2006 OPEN FORUM Abstracts
NEW PULSE OXIMETRY SENSORS WITH LOW SATURATION ACCURACY CLAIMS - A CLINICAL EVALUATION
Cox PN, MD;
Hospital for Sick Children, Toronto, Canada
Introduction - Despite the recognized
inaccuracies of pulse oximetry at oxyhemoglobin saturations below 70%, pulse
oximetry has been recognized as a standard monitoring tool for patients in the
operating room and the intensive care unit. Despite advances in pulse oximetry
technology in other areas such as motion resistance and low perfusion
sensitivity, monitoring low oxyhemoglobin continues to be a problem. Recently,
both Masimo and Nellcor have introduced sensors with accuracy claims for
saturation levels below 70%. Patients with congenital cyanotic cardiac lesions
(CCCL) have low oxyhemoglobin saturation levels. Careful maintenance of these
low saturations, within very narrow limits, is required to ensure adequate
cardiac output and peripheral perfusion prior to surgical correction of certain
CCCL. For this reason, these patients present specific problems for pulse
oximeters. We set out to test the accuracy of a traditional pulse oximeter
sensor and these new sensors in CCCL patients in the Critical Care unit. Methods
- Following IRB approval, patients with CCCL were studied while in the ICU.
Monitoring for the postoperative care of these patients was routine, and
included our standard pulse oximeter sensor, (Masimo LNOP). In addition to the
standard sensor, a Masimo LNOP Blue sensor attached to a Masimo SET Radical
pulse oximeter and a Nellcor Max-I attached to a Nellcor N600 pulse oximeter
equipment with LoSat was placed on the thumb of the left hand, or the great toe
of either foot, as recommended by the manufacturer. Data from all 3 pulse
oximetry sensors was recorded on a laptop computer. Arterial blood gases (ABG),
including CO-oximetry (SaO2), the gold standard, were obtained as
clinically indicated. The time the ABG was obtained was noted in the computer
record. SpO2, from the three oximeters, and SaO2 were
compared using linear regression analysis, the Bland Altman technique of
calculating bias and precision and the ARMS, an accuracy statistic
used by the FDA. Additionally, paired t-testing was used to compare the ARMS (accuracy) from each of the three
sensors. Results - A total of 8 patients (4 males) were studied. The
mean (+ SD) age and weight were 33 (+ 34) days and 5.8 (+
2.8) kgs, respectively. A total of 41 ABGs (mean + SD = 5 + 2.7
per patient) were obtained. The mean (+ SD) and range of the SaO2 was 72.5% (+ 7.6%) and 85% - 56.1%. The bias, precision and the
regression analysis are presented in Table. Discussion - Accurate pulse
oximetry monitoring provides a valuable clinical tool. Despite advances in
technology, only the new Masimo Blue sensor demonstrates acceptable accuracy as
demonstrated by a smaller bias, precision, and ARMS.
| SaO2 (CO-oximetry) | Masimo SET Radical with Blue Sensor | Nellcor N600 with Lo-Sat and Max-I sensor | LNOP sensor | |
| Mean (+SD) % | 72.5 (7.6) | 70.6 (8.1) | 75.9 (7.0) | 74.3 (7.0) |
| Range % | 85 - 56.1 | 87 - 52 | 89 - 61 | 91 - 57 |
| Bias | - | -1.91 | 3.38 | 1.82 |
| Precision | - | 3.32 | 4.60 | 6.31 |
| ARMS | 3.83 * | 5.71 | 6.57 | |
| R2 value | - | .832 | .638 | .399 |
| Regression equation | - | = 0.854 + 0.962(x) | = 32.178 + 0.603(x) | = 32.264 + 0.580(x) |
Table
- The bias, precision, ARMS and regression analysis for the new sensors
with low saturation accuracy claims, and the LNOP sensor in 8 children with
congenital cyanotic cardiac lesions. Paired t-test of the ARMS shows
a significant difference between the Masimo LNOP Blue and the other sensors, p
< 0.001.