Evaluation of a transcutaneous carbon dioxide monitor in severe obesity

Intensive Care Med DOI 10.1007/s00134-008-1078-8 PHYSIOLOGICAL AND TECHNICAL NOTES Mauro Maniscalco Anna Zedda Stanislao Faraone Pierluigi Carratù Matteo Sofia Evaluation of a transcutaneous carbon dioxide monitor in severe obesity Received: 2 July 2007 Accepted: 27 February 2008 Springer-Verlag 2008 Present address: M. Maniscalco, Largo delle Mimose 1, 80131 Napoli, Italy M. Maniscalco ( ) A. Zedda S. Faraone Hospital S. Maria della Pietà, Section of Respiratory Medicine, Via San Rocco n 9, CAP 80026 Casoria, Naples, Italy e-mail: mauromaniscalco@hotmail.com Tel.: +39-081-7411457 Fax: +39-081-7411457 M. Maniscalco P. Carratù M. Sofia University Federico II of Naples, Department of Respiratory Medicine, AO Monaldi, Via Leonardo Bianchi n 1, CAP 80131 Naples, Italy Abstract Objective: To determine the reliability of estimating arterial CO 2 pressure (PaCO 2 )usingarecently introduced transcutaneous CO 2 pressure (PtcCO 2 )monitorin severe obese patients. Design: Observational and interventional study. Setting: District hospital with respiratory ward and bariatric surgery unit. Patients and methods: PtcCO 2 was measured in 35 obese patients with varied pathology, including chronic obstructive pulmonary disease, obstructive sleep apnea syndrome and hypoventilation syndrome. Ten minutes after the probe had been attached to an earlobe, PtcCO 2 was recorded immediately before arterial blood sampling. The PtcCO 2 and PaCO 2 values obtained with two methods were compared by Bland Altman analysis. In a subgroup of 18 obese patients with chronic obstructive pulmonary disease and/or obstructive sleep apnea syndrome with moderate to severe hypercapnia both PtcCO 2 and PaCO 2 were re-evaluated during continuous positive airways pressure (CPAP) or bi-level positive airway pressure (Bi-PAP) treatment. Results: The mean difference between PaCO 2 and PtcCO 2 was 1.4 mmhg, and the standard deviation of the difference was 1.3 mmhg. Bland Altman analysis showed generally good agreement between the two methods with a 95% limit of agreement of 4 to 1.1. The agreement between methods did not significantly change before and during cpap or Bi-PAP treatment in hypercapnic patients. Conclusions: The accuracy of estimation of PaCO 2 by transcutaneous monitoring was generally good in comparison with standard arterial blood gases examination. The device appears to be promising for use in obese patients to evaluate abnormalities in their alveolar ventilation. Keywords Transcutaneous carbon dioxide Non-invasive monitoring Obese Arterial blood analysis COPD OSAS Introduction Alveolar ventilation is assessed by the measurement of the partial pressure of carbon dioxide (PaCO 2 )inarterial blood. The gold standard technique for the measurement of PaCO 2 is represented by direct analysis of arterial blood gases, but it is invasive, intermittent and may be unpleasant. The measurement of transcutaneous carbon dioxide (PtcCO 2 ) represents an alternative method for non-invasive estimation of arterial PCO 2 [1]. Recently, PtcCO 2 monitors have been used in infants and in adult patients showing a good agreement between PtcCO 2 and PaCO 2 [2 4]; however, only few data are available in severely obese patients where excessive adipose tissue beneath the skin may possibly affect the accuracy of PtcCO 2, especially when forearm has been used as the site of electrode application [5]. Furthermore, arterial puncture procedure may be unpleasant or difficult in severe obesity, especially when repeated examinations are required in patients with episodes of acute disease

exacerbation. Moreover, comparative data acquired during moderate to severe hypercapnia are lacking. We therefore determined the reliability of estimating PtcCO 2 by comparing a transcutaneous carbon dioxide monitor and PaCO 2 in unselected adult severely obese patients. Methods Thirty-five obese patients were enrolled in the study. Their mean age was 52.4 years (age range 31 74 years) and mean BMI was 43.7 kg/m 2 (BMI range 39 52 kg/m 2 ). All patients were in-patients in a respiratory ward that required arterial blood gas analysis for clinical reasons. This study was approved by the Local Research Ethics Committee and informed patient consent was obtained from patients. The PtcCO 2 was measured in the supine position by a combined SaO 2 /PtcCO 2 monitor (TOSCA, Linde Medical Systems, Basel, Switzerland) [6]. The PtcCO 2 was measured (along with SpO 2 ) via a sensor attached by a low-pressure clip to an earlobe. The sensor probe heats the earlobe to 42 C to enhance blood flow. After automated calibration, the TOSCA sensor was attached to an earlobe to monitor PtcCO 2 ; after 10 min, PtcCO 2 was recorded immediately before arterial blood sampling for blood gas analysis with patient in the same posture following standard recommendations [7]. In particular subcutaneous administration of local anaesthetic (xylocaine) was used to minimize the pain during the procedure. PaO 2 and PaCO 2 were measured using a gas analyzer (Instrumentation Laboratory, GEM Premier 3000, Milan, Italy) and expressed in millimetres of mercury (mmhg) according to the manufacturer s recommendations. In a subgroup of 18 patients affected by chronic obstructive pulmonary disease and/or obstructive sleep apnea syndrome with moderate to severe hypercapnia, both the measurements of arterial blood gas analysis and PtcCO 2 were repeated after the use of non-invasive continuous positive airways pressure (CPAP) or bi-level positive airway pressure (Bi-PAP). Furthermore, the mean number of puncture for one arterial blood gas and the time consumption for arterial blood sampling were recorded in ten patients. Statistical analysis Data are presented as mean ± standard deviation. We assessed the level of agreement between PaCO 2 and PtcCO 2 and reproducibility by Bland Altman analysis [8]. A linear regression equation for PaCO 2 and PtcCO 2 was performed. Analysis of variance (ANOVA) test for repeated measurements was used for comparing the difference of PaCO 2 from the blood samples and the corresponding PtcCO 2 for each patient treated by CPAP or Bi-PAP. Results The patients had a wide variety of diagnoses, including chronic obstructive pulmonary disease, obstructive sleep apnea syndrome, hypoventilation syndrome and respiratory failure (Table 1). Ten patients had an arterial ph < 7.35, all with chronic obstructive pulmonary disease and/or obstructive syndrome apnea syndrome. Bland Altman analysis showed generally good agreement between the methods with a mean difference of Fig. 1 Bland Altman plot of difference between paired measurements of arterial PCO 2 (PaCO 2 ) and transcutaneous CO 2 (PtcCO 2 ) and their average in 35 severe obese patients Table 1 Baseline characteristics of patients. COPD, chronic obstructive pulmonary disease; OSAS, obstructive sleep apnea syndrome; RF, respiratory failure; OHS, obesity hypoventilation syndrome; BMI, body mass index; PaO 2, arterial O 2 PaCO 2,arterialCO 2 pressure pressure; Diagnosis Number Gender Age (years) BMI (kg/m 2 ) ph PaO 2 (mmhg) PaCO 2 (mmhg) COPD 7 7 men 46.4 ± 5.2 44.1 ± 4.5 7.39 ± 0.03 63.4 ± 8.7 49.6 ± 6.5 OSAS 12 5 men 53.8 ± 8.7 42.5 ± 2.8 7.41 ± 0.02 84.0 ± 4.8 40.7 ± 1.8 RF 13 9 men 56.1 ± 8.2 44.8 ± 2.1 7.34 ± 0.02 48.3 ± 5.7 61.2 ± 8.8 OHS 3 2 men 44.7 ± 4.5 42.7 ± 2.3 7.41 ± 0.02 76.3 ± 2.5 46.5 ± 1.5

(p < 0.001; Fig. 3). The mean number of punctures for allowing one arterial blood gas sample was 2.6 (range 1 5) and the time consumption for arterial blood sampling was 4.1 min (range 90 s to 11 min). Fig. 2 Linear regression analysis between arterial CO 2 pressure (PaCO 2 ) and transcutaneous CO 2 (PtcCO 2 ) in 35 severe obese patients Fig. 3 Linear regression analysis of the change of PaCO 2 observed between the serial blood samples and the corresponding change in PtcCO 2 in 18 hypercapnic obese patients during the treatment with CPAP or Bi-PAP 1.4 mmhg, a standard deviation of the difference of 1.3 mmhg and a 95% limit of agreement of 4 to 1.1 (Fig. 1). Eleven of 55 measurements showed a PCO 2 difference of > 2 mmhg. Linear regression between PaCO 2 and PtcCO 2 showed an r 2 of 0.98 (Fig. 2). No technical problems were experienced. In the 18 hypercapnic patients treated by CPAP or Bi-PAP, Bland Altman analysis showed generally good agreement between the two methods across the range of PaCO 2 values recorded. The change over time between the two measurements did not differ between the two methods (p = 0.47). Linear regression analysis of the change of PaCO 2 observed between the serial blood samples and the corresponding change in PtcCO 2 showed an r 2 =0.93 Discussion We evaluated the accuracy of a transcutaneous carbon dioxide sensor for non-invasive estimation of arterial carbon dioxide in severely obese patients. Our results show generally good agreement between PtcCO 2 measured by transcutaneous electrode and PaCO 2 measured from sampling arterial blood. Measurements of PtcCO 2 are less informative than arterial blood gases, as isolated elevated CO 2 levels give an incomplete picture, for instance in patients with chronic obstructive pulmonary disease. Measurements of PtcCO 2 should be used in conjunction with arterial blood gas measurement, particularly for following trends after an initial arterial sample has been taken, because it provides a continuous measurement and reduces the need for frequent invasive sampling of arterial blood. Similar to a previous study by Parker and Gibson [2] we observed a discrepancy of > 2 mmhg between PtcCO 2 and PaCO 2 in 11 among 55 paired measures. As hypothesised in that study, a possible explanation might be due to the heated sensor. In fact, capnometer uses heated sensors to improve local perfusion [9]. The sensor temperature to improve local perfusion (capillary arterialisation) also increases local production of carbon dioxide in the tissue, which leads to falsely high measurements [5]. For this reason, Severinghaus s temperature-correction factor, in combination with a metabolic constant, is established in the TOSCA as an automatic mode. Since this formula was estimated under the conditions of a sensor working at approximately 44 C and being placed at the forearm. the effectiveness should be re-estimated under the conditions of 42 C and at the earlobe with a clamp, which produces substantial pressure on the skin [10]. We should also consider that a systematic slightly higher value was evident with PtcCO 2 in comparison with PaCO 2. This finding in our adult severely obese patients is in agreement with the study by Bernet-Buettiker et al. [4] who compared arterial PCO 2, capillary CO 2 and PtcCO 2 in 30 newborns. Both PtcCO 2 and capillary CO 2 values revealed a negligible bias. The data of the present study in obese could support the concept that PtcCO 2 may represent PaCO 2 of the capillary bed with a slightly higher value than CO 2 in arterial blood. Likewise, as recently reported [11], the clinical usefulness of PtcCO 2 measurement is significantly dependent on the confidence limits of agreement (e.g. 1 kpa) between different methods suggesting that PtcCO 2 measurement cannot replace conventional blood gas analysis. In contrast to Parker and Gibson s study [2], in which the studied subjects were clinically stable, in our study

we evaluated also patients with arterial ph less than 7.35 and a wider range of hypercapnia. We did not observe any differences between data acquired in clinically stable patients and in patients with acute respiratory failure. None of our patients demonstrated severe hemodynamic abnormalities requiring catecholamine infusion. It is noteworthy that Rodriguez et al. [12] recently evaluated PtcCO 2 measurement in critically ill adult patients in an intensive care unit and found that only major cutaneous vasoconstriction significantly influenced the accuracy of PtcCO 2 measures, whereas catecholamine, respiratory support or hypothermia did not exert substantial interference. Obese patients would benefit most from using transcutaneous carbon dioxide measurement for several reasons, including the high prevalence of alveolar ventilation abnormalities and the consequent indication to a formal assessment of PaCO 2 arterial levels; however, excessive adipose tissue beneath the skin may possibly affect the arterial blood sampling itself. In our study we used lidocaine infiltration to minimize the pain from the puncture. It has been reported that arterial puncture with prior infiltration of local anaesthetic is the least painful procedure among those studied, and the use of local anaesthesia is indicated whenever conventional arterial puncture is required [13]. However, as recently reviewed, except among anaesthesia providers, the use of a local anaesthetic before arterial puncture is not universal, contrary to the standard of practice [14]; thus we could believe that for an awake severe obese patient the experience of single or repeated attempts to arterial blood sampling in real life might be more unpleasant in comparison with non-obese people, making the PtcCO 2 measurement a reasonable choice. We chose to make one baseline comparison of PaCO 2 and PtcCO 2 because a large number of patients in the study required only a single arterial blood sample. Noticeably in a small group of obese patients treated with CPAP or Bi- PAP and then re-examined with both methods, the same agreement between PaCO 2 measured by gas analyzer and the PtcCO 2 was found. Our data are at variance with another interventional study [15], which has shown reduced correlation between arterial and PtcCO 2 measurement when patients were evaluated from a steady-state to a dynamic phase consisting of high-frequency ventilation through a bronchoscope. In the present study non-invasive nasal ventilation using CPAP or Bi-PAP was administrated to obese patients as we routinely perform to reach a new steady state in blood CO 2 levels. This may explain the greater correlation we found both before and after the intervention. Moreover, the results of study in which repeated measures were taken at different time points are in agreement with the study by Janssens et al. [16]: They found that a change in PtcCO 2 was detected within 60 s, and the trend accurately reflected the change in PaCO 2 when a ventilatory event was induced by initiating or interrupting non-invasive positive pressure ventilation in hypercapnic patients. However, further studies examining the accuracy of the TOSCA sensor when used continuously over an extended period are required. We conclude that the accuracy of the transcutaneous carbon dioxide sensor used in obese patients with variable levels of hypercapnia in routine respiratory practice is satisfactory in comparison with standard PaCO 2 measurement from arterial blood sampling with the device, which is promising for its use in this setting. References 1. Franklin ML (1995) Transcutaneous measurement of partial pressure of oxygen and carbon dioxide. Respir Care Clin N Am 1:119 131 2. Parker SM, Gibson GJ (2007) Evaluation of a transcutaneous carbon dioxide monitor ( TOSCA ) in adult patients in routine respiratory practice. Respir Med 101(2):261 264 3. Bendjelid K, Schutz N, Stotz M, Gerard I, Suter PM, Romand JA (2005) Transcutaneous PCO2 monitoring in critically ill adults: clinical evaluation of a new sensor. Crit Care Med 33:2203 2206 4. Bernet-Buettiker V, Ugarte MJ, Frey B, Hug MI, Baenziger O, Weiss M (2005) Evaluation of a new combined transcutaneous measurement of PCO2/pulse oximetry oxygen saturation ear sensor in newborn patients. Pediatrics 115:e64 e68 5. Griffin J, Terry BE, Burton RK, Ray TL, Keller BP, Landrum AL, Johnson JO, Tobias JD (2003) Comparison of end-tidal and transcutaneous measures of carbon dioxide during general anaesthesia in severely obese adults. Br J Anaesth 91:498 501 6. Dullenkopf A, Bernardo SD, Berger F, Fasnacht M, Gerber AC, Weiss M (2003) Evaluation of a new combined SpO2/PtcCO2 sensor in anaesthetized paediatric patients. Paediatr Anaesth 13:777 784 7. Williams AJ (1998) ABC of oxygen: assessing and interpreting arterial blood gases and acid-base balance. Br Med J 317:1213 1216 8. Bland JM, Altman DG (1986) Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1:307 310 9. Eberhard P, Gisiger PA, Gardaz JP, Spahn DR (2002) Combining transcutaneous blood gas measurement and pulse oximetry. Anesth Analg 94:S76 S80 10. Severinghaus JW, Peabody JL, Thunstrom AM, Stafford M (1978) Transcutaneous oxygen measurements: calibration analysis and electrode stabilization methods. Acta Anaesthesiol Scand Suppl 70:180 182 11. Bolliger D, Steiner LA, Kasper J, Aziz OA, Filipovic M, Seeberger MD (2007) The accuracy of non-invasive carbon dioxide monitoring: a clinical evaluation of two transcutaneous systems. Anaesthesia 62:394 399 12. Rodriguez P, Lellouche F, Aboab J, Buisson BC, Brochard L (2006) Transcutaneous arterial carbon dioxide pressure monitoring in critically ill adult patients. Intensive Care Med 32:309 312

13. Giner J, Casan P, Belda J, Gonzalez M, Miralda RM, Sanchis J (1996) Pain during arterial puncture. Chest 110:1443 1445 14. Hudson TL, Dukes SF, Reilly K (2006) Use of local anesthesia for arterial punctures. Am J Crit Care 15:595 599 15. Eberhard P (2004) Comparison of transcutaneous and end-tidal CO2- monitoring for rigid bronchoscopy during high-frequency jet ventilation. Acta Anaesthesiol Scand 48:260 16. Janssens JP, Howarth-Frey C, Chevrolet JC, Abajo B, Rochat T (1998) Transcutaneous PCO2 to monitor noninvasive mechanical ventilation in adults: assessment of a new transcutaneous PCO2 device. Chest 113:768 773