FiSi).

A. B. BAKER, M.B., B.S., D.PHIL., F.A.N.Z.C.A., F.R.C.ANAES. ; A.MCGINN; C.JOYCE, M.B., CH.B., F.A.N.Z.C.A.; Department ofAnasthesia and Intensive Care, Otago University, Dunedin, NewZealand. Accepted for Publication: October 9, 1992.

Correspondence to A.B.B.

260 BRITISH JOURNAL OF ANAESTHESIA

TABLE I. Volunteer characters. CC = Closing capacity; % N2lnc. = percentage nitrogen increase of the phase 111 slopeper litre of expired gas; FEF25% = forced expiratory flow at 25% VC. *Only consistent smoker who had ceased

3 months before the study; -\sinus problems; ^exercise-induced asthma

Subject

M.D.J.W.*DM.A.P.fDP.A.C.S.G.SA.R.CD.C.J.J.B.S.L.A.W.I.O.L.F.M.N.J.s.

SUBJECTS

Sex

MMFMMFFFMMFFMFFFF

AND

Age(yr)

3524322128213120442749294241405723

METHODS

Weight(kg)

75.072.060.060.065.575.081.051.058.078.075.052.087.080.066.051.054.0

Height(cm)

180.0163.0160.0170.0170.5173.5166.5155.0168.0180.0171.0158.5174.0162.5168.5157.5170.0

VC

4.544.903.703.704.704.194.143.593.865.694.003.855.284.013.913.033.86

~'JtJ

FEV, CC

3.89 2.093.723.263.544.16 :3.66 (3.442.823.27 :

.83

.79

.552.08).91.72.05

2.035.13 2.403.05 2.392.963.93 :3.022.942.193.47

.122.14.75.80.97—

% N, Inc.

1.060.521.170.990.780.641.421.230.610.731.541.161.822.211.661.831.48

FEF25 %

3.723.832.713.153.944.123.142.103.424.321.912.513.353.582.272.552.78

All subjects were informed volunteers and the studywas approved by the Human Ethics Committee ofDunedin Hospital. A limited respiratory history waselicited from the volunteers with respect to smoking,asthma or other chronic respiratory illness, and toensure that they did not have an acute respiratoryillness. No subject was excluded from the study. Aseries of baseline respiratory function studies wereperformed with the subjects seated in a pressurebody plethysmograph (constant volume—Du Boistype) (Morgan No. 090, P. K. Morgan Ltd, Kent,England) connected to a Gould X-Y plotter. Vitalcapacity (VC), forced expired volume in 1 s (FEV,),FRC, expiratory reserve volume (ERV), RV, closingcapacity (CC) using oxygen as the marker gas,percentage nitrogen increase of the phase III slopeper litre of expired gas (% Ns Inc.) between 750 mland 1250 ml expiration, and forced expiratory flow at25% vital capacity (FEF 25%) were measured asdescribed by Cotes [25] (table I). The closingcapacity studies were performed first using a MedScience wedge spirometer and Godart ionizationnitrogen analyser; the FEVj studies followed, usingthe wedge spirometer; and then the other studiesusing the body plethysmograph and pneumo-tachograph. All measurements except single breathnitrogen washout were a mean of three separatereplicates.

The study was designed as a factorial analysis and16 of the 17 subjects completed all the studies, withone subject completing only four of the five gascompositions because of relocation to another city.The gases were allocated randomly to all subjectsexcept that the 25% oxygen-75% nitrogen com-bination was introduced after commencement ofthe study and the first 10 subjects breathed this gascombination after the other four combinations. Allother gas combinations were breathed in randomorder, with at least 1 week between different gasmixtures. With the subject seated in the bodyplethysmograph a set of baseline studies was taken as

FIG. 1. Recording of the restricted breathing pattern, showing thereduction in FRC and tidal breathing during the last 5 min of the

experimental period as recorded from the "Respitrace".

described previously (control data). After thesecontrol studies the subject breathed the selected gascombination from a 300-litre Douglas bag connectedto a non-rebreathing circuit for a 15-min period. TheDouglas bag contained one of the following gascombinations: 100% oxygen; 50% oxygen-50%nitrogen; 30% oxygen-70% nitrogen; 25% oxy-gen-75% nitrogen; 21% oxygen-79% nitrogen.

The Douglas bag was placed outside the bodyplethysmograph, with the subject seated inside theplethysmograph. During the last 5 min of this 15-min period, the body plethysmograph was pressur-ized to approximately 5-10 cm H2O greater thanatmospheric for each subject, to achieve a reductionin the FRC of one tidal volume. There was noreduction in tidal volume with this reduction in FRC(fig. 1) or any difference in CC following airbreathing during this manoeuvre (means 1.727 (SD0.45) litre at control and 1.801 (0.52) litre after the15-min experimental breathing; paired t test prob-ability of 0.36). The pressurization was producedfrom a high pressure air source through a reductionvalve; initially at a rapid flow rate to obtain thedesired pressure and then at a low flow to maintainthe pressure against a slow leak. The change in FRCwas estimated using a "Respitrace" impedanceplethysmograph (Ambulatory Monitoring Inc.,

OXYGEN CONCENTRATIONS AND LUNG VOLUMES 261

ER; vc

1000 '

10s

FIG. 2. Recording of the measurement of FRC using the body plethysmograph and the panting manoeuvre. Lungvolume measurements are also recorded showing measurement of ERV before VC. These measurements were

repeated for three replicates.

TABLE II. Procedure for measurements

(1) FRC: lung volumes (tidal breathing followed by full expiration from FRC to give RV, then tidalbreathing followed by a full inspiration, then a complete expiration to give VC)—control data.(2) FEV,: flow-volume loops for FEF25 %; closing volume—single breath oxygen technique from FRC.(3) Random oxygen mixture breathing for 15 min—first 10 min at normal FRC with the last 5 min from arestricted lung volume with pressurization of the body plethysmograph.(4) Breathing after gas mixture—immediately after cessation of pressurization and breathing room air—repeat(1)—experimental data.(5) Lungs fully inflated—five deep breaths of room air, and 5 min walking around room.(6) Repeat (1)—recovery data

N.Y., U.S.A.) connected via two bands to the lowerthoracic cage and the upper abdomen throughout thestudy. The calibration of the Respitrace was pro-vided on each occasion by the subject's own tidalvolume as measured by the wedge spirometer.

At the end of this 15-min period, the pressure inthe plethysmograph was returned to atmospheric.The subjects, without taking any deep breaths,allowed their end-expiration to return to a restingFRC while remaining in the body plethysmograph.The subject's FRC was then measured breathingroom air and, before any deep breaths were taken,other lung volumes were measured with ERV beingmeasured first from FRC and second from total lungcapacity (TLC) as part of VC (fig. 2). This order waschosen so that the non-expanded RV could beinvestigated. The residual volumes used in thisstudy were calculated using the ERV measured fromFRC. Exhalation from FRC does not expand thechest before ERV is measured, whereas exhalationfrom TLC does. The tests were repeated for threereplicates. As each replicate included a breath toTLC, there would be some lung expansion beforethe second and third replicates. These studies weretermed the experimental data.

After each of these tests, the subjects took fivemaximal inspirations breathing room air, walkedaround for 5 min and lung volumes were againmeasured (recovery data). The procedure formeasurements is summarized in table II.

Statistical analysis was by analysis of variance(completely crossed analysis), paired t test and linearregression using the MASS statistical analysis pro-

gram (Westat Associates Pty Ltd, P.O. Box 247,Nedlands, Australia 6009) and a Macintosh SEcomputer. The statistical analysis using analysis ofvariance was complex because of the fact that each ofthe control studies was not randomized because thesubjects were always breathing room air at thistime. On each occasion the control measurement wasessentially the same volume in the same subjectunder the same conditions, the only variable beingthe passage of time. The analysis of variancecomputation therefore allowed for this fact by usingthe difference from control for experimental andrecovery studies. The analysis refers to "breathingstate" for differences between control, experimentaland recovery states.

RESULTS

Tables III and IV list the individual results andmeans (SD) values for FRC and RV for the five gasmixtures. Tables V and VI show the analysis ofvariance statistics for the FRC and RV data,demonstrating significant relationships (P < 0.001)for the FRC results for FIO J , breathing state and forsynergism between these two; for RV the resultswere significant (P < 0.001) for FIO J but did notachieve statistical significance for the breathing state.The significant differences between subjects are tobe expected by such an analysis. The reproducibilityof the measurements in the context of these studiesover at least 5 weeks may be gleaned from analysis ofthe five replicate control studies breathing room airas the volunteers presented for their random studies.

Subject

M.D.J.W.D.M.A.P.D.P.A.C.S.G.A.R.CD.C.J.J.B.S.L.A.W.I.O.L.F.M.N.J.S.MeanSD

Subject

M.D.J.W.D.M.A.P.D.P.A.C.S.G.A.R.C D .C.J.J.B.S.L.A.W.I.O.L.F.M.N.J.S.MeanSD

C

3.2552.7632.7982.8333.5322.0802.6022.0273.6933.3543.1091.8023.4222.3532.6422.9982.5192.8100.549

C

1.2721.3631.5321.6001.7320.6801.4020.9272.2772.1382.0090.8021.8221.4871.1421.7650.9861.4700.464

100%

E

3.0372.6312.3662.1733.3281.8482.1741.8993.1433.0862.9101.7492.6442.2972.4422.9272.5452.5400.482

100%

E

1.0371.2641.3001.0071.6280.7481.1490.7661.8432.0861.8440.8161.2781.4311.2421.8610.8451.3000.423

TABLE III.

R

3.2112.6312.4862.7453.4191.9572.3121.9573.3783.1852.9621.7743.1532.3422.6712.9452.5842.6900.507

TABLE IV.

R

1.1111.3981.3531.5451.6860.6241.1120.8572.0122.0521.6960.7741.7531.5761.2711.6790.9511.3800.425

Subject FRC details for different oxygen

•

C

3.4312.7432.6652.8013.7762.1742.5911.9283.0503.6053.0331.8913.2342.2752.4462.9852.5992.7800.549

50%

E

3.0722.5582.3322.5793.3992.0732.5231.7542.9593.4562.8241.8993.0372.0152.3162.5512.5462.5803.497

R

3.3832.6542.4502.7083.4692.1222.5631.8242.9673.5752.8401.9313.1522.2752.4092.7372.5862.6800.511

Subject R V details for different oxygen

C

1.698

50%

E

.4061.043 0.8581.5651.4402.076

.432

.441

.8190.574 0.6071.325 .3230.828 0.6241.6501.9721.9670.925 (1.8681.5091.3801.8851.0661.4600.446 (

.593

.790

.7583.7331.5371.3151.3161.5511.0461.3003.395

R

1.6171.1941.1841.3752.103

concentrations. C -

C

3.7892.5462.7462.8953.3852.3922.6701.8872.8953.6112.8191.8623.4792.3362.7952.7222.5302.7903.543

30%

E

3.3412.5462.6832.8753.4062.1462.5141.6772.8103.5362.8191.8283.3412.2862.6412.5112.4782.6700.530

concentrations. C =

C

.856

.246

.346

.665

.4190.689 0.8921.330 L.4700.791 0.7211.7011.7751.8400.7981.6521.6091.2091.8041.0201.3900.415

.5951.9111.6863.7621.6791.4031.5621.6893.8971.4003.376

30%

E

1.6411.2661.4801.4421.5060.9461.4480.5771.5102.0361.5530.6781.8081.3861.3411.5451.1451.3700.370

= Control

R

3.3882.5142.7012.8753.3192.2792.5271.7442.8423.5612.7861.9003.3862.3482.6562.7222.5042.7100.505

Control,

R

1.5551.2481.3351.6421.4530.8791.5270.6781.5091.0951.6530.6001.8531.5481.4561.7221.1381.3500.361

; E = experimental,

C

3.5012.9562.7142.6403.706

—2.6422.2473.0483.7633.2771.9262.9392.1672.3282.7202.3872.8100.548

E = experimental;

C

1.935 :1.8561.3811.1741.906

—1.4760.9811.5482.0302.1440.926 (1.3731.5340.8281.4541.021 (1.4700.415 (

K =

25%

E

3.4083.0642.7602.6403.706—

2.7932.3343.0023.4342.9592.0783.0862.1082.4352.8172.4242.8203.471

R =

2 5 %

E

!.1O8.798.360.207.906—.560.068.469.834.959

3.978.453

1.2421.1021.7173.9241.4803.376

recovery

R

3.3192.9362.6702.5843.734

—2.7312.3002.9853.4103.3212.0163.1232.0802.6332.7362.2412.8000.499

recovery

R

1.9191.7361.3041.1842.134

—1.631

(

.670

.552

.8102.055).866.557.447.433.603

0.9411.5500.357

C

3.4082.0322.7352.8973.4772.1162.5752.0312.6893.7342.7932.0913.3372.1012.4892.6852.3782.6800.545

C

1.5080.7341.6021.3311.5970.4191.3750.6651.3562.1011.8271.0581.6371.4671.2561.6190.6781.3100.457

2 1 %

E

3.5032.6862.6462.8773.5052.0622.5872.0502.7963.6392.8251.9833.1492.0682.6192.6422.4572.7100.516

2 1 %

E

2.2731.1861.5801.2671.745

R

3.4792.3012.6762.8973.4772.1992.5321.9712.7343.5052.7631.9833.1492.2732.4472.5322.4452.6703.494

R

.7081.068.510.697

1.7770.479 0.4591.387 1.4320.717 0.6381.4962.0731.7590.817 <1.4831.4351.3191.5920.6911.3700.485

.434

.9051.7303.7831.6491.3731.3141.4993.8121.3403.435

262

aH00

X<—i

Oc

o

w00

or>

OXYGEN CONCENTRATIONS AND LUNG VOLUMES 263

TABLE V. Analysis of variance: FRC

Strata and effects

Oxygen concentrationBreathing stateSubjectsOxygen-breathingOxygen-subjectsBreathing-subjectsResidualTotal

Sum ofsquares

1.04630.09060.66490.21512.06950.13150.42264.6405

d.f.

41

164

641662

167

Meansquare

0.26160.09060.04160.05380.03230.00820.0068

F ratio

38.3813.306.107.894.741.21

P

< 0.0010.001

< 0.001< 0.001< 0.001

0.289

TABLE VI. Analysis of variance: RV

Strata and effects

Oxygen concentrationBreathing stateSubjectsOxygen-breathingOxygen-sub j ectsBreathing-subjectsResidualTotal

Sum ofsquares

0.87640.05110.79630.11512.91800.53721.07496.3690

d.f.

41

164

641662

167

Meansquare

0.21910.05110.04980.02880.04560.03360.0173

F ratio

12.642.952.871.662.631.94

P

< 0.0010.0870.0020.170

< 0.0010.033

4 -,

-12Control Experimental Recovery

FIG. 3. Percentage change in FRC (mean, SEM) from control for the experimental situation after the gas mixture wasbreathed for 15 min, and the recovery situation after five deep breaths of room air. Oxygen concentrations: 21 % ( • ) ;

25% (O); 30% (# ) ; 50% ( • ) ; 100% (D)- Paired t test: *P<0.05; **P<0.01; ***P < 0.001.

<J

14 -12 -10 -8 -6 -4 -2 -0 -i

- 2 -

—** —- 6 :

- 8 -1 0 -

1 2 -

14

. — j

I

!

* * ^

I1• 1 • •

Control Experimental Recovery

FIG. 4. Percentage change in RV (mean, SEM) from control for the experimental situation after the gas mixture wasbreathed for 15 min, and the recovery situation after five deep breaths of room air. Oxygen concentrations: 21 % ( • ) ;

25% (O); 30% (» ) ; 50% ( • ) ; 100% ( • ) . Paired t test: *P < 0.05; **P < 0.01; ***P < 0.001.

264 BRITISH JOURNAL OF ANAESTHESIA

4 i2 -0 -

-2 -- 4 •

-6 --8 -

-10 --12

21 25 30 50

Oxygen concn(%)

100

FIG. 5. Regression graph of the percentage decrease in FRC(mean, SEM) from control that occurred with increasing concen-

trations of oxygen (note log scale).

The within-replicate SD was 0.163 litre for FRC,which compares favourably to the preliminaryreplicate procedures on the same day of 0.097 litrepublished for our previous study [19].

Figures 3 and 4 show the graphical representationof % volume change (SEM) from the control readingsfor the FRC and RV data. Because the analysis ofvariance is significant, paired t tests were used toassess the statistical significance of the differencebetween the means for both FRC and RV. Therewere statistically significant volume reductions inFRC for Fio , of 100%, 50% and 30% at both theexperimental and recovery stages. There were stat-istically significant volume reductions in RV for FiOjof 100 % and 50 % at the experimental stage, and for

-eas

e i

o<D

Q

12 -10 -8 -6 -4 -2 -0 -

-2 -

-4 --6 -- 8 -

- 1 0 --12 --14 --16

21 25 30 50

Oxygen concn(%)

100

FIG. 6. Regression graph of the percentage decrease in RV (mean,SEM) from control that occurred with increasing concentrations of

oxygen (note log scale).

100% oxygen at the recovery stage. These volumereductions at recovery are less than those immedi-ately after cessation of restricted breathing (theexperimental state), but show that some effectspersist even after efforts are made to re-expand thelungs.

Figures 5 and 6 show the means (SEM) for FRCand RV vs the logarithm of FiOf. There was a lineardecrease in lung volumes with greater FIO J , which isconfirmed as statistically significant at the 5 % levelby regression analysis for both FRC and RV (tablesVII, VIII). This statistical test was used because itevaluates best any effects of increasing FiOi.

TABLE VII. Regression analysis of variance: FRC. XI = Log of % oxygen concentration; X2 = % change from control

VariationSum ofsquares d.f.

Meansquare F ratio

Regression 0.26317444 1Residual 0.03768036 3Total 0.30085480 4Multiple R 0.9353Adj. R Sq (Fisher's A-Sq) 0.8748

0.263174440.01256012

20.953 0.018

Regression equation: XI -

Variable

X2Constant

= 1.386-0.054 JT2

TABLE VIII. Regression analysis of variance:

Variation

Regressioncoefficients

-0.054471.38563

RV.X1 = Log

Sum ofsquares

SE

0.011900.44665

of % oxygen

d.f.

t

-4.577

concentration

Meansquare

P

0.010

; X2 = %

F ratio

change from control

P

Regression 0.24435559Residual 0.05649921Total 0.30085480Multiple R 0.9012Adj. R Sq (Fisher's A-Sq) 0.8122Regression equation: XI = 1.492-0.033 X2

0.244355590.01883307

12.975 0.035

VariableRegressioncoefficients

X2Constant

-0.033101.49170

0.009190.44696

-3.602 0.023

OXYGEN CONCENTRATIONS AND LUNG VOLUMES 265

For both FRC and RV, there was a reduction ofapproximately 10 % in lung volume after breathing100% oxygen, and lesser reductions with increasing•FINJ. There were small mean increases in lungvolumes after both 21 % and 25 % oxygen mixtures,although these were not significantly different fromthe control values on paired t testing, and may havebeen a chance finding.

DISCUSSION

These results confirm both the reduction in lungvolumes shown in our previous studies [19, 24] andthe results from other studies in which subjectsbreathed 100% oxygen. This study confirms thework by Green [13], who showed that addition of20% and 40% nitrogen to oxygen reduced thetendency for alveolar collapse to occur. The presentstudy was more detailed, and showed that there is alog linear effect produced by adding nitrogen to anoxygen mixture up to a concentration of 75 %nitrogen. With the addition of increasing nitrogen tothe gas mixture, the reduction in lung volumescaused by breathing oxygen at a reduced lung volumewas lessened, and finally neutralized.

When lung expansion with deep breaths wasencouraged towards the end of each study, severalvolunteers noticed sharp chest pain which limitedtheir deep breaths. This confirms other experiences[6-10]. Recently, two studies [26,27] have shownlimitation in VC measurements in volunteers whobreathed oxygen at reduced lung volume caused byhigh gravitational forces. With progressive VCmeasurements, VC returned in steps to the controlvalue. Chest pain was noted at the limits of chestexpansion reached during each VC measurement.This chest pain may be caused by alveolarre-expansion after atelectasis. When atelectasis hasoccurred, greater forces are necessary to re-expandthe collapsed alveoli [28,29]. The shear forcesassociated with reopening the alveoli may be thecause of the sharp pain noticed by some volunteers.

Our results confirm that there was a decrease inlung volumes which persisted at least for a shorttime, and was still present after the subjects hadmade attempts to expand the lungs. The duration ofthis reduction was not studied, although this isimportant in applied situations. The time elapsing inthe genesis of such reduction in lung volumes mightalso be important. Longer times might predispose tomore atelectasis in other alveoli with marginal V/Qratios, and might also imply that greater expansionforces are required to re-expand increased atelectaticareas of lung. The time elapsed before any efforts aremade to re-expand the lungs may be important, aslonger periods of unresolved atelectasis may alsorequire greater expansion forces for re-expansion.

Studies using chest strapping have shown thatchanges in pulmonary mechanics, including in-creased elastic recoil of the lung, are caused solely bybreathing at low lung volume [30—34]. This makes itunlikely that the reduction in lung volumes wascaused by an increase in elastic recoil of the lung[19].

The small though non-statistically significantincrease in FRC and RV after this restrictedbreathing exercise with the increased nitrogen mix-tures of 79% and 75%, may be a chance finding,but it may be that breathing with a reduced FRCfollowing pressurization of the body plethysmographinduces some changes in ventilatory mechanics (suchas increased elastic recoil of the chest wall) whichthen produce larger ventilatory volumes immediatelyon cessation of the restricted breathing manoeuvre.If this is so, then the effect of increased FiOt is evengreater than it appears at first.

The variation between volunteers in the responseto breathing the gas mixtures at a reduced lungvolume mimics our earlier studies, and also that ofNunn and colleagues [10]. Such variation mayresult from different V/Q distribution, particularlyat the small ratios in normal healthy young volun-teers. It is possible that variation in collateralventilation may also be a factor [35]. This biologicalvariation may explain the results of Turaids, Nobregaand Gallagher [23] who found 5 % nitrogen pre-vented atelectasis in one subject.

Dantzker, Wagner and West [36] have shown acritical V/Q effect, during steady state, on thosealveoli with V/Q ratios between 0.002 and 0.08 whenFIO J beween air and 100% oxygen are considered.In particular, for a difference in FiOi which isdemonstrated at 30% but not at 25%, Dantzker'sgroup [36] predict that V/Q ratios of 0.003-0.004 arecritically important. An increase in FiOi shifts thecritical V/Q to greater ratios which may lead to ab-sorption atelectasis at greater V/Q ratios. Becausethe concept of critical V/Q depends on a balancebetween reduced ventilation and FiOl it is possiblethat with differing ventilation there will be differentconcentrations of nitrogen in oxygen which neutral-ize the effect on reducing lung volumes. The timeover which any reduced ventilation or increased FiOtacts is also important because time is required forcollapse to occur. In situations close to the criticalV/Q ratio where alveolar ventilation is just insuf-ficient to maintain alveolar volume, it will take sometime for collapse to occur, and probably longer thanthe time available in this study. Nevertheless, thisstudy did show a reduction in lung volumes withFiOt of 30 % or greater. Also, a reduction in lungvolume may result in a decrease in ventilation toalveoli already at a relatively low volume because ofthe compliance effect at low lung volumes. Thiscould further exacerbate critical V/Q ratios, leadingto absorption atelectasis.

The combination of increased FIO J and reducedlung volume has been demonstrated in this study toproduce a decrease in FRC and RV. We postulatethis lung volume reduction to be caused by ab-sorption atelectasis in alveoli with small V/Q ratios.When the V/Q ratio is less than the critical V/Qratio, the smaller the V/Q the more rapidly completealveolar collapse occurs. With the relatively shortduration of this study, complete alveolar collapsemay not have occurred in the time available, inalveoli with V/Q ratios less than but near to thecritical V/Q ratio. Thus it is likely that the thresholdV/Q ratios at which nitrogen prevents collapse in

266 BRITISH JOURNAL OF ANAESTHESIA

this study are smaller than the critical V/Q ratiospostulated by Dantzker, Wagner and West [36].

The reduction in lung volumes noted in this studyis important in at least three areas of appliedrespiratory physiology: aviation with oxygen breath-ing and high ^-forces [11-14]; diving medicine withincreased oxygen breathing and reduction in FRC[15-17]; and in anaesthesia. During anaesthesia thereis a reduction in FRC, and increased concentrationsof oxygen are often administered during the recoveryphase when the patient may be hypoventilating as aresult of residual anaesthesia, incomplete antagonismof neuromuscular paralysis, opioid administration orpain. There is likely to have been a prolonged periodduring anaesthesia where the patient breathed a gasmixture of oxygen and perhaps the more solublenitrous oxide. These results indicate that at least75 % nitrogen or other inert gas of low solubilitywould need to be added to the oxygen mixture toremove the effect totally. The addition of 50%nitrogen, however, does limit the reduction in lungvolumes by approximately 50 % and may still proveuseful in those situations in which FiOi greater than30% are indicated to prevent hypoxia. Although10% may be considered to be a small decrease inFRC or RV, it may be clinically critical in a patientwith reduced initial lung volumes as a result of pre-existing disease.

ACKNOWLEDGEMENTS

The authors thank Drs Malcolm Faddy and Tony Morton forhelp with statistical analysis.

REFERENCES1. Mendelssohn A. Der Mechanismus der Respiration und

Cirkulation oder das cxplicint Wesen der Limgenhyperamten.Berlin: B. Behrs Buchhandlung, 1845; 384.

2. Herausgeber V. Die Ursachen und die Beschaffenheit derjen-igen Veranderungen, welche das Lungenparenchym nachDurchschneidung der Nn. Vagi erleidet. Inhalt II in TraubeL. Beitrage zur Experimentellen Pathologic und Physiologic1846; 1: 65-200.

3. Lichtheim L. Versuche uber Lungenatelektase. Archiv furExperimentellen Pathologic und Pharmacologic 1879; 10:54-100.

4. Coryllos PN, Birnbaum GL. Studies in pulmonary gasabsorption in bronchial obstruction. American Journal ofMedical Sciences 1932; 183: 317-359.

5. Dale WA, Rahn H. Rate of gas absorption during atelectasis.American Journal of Physiology 1952; 170: 606-615.

6. Burger EJ, Macklen PT. The effect on the lungs of breathing100% pure oxygen near residual volume. Federation Pro-ceedings 1966; 25: 566.

7. DuBois AB, Turaids T, Mammen RE, Nobrega FT.Pulmonary atelectasis in subjects breathing oxygen at sealevel or at simulated altitude. Journal of Applied Physiology1966; 21: 828-836.

8. Nunn JF, Coleman AJ, Sachithanandan T, Bergman NA,Lews JW. Hypoxaemia and atelectasis produced by forcedexpiration. British Journal of Anaesthesia 1965; 37; 3-12.

9. Prys-Roberts C, Nunn JF, Dobson RH, Robinson RH,Greenbaum R, Harris RS. Radiologically undetectable pul-monary collapse in the supine position. Lancet 1967; 2:399-401.

10. Nunn JF, Williams IP, Jones JG, Hewlett AM, HulandsGH, Minty BD. Detection and reversal of pulmonaryabsorption collapse. British Journal of Anaesthesia 1978; SO:91-99.

11. Ernsting J. Some effects of oxygen breathing in man.Proceedings of the Royal Society of Medicine 1960; S3: 96-98.

12. Glaister DH. Acceleration atelectasis. In: The Effects of

Gravity and Acceleration on the Lung, chapter 8. AGARDMonograph No. 133, 1970; 121-149.

13. Green ID. The pathogenesis and physioloical effect ofpulmonary collapse induced by breathing oxygen and in-creased gravitational force (Ph.D. Thesis.) London: LondonUniversity, 1967.

14. Langdon DE, Reynolds GE. Postflight respiratory symptomsassociated with 100 per cent oxygen and G-forces. AerospaceMedicine 1961; 32: 713-718.

15. Balldin UI, DahJback GO, Lungren CEG. Changes in vitalcapacity produced by oxygen breathing during immersionwith the head above water. Aerospace Medicine 1971; 42:384-387.

16. Dahlbach GO, Balldin UI. Positive-pressure oxygen breath-ing and pulmonary atelectasis during immersion. UnderseaBiomedical Research 1983; 10: 39-44.

17. Dahlbach GO, Balldin UI. Pulmonary atelectasis formationduring diving with closed-circuit oxygen breathing apparatus.Undersea Biomedical Research 1985; 12: 129-137.

18. Baer R, Dahlbach GO, Balldin UI. Pulmonary mechanicsand atelectasis during immersion in oxygen-breathingsubjects. Undersea Biomedical Research 1987; 14: 229-240.

19. Baker AB, Restall R. Changes in residual volume followingoxygen breathing. British Journal of Anaesthesia 1983; 55:817-824.

20. Dery R, Pelletier J, Jaques A, Clavet M, Houde J. Alevolarcollapse induced by denitrogenation. Canadian AnaesthetistsSociety Journal 1965; 12: 531-544.

21. Garfinkel F, Fitzcrgerald RS. The effect of hyperoxia,hypoxia and hypercapnia on FRC and occlusion pressure inhuman subjects. Respiration Physiology 1978; 33: 241-250.

22. Rahn H, Farhi LE. Gaseous environment and atelectasis.Federation Proceedings 1963; 22: 1035-1041.

23. Turaids T, Nobrega FT, Gallagher TJ. Absorpu'onal atel-ectasis breathing oxygen at simulated altitude: preventionusing inert gas. Aerospace Medicine 1967; 38: 189-192.

24. Smith JG. The effect of breathing high inspired oxygen atreduced lung volume (B.Sc. Thesis). Otago: University ofOtago, 1986; pp. 63.

25. Cotes JE. Lung Function—Assessment and Application inMedicine, 3rd Edn. Philadelphia: Lippincott, 1975; 118-121,170-174.

26. Haswell MS, Tacker WA, Balldin UI, Burton RR. Influenceof inspired oxygen concentration on acceleration atelectasis.Aviation Space and Environmental Medicine 1986; 57:432^37.

27. Tacker WA, Balldin UI, Burton RR, Glaister DH, Gilling-ham KK, Mercer JR. Induction and prevention of ac-celeration atelectasis. Aviation Space and EnvironmentalMedicine 1987; 58: 69-75.

28. Levine BE, Johnson RP. Effects of atelectasis on pulmonarysurfactant and quasi-static lung mechanics. Journal of AppliedPhysidlogy 1965; 20: 859-864.

29. Mead J, Whittenberger JL, Radford EP jr. Surface tension asa factor in pulmonary volume-pressure hysteresis. Journal ofApplied Physiology 1957; 10: 191-196.

30. Caro CG, Butler J, DuBois AB. Some effects of restriction ofchest cage expansion on pulmonary function in man: anexperimental study. Journal of Clinical Investigation 1960;39: 573-583.

31. Douglas NJ, Drummond GB, Sudlow MF. Breathing at lowlung volumes and chest strapping: a comparison of lungmechanics. Journal of Applied Physiology 1981; 50: 650-657.

32. Manco JC, Hyatt RE. Relationships of air trapping toincreased lung recoil pressure induced by chest cage re-striction. American Review of Respiratory Diseases 1975; 111:21-26.

33. Stubbs SE, Hyatt RE. Effect of increased lung recoil pressureon maximal expiratory flow in normal subjects. Journal ofApplied Physiology 1972; 32: 325-331.

34. Sybrecht GW, Garrett L, Anthonisen NR. Effect of cheststrapping on regional lung function. Journal of AppliedPhysiology 1975; 39: 707-713.

35. Henderson R, Horsfield K, dimming G. Intersequentialcollateral ventilation in the human lung. Respiration Physi-ology 1969; 6: 128-134.

36. Dantzker DR, Wagner PD, West JB. Instability of lung unitswith low VK/Q ratios during O, breathing. Journal of AppliedPhysiology 1975; 38: 886-895.