Unlike other intensive care units, the NICU cares for some otherwise healthy premature infants who need support to grow and develop. These patients have a high risk of acquiring a severe infection during their months-long hospital admission. Sepsis in premature infants can have life-long or life-limiting consequences and damages the developing brain. In the NICU, we have the opportunity to see normal, healthy vital sign patterns transition gradually or suddenly into abnormal patterns as an illness develops. Some abnormal patterns are impossible to see on the standard bedside monitor.
At UVA, researchers developed, validated, and commercialized a predictive monitoring algorithm and display, called the HeRO Score, to warn clinicians of abnormal heart rate characteristics that indicate an increased risk of imminent sepsis. Since then, the UVA research team discovered patterns in pulse oximetry data that add to heart rate characteristics to warn clinicians of an imminent cardiorespiratory deterioration. Led by Neonatologists Drs. Brynne Sullivan and Karen Fairchild, the research team is working in a multi-center collaboration to validate and display novel cardiorespiratory algorithms to detect early signs of sepsis using non-invasive data from the bedside monitor.
Recently, Dr. Brynne Sullivan led a team evaluating clinical and vital sign changes surrounding the time of blood cultures diagnosing sepsis in very low birth weight infants at three collaborating NICUs. In this study, the authors found that an increase in apnea events was the most common reason for ordering a blood culture, and half of the blood cultures prompted by this clinical sign were positive. Furthermore, the cross-correlation of heart rate and oxygen saturation (SpO2), a mathematical way to quantify apnea with bradycardia and desaturation, increased near the time of blood culture. They developed and analyzed multivariable models and showed that adding calculated vital sign features, including cross-correlation of heart rate and SpO2, improves sepsis risk prediction compared to models using demographic risk factors or clinical signs. These results align with a previous study where Dr. Sullivan and colleagues showed that including physiologic data in predictive models for risk stratification during the first week after birth in preterm infants adds information to demographics on the risk of later morbidity and mortality.
As we work to improve care for patients with sepsis, we have also implemented a neonatal Sequential Organ Failure Assessment (nSOFA) score that predicts mortality during sepsis treatment as a measure of illness severity. Dr. Sullivan recently collaborated with several other NICUs across the country analyzing HeRO and nSOFA that showed the combined utility of these scores for predictive monitoring and mortality risk stratification. Knowing when infants are getting sick before showing any symptoms and, once they are sick, what their risk is can undoubtedly help clinicians treat infants and counsel families appropriately.
The UVA NICU is at the forefront of using predictive monitoring. Earlier detection of infections is key to preventing life-changing or life-limiting complications. With this kind of advanced monitoring technology, we hope to see improved outcomes for our smallest patients.
For more information, please see:
Filed Under: Features
Tags: jrs3yc