A long established measure of the degree and duration of fetal hypoxia has been blood pH, drawn either from the scalp or the cord, a practice carried into newborn care with pH analysis of heel puncture blood. Because of the stability of this value, factors such as the length of sampling time, the degree of exposure to air, and mixing of blood sources (arteriolar, capillary and venous) have little affect on measurement accuracy. However, of relevance to the topic at hand, blood gases (PaO₂ and PaCO₂), blood CO-oximetry (SaO₂), and pulse oximetry (SpO₂ and pulse rate) values are highly volatile in neonates and affected by many variables. In addition, blood analyzers are technically sophisticated devices, requiring a great degree of skilled maintenance to be kept in optimal working order. Conversely, a pulse oximeter requires little sophistication to operate, and the typical sensor is quite intuitive to place. Pulse oximetry is largely trouble-free, and requires minor user interface (other than the occasional changing of the sensor site). Perhaps the simple nature of pulse oximeter operation is cause for the lack of end-user understanding and results in badly interpreted data.^1-4

The clinical application of blood gas analysis (BGA) and pulse oximetry (PO) is regulated in the United States by the Joint Commission on Accreditation of Healthcare Organizations (JCAHO). However, JCAHO requires compliance with the guidelines of other groups, primarily the Clinical and Laboratory Standards Institute (CLSI) and its predecessor NCCLS, prior to accrediting an institution. Relevant to this discussion, these guidelines focus on measurement via analytical instruments, such as analyzers for measurement of blood CO-oximetry, gas and pH.^5-8 These guideline documents primarily describe principles for collecting, handling, transporting, and analyzing arterial blood specimens, as well as instrument quality control matters, with the aims of reducing collection hazards, ensuring integrity of the arterial specimen, and achieving reliable data.

The standards emphasize that PO SpO₂ values are not a direct measure of the percent of arterial hemoglobin saturated with oxygen.^9-11 The current pulse oximeters in neonatal use cannot differentiate nor quantify dyshemoglobins. It is important to note that the calculated SaO₂ value reported via a blood gas analyzer is not a reference for SpO₂. Rather, only the functional SaO₂ value from CO-oximeter analysis of an arterial blood specimen can provide a reference for SpO₂. The CO-oximeter also provides a measure of the content of fetal versus adult hemoglobin. Table 1 reviews the user interventions and rationale associated with optimizing accuracy in comparing SaO₂ and SpO₂ values in neonates.

Download Optimizing comparison of a CO-oximetry SaO₂ value with pulse oximetry (PO) SpO₂ values in neonates.

The accuracy of PO SpO₂ values is derived by testing healthy adult human volunteers under controlled laboratory conditions (CLC). In CLC, every effort is taken to isolate and limit variables, and to seek steady-state conditions during which reference data can be synchronized. In the NICU, when the infant is the least stable, concurrence of BGA and PO data is most desired and expected. Table 2 details why the NICU environment is less conducive to accurate data acquisition than the CLC used by PO manufacturers. It is safe to assume that in contrast to the NICU, the collection of this data in the delivery room setting would be far more difficult and prone to patient and caregiver variables.

Download Comparing SaO₂ to SpO₂ data in an NICU versus that from controlled laboratory conditions.
Note: Author Comments in Parentheses

New guidelines for neonatal resuscitation have been promulgated.¹² Care of the sick newborn requires the best instrumentation complemented by thorough clinical vigilance. Pulse oximetry data derived immediately post-delivery have prompted comparative interventional studies.¹³ Lack of PO was a contributing factor of poor outcome during infant sedation.¹⁴ Some have suggested close-looping SpO₂ values to maintain normoxemia in mechanically ventilated infants.^15,16 With this degree of reliance on basic and advanced newborn care, PO accuracy is essential. Unfortunately, in clinical use, the accuracy of SpO₂ values in neonates was found appreciably worse than manufacturers’ claims. Caregivers should recognize multiple sources of compounding error in the clinical setting that can corrupt correlation of SaO₂ to SpO₂ values versus the claimed accuracy data of PO manufacturers. There is dissimilarity in populations (the ill neonate versus healthy adult) and an SpO₂ plateau at the time of blood sampling may not occur in the NICU, whereas SpO₂ plateaus are an expectation in the laboratory setting. Without controlling such influences, comparison between blood analysis and pulse oximetry values is fraught with error and the potential for misinterpretation. Owing to an avoidance of invasive risk, SpO₂ data are often relied upon in conditions where direct arterial access for corroboration of SpO₂ values does not exist, so the data go unchecked.

At the moment, it appears that while pulse oximetry is included in the updated resuscitation guidelines, clinical use data reveal oximeters currently used in neonatal care can lack the precision claims of their manufacturers. Given that, caution may be warranted in using pulse oximetry for a discrete value of SpO₂. Rather, a trend of stability or declining and increasing SpO₂ values may prove more beneficial to clinical care. Suggestions and rationale for optimizing blood gas and pulse oximetry data were included in this review to provide a common denominator for future comparative studies and maximize accuracy in the assessment of oxygenation in the acute care of newborns.

References:

The guidelines are based on a review of the latest scientific evidence, and affirm assessment of oxygenation status and management of supplemental oxygen therapy via pulse oximetry. Lacking, however, is specificity for resuscitation with airway concentrations of 100% oxygen or less, as well as adequate caution regarding the subjectivity of pulse oximetry data, and the need for performance optimization in the resuscitation setting (where challenges to instrument accuracy can be extreme).

The authors captured a snapshot of current NRP practices across the Unites States. Among their results, pertinent to the subject of this review, is that 52% of program directors used pulse oximetry, with 23% indicated that they had useful readings within 1 minute of placing a sensor. The authors remark that while the time to data acquisition may vary, pulse oximetry is useful for monitoring subsequent care of infants and is essential if clinicians wish to use a blender and to provide <100% oxygen. They further comment that their experience in evaluating neonatal resuscitation suggests that infants spend far more time in the resuscitation area than is anticipated, and the use of blenders and oximeters in such circumstances can reduce unnecessary exposure to excessive supplemental oxygen. While they suggest that in the delivery room the oximeter should be set to its lowest averaging time and highest sensitivity, the rationale and consequences of such settings are not discussed. Moreover, the shortcomings of pulse oximetry are not mentioned nor is reference made to means of reducing error.

The authors conclude that substantial variations exist in neonatal resuscitation practices, some of which are not addressed in existing guidelines. They recommend that future guidelines include the use of blenders, oximeters, continuous positive airway pressure/PEEP, and plastic wrap during resuscitation. Further, the authors express their hope that the widespread dissemination of these survey results will encourage re-evaluations of practice efficacy, so that future resuscitation practices will be evidence based.

This wide-ranging study by Gerstmann et al sought to analyze available data recorded during the process of providing routine neonatal care. They performed a historical evaluation, a prospective evaluation, a procedural evaluation, a verification evaluation, and an evaluation of corroborating data from NICUs at several institutions.

Of particular note is their determination that a calculated SaO₂ via a blood gas analyzer was not appropriate for comparing SpO₂ accuracy; rather, they measured functional SaO₂ with the corresponding pre-blood draw baseline SpO₂. Multiple brands of CO-oximeters (ABL, Beyer, and Instrumentation Laboratory) and pulse oximeters (Datex-Ohmeda, Masimo, Nellcor and Spacelabs) were used. In order to contrast the neonatal performance of the models of pulse oximeter for which data were available, operational performance was defined as the frequency with which, at any given SaO₂ value, the corresponding SpO₂ reading was within a manufacturer’s specified accuracy for neonates (typically, ±3 digits (1 SD) between 70% and 100% saturation).

Nearly 32 thousand data sets of SaO₂ to SpO₂ values collected during routine care were compared. While the range of SaO₂ was from 57% to 100%, most data were between 85% to 100%. As a point of measurement consistency between the two brands of CO-oximeters, 52 arterial blood samples were drawn from 10 neonates and run simultaneously on both CO-oximeters. The average difference in SaO₂ values was 0.07, which proved to be not statistically significant. However, none of the pulse oximeters provided a consistent bias of SpO₂ values across the range of SaO₂ of 70% to 100%. While SpO₂ accuracy was device dependant between an SaO₂ of 92% to 97%, performance declined for all manufacturers above and below this range. Indeed, a rapid drop in operational performance was seen in all pulse oximeters as SaO₂ decreased. As an example, at an SaO₂ of 80%, SpO₂ will typically read approximately 90%.

This study is the largest collection of neonatal data comparing SaO₂ and SpO₂ values. Notably, the period of data collection was nearly concurrent with the survey of neonatal directors by Leone et al (as above). Gerstmann et al suggest that the poor SpO₂ accuracy seen in their data seems at a level inconsistent with pulse oximetry’s perceived importance, role and use in the NICU. Elaborating on the results, concerns of both over- and under- treatment were raised. They conclude, pending improvement in SpO₂ accuracy, that adjustments to supplemental oxygen and ventilator settings in the NICU patient “must be based on and re-evaluated by arterial blood analysis.” Further, the authors demonstrated that some variables were not evident, e.g. the difference in CO-oximeter brand.

It must be noted that while the multi-center, multi-caregiver design of this study doubtless introduced inconsistencies, those inconsistencies accurately replicate the environment typical to the daily care of neonates in centers across the Nation.

LAST MONTH�S Q & A    March 2006 - Volume 3 - Issue 7

Last issue we reviewed the clinical application of blood gas analysis and pulse oximetry, discussed the new guidelines enforcing the use of these technologies, and compared PO manufacturers' accuracy claims vs the reality of caring for the sick newborn.