1038 Capnometry and the paediatric laryngeal I.A. Spahr-Schopfer MD, B. Bissonnette MI~ mask airway E.J. Hartley MD

The (LMA), an alternative to tracheal than PaC02 (P < 0.05) In addition, use of the LMA does intubation in certain situations, has gained popularity in recent not prevent the hypercapnia associated with halothane anaes- years. Initially designed for use in adults it has now become thesia in children breathing spontaneously. available in suitable sizes for paediatric anaesthesia. The ob- jectives of this study were to identify the preferred site of sam- De plus en plus utilisd depuis quelques anndes, le masque la- piing the end-tidal (PETCOz) with the LMA ryngd permet dans certains cas d~viter l~ntubation. Destind and to determine the accuracy of this recording when compared d'abord aux adultes, il est maintenant disponible en plusieurs with arterial C02 (PaCOz). We studied 30 healthy children, grandeurs pour usage pddiatrique. Cette dtude vise d trouver age one to.['we years and weighing between 10 and 25 kg un- sur le masque laryngd le site d~chantillionnage iddal pour le dergoing minor requiring mask anaesthesia. In each gaz carbonique tdldexpiratoire (PETCOz) et dvaluer la prdcision case, after induction of anaesthesia, the LMA was inserted de cette mesure comparativement au C02 andriel (PaC02). under direct vision to eliminate the possibility of epiglottic air- Trente enfants agds d'un d cinq ans et pesant de 10 ?t 25 kg, way obstruction. The fresh gas flow was provided by a Jackson programmds pour une intervention chirurgicale mineure avec Rees modification of an Ayre's T-piece and was determined anesthdsie au masque, font pattie de l~tude. Dans chaeun des according to the following formula: 3 • (1000 + (100 • body cas, le masque laryng~ est ins~rd sous vision directe pour dli- weight)) LPM. pressure, ECG, 02 saturation, temper- miner toute possibilitd d'obstruction par l~piglotte. Les gaz frais ature and end-tidal gas concentrations were recorded. The meas- sont fournis par un tube en T d'Ayre modifid par Jackson ures of peak PETC02 were taken at pre-determined distances Rees avec un ddbit calculd selon la formule suivante: 3 X (1000 from the elbow connector down the LMA shaft. During the + (100 X poids)) L" min -1. La pression artdrielle, I'ECG, la sampling sequence an arterial blood sample was taken for gas saturation, la tempdrature et les concentrations tdldexpiratoires analysis. The PaC02 was 63.5 + 9.3 mmHg (mean -6 SD) sont enregistrdes. Les dchantillons de PETCO2 sont prdlevdes At any given sampling site, mean PETCO2 values were less than distances prdddtermindes du raccord le long de la tubulure PaCO2 (P < 0.05). However, in eight patients P~'C02 values fixde au masque laryngd. En m~me temps que la capnomdtrie, measured at the distal site were higher than the PaC02 (negative on prdlbve un echantillon de sang pour l'analyse des gaz ar- P(a-ET)C02 gradients) The results of this study suggest that tdriels. La PaCO2 est de 63,5 -6 9,3 mmHg (moyenne -6ET). when the LMA is used in ctu'ldren undergoing minor surgery A chacun des sites d'dchantillonnage gazeux, les valeurs who are spontaneously breathing halothane, the PErC02 values moyennes de PelC02 sont moindres que la PaC02 (P < 0,05). obtained from different sites underestimate the value of the Cependant, chez huit patients, au site le plus distal, les valeurs PaC02. The preferred site for measuring PerCO~ in these chil- de PETC02 ddpassent la PaC02 (gradient P(a-eT)COe ndgati~. dren is the distal end of the shaft although this value is less Les rdsultats de cette dtude sugg~rent que lt,~sage du masque laryngd, chez l'enfant soumis ~ une chirurgie mineure sous anesthdsie gt l~alothane en ventilation spontande, procure des Key words mesures de Pro'CO2 qui sous-estiment la PaCO~ Chez ces en- ANAESTHESIA"paediatric; fants, il est prdfdrable de prdlever les dchantiUons de PETCO2 CARBON DIOXIDE: end tidal; la partie distale de la tubulure fixe du masque bien que cene EQUIPMENT: laryngeal mask; valeur soit moins dlevde que la PaCO2 (P < 0,05) L'utilisation MEASUREMENTTECHNIQUES: capnometry. du masque laryngd ne prdvient pas non plus lfiypercapnie as- From the Department of Anaesthesia, The Hospital for Sick socide ~ l'anesth~ise d l~udothane en ventilation spontande. Children, 555 University Avenue, Toronto, Ontario, Canada. Presented in part at the 49th annual meeting of the CAS, June 1992, Toronto, Ontario, Canada. Capnometr~ and are required in patients Address correspondence to: Dr. E.J. Hartley, Department of whose are intubated during . Anaesthesia, The Hospital for Sick Children, 555 University An accurate measurement of end-tidal CO2 (PETCOz) is Avenue, Toronto, Ontario, Canada MSG IXH. difficult to determine in infants and children due to a

CAN J ANAESTH 1993 / 40: 11 / pp 1038-43 Spahr-Schopfer et al.: LMA AND CAPNOMETRY 1039 combination of high fresh gas flows and small tidal vol- umes. 2 Using a mass spectrometer with a continuous sampling technique, Kaplan and Paulus demonstrated that fresh gas flows continuously through the Mapleson F circuit (such as a Jaekson-Rees' modification of Ayre's T-piece) during inspiration and expiration. 3 They sug- gested that the fresh gas flow could contaminate the ex- haled gases, dilute them and lead to inaccurate estimates of end-tidal gas concentrations. Fletcher has suggested that the P(a-ET)CO2 gradient is lower at large tidal volumes. 4 He concluded that the presence of a large P(a-ET)CO2 gradient indicates a low efficiency of ven- tilation similar to that observed at small tidal volumes. FIGURE 1 A size #2 laryngeal mask airway was used. The elbow These factors produce dilution of the P~rCO2 and a con- connectorwas modifiedand the modificationconsisted of two ports sequent underestimation of the arterial carbon dioxide wh[eh allowed placementof a flexiblefibreoptie and an ETCO-z (PaCOg. Several studies have suggested sampling catheter. Samp[ingsites are as follows;E = elbow; 1,2,3 = alternatives to circumvent this problem. 5-7 Sampling the 4.5, 6, 11 r from the elbow; D -- distal (17 em fromthe elbow);T = PETCO2 at the distal end of the endotracheal tube was tracheal (3 cm into the ). shown to improve the accuracy of PErCO2 in estimating PaCO2 in small infants of <12 kg. s Recently, a labo- ratory study visualizing the mixing of fresh gas and ex- 0.02 mg-kg -t iv anaesthesia was induced with either pired gas in the Mapleson D circuit showed that, re- thiopentone 5 mg. kg -~ /v or halothane (in increments gardless of fresh gas flow rates, the site of mixing is of 0.5% up to 3%) by mask. Anaesthesia was maintained always proximal to the point of narrowing of the endo- with halothane and 70% N20 in 05. Spontaneous res- connector. 8 It confirms the result of two piration was allowed in all cases. previous investigations wb_ieh suggested that PETCO~ A size #2 laryngeal mask airway was inserted under sampling immediately below the connector with a needle direct vision to ensure its correct placement. The elbow inserted through the endotracheal tube wall was sufficient connector (Gybeck-Dryden Corporation, Indianapolis, to estimate the PaCO2. 9-~~ IN) was modified by the addition of two ports which Because the laryngeal mask airway has a large shaft, allowed placement of a flexible fibrcoptic it is possible that the dilutional effect caused by the fresh and an ETCO2 sampling catheter (Figure 1). The fresh gas flow may be more important than that which is re- gas flow was provided by means of a Mapleson F circuit ported with an endotraeheal tube. This study was there- (Jackson-Rees' modification of an Ayre's T-piece) and fore designed to determine the preferred site of sampling was determined according to the formula: 3 • (1000 + of end-tidal COs and the accuracy of this site as an es- (100 X body weight in kg)) LPM. II If rebreathing was timate of PaCO2 in anaesthetized infants and children observed on the capnometer the fresh gas flow was in- breathing spontaneously through a laryngeal mask air- creased until it could be eliminated. The end-tidal carbon way. dioxide partial pressures were recorded at steady-state by a Puritan-Bennett/Datex 255 infrared sidestream cap- Methods nometer (Datex Instrumentarium Corp., Helsinki, Fin- After approval by the Human Subject Review Committee land) calibrated before each study with dry gas of known and obtaining informed parental consent, 30 unpremed- composition (5% CO2 and 36% N20). All values of itated healthy infants and children, ASA physical status P~7CO2 were corrected for BTPS. 12 The measurements I and II, weighing between 6.5 and 25 kg were enrolled of peak end-tidal CO2 were taken in random order at in the study. The infants and children were fasted for predetermined distances from the elbow connector down three hours. All patients were scheduled for elective minor the LMA shaft using a #19 standard-wire gauge Deseret surgery requiring general mask anaesthesia and lasting Intracath (Beckton Dickinson Deseret Medical, Sandy, less than one hour. Patients with abnormal airway anat- UT) inserted through the lateral Luer lock port at the omy or known to have cardiopulmonary disease were elbow connector. If was cyclical the highest excluded. All children were supine and were monitored PExCO2 was recorded for analysis. Samples were taken with a blood pressure cuff, ECG, transcutaneous oxy- proximally at the elbow connector and distally from the gen saturation, and end-tidal gas monitor and an axillary elbow connector at levels of 4.5 cm, 6 cm, 11 cm, 17 temperature probe. Following administration of cm (the distal end of the laryngeal mask airway) and 1040 CANADIAN JOURNAL OF ANAESTHESIA

20 cm (3 cm into the trachea) (Figure I). During the 80 sampling sequence an arterial blood gas was drawn from each patient by radial stab. and tidal volumes were measured by means of a Wright Pediatric ~ 70 Respirometer (Ohmeda, Mississauga, Ontario). E E ~'60 Statistical analysis 8 All demographic and parametric data are expressed as a. mean + standard deviation (SD). The normally distrib- 50 uted data were analyzed with repeated-measures ANOVA and the Student-Neuman-Keuls tests for multiple com- 40 parisons. P < 0.05 was accepted. The degree of agree- Elbow i :Z :~ Dist Trach PaC02 ment between measurements at the proximal and distal sites and arterial PCO2 was calculated using the method FIGURE 2 The mean (+ SD) for end-tidal CO 2 measurements of Bland and Altman. ~3 We decided, a priori, that a clin- obtained at predetermined distances from the elbow connector down to the distal end of the laryngeal mask airway and within the trachea. ically acceptable range of deviation from the "true" CO2 *P < 0.05 represent each site compared to PaCO 2. value (PaCOz) would be + 5 mmHg, i.e., a PE'rCO2 measurement in the range of 35-45 mmHg would be acceptable for a PaCO2 value of 40 mmHg. Subsequently, 90 from prospectively collected PETCO2 measurements, the 95% confidence interval between paired CO2 measure- ments was calculated by the mean 5:2 SD. Assuming 80 a normal distribution, 95% of the differences between paired CO2 would fall within the range of mean + 2 "r 70 SD and mean - 2 SD. If this calculated 95% confidence E v interval falls within the previously set clinically acceptable e~ 60 range, the degree of agreement between the two methods O O is considered to be good. Coefficients of determination I--- 50 (r 2) for the PETCO2 recorded at the proximal and distal tu site were compared to PaCO2 using a linear regression analysis. Both coefficients of determination were com- 40 [] D pared using an unpaired t test. The two different analyses were chosen to show that the coefficient of determination 3D measures the strength of relationship between the two i ! i i i l ~"-'I 30 40 50 60 70 80 90 variables, not the agreement between them. PaC02 (mmHg)

Results FIGURE 3 PETCO2(mmHg) (B distal and 17 proximal) The age and weight of the children were 3.9 5:1.5 yr measurements are plotted against the PaCO 2 (mmHg). The coefficient and 16.6 "4- 3.6 kg respectively. The mean respiratory of determination (r 2) for distal and proximal were 0.5 and 0.2 rate was 45 + 8- min -t and the mean was respectively. 2.2 + 0.7 ml- kg -I. breath -I. At the distal and tracheal sampling sites, rebreathing of 4-5 mmHg CO2 was noted (P < 0.05) (Figure 3). The paired mean difference be- which could not be eliminated by increasing the fresh tween the airway capnometry CO2 and the PaCO2 for gas flow. The preset fresh gas flow was sufficient to avoid the proximal site was 9.34 + 9.9 mmHg with 95% con- rebreathing at any of the other sampling sites. fidence interval 2.94 to 15.74 mmHg (Figure 4). Figure The mean PaCO2 measured by a single arterial stab 5 shows the same graphical analysis for distal PETCO2. during the sampling sequence was 63.5 + 9.3 mmHg. The mean difference was 2.02 • 6.9 mmHg, 95% con- Measured PETCO2 rose progressively as sampling became fidence interval from -2.46 to 6.50 mmHg. There was more distal from the elbow (Figure 2). Mean PETCO2 no correlation between the weight of the children and values measured at each sampling site were less than the the gradient between arterial and end-tidal PCO2 values PaCO2 (P < 0.05). The coefficient of determinations (r 2) whether measured at the proximal or distal sites. for the proximal (E and the distal (D) sites (Figure 1) One child developed a brief episode of laryngospasm were 0.20 and 0.50 respectively and they were different during LMA insertion which was treated with positive Spahr-Schopfer et al.: LMA AND CAPNOMETRY 1041

A mean + 2 SD I o i I O

i | Clinical --=O= 10 range

o ~E O ! mean - 2 SD

-20 40 50 60 78 80

Average PaCO2 FIGURE 6 Capnogram. FIGURE 4 The differencesin paired proximalend-tidal CO2 measurements are plotted against PaCO 2. The mean, mean + 2D and piing site appear to influence the accuracy of end-tidal mean - 2SD values are shown in the graph. capnometry. ,4 It has been shown that in children weigh- ing more than 12 kg with tracheas intubated and ventilated by intermittent positive pressure ventilation 2~ using the same partial rebreathing circuit as in this study, mean + 2 SO both distal and proximal samples approximated PaCO2 values.5 In a study using single breath end-tidal CO2 as

I i a means of estimating arterial PCO2 in infants and chil- r162 i 41' 9 9 I clinical 9 I dren, it was suggested that proximal measurements of 0"0, I ql ,nean range ._r o~' ii I I m~E PETCO2 accurately estimate PaCO2 provided that the 8~.E C~ fresh gas flow is interrupted during the period of meas- -1( urement. 9 Recently, in a laboratory study, it was reported mean - 2 SD that using a 3.5 mm endotracheal tube and fresh gas flows as large as 20 L. min-l delivered by a Mapleson -2( 50 D circuit, dilution of the expired gases by the fresh gas Average PaC02 (mmHg) flow occurred only as far as the connector, s In another laboratory investigation using intubated rabbits and a FIGURE 5 The differencesin paired distal end-tidal CO2 Mapleson D circuit, it has been shown that PETCO2 measurements are plotted against PaCO 2. The mean, mean + 2D and mean - 2SD values are shown in the graph. values are accurate if measured distally to the connector regardless of the mode of ventilation, l0 There are several factors that may explain the differ- pressure ventilation. There were no other episodes of oxy- ences between the results of this and previous studies. gen desaturation due to placement or removal of the la- Firstly, the internal diameter of the shaft of the LMA ryngeal mask airway. There were no complications as (7 mm) is larger than an endotracheal tube used in a a result of this study. child of the same size. The LMA also maintains an ar- tificial "pharyngeal" cavity which is bypassed by tracheal Discussion intubation. ~5 Secondly, the patients in this study were The results of this study show the PETCO2 values meas- breathing spontaneously with a combination of high res- ured at different sites in the LMA are less than the PaCO2 piratory rates and low tidal volumes. This pattern of res- in children breathing halothane spontaneously (P < piration increases the dilutional effect of the fresh gas 0.05). The best approximation of the PaCO 2 is the flow due to the smaller expired volume of gases. Thirdly, PETCO2 value measured at the distal end of the shaft a rapid respiratory rate causes inaccuracy in capnometer of the LMA but the correlation remains poor (coefficient readings. 16 The respiratory rates recorded in this group of determination (r 2) = 0.5). of patients varied from 40 to 70 breaths per min. This The use of capnometry is recognized as a means of was in most cases fast enough to modify the recording estimating PaCO2 in adults. In infants and children, how- pattern and affect phase III (plateau phase) of expira- ever, PETCO2 measurement is more difficult and its value tion. ]7 At high respiratory rates no good plateau phase as a means of estimating PaCO2 levels remains contro- could be observed (Figure 6). This would lead to an un- versial. Both the type of anaesthetic circuit and the sam- derestimation of the real PETCO2. 1042 CANADIAN JOURNAL OF ANAESTHESIA

The eight negative PaCO2-PETCO2 are more difficult fluencing capnography in the Brain circuit. J Clin Monit to explain. In the case of cyclical respiration we took 1985; 1: 6-10. the highest PETCO2 for analysis. We may have overes- 3 Kaplan RF, Paulus DA. Error in sampling of exhaled timated the PETCO2 as it has been shown that the PaCO2 gases (Letter). Anesth Analg 1983; 62: 955-6. is close to the time-weighted average of alveolar CO2, 40lsson SG, Fletcher R, Jonson B, Nordstrom L, Prakash which is lower than the PETCO2 in cyclical respiration. ~s O Clinical studies of gas exchange during ventilatory sup- The other explanations for negative a-ETCO2 gradients port - a method using the Siemens-Elema CO2 analyser. (low respiratory rate, big tidal volume, sloping phase III) Br J Anaesth 1980; 52: 491-9. do not apply in children with rapid shallow respirations. ]9 5 Badgwell JM, McLeod ME, Lerman J, Creighton RE. The hypercapnia observed in our patients after ap- End-tidal PCO2 measurements sampled at the distal and proximately 30 min of spontaneous ventilation with an proximal ends of the endotraeheal tube in infants and chil- inspired halothane concentration of 2% may be explained dren. Anesth Analg 1987;66: 959-64. as follows. is produced by a concentra- 6 Badgwell JM, Heavner JE, May WS, Goldthorn JF,, Ler- tion of halothane in excess of twice the minimal alveolar man J End-tidal PCO2 in infants and children concentration)~ In a recent study investigating minute ventilated with either a partial rebreathing or a non- ventilation during mask halothane anaesthesia in infants rebreathing circuit. 1987; 66: 405-10. and children it was shown that halothane reduces minute 7 Hillier SC, Badgwell JM, McLeod ME, Creighton RE, volume, tidal volume and increases breathing frequency. 2J Lerman J Accuracy of end-tidal PCO2 measurements In addition, the magnitude of the produced using a sidestream capnometer in infants and children ven- by the larger internal diameter of the laryngeal mask dated with the Sechrist infant . Can J Anaesth airway compared with the small tidal volume of spon- 1990; 37: 318-21. taneously breathing children resulted in rebreathing at 8 Halpern L, Bissonnette B. Visualizing the mixing of fresh both the distal and tracheal sampling sites. gas and expired gas in the Mapleson D circuit: a laboratory Several studies have identified sources of inaccuracy study. Anesthesiology 1991; 75: A421. in measurements of PETCO2 in children. These have been 9 Bissonnette B, Lerman J Single breath end-tidal CO2 esti- avoided in the present study. Thus, the sampling aspi- mates of arterial PCO2 in infants and children. Can J ration rate of the end-tidal gas monitor was 150 Anaesth 1989; 36: !10-2. ml- min-l which has been shown to ensure an accurate 10 Rich GF, Sullivan MP, Adams JM. Is distal sampling of estimate of the alveolar CO: content without entrainment end-tidal CO2 necessary in small subjects? Anesthesiology of fresh gas.22 Severinghaus pointed out that the avail- 1990; 73: 265-8. ability of water-permeable catheters has led to an incor- 11 Rose DK, Byrick R J, Froese AB. Carbon dioxide elimina- rect calibration practice. ]2 Our PETCO2 values have been tion during spontaneous ventilation with a modified Maple- corrected for BTPS. son D system: studies in a long model. Can Anaesth Soc J In conclusion, the results of this study show that chil- 1978; 25: 353-65. dren under halothane anaesthesia and breathing spon- 12 Severinghaus JI,E. Water vapor calibration errors in some taneously through a laryngeal mask airway become hy- capnometers: respiratory conventions misunderstood by percapnic. The anaesthetist may not be aware of the level manufacturers? Anesthesiology 1989; 70: 996-8. of hypercapnic as the PETCO2 measurements do not 13 Bland JM, Ahman DG. Statistical methods for assessing correlate well with PaCO2. The most accurate sample agreement between two methods of clinical measurement. site for PETCO2 sampling is within the trachea. However, Lancet 1986; 307-10. using this site defeats the purpose for using a laryngeal 14 Kirpalani H, Kechagias S, Lerman J. Technical and clini- mask airway which is to provide a satisfactory airway cal aspects of capnometry in neonates. J Med Eng Technol without trespassing beyond the glottis. 1991; 15: 154-61. 15 Nunn JE Applied Respiratory Physiology. 2nd Ed. Lon- Acknowledgement don: Butterworths & Co Ltd., 1977. The authors would like to thank Dr. J.E.S. Relton for 16 Sasse FJ. Can we trust end-tidal carbon dioxide measure- his advice in the preparation of this manuscript. ments in infants? J Clin Monit 1985; 1: 147-8. 17 Fletcher R, Jonson B. Prediction of the physiologicaldead References space/tidal volume ratio during anaesthesia/IPPV from 1 Guidelines to the practice of anaesthesia as recommended simple pre-operative tests. Acta Anaesthesiol Stand 1981; by the Canadian Anaesthetists' Society. Suppl Can J 25: 58-62. Anaesth 1993; 40: 10. 18 /5"J~,erJ. Blood-gas equilibrium of carbon dioxide in lungs: 2 Gravenstein N, Lampotang S, Beneken JEW Factors in- a conlinu2ng controversy. J Appl Physiol 1986; 60: 1-8. Spahr-Schopfer et aL: LMA AND CAPNOMETRY 1043

19 Fletcher R, Jonson B. Deadspaee and the single breath test for carbon dioxide during anaesthesia and . Effects of tidal volume and frequency of respi- ration. Br J Anaesth 1984; 56: 109-19. 20 Wren WS, Allen R Synnott A, O'Keeffe D, O'Griofa R Effects of halothane, isoflurane and enflurane on ventila- tion in children. Br J Anaesth 1987; 59: 399-409. 21 Brown KA, Bissonnette B, Holtby H, Ein S, Shandling B. Minute ventilation during mask halothane anaesthesia in infants and children. Can J Anaesth 1993; 40; 2:112-8. 22 Gravenstein N. Capnometry in infants should not be done at lower sampling rates (Letter). J Clin Monit 1989; 5: 63-4.