Introduction: It is common clinical practice to use pulse oximetry (SpO2) as a continuous, non-invasive measurement of systemic oxygen content. There has been recent literature discussion of how this method may not be accurate regarding certain races, and ethnicities. There is also recent literature describing how it may not be accurate in the pediatric congenital heart population. We aimed to determine if the discrepancies were present in the pediatric congenital heart population based on race as well as cyanotic or non-cyanotic lesions using an expected oxyhemoglobin dissociation curve as a surrogate. Methods: In this study we retrospectively analyzed 29 patients over 39 encounters reviewing 54 arterial blood gases and the SpO2 that correlated with time of lab draw. We then compared the values and plotted them along an oxyhemoglobin dissociation curve. The patient’s lesion was categorized as cyanotic or non-cyanotic as well as repaired or not repaired. A maximum of two samples per patient per encounter were used. The patient’s weight, race, ethnicity, and age were also collected for further data analysis as able. The primary correlation was assessed based on the expected oxyhemoglobin dissociation curve. When applicable, pre-ductal saturations were used. Artificial intelligence was used to assist with preliminary statistical analysis. Results: In our analysis it was found using the Fishers exact test that cyanotic lesions are about 9 times more likely than non-cyanotic lesions to have an incongruent SpO2 and partial pressure of oxygen (PaO2) (OR 9.13, p< 0.05). This held across all racial and ethnic groups. It was also consistent across age groups, specifically infants < 30 days compared to all others. When analyzing several older patients that had their lesions corrected there seemed to be no statistical difference in their congruence. Conclusions: Cyanotic, uncorrected pediatric congenital heart lesions show a significantly increased risk of having incongruent SpO2 and PaO2 measurements suggesting that care must be taken in these patients to not rely solely on pulse oximetry when assessing overall oxygenation. No significant racial or ethnic variation in this trend was noted. An increased number of patients and lesions is required to further delineate correlations.
Hayes et al. (Sun,) studied this question.