A Combined Ear Sensor for Pulse Oximetry and Carbon Dioxide Tension Monitoring: Accuracy in Critically Ill Children

Abstract

Pulse oximetry is a noninvasive technique used to monitor arterial oxygen saturation (Sao2) (1). Monitoring of ventilation is as important as the assessment of oxygenation. A major disadvantage of the single use of pulse oximetry is the lack of information about ventilation (2,3). Determining the arterial carbon dioxide tension (Paco2) status still requires periodic blood sampling, a discontinuous and invasive technique. Capnography of end-tidal gas is used as an estimate of Paco2. Whereas capnography may result in an unreliable estimation of Paco2(4,5), monitoring of cutaneous carbon dioxide tension (Pcco2) offers a reliable evaluation of newborn, infant, and adult Paco2 status (6–8). A combined sensor providing noninvasive and continuous estimation of Sao2, Paco2, and pulse rate at the earlobe has been developed. The purpose of this study was to evaluate the accuracy of this new sensor in critically ill children and to compare the values with simultaneously measured Spo2 (conventional pulse oximeter), Sao2 (hemoximeter), and Paco2 (blood-gas analyzer) values. Methods Written, informed consent was obtained from the parents, and the investigative protocol was approved by the local ethics committee. Twenty consecutive critically ill children were studied during intensive care treatment. Each patient was studied once. During an 8-h study period, combined Spo2/Pcco2 ear sensor data and conventional pulse oximeter data were recorded. After 2 and 8 h, arterial blood samples were drawn and analyzed independently. The combined Spo2/Pcco2 ear sensor (Kontron Instruments, Zürich, Switzerland) integrates the basic elements of a pulse oximetry saturation sensor (6) and of a Severinghaus-type carbon dioxide tension sensor (9,10). The sensor is warmed to a surface temperature of 42°C. The sensor is placed at the earlobe with a dedicated ear clip. Spo2 values are available immediately, and Pcco2 values are available after a typical onset time of 3 min (11). Automatic recalibration is performed every time the sensor is placed on the sensor docking station between measurements. Conventional pulse oximetry was performed with a Nellcor N-100 pulse oximeter (Nellcor Inc., Pleasanton, CA). The sensor was applied to the right index finger (12). Sao2 was measured with a hemoglobin analyzer (OSM 2 hemoximeter; Radiometer, Copenhagen, Denmark). Paco2 was measured with a blood-gas analyzer (AVL 995; AVL, Schaffhausen, Switzerland). Statistical analysis consisted of 1) linear regression with standard error of the estimate (sdee), 2) Bland-Altman analysis (13), and 3) χ2 analysis with Yates’ correction by using a contingency table. Results Forty data sets were obtained from 20 critically ill children. Each data set compared combined ear sensor values, conventional pulse oximeter values, and independent arterial values (hemoximeter and blood-gas analyzer). The median age of the patients was 10.5 yr (range, 3–15 yr), and median weight was 29.0 kg (range, 14–66 kg). The measurement range was 80.5%–98.0% for Sao2 and 30.0–46.5 mm Hg for Paco2. For oxygen saturation, linear regression analysis revealed the equations Spo2 (combined ear sensor) = 0.92 × Sao2 + 8.58% (sdee, ±1.94%) and Spo2 (conventional pulse oximeter) = 0.82 × Sao2 + 18.32% (sdee, ±2.87%). Equation: Spo2 =a × Sao2 +b (a and b: slope and intercept of the regression line, sdee). With Bland-Altmann analysis, comparing the Spo2 (combined ear sensor) and the arterial Sao2 values, the mean difference was +1.54%, with limits of agreement of +5.36% to −2.28% (±2 sd). Comparing the Spo2 (conventional pulse oximeter) and the arterial Sao2 values, the mean difference was +1.64%, with limits of agreement of +7.46% to −4.18% (±2 sd) (Fig. 1).Figure 1: The relationship is shown between the mean oxygen saturation (Mean So2) and the difference between the combined ear sensor Spo2 (○ and arterially measured Sao2. The relationship is shown between the mean So2 and the difference between conventional pulse oximeter Spo2 (×) and arterially measured Sao2. Mean differences and limits of agreement (±2 sd) are shown.The absolute Spo2 (combined ear sensor) value deviated ≤2% from the arterial Sao2 value in 32 (80%) of the 40 data sets. The absolute Spo2 (conventional pulse oximeter) value deviated ≤2% from the arterial Sao2 value in 26 (65%) of the 40 data sets. For carbon dioxide tension, linear regression analysis revealed the equation Pcco2 (combined ear sensor) = 0.96 × Paco2 + 1.88 mm Hg (sdee, ±2.85 mm Hg). Equation: Pcco2 =a × Paco2 +b (a and b: slope and intercept of the regression line, sdee). With Bland-Altman analysis, comparing the Pcco2 (combined ear sensor) and the arterial Paco2 values, the mean difference was +0.23 mm Hg, with limits of agreement of +5.78 to −5.33 mm Hg (±2 sd) (Fig. 2). The absolute Pcco2 (combined ear sensor) value deviated ≤4.5 mm Hg from the arterial Paco2 value in 33 (83%) of the 40 data sets.Figure 2: Relationship is shown between the mean carbon dioxide tension (Mean Pco2) and the difference between the combined ear sensor cutaneous carbon dioxide tension (Pcco2) (•) and arterially measured Paco2. Mean difference and limits of agreement (±2 sd) are shown.Discussion The purpose of developing a combined oxygen saturation and carbon dioxide tension ear sensor was to improve the efficiency of respiratory monitoring. We assessed the accuracy of a new combined ear sensor. Correlation between Spo2 of the combined ear sensor and arterial Sao2 values was better than between Spo2 of the conventional pulse oximeter and arterial Sao2 values. The mean differences in the arterial Sao2 measurements were similar for the combined ear sensor and the conventional pulse oximeter. However, the combined ear sensor demonstrated a better precision than the conventional pulse oximeter when compared with arterial Sao2 values. The increased precision for the combined ear sensor is most probably due to the warmed measurement site. Warming improves perfusion of the tissue (14), and this results in a stronger signal. The conventional pulse oximeter freezes the display up to 60 seconds during low perfusion or when the signal is lost, to reduce false alarms (12). Especially in critically ill patients, this approach results in less measurement precision and increases the risk of missing true hypoxemia. Skin marking due to warming to 42°C either did not exist or was only transient. This finding is similar to those in other studies (7). The correlation and mean difference between the Pcco2 of the combined ear sensor and arterial Paco2 were clinically acceptable. We conclude that the two methods of estimating the patient’s carbon dioxide tension status may be used interchangeably. A key advantage of the combined ear sensor is the possibility of monitoring the efficiency of mechanical ventilation on-line. This is done by analyzing the difference between the Pcco2 of the combined ear sensor and the end-tidal Pco2(15). Noninvasive, continuous measurement of Pcco2 will detect hyperventilation situations fast and reliably. Monitoring oxygen saturation and carbon dioxide tension simultaneously allows the differential diagnosis of hypoxemia versus apnea. Overall, we conclude that this new combined Spo2/Pcco2 measurement technology has the potential to replace conventional pulse oximetry. The authors thank Dr. Serge Kocher, SenTec Inc., for help with manuscript preparation.