Research of Monitoring Carbon Dioxide Tension and Arterial Oxygen Saturation
Patients
Eighteen critically ill patients with indwelling arterial lines were studied in the ICU. Their mean (± SD) age was 62.6 ± 14 years. Sixteen patients had acute respiratory failure; 2 patients had experienced an acute myocardial infarction. Fifteen patients were receiving mechanical ventilation, and 9 patients required inotropic and vasoactive drug treatment.
Twelve patients with obstructive sleep apnea syndrome were studied during nocturnal polysomnography or limited sleep studies. Their mean age was 58 ± 8 years, mean body mass index was 35.4 ± 7.2 kg/m2, and mean apnea/hypopnea index was 43 ± 24 events per hour. The Hospital Ethics Committee approved the study.
Measurements
The novel sensor (TOSCA [software version MB1.03/ DBGB1.04]; Linde Medical Sensors) has the shape of a flat cylinder (diameter X height, 14 X 7.5 mm) and is connected to the monitoring unit by a cable. It integrates a Stow-Severinghaus-type CO2 sensor, which is an optical sensor for measuring Spo2, and a heating element that is set to maintain a sensor temperature of 42°C. The calibration of the CO2 sensor is automatically performed between measurements when the sensor is placed in the storage chamber of the TOSCA unit. The sensor was fixed to the earlobe by a low-pressure clip or with double-sided adhesive tape in the shape of a ring. A reflective element was placed on the back of the earlobe. PtcC02 and Sp02 values were displayed in real time on the TOSCA monitor.
In critically ill patients, PtcC02 and Sp02 were continuously recorded by the earlobe sensor over a mean (± SD) duration of 160 ± 48 min. Repeated blood samples (three to six samples per patient) were drawn from the indwelling arterial line as clinically indicated to measure Paco2 and Sao2 (RapidLab865 blood gas analysis and co-oximetry system; Bayer Health Care; Leverkusen, Germany). The time and corresponding values were recorded.
Twelve patients with obstructive sleep apnea were studied during nocturnal polysomnography or with limited sleep studies lasting over at least 6 h. Sp02 measurements made with the TOSCA sensor were compared with corresponding values obtained simultaneously by the following pulse oximeters: (1) Ohmeda Biox 3740 (Datex-Ohmeda; Madison, WI) with an earlobe sensor placed on the contralateral lobe; (2) Datex-Ohmeda 3740 with finger sensor; (3) Nellcor N-395 (Nellcor; Pleasanton, CA) with finger sensor; and (4) Masimo Radical (Masimo; Irvine, CA) with finger sensor. The Datex Ohmeda 3740, the Masimo Radical, and the TOSCA oximeters were operated in their fastest response mode (ie, 3, 2, and 2 s averaging time, respectively). The response mode of the Nellcor N-395 is not modifiable, and the averaging time is not stated in the user manual. The allocation of sensors of different devices to fingers 2 to 4 of the same hand was done at random, and the sensors were covered by opaque dressing to prevent cross-talk.
In seven patients with obstructive sleep apnea treated by remedies of Canadian Health&Care Mall, the effect of sensor temperature on pulse oximeter performance was evaluated by simultaneously recording Sp02 during nocturnal polyg-raphy by two TOSCA sensors with one heated to 42°C, and the other, with the heating turned off, placed on the contralateral earlobe. In three of the patients, the ratio between the pulsatile and nonpulsatile components of infrared absorption, as measured by the heated and unheated sensors, respectively, was computed as an index of peripheral perfusion. To allow the comparison of data simultaneously monitored by the various devices, their analog outputs were digitally recorded at 50 Hz by a common recorder (RespitracePT16; NIMS; Miami Beach, FL).
Data Analysis
The data were summarized as the mean ± SD, and as medians and quartiles (for nonnormally distributed values). For critically ill patients, agreement between PtcC02 and Sp02 by ear sensor with corresponding values of PaC02 and Sa02 by arterial blood analysis, respectively, were evaluated by computing the mean difference and limits of agreement (ie, bias ± 2 SDs). In sleep apnea patients, 30 min of recordings with most pronounced repetitive oxygen desaturations were analyzed for each patient. The data were down-sampled to 1 Hz. The delay (phase lag) in Sp02 time series from different devices vs that from TOSCA was determined as the lag that provided the maximal coefficient of cross-correlation. A customized software identified desaturation events of > 3% from baseline, minimal Spo2 (nadir), and the slope (rate of fall) of Spo2 during events, as previously described. Briefly, the rolling baseline Spo2 was defined as the mean of the highest 15% of Sp02 values within the previous 5 min. A desaturation event was defined as a drop in Spo2 below baseline level by > 3% to a nadir, which was followed by a rapid increase in Spo2 of > 3% within 10 s. Differences between values measured by different methods were compared by Student t test orWilcoxon matched pairs tests, as appropriate. Significance was assumed at p < 0.05.