Arch Dis Child: first published as 10.1136/adc.59.6.542 on 1 June 1984. Downloaded from

Archives of Disease in Childhood, 1984, 59, 542-547 Analysis of alveolar ventilation in the newborn

K SANDBERG, B A SJOQVIST, 0 HJALMARSON, AND T OLSSON

Department of Paediatrics I, University of Goteborg and Research Laboratory of Medical Electronics, Chalmers University of Technology, Goteborg, Sweden

SUMMARY Twelve healthy term infants were examined at the median ages of 21/2 and 26 hours. Their alveolar ventilation, efficiency of ventilation, functional residual capacity, and lung nitrogen elimination patterns were studied by means of a computerised nitrogen wash out method. The results showed that alveolar ventilation and functional residual capacity increased over the period studied. At the same time effective dead space decreased leaving minute ventilation unchanged. Distribution of ventilation did not change.

In one critical phase at birth the lungs of the interrupted and repeated. A second test was not newborn take over the gas exchange function from performed for at least 15 minutes to allow for gas the placenta by establishing effective ventilation. In equilibration. The nitrogen concentration in the face the healthy infant sufficient function is achieved mask and the respiratory flow rate were sampled by within the first few breaths and steady state condi- a computer for subsequent calculations. The analy- tions are present after a few hours.' sis of the nitrogen wash out curve was performed on The efficiency of the ventilation obtained de- a PDP 11/40 computer and the software was written pends mainly on the volume of the ventilated in FORTRAN. airspace (functional residual capacity), the dead The total expired nitrogen volume during wash space, the distribution of ventilation, and the gas out was determined by integrating the product of the mixing efficiency within the lungs. These factors expired ventilatory flow rate and the nitrogen have not been studied in combination in newborn concentration in the expired air after compensation infants. In this investigation a computerised analysis for equipment delay and plethysmograph character- http://adc.bmj.com/ of nitrogen wash out curves was used to calculate istics. The functional residual capacity was obtained functional residual capacity and alveolar ventilation by dividing the total expired nitrogen volume after and to analyse gas mixing over the first day of life in the beginning of oxygen breathing by the end healthy, term infants. expiratory nitrogen concentration before the ni- trogen wash out. The method was designed to give Methods the lung volume in BTPS. The reproducibility of the functional residual A detailed description of the method is given capacity estimation was tested by means of a on September 27, 2021 by guest. Protected copyright. elsewhere.2 3 For the examination each infant was mechanical lung model inserted into the plethysmo- placed in the supine position in a 'face out' volume graph. This has been described in detail elsewhere.2 displacement body plethysmograph, with a pneumo- The known volume of the mechanical lung fell tachograph (Fleisch 1) in the wall. When the child within the 95% confidence interval of the estimated had adapted to the plethysmograph, respiratory volume when 10 repeated estimations were per- frequency, tidal volume, and minute ventilation formed. The coefficient of variation was 2-9%. were calculated. After this a face mask (dead space No correction was made for the elimination of 2 ml) was gently placed over the infant's face and tissue nitrogen in the estimation of functional mouth. A nitrogen analyser (Hewlett Packard residual capacity. According to Groom et al4 and model 47302A) and a system of tubings were Lundins one minute of oxygen breathing, which was connected to the face mask. During an expiration the approximate time for a nitrogen wash out test in the infant's breathing air was instantaneously this study, would add approximately 0-7 ml N2/kg changed to 100% oxygen and a nitrogen wash out body weight to the expired N2 volume implying an was performed until the end expired nitrogen overestimation of functional residual capacity by concentration was below 2%. If the infant was only about 1 ml/kg body weight. disturbed or if gas leaks were observed the test was A total lung nitrogen volume elimination curve 542 Arch Dis Child: first published as 10.1136/adc.59.6.542 on 1 June 1984. Downloaded from

Analysis of alveolar ventilation in the newborn 543 was constructed with breath number on the abscissa In the analysis of the nitrogen elimination pattern (Fig. 1). This volume elimination curve was used for the best estimate of the observed elimination curve the further analysis of the nitrogen elimination was sought by fitting one, two, and three component pattern. In the ideal case the nitrogen elimination models to the linear curve by using the z transform,6 was assumed to be described by a single exponential multiple linear regression, least square fit, and F test function: of the residuals. The estimated curve was considered VLn=VLOe an=VLOWn (equation 1) to fit the original nitrogen elimination curve well (n=breath number, VLn=nitrogen volume remain- when the difference in curve areas-both curves ing in the lung after n breaths, VL0=nitrogen having the first point in common-was less than 5%. volume in the lung before the elimination, if the difference was more than 5% the wash out a=constant) where e-a=W is the dilution factor curve was considered insufficiently described by our defined as W=functional residual capacity/ exponential models. (functional residual capacity+tidal volume-dead The value of tidal volume minus dead space was space). obtained from the dilution factor in the best single Deviations from the ideal nitrogen elimination exponential fit to the wash out data according to pattern were assumed to be described by a sum of equation 1. As functional residual capacity and tidal exponential components with different dilution fac- volume were known, dead space could be estimated. tors. In this case the nitrogen elimination can be This estimated dead space is, by definition, the described as virtual part of the tidal volume not participating in VL, = F1VLOWL +F2VLOW2 +...... +FkVLOWkn the gas exchange between the inspired and alveolar (equation 2) gas, referred to as an effective dead space.7 In the elimination course dead space where VLO=FlVLO+F2VLO+...... +FkVLO and single exponential F1,F2. and Fk, are fractions of the total nitrogen may easily be estimated from the dilution factor. In lung volume. W1, W2...... and Wk are the dilution the multiple exponential course, however, the dead factors for the different exponential components in space changes continuously during nitrogen elimina- the elimination pattern. tion and cannot easily be described by a single value.

100 http://adc.bmj.com/

- E on September 27, 2021 by guest. Protected copyright.

g e

Breath No Fig, 1 Example ofthe nitrogen volume elimination curve with percentage ofthe total expired nitrogen volume on the ordinate and breath number on the abscisse. Arch Dis Child: first published as 10.1136/adc.59.6.542 on 1 June 1984. Downloaded from

544 Sandberg, Sjoqvist, Hjalmarson, and Olsson In these cases dead space was estimated from the dilution factor achieved by the best single exponen- tial least square fit to the data (equation 1). In this way the estimated dead space is the one obtained in LQJ q a g. .q a .3. an ideal system having the same lung volume and -w-- XN ZC5 4 ., z:_ -- 14 IC = tidal volume as the lung examined and represents a = r- WI, .: Z 7.c tr,, - OC rz. 7 weighted value of the continuously changing dead space. The alveolar ventilation was calculated as (tidal volume minus dead space)xrespiratory frequency At and the effective breath fraction as tidal volume v+,X- X~ ell minus dead space/tidal volume. The nitrogen clear- ance was calculated as the necessary ventilation for X dilution of the lung nitrogen to the end tidal nitrogen concentration of 2% divided by the calcu- _I lated functional residual capacity. In the statistical analysis Student's paired t test, 9 .5 Fisher's exact test, and analysis of covariance were A It a: used. Ninety five per cent confidence intervals for 0 the mean differences were calculated. Results are presented as mean (SEM). 11

.< Patients .,

0- -.-.cl Twelve vaginally delivered, healthy, term infants Oc W< ON Nt C CsC 11 were studied during the neonatal period. The I in median gestational age and range were 40 5 (38 to 42) weeks and the median birthweight and range Cr C' ., were 3-64 (2.95 to 4.15) kg. All infants were studied 'ii~ .5 r. twice at a median age and range of 2½/2 (2 to 4) and 0 http://adc.bmj.com/ 26 (21 to 33) hours. Six infants were boys and 6 were r-oNocNo r.n girls. No infant showed signs of respiratory disease (respiratory frequency greater than 60/minute, 11 cyanosis, retractions, or grunting). Nor were there E signs of perinatal asphyxia, cardiac disease, or o infection. The neonatal period was uneventful in all * infants. The investigation was approved by the local rz~X>~>'s > I; ethical committee and informed maternal consent 72 C) on September 27, 2021 by guest. Protected copyright. was obtained before the examination in all cases. _t IC I CI u Results '0 SZ

The ventilatory parameters before and during the C'N ~ wC> nitrogen wash out are presented in Table 1. There .,. s 'A was an increase in minute ventilation during the 'No nitrogen wash out measurement compared with that .e1 found when the baby was sleeping in the plethysmo- C)eu graph without the face mask. There were, however, i no significant increases in minute ventilation, tidal C%) volume, and respiratory frequency between 21/2 and H~CO r. 26 hours of age. During this time the functional residual capacity increased significantly from mean ._o. (SEM) 65 (5-4) ml to 81 (4-4) ml (P<0-05) and the u alveolar ventilation also increased significantly from V 383 (32) ml/min to 511 (39) ml/min (P<0c05). "I Arch Dis Child: first published as 10.1136/adc.59.6.542 on 1 June 1984. Downloaded from

Analysis of alveolar ventilation in the newborn 545 During wash out the dead space decreased from (SEM) 0-51 (0-02) and 0-60 (0-02) respectively 7-8 (0-7) ml at 21/2 hours to 7-0 (0-6) ml at 26 hours (P<0-02). of age. At both examinations, however, there was a The analysis of the nitrogen elimination curves significant correlation between dead space and tidal (Table 2) showed single exponential courses as the volume (P<0(01) (Fig. 2). When the variance in best fit in 7 of 12 infants at 21/2 hours and 5 of 12 dead space related to differences in tidal volume was infants at 26 hours of age. This difference was not compensated for by means of covariance analysis, statistically significant. The other infants showed a the difference in dead space at 21/2 and 26 hours of two exponential elimination curve as the best fit in 1 age was significant, (P<0.05). of 12 and 4 of 12 infants at the two examinations The ventilation efficiency and the nitrogen respectively. No infant had a three exponential wash elimination pattern are presented in Table 2. There out pattern. In four curves at 2½/2 hours of age and was significant improvement in nitrogen clearance three curves at 26 hours of age the exponential from mean (SEM) 9-6 (0-3) to 8-5 (0-2) between 2½/2 estimates deviated more than 5% from the original and 26 hours of age (P<0.01). The effective breath data. Thus, these curves could not be described fraction also increased significantly during this time sufficiently by our nitrogen elimination models. The from mean (SEM) 0-45 (0-02) to 0-53 (0-02) individual alveolar dilution factors, the lung frac- (P<0-01). When the equipment dead space was tions, and the deviations are presented in Table 3. In excluded in the calculation the values were mean Table 3 Nitrogen elimination curve analysis

Itnfant Age Alveolar dtiluitioto iacttors Lutntg fra( tiotnts Deviteao 15 No (ht) > 5'S. . W WI WI F, F, I ~~~~~~~2 (0-934 26 (0-884 (0-238 (0-898 (0-08 01-92 10- 0 0)-924 + 24 ()-9(K( VD (ml) 3 2 (0-833 (0-723 (0-889 (0.4 18-59 0 0 I-0100 24 (0-876 (0-362 (0-9(02 (0-15 4 22½ 0-892 5- 26 ()(9(1) 5 21' 0-921 0 23 (-866 6 3 0-890)

28 019 15 http://adc.bmj.com/ 7 2 0)-914 0- (0-874 ((}-9(1 (1-899 11-I.t 01-87 + lb 15 20 2525 *b 8 ~~~~~~~~214 (0-857 VT (ml) 33 (0-902 ()-t71 () X94 ()-Ir~~~~~~~~ 9 3 (0-847 (1.497 (1-863t ((-((9 (1-91 + Fig. 2 The correlation between tidal volume (VT) and 29 0-8901 corresponding dead space (VO)) at 2'112 (0) and 26 (O) hours 10 3 (1-X89 25 (0-897 of age. Linear regression line at 2''2 hours of age (1): 3 (1-911 VI=-4 06+0 66 VT (r= 080, P<( O1) and at 26 hours of 28 0-898 age (11): V)=-3-18+0 53 V-,- (r=O.79, P-88

Table 2 Efficiency of ventilation and nitrogen elimination pattern

Age Ditferentce 2' Ii 26 hi M(at9 c(tofittdentce interval Meati (SEM) Meat, (SEM) Nitrogcn clea.ranice 9-6 ((1 3) 8 5 ((0 2) -2 (0 4 to 9 Eftcctive hre.ath tr.actioni ((V-o-Vl)/V1-) (0-49 (0()2) () 3 ((02))00 -08 -0- 13 to -( 03 Effcctivc hrcath fraction' )-51 0(0)2) (0-6(0 ((002) -0-08( -(-1 to -0(02 Nitrogcn climinnatiott pattcrn: One expoitenititl 7 § Two cxpottcittial 4 § Thrcc exponenitial () Other 4 3 9P<0)-0)i* P<(0-(02 'Equipment de.Id space (2 ml) excluded. Ditffcrencc not significant (Fisher's exaict test, P>.-1(5). Vt-tidal volume- Vl0=dcad space. Arch Dis Child: first published as 10.1136/adc.59.6.542 on 1 June 1984. Downloaded from

546 Sandberg, Sj6qvist, Hjalmarson, and Olsson the infants with a two exponential model as the best efficiency, measured as nitrogen clearance and fit one of the components was small and fast in all effective breath fraction, improved significantly cases but one. between 21/2 and 26 hours of age. This also means that the effective dead space decreased during this Discussion time. The dead space value measured in this way is not, however, a static volume but a function of other Although the nitrogen wash out method has not ventilatory variables. We found a significant correla- been used frequently on newborns,8 it has often tion between dead space and tidal volume at both been used on adults to study the distribution of 21/2 and 26 hours of age-a correlation that has been ventilation and to measure the functional residual found previously in neonates and in adults. 17 lX This capacity.1(' We have designed a nitrogen wash out means that comparison between dead space at method which minimises interference with the in- different measurements can only be made in connec- fants' breathing2 and in addition, analyses objec- tion with the corresponding tidal volume. When, in tively the nitrogen volume elimination curve without this study, analysis of covariance was used to using semilogarithmic plots.3 A similar approach has eliminate variance due to differences in tidal volume been used by Cumming et al;1 12 they, however, between the examinations, the difference in dead analysed the efficiency of ventilation from a repre- space between 21/2 and 26 hours of age became sentative litre of the lung volume while our esti- obvious (Fig. 2). mates of this and the nitrogen elimination pattern Dead space as a function of other ventilatory are based on the total expired nitrogen volume. This variables can be understood from recent studies on study method has been used to investigate ventila- gas mixing in the human lung. '9 Ventilatory flow tory adaptation during the first day of life in healthy, rate and lung expansion have implications for the vaginally delivered term infants. diffusion between inspired and alveolar gas and There is one study only, that of Prod'hom et al,'3 consequently affect the effective dead space in the that investigates the development of alveolar ven- lung. From animal studies it is known that the tilation during the first day of life starting at 1 hour newborn lung is hyperhydrated.2(k22 It is also known of age, and in this no changes in total ventilation and that lung oedema preferentially locates in the alveolar ventilation between 1 and 24 hours of age interstitial tissue round airways and vessels.23 Accu- were found. The infants studied were delivered by mulation of interstitial lung fluid during the first caesarean section and the mothers had diabetes hours of life may affect the total cross sectional area of these infants had of neonatal of the small airways in the lung. As the contact area mellitus. Some signs http://adc.bmj.com/ asphyxia, some had tachypnoea during the first between inspired and alveolar gas is im ortant for hours of life, and they were also slightly premature the efficiency of gas mixing by diffusion ,) this may with gestational ages ranging from 35 to 37 weeks. explain the reduced ventilatory efficiency found at As asphyxia and caesarean section are factors 21/2 hours of age compared with later. Since total affecting ventilation in neonates'4 these infants can ventilation was unchanged during this time con- hardly be regarded as 'normal'. tinuous drainage of interstitial fluid with an increase All infants in our study were term, vaginally in the total cross sectional area of the small airways delivered, and without perinatal asphyxia. We may explain the resulting improvement in ventila- on September 27, 2021 by guest. Protected copyright. found a significant increase in alveolar ventilation tory efficiency found in this study. The significant between 21/2 and 26 hours of age while the total increase in functional residual capacity is an indi- ventilation was unchanged. The increase in alveolar cation of the same process. ventilation parallels the increase in oxygen con- During the period after birth when resorption of sumption during the first day of life found in interstitial lung fluid occurs it could be assumed that previous investigations.'5 16 In a study of basal the inspired gas is unevenly distributed into the lung oxygen consumption in term infants at the ages of 3 and that some areas of the lung are poorly ventilated and 24 hours, Hill et al found an increase in oxygen in comparison with others. This may lead to parallel consumption of 38%.16 After the age of 24 hours inhomogeneity in the ventilation of the lung with only a slight further increase was found. This different lung compartments ventilated at varying increase in oxygen consumption is in good agree- rates. In previous studies of nitrogen wash out ment with our finding of an increase in alveolar curves great efforts have been made to analyse ventilation of 33% between the ages of 21/2 and 26 whether lung compartments are ventilated in paral- hours. lel or whether stratified gas inhomogeneity exists.24 If alveolar ventilation increases under conditions Recent model analyses, however, have shown the of unchanged total ventilation, ventilation efficiency complexity of gas mixing in the lung.25 Evidently must have improved. We found that ventilation concentration gradients normally develop and Arch Dis Child: first published as 10.1136/adc.59.6.542 on 1 June 1984. Downloaded from

Analysis of alveolar ventilation in the newborn 547 change during the whole breathing cycle throughout Lundin G. Nitrogen elimination during oxygen breathing. Acda Physiol Scand 1953;suppl 111:130-43. the lungs and interpretation of nitrogen elimination Terell TJ. Introduction to digital filters. London: Macmillan curves in terms of parallel or stratified inhomogen- Press, 198t). eity has been questioned. '9 Astonishingly, in spite of 7Cumming G, Semple SJ. Disorders of the respirators' system. this complexity in the distribution of alveolar gas Oxford: Blackwell Scientific Publications, 1973. during oxygen 8Hanson JS, Shinozaki T. Hybrid computer studies of ventilatory concentrations, nitrogen elimination distribution and lung volume. Pediatrics 1970;46:9()0-14. breathing often follows a single exponential course Strang LB, McGrath MW. Alveolar ventilation in normal and may be expressed by a single model. In this newborn infants studied by air wash-in after oxygen breathing. study we found single exponential elimination pat- Clin Sci 1962;23:129-39. infants at both examinations. Fowler WS, Cornish ER, Seymour SK. Lung function studies terns in half of the VIII. Analysis of alveolar ventilation by pulmonary N2 clearance Some of the other infants' curves could be described curves. J CliGi Invest 1952;31:40-}5(. using the sum of two exponential curves but in some Cumming G. Jones JG. The construction and repeatability of curves exponential fitting was not possible at all lung nitrogen clearance curves. Respir Phvsiol 1966;1:238-48. When the curve 12 Cumming G. Gas mixing efficiency in the human lung. Respir according to our model and criteria. Physiol 1967;2:213-24. could be described by the sum of two exponential 3 Prod'hom S. Levison H, Cherry RB, Drorbaugh JE, Hub- components the additional component was often bell JP, Smith CA. Adjustment of ventilation, intrapulmonary small and fast (Table 3). There were no significant gas exchange, and acid-base balance during the first day of life. the Normal values in well infants of diabetic mothers. Pediatrics differences in elimination patterns between 1964;33:682-93. examinations (Table 2); consequently no detectable 4 Hjalmarson 0, Krantz ME, Jacobsson B, Sorensen SE. The change in significant parallel ventilation due to importance of neonatal asphyxia and caesarean section as risk elimination of interstitial lung fluid could be found. factors for neonatal rcspiratory disorders in an unselected elimination population. Acta Paediatr Scantd 1982;71:403-8. The presence of a single exponential I5 Karlberg P. Determination of standard energy metabolism pattern in about half the infants is in agreement with (basal metabolism) in normal infants. Acta Paediatr Scand findings in healthy adults."' 1952:suppl 89. In summary, this study has shown that adaptive Ih Hill JR, Rahimtulla KA. Heat balance and the metabolic rate of changes occur in healthy, term infants with increases newborn babies. J Phvsiol 1965;180:239-65. " Chu J, Clemets JA, Cotton EK, Klaus MH, Sweet AY. in alveolar ventilation, efficiency of ventilation, and Tooley WH. Neonatal pulmonary ischemia. Part 1: Clinical and functional residual capacity between 2 and 4, and 21 physiological studies. Pediatrics 1967;40:7t)9-82. and 33 hours of age. lx Wolff CB, Garratt RC. Measurement of dead space ventilation. J Appl Physiol 1982;53:297-8. 9 Engel L. Gas mixing within the acinus of the lung. J Appl

Phvsiol 1983;54:609-18. http://adc.bmj.com/ This study was supported -by the Swedish Medical Research 2(1 Aherne W, Dawkins JR. The removal of fluid from the Council, Project No. B180-19X-)5703, and The Mcdical Faculty; pulmonary airways after birth in the rabbit, and the effect on this University of Gotehorg. Sweden. of prematurity and prenatal hypoxia. Biol Neonate 1964;7: 214-29. References 21 Adams FH, Yanagisawa M, Kuzela D, Martinek H. The disappearance of fetal lung fluid following birth. J Pediatr Karlberg P. The adaptive changes in the immediate postnatal 1971;78:837-43. period, with particular reference to respiration. J Pediatr 22 Bland RD, McMillan DD, Bressack MA, Dong L. Clearance of 1960:56:585-604. liquid from lungs of newborn rabbits. J AppI Phvysiol

2 Sjoqvist BA, Sandberg K. Hjalmarson 0. Olsson T. Assessmnentt 1980;49:171-7. on September 27, 2021 by guest. Protected copyright. of lutng volumnes in newborn in.fants bY a computer assisted 23 Staub NC, Nagano H, Pearce ML. Pulmonary edcma in dogs, plethysmnographic ntitrogetn wash-out mnethod. (Technical Report especially the sequence of fluid accumulation in lungs. J Appl 4:83) Goteborg, Sweden: Research Laboratory of Medical Physiol 1967;22:227-40. Electronics, Chalmers University of Technology. t983. 24 Fowler WS. Intrapulmonary distribution of inspired gas. Phvsiol 3 Sjoqvist BA. Sandberg K, Hjalmarson 0, Olsson T. Analysis of Rev, 1952;32:1-20. alveolar venitilationi; a mnethod suit(ablefor clinical use in newborn 25 Paiva M. Engel LA. Pulmonary interdepcndencc of gas trans- inifanits. (Technical Report 5:83) Gdteborg, Sweden: Research port. J AppI Phv.siol 1979;47:296-3)5. Laboratory of Medical Electronics. Chalmers University of Tcchnology. 1983. Correspondencc to Dr K Sandbcrg, Departmcnt of Paicdiatrics 1. 4 Groom AC. Morin R, Fahri LE. Determination of dissolved N2 East Hospital, S-41685 G6teborg. Swcden. in blood and investigation of N2 wash-out from the body. J Appl Phsvsiol 1967;23:706-12. Received 20) February 1984