Laryngeal mask airway (LMA) is useful for management of difficult airways [1-5], can be used to assist fiberoptic intubation [1,6], and may facilitate airway management in patients with cervical spine injuries . However, when a relatively new device is introduced into clinical practice, care should be taken to ensure that previously held assumptions continue to hold true for the newer alternative. End-tidal carbon dioxide (PETCO2) may be used to monitor ventilation in adults and approximates PaCO2 when measured in patients ventilated via endotracheal tube (ETT) . During spontaneous ventilation in adults, the mean difference between PaCO2 and PETCO2 measured via the LMA is similar to that measured via the ETT . Unfortunately, PETCO2 measured via the LMA does not accurately reflect PaCO2 in spontaneously breathing children . No data are available regarding the relationship between PaCO2 and PETCO2 measured via the LMA in the pediatric population during controlled ventilation. We have, therefore, compared the differences between PaCO2 and PETCO2 measured via the LMA to that measured via the ETT during controlled ventilation in a pediatric population receiving general anesthesia.
With institutional review board approval and parental consent, we studied 22 ASA grade I and II patients, between the ages of 4 mo and 16 yr, undergoing surgical procedures requiring both general anesthesia and arterial cannulation. Exclusion criteria were history of cardiac or pulmonary abnormalities, abnormal airway anatomy, or any condition which increased risk of regurgitation of gastric contents.
Subjects were studied supine after induction of general anesthesia. After establishing adequate depth of anesthesia, a muscle relaxant was administered (atracurium 0.5 mg/kg or vecuronium 0.1 mg/kg). An appropriately sized LMA was inserted according to the manufacturer's instructions and proper position was subsequently confirmed with direct laryngoscopy. Mechanical ventilation was then initiated with tidal volume of 8-10 mL/kg and fresh gas flow of 4-6 L/min using a nonrebreathing circle system without any bacterial filter. Tidal volume was measured with a Model 5400 volume monitor (Ohio Medical Products, Madison, WI) on the distal end of the expiratory limb of the circuit.
Peak airway pressure was maintained under 20 cm H2 O and the absence of a leak around the LMA was confirmed by auscultation over the neck. Arterial cannulation was then performed. The hemodynamic status was monitored and maintained within 10% of the base values. During placement of the arterial catheter, PETCO2 was allowed to stabilize (minimum 5 min), after which an arterial blood sample was obtained and the PETCO2 noted. The LMA was then removed, the trachea intubated, and correct endotracheal position of the tube confirmed by capnography and auscultation of the chest (to exclude endobronchial intubation). The prior ventilatory variables were resumed and after a minimum of 5 min of ventilation via the ETT, a second PaCO2 was obtained and PETCO2 noted. The average time interval between the two pairs was 16 min. During this period no invasive procedures, including urinary catheterization and central line placements, were performed and stable hemodynamic variables and anesthetic depth were maintained.
The PETCO2 was monitored with a Nellcor N-2500 capnograph (Nellcor, Inc., Hayward, CA), calibrated prior to the beginning of each case. Sampling of expiratory gas was performed at the elbow connector of the anesthesia circuit.
The data were analyzed using the Bland and Altman method  for assessing agreement between two methods of clinical measurement. Bias, defined as the mean difference between values, and precision, defined as the SD of the bias, were determined for PaCO2 and PETCO2 pairings for both LMA and ETT. The root mean squared error (RMSE), proposed as a measure of overall agreement when comparing two devices, was calculated as the square root of the average squared bias .
Twenty-two patients (9 females, 13 males), weighing from 6.9 to 79.9 kg (mean, 33 kg) and between 4 mo and 16 yr of age (mean = 7 yr, 4 mo) were studied. Surgical procedures included 12 craniectomies or craniotomies, 7 spinal fusions, and 1 each cervical decompression, tethered cord repair, and pelvic osteotomy.
(Table 1) summarizes the relationship between PaCO2 and PETCO (2) samples obtained during ventilation via either the ETT or the LMA. The mean PaCO2 and PETCO2 difference was less than 4.5 mm Hg. As would be expected, PaCO2 was greater than PETCO2 in most cases. On only one occasion did PETCO (2) exceed PaCO2, suggesting either measurement error or improper timing of sample acquisition.
Performance characteristics of LMA and ETT relative to PaCO2 and PETCO2 were evaluated. Bias, precision, and RMSE are shown in Table 1. All data were obtained during steady-state conditions. Measurement error was calculated as the difference between PaCO2 and PETCO2 with both LMA and ETT. Both LMA and ETT had comparable performance as measured by bias and precision (4.2 +/- 3.66 for LMA and 4.0 +/- 3.42 for ETT). The RMSE was 0.85 for ETT and 0.89 for LMA. The 95% confidence range of the differences of PaCO2 and PETCO2 was -3.14, 11.5 (LMA), and -2.84, 10.84 (ETT). All values are in millimeters of mercury unless otherwise noted.
The difference plots of measurement error (LMA vs ETT) are shown in Figure 1. The confidence limits based on total variability are shown and are plotted relative to the overall bias exhibited by each airway device. The PaCO2 and PETCO2 had significant differences in each group but were comparable as measured by bias, precision, and RMSE and had no significant intergroup differences. For both the ETT and the LMA there was a tendency for the higher PaCO2 to have greater underestimation by the end-tidal value. Similarly, the relationship between the weight of the patients and its effects on PaCO2 and PETCO2 for both the groups was studied Figure 2 A and B which suggests that there was no correlation between the weight of the patient and difference between PaCO2 and PETCO2 for the weight ranges studied.
These data suggest that, in mechanically ventilated, anesthetized pediatric patients weighing more than 10 kg PETCO2 measured via the LMA or the ETT correlates well with PaCO2.
The use of a within-patients design and inclusion of only healthy children excluded multiple confounding variables, at the cost of limiting a generalized application of the conclusions. Furthermore, we did not randomize the order of observations, systematically making the LMA measurements first, in order to avoid tracheally intubating half the patients twice. With the limited sample size it is not possible to compare different values in infants, toddlers, young children, and adolescents and it restricts our observations to the pediatric population in general.
We collected all end-tidal carbon dioxide samples at the elbow connector of the pediatric nonrebreathing circuit (circle system) with unidirectional valves separating the inspiratory and expiratory phase. Our findings are different from those of Badgwell et al.  who reported limited value of sample collection at the elbow connector in patients weighing less than 12 kg; however, this was observed in patients ventilated via partial rebreathing circuits (Jackson-Rees modification). Badgwell et al.  also showed that measurements sampled at the proximal end of the ETT accurately predict PaCO2 in infants weighing less than 8 kg when ventilated with a nonrebreathing system.
In contrast to our results, Spahr-Schopfer et al.  reported a poor correlation between PETCO2 and PaCO2 in children spontaneously breathing via the LMA during general anesthesia. We believe the difference between our results and theirs can be attributed to our use of controlled ventilation and a nonrebreathing circle system, which minimized variation in alveolar ventilation and rebreathing of expired gases during observation intervals.
Use of controlled ventilation via the LMA has been reported in both pediatric and adult populations [5,14-16]. This practice may place patients at increased risk for aspiration of gastric contents, as gases may be insufflated into the stomach if inspiratory pressures are not limited [5,17]. We maintained peak airway pressures below 20 cm H2 O to reduce this risk.
We expected the increased dead space within the LMA to increase the error of PETCO2 measurements with smaller patients, for whom dead space to tidal volume ratios are higher. With this limited sample size and variability in weight, it is not possible to establish any significant relationship of weight with PETCO2 and PaCO2 in either of the groups, but the trend in data in our study suggests that there is no correlation of weight with PETCO2 and PaCO2 with either the LMA or the ETT. However, this issue needs to be investigated further in infants weighing less than 10 kg. The slightly greater PaCO2-PETCO 2 difference observed during ventilation via the LMA supports the assumption that the dead space to tidal volume ratio was increased by the LMA, but this difference was neither statistically significant nor clinically relevant for the ventilatory variables used in our study. This may not necessarily be true if other ventilatory modalities are used.
It has been suggested that smaller children may become hypercapnic during spontaneous ventilation via the LMA . We agree that this may be a valid concern. Even with mechanical ventilation, two of our patients had unexplainable high PaCO (2) (58 and 76 mm Hg). Prevention of hypoventilation and evolution of unrecognized hypercapnia requires the same vigilance and monitoring during either mechanical or spontaneous ventilation via the LMA or the ETT in infants and children.
In summary, in pediatric patients weighing more than 10 kg who are mechanically ventilated via the LMA, PETCO2 is as accurate an indicator of PaCO2 as that measured via the ETT, and capnometry may be used to evaluate the adequacy of ventilation.
The authors wish to thank Richard Rivers, MD, Denham S. Ward, MD, PhD, and Ronald A. Gabel, MD, for assistance with data analysis and review of the manuscript.
1. Pennant J, White P. The laryngeal mask airway. Anesthesiology 1993;79:144-63.
2. Smith I, Joshi G. The laryngeal mask airway for outpatient anesthesia. J Clin Anesth 1993;5(Suppl 1):22S-8S.
3. Efrat R, Kadari A, Katz S. The laryngeal mask airway in pediatric anesthesia: experience with 120 patients undergoing elective groin surgery. J Pediatr Surg 1994;29:206-8.
4. McEwan A, Mason D. The laryngeal mask airway. J Clin Anesth 1992;4:252-7.
5. Wilson I. The laryngeal mask airway in pediatric practice [editorial]. Br J Anaesth 1993;70:124-5.
6. Brimacombe J. Laryngoscopy through the LMA--a useful skill to acquire [letter]. Anaesth Intensive Care 1992;20:535.
7. Waldvogel V, Schneck HJ, Felber A, Hundelshausen BV. Anasthesierelevante besonderheiten der laparoskopie--wertigkeit der kapnometrie. Anaesthesiol Reanim 1994;19:4-10.
8. Hicks I, Soni N, Shephard J. Comparison of end-tidal and arterial carbon dioxide measurements during anaesthesia with the laryngeal mask airway. Br J Anaesth 1993;71:734-5.
9. Spahr-Schopfer IA, Bissonnette B, Hartley EJ. Capnometry and the pediatric laryngeal mask airway. Can J Anaesth 1993;40:1038-43.
10. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307-10.
11. Sheiner LB, Beal SL. Some suggestions for measuring predictive performance. J Pharmacokinet Biopharm 1981;9:503-12.
12. Badgwell JM, McLeod ME, Lerman J, Creighton RE. End-tidal PCO2
measurements sampled at the distal and proximal ends of the endotracheal tube in infants and children. Anesth Analg 1987;66:959-64.
13. Badgwell JM, Heavner JE, May WS, Goldthorn JM, et al. End-tidal PCO2
monitoring in infants and children ventilated with either a partial rebreathing or a non-rebreathing circuit. Anesthesiology 1987;66:405-10.
14. Devitt JH, Wenstone R, Noel AG, O'Donnell MP. The laryngeal mask airway and positive pressure ventilation. Anesthesiology 1994;80:550-5.
15. Graziotti PJ. Intermittent positive pressure ventilation through laryngeal mask airway: is a nasogastric tube useful. Anaesthesia 1992;47:1088-9.
16. Brain AIJ. The intavent laryngeal mask: instruction manual. Berkshire, UK: Brain Medical, 1992.
© 1996 International Anesthesia Research Society
17. Winans CS. Manometric asymmetry of lower oesophageal high-pressure zone. Am J Dig Dis 1977;22:348-54.