549 - Evaluating Accuracy of Electrolyte and Blood Gas Level Point of Care Testing among Inborn Infants in the Neonatal Intensive Care Unit

Publication Number: 549.4604

Isha V. Parikh, McGovern Medical School at the University of Texas Health Science Center at Houston, Sugar Land, TX, United States; Sepideh Saroukhani, McGovern Medical School at the University of Texas Health Science Center at Houston, Houston, TX, United States; Matthew Rysavy, McGovern Medical School at the University of Texas Health Science Center at Houston, Houston, TX, United States; Mar Romero-Lopez, University of Texas Health Science at Houston, Houston, TX, United States; Brian Chang, UT Health Houston, McGovern Medical School, Houston, TX, United States; Lindsay F. Holzapfel, McGovern Medical School, University of Texas Health Science Center at Houston, Houston, TX, United States

Background: Point-of-Care Testing (POCT) offers rapid laboratory results with minimal blood volume, making it ideal for critically ill infants with a limited blood volume and rapid result turnaround. While validated in adult and pediatric populations, POCT is not well-studied in infants.

Objective: We aim to evaluate the accuracy of the POCT GEM Premier 5000 blood gas analyzer in infants compared to a single center central laboratory (CL) testing to improve care efficiency in the NICU.

Design/Methods: In this retrospective, cohort study, all inborn infants in a large academic neonatal intensive care unit between September 2022 and May 2024 were included if they received paired POCT and CL testing within one hour in the first 14 postnatal days. Measurements for sodium, potassium, chloride, glucose, hemoglobin, and hematocrit were compared between methods. To assess agreement, the mean difference and 95% interval of difference along with limits of agreement between POCT and CL test pairs were measured and tested using paired-t-test. Bland-Altman analyses quantified and visualized agreement and bias between POCT and CL test values. Similar analyses were performed stratified by gestational age, birthweight, and illness severity as potential modifiers.

Results: In the 20-month period, 118 inborn patients were included with birthweight 2115 ± 1043g (mean ± SD) and gestational age 33.1 ± 5.3 weeks. Our sample was 61% male with 75% delivered through C-section. Table 1 shows that on average POCT sodium was 6.0 mEq/L lower than CL (n = 609, p < 0.001); POCT potassium was 0.1 mEq/L lower than CL (n = 657, p < 0.001); POCT chloride was 4.4 mEq/L lower than CL (n = 593, p < 0.001); POCT glucose was 3.5 mg/dL lower than CL (n = 593, p < 0.001); POCT hemoglobin was 0.04 g/dL higher than CL (n = 325, p = 0.08); and POCT hematocrit was 0.6 higher than CL (n = 428, p < 0.001). Figure 1 shows the differences between POCT and CL, which were similar across all lab value ranges except for potassium and glucose, which had increased variability at the extreme values. There was no apparent modification in agreement by gestational age, birthweight, or illness severity (p>0.05).

Conclusion(s): This study demonstrates systematic differences between POCT and CL measurements in NICU patients, with POCT consistently reading lower for electrolytes and glucose but showing excellent agreement for hemoglobin. Implementation of POCT may allow clinicians to limit daily laboratory blood volumes and provide more rapid results for expedited treatment.

Table 1. Comparison of mean POC and mean CL values


Figure 1. Bland Altman Plot of CL vs POC


Table 1. Comparison of mean POC and mean CL values


Figure 1. Bland Altman Plot of CL vs POC