Attempts at estimating mixed venous carbon dioxide tension by the single-breath method. H. Ohta, 0. Takatani, T. Matsuoka

Size: px

Start display at page:

Download "Attempts at estimating mixed venous carbon dioxide tension by the single-breath method. H. Ohta, 0. Takatani*, T. Matsuoka*"

Marvin Briggs
5 years ago
Views:

1 Er Respir J 1989,2, ~95 TECHNICAL NOTE Attempts at estimating mixed venos carbon dioxide tension by the single-breath method H. Ohta, 0. Takatani*, T. Matsoka* Attempts at estimating mixed ve1ws carbon dioxide tension by the singlebreath method. H. Ohta, 0. Takatani, T. Matsoka. ABSTRACT: The single-breath method was originally proposed by KIM et al. [1] for estima.ting the blood carbon dioxide tension and cardiac otpt. Its reliability has not been proven. The pre.sent stdy was ndertaken, sing dogs, to compare the mixed venos carbon d ioxide tension (P9co ) 1 calclated by the single-breath method with the Pvco 1 meac;red In mixed venos blood, and to evalate the Inflence of variations In the exhalation dration and the volme of expired air sally discarded from comptations as the deadspace. Among the exhalation drations of 15, 30 and 45 s tested, the 15 s dration was fond to be too short to obtain an analyzable O, crve, bt at either 30 or 45 s, the calclated vales of PVC0 2 were comparable to the measred Pvcol. A significant agreement between calclated and was obtained when the expired gas with less than 22 was considered as deadspace gas. Er Respir J., 1989, 2, The Third Dept of Internal Medicine, National Defence Medical College, Japan. Correspondence: Dr H. Ohta, The First Dept of Internal Medicine, Shinsh University School of Medicine, Asahi, Matswnoto, 390, Japan. Keywords: Blood gas estimation; deadspace; exhalation dration; expired gas analysis; mixed venos blood carbon dioxide tension. Received: September, 1988; accepted after revision September 23, The single-breath method was developed by KIM et al [1], originauy for the prpose of estimating the mixed venos and arterial carbon dioxide tension and cardiac otpt. The procedres described by them are as follows: take aliqots of expired gas dring a prolonged single exhalation, and measre their carbon dioxide tension (Pco ) 2 and oxygen tension (Po ). 2 Plot the against the Po 2 for each sample, and draw a smooth crve throgh these points. Pt the slope of the crve at a given point into the alveolar eqation with corresponding vales of Po 2 and to calclate instantaneos R (Rinst). Then plot the alveolar carbon dioxide tension (PAco ) 2 against the Rinst, connect the points with a straight line, and read the vales of mixed venos carbon dioxide tension (PVco ) 2 and (PAco ) 2 at appropriate Rinst vales. Many investigators have assessed this method, and some have demonstrated a good agreement between the vales calclated by the single-breath method and those obtained by blood gas analyses [2-4]. Others have failed to show sch an agreement [5]. In practising the single-breath method, there have been differences in e.g. dration of exhalation, choice of series deadspace volmes and the volme of expirate. No standard procedre has been established. In this stdy, we sed the mixed venos carbon dioxide tension as a standard for a comparison prpose and stdied the effects of variations in the exhalation dration and the choice of series deadspace on the reslts of the single-breath method. We sed dogs as an experimental model. Methods The experimental set-p is illstrated in figre I. Five mongrel dogs (mean weight 12 kg) were anaesthetized with pentobarbital sodim (30 mg kg- 1, i.v.). A tracheal cannla was inserted and pancronim bromide (0.08 mg kg- 1, i.v.) was administered. The dogs were artificially ventilated with air sing an animal ventilator. C02[S] X-Y recorder MASS SPECTROMETER \ 0 2 Fig. I. - Schematic drawing of experimental set-p.

ESTIMATING MIXED vt:nn'js CARBON DIOXIDE TENSION 91 The gas in the tracheal cannla was continosly analysed by a mass spectrometer (Centronic 200MGA, time Jag 500 ms) which was calibrated with

2 ESTIMATING MIXED vt:nn'js CARBON DIOXIDE TENSION 91 The gas in the tracheal cannla was continosly analysed by a mass spectrometer (Centronic 200MGA, time Jag 500 ms) which was calibrated with standard gas mixtre before and after the sample analysis. Blood gas tension was detennined sing a blood gas analyser (IL813). A Swan-Ganz catheter was passed throgh a femoral vein and its tip was positioned in the plmonary artery to sample mixed venos blood. The procedre sed to obtain prolonged exhalation was similar to that of Mm-IAMMED and HAJ:NswoRTH [3J; when the end-tidal and heart rate became constant, the dog was disconnected from the respirator. and the lngs were deflated passively and then inflated manally, sing a 2 l syringe, with an air volme twice the tidal volme. The plnger was then withdrawn immediately, bt as slowly and continosly as possible, to the initial resting end-expiratory lng volme. Since the expiratory flow rates were not fixed between the procedres, the different exhalation drations corresponded to differing expiratory flow rates. The concentrations of 0 2 and C0 2 measred by the mass spectrometer were recorded on a X-Y recorder (Riken Denshi, F-42CP) giving the original 0 2 crve. From this crve, 50 points were read manally and, after eliminating the points which corresponded to the deadspace, a smooth crve, expressed as a qadratic regression, was drawn throgh these points sing the least-sqares techniqe. This regression crve was sed for sbseqent analysis. The first derivative, which is eqivalent to the slope (s) of the tangent of the crve, was determined at varios given portions of the crve. Using these s-vales, the Rinst was obtained by the following eqation [1): Rinst = (s-fo 2 s-fco 2 )/(l-fo 2 s-fco 2 ) where Fo 2 and Fco represent the expired gas fraction. The was then piotted against the Rinst, and a straight line was drawn throgh the plotted points. The at Rinst=0.32 was assmed to be the PVco 2 (calclated Pvco ). 2 Mixed venos blood was withdrawn jst before the single-breath exhalation, and the of the blood was detennined with the blood gas analyser (measred PVco ) 2 to compare with the calclated PVco 2 The tidal volme and respiratory freqency of the respirator were varied between each single-breath exhalation in order to obtain varios vales of PVcor The exhalation drations sed were 15, 30 and 45 s. There are several ways to define the deadspace. We adopted the method in which the deadspace was defined by the vales of in expired air, since this method had been employed by many other investigators [2, 6, 7] and is easy to perform. The ct-off vales for chosen were 10, , 25, and 30. Figre 2 shows an example of the 0 2 crve and ct-off C0 2 levels which define the deadspace. Within 50 points, read manally, there were almost no points which precisely corresponded to the ct-off vales of deadspace, so we adopted the point which was nearest to the ct-off vales of dead space. At exhalation drations of 30 and 45 s, scattergrams of measred PVco 2 vs calclated Pvco 2 were constrcted for the five different deadspaces defined by varios in expired air as mentioned above. A regression line was drawn on each scanergram, providing five regression lines at each exhalation dration. We assmed that a slope closer to 1.0 and an intercept on the Y -axis closer to zero indicated the closer agreement between the calclated PVco 2 and the measred PVco 2 Using the vales of the slopes and intercepts of these regression lines, the best condition for estimating PVco was fond by regression analysis. In this stdy we selected the data by se of the Smimov test. Poo, Po 2 Fig An example of ihe 0 crve in which dration of exhalation is 30 s. Each vale ol Pco 1 means the level of assmed deadspace volme. Reslts Since observed 0 2 crves were nearly straight at an exhalation dration of 15 s and analysable 0 2 -C0 1 crves cold hardly be obtained, the reslts achieved at this exhalation dration were not inclded in frther analysis. Figre 3 shows the reslts obtained at a constant exhalation dration of 30 s and varios deadspaces. For each deadspace volme defined by a level of 10, 15, 20, 25 or 30, there was a significant (r= , p<0-01) positive correlation between the measred and the calclated Pvco 2 vales. The reslts obtained at an exhalation dration of 45 s are shown in figre 4. Here again, measred and calclated PVco 2 vales were significantly (r= , p<0.01) correlated with each other. The relationship between the slope of the regression line and the level of presmed deadspace air for the 30 s dration is shown in figre SA, and the relationship between the Y -intercept of the regression line and the level in figre 5B. Regression analysis indicated that both relationships were first-order regression (p<0-05) and the best approximation was obtained when the ctoff for the deadspace was assmed to be 22, namely at the slope and the intercept being close to 1 and 0, respectively. The relationship between the slope and the level and that between the intercept and the Pco, 2 at the 45 s exhalation dration, are shown in figre 6A and

3 92 H. OHTA, 0. TAKATANI, T. MATSUOKA IDO A I> y:2.29x-34.n o" I~ "i 60 r:0.88-1'0 "S 40 "B 20 i -1'0 "S iu y:1.56x r: g"' I> 40 '0 Cll Cii "S 20 iu c y =0.97x r =0.93 "'---L-----' ='- 60 Is"' ~ 40 i Cii "5 20 "B D y = 0. 79x r = l "i -1'0 '3 iu E 60 y = 0.43x ,6,..._0. 0 Fig Relationship between and calclated P9cot at exhalation dration 30 s. A: of deadspace gas is assmed <10, 8: <15 ; C: <20 ; D: <25 ; E: <30. Each case shows significant correlation (p<o.oi). 100 A 8 g"' I> 8 i 60 Y=2.01x I~ as r=0.89 i Y=1.4x '3 j! 40 Cii 40 r=0.9 ~ "S j! 20 ~ c D E y =0.84x y = D. 70x I> ~ r = 0.87 I~ i i ea 1'0 'B "S -ea "S - - i "B "B Fig. 4. -Relationship between measred PVco 2 and calclated PVco 1 at exhalation dration 45 s. A: Pco 1 of dead space gas is assmed <10, B: <15 ; C: <20 ; D: <25 ; E: <30. Each case shows significant correlation (p<o.ol). "B

4 ESTIMATING MIXED VENOUS CARBON DIOXIDE TENSION 93 Slope A Intercept ~ Fig Relationship between Pco 1 level of deadspace gas and slope and intercept of regression line (represented in fig. 2). Exhalation dration is 30 s. A: Relationship between level of deadspace and slope of regression line. 8: Relationship between level of deadspace and intercept of regression line. A Intercept 8 Slope Fig Relationship between Pco 1 level of deadspace gas and slope and intercept of regression line (represented in fig. 3). Exhalation dration is 45 s. A: Relationship between level of deadspace and slope of regression line. B: Relationship between level of deadspace and intercept of regression line. figre 6B, respectively. As in the experiment at the 30 s exhalation dration, the significant first-order regression line (p<0.05) cold be drawn for both relationships and the best approximation was obtained when the ct-off for deadspace was assmed to be less than 22. Discssion Althogh all previos investigations of this method were based on the same principle, described by KrM et al. [1], there are some differences in detailed procedres. The main cases that lead to varios appraisals of this method are differences in the manner of smoothing the data sampled from the 0 2 crve,of determining the eqation of the crve, and of defining the dead space, and dration of exhalation. In data sampling from the 0 2 diagram KrM et al. [1) discarded any points which deviated from the crve by more than ±0.35 in either 0 2 or C0 2, and if more than 3 ot of 7 points were ths rejected the whole experiment was discarded. GILBERT and AcHINcLoss [8] discarded the crve if more than one point of deviated by more than 0.2% from the crve. CHEN et al. [2] rejected any points that deviated from regression crve by more than 0.7 and compted the entire regression again. If more than 25% of

5 94 H. OHTA, 0. TAKATA"'I, T. MATSUOKA the points were rejected, they discarded the whole breath. On the other hand, HLAsTALA et al. [5], BoERER et al. [6] and MoHAMMED and eo-workers [3,4] smoothed the crve withot rejecting any points, as we also did in this stdy. With regard to the eqation for the 0 2 crve, BoERER et al. [6] and CHEN et al. [2) sed qadratic regression, HLAsTALA et al. [5) and GR0NLUND [9] employed third-order polynomial eqations, and lnman et al. (7) sed an exponential eqation. GILBERT and AUCHJNCLOSS [8), MOHAMMED and eo-workers [3, 4] and BUDERER et al. [6] did not describe their method of approximation. GWNLUND [9] sggested, in his simlation stdy, that minte differences in the approximation crve led to large variations in the calclated reslts. We fitted the data by qadratic regression becase this polynomial order reslted in the least residal sm of sqares by the least-sqares techniqe, indicating this regression to be the best fitted approximate polynomial. A similar problem occrs in fitting a polynomial for the R- crve. CHEN et al. [2] sed qadratic regression and obtained calclated reslts from the tangent line drawn at R=0.38. We fitted regression lines by the leastsqares techniqe. The manner of discarding the early portion of expired air which contains predominantly deadspace air varied with athors. According to KIM et al. [1], the first part of the expiration was delivered to a 500 ml bag, ths eliminating the deadspace gas fl'om the sampling. GILBERT and AcHINCLOss [8] did not describe how the deadspace gas was treated. lllastala et al. (5] measred the deadspace dring each tidal breath, which was eqivalent to the anatomical deadspace, sing C0 2 as an indicator. BUDERER et al. (6), CHEN et al. (2) and INMAN et al. [7] assmed Lhe dead space as the part of the 0 2 crve corresponding to the less than 30, whereas MoHAMMED and eo-workers [3, 4] eliminated the first 2 s of the expiration. Althogh the manner of HLAsTALA et al. [5] seems reasonable, their method can be performed only when tidal breathing is adopted and might be impossible for prolonged exhalation. ln this stdy, we obtained the best calclated vales with the deadspace defined as the part of the 0 2 crve corresponding to less than 22. The exact volme of the deadspace to be eliminated from the expiration has not been examined critically. With regard to the dration of exhalation, a dration over 10 s was sed by K:!M et al. [1] and MoHAMMED and eo-workers [3, 4], whilst CHEN et al. [2] adopted precisely 10 s dration. HL.AsTALA et al. [5] and GR0NLUND [9] demonstrated that the longer the dration, the smaller the deviation of vales of cardiac otpt. In or stdy, the drations of exhalation of 15, 30 and 45 s were examined. With 15 s dration it was impossible to calclate the slope of the tangent. On the other hand, with 30 or 45 s dration, the 0 2 crve cold be compted and similar reslts were obtained at either dration. This indicates, therefore, that the single-breath method reqires appropriately prolonged exhalation. Sch drations as 30 and 45 s are applicable only in experimental animals, bt apparently cannot be applied directly to hman sbjects. The dration of exhalation sed in or stdy was extremely long compared with previos stdies in dogs [3, 4]. This may be attribted to the fact that a tracheal tbe inserted orally in a small dog may increase the deadspace. Dring exhalation longer than 20 s, recirclation occrs and this may affect the composition of expired air [10, 11]. Hence, the appropriate dration of exhalation shold also be considered in order to attenate the effect of recirclation. Since the alveolar gas eqation is sed for analysis in the single-breath method, it is highly possible that diseased lngs with ventilation/perfsion (V/Q) nevenness may ~iv.e erroneos reslts [12]. Compartment with a large V /Q may be attribted to the case of errors in the single-breath method [13]. Acknowkdgement: The athors wish to thank Dr Nagata for his helpfl sggestions in this stdy. References 1. Kim TS, Rahn H, Farhi LE. - Estimation of tre venos and arterial by gas analysis of a single breath. J Appl Physiol, , Chen H, Silverton NP, Hainsworth R. - Evalation of a method for estimating cardiac otpt from a single breath in hmans. J Appl Physiol: Respirat Environ &rercise Physiol, 1982, 53, Mohanned MMJ, Hainsworth R. -Evalation sing dogs of a method for estimating mixed venos and arterial Pco 1 from a single breath. J Appl Physiol: RespiraJ Environ Exercise Physiol, 1981, 50, Mohammed MMJ, Wood LM, Hainsworth R. - Evalation sing dogs of a method for estimating cardiac otpt from a single breath. J Appl Physiol: Respirat Environ Exercise Physiol, 1981, 50, Hlastala MP, Wranne B, Lenfant CJ. - Single-breath method of measring cardiac otpt - a re-evalation. J Appl Physiol, 1972, 33, Bderer MC, Rmmel JA, Sawin CF, Maldin DF. - Use of the single-breath method of estimating cardiac otpt dring exercise-s1ress testing. Aerospace Med, 1973, 44, Inman MD, Hghson RL, Jones NL. - Comparison of cardiac otpt dring exercise by single-breath and C0 1 -rebreathing methods. J Appl Physiol, 1985, 58, Gilbert R, Achincloss lli Jr. - Comparison of single breath method and indicator diltion measrement of cardiac otpt. J Appl Physiol, 197-0, 29, Grfllnlnd I. - Errors de to data redction in singlebreath method for measrement of plmonary blood flow. J Appl Physiol: Respirat Environ Exercise Physiol, 1982, 52, Ashton CH, McHardy GIR. - A rebreathing method for determining mixed venos dring exercise. J Appl Physiol, 1963, 18, Jones NL, Campbell EJM, McHardy GIR, Higgs BE, Clode M. - The estimation of carbon dioxide pressre of mixed venos blood dring exercise. Clin Sci, 1967, 32, Otis AB. - Qantitative relationship in steady-state gas exchange. ln: Handbook of Physiology. 3: Respiration. Vol. 1. W.O. Fenn and H. Rahn eds, American Physiological Society, Washington DC, 1964, pp Ok H, Ohi M, Kno K.- Mixed venos carbon dioxide tension. Koky, 1985, 4,

6 ESTIMATING MIXED VENOUS CARBON DIOXIDE TENSION 95 Essai d'estimation de la d sang veinex mill par la mithode "single breath". H. Ohta, 0. Takatani, T. Matsoka. La m.sthode "single breath" a ete proposee par Kim, et al, por estimer la tension sangine en C0 2 et le debit cardiaqe. Sa valer n'a pas encore ete provee. L'etde a ete entreprise por comparer, chez les chiens, la Pvco 2 calclee par la methode "single breath" avec la P\lco 2 mesree dans le sang veinex mele, et por evaler!'inflence des variations de la dree d'expiration celle volme de gaz expire, habitellement rejete des calcls comme espace mort. Parmi les dr.ses d'expiration de 15, 30 et 45 secondes, qi ont ete testees, celle de 15 secondes s'avere trop breve por obtenir ne corbe 0 2 analysable, mais, qe ce soit a 30 o a 45 secondcs, les valers calclees de Pvco 2 sont comparables ax valers mesrees. Une similitde significative entte les valers calctees et mesrees de Pvcol. a done ete obtene lorsqe le gaz expire ayant moins de 22 de Paco 2 a ete considere comme espace mort. Er Respir J., 1989, 2,

Utilization and Timing of Signal Change Interval

86 TRANSPORTATION RESEARCH RECORD 1114 Utilization and Timing of Signal Change Interval FENG-BOR LIN, DONALD COOKE, AND SANGARANATHAN VIJAYAKUMAR The problem or timing the signal change Interval has received