Continuous Non-invasive BP monitoring in Neonates
by Anoop Rao
Coauthors: Anoop Rao*(1), Fatima Eskandar-Afshari (2), Ya'el Weiner (3), Elle Billman (3), Eric Helfenbein (4), Thomas Roxlo (5), Junjun Liu (5), Alan Walendowski (5), Arthur Muir (5), Weyland Leong (5), Siddharth Siddharth (5), Alexandria Joseph (6), Archana Verma⁶, Chandra Ramamoorthy⁶, Anita Honkanen⁶, Noa Sella¹, William Rhine¹, Xina Quan (5) and Zhenan Bao (3) 1) Lucile Packard Children’s Hospital, Stanford CA, 2) Division of Neonatology, LAC+USC Medical Center, Keck School of Medicine of the University of Southern California, Los Angeles, CA, United States. 3) Stanford University, Stanford, CA; 4) Philips Healthcare, Sunnyvale, CA; 5) PyrAmes Inc., Cupertino, CA 6) Stanford University School of Medicine, Department of Anesthesia Funding: FDA, Stanford Maternal and Child Health Research Institute and NIH R43 Grants
Medical Devices & Digital Health
Continuous monitoring of arterial blood pressure (BP) is vital for assessing and treating cardiovascular instability in a sick infant. Currently, invasive catheters are inserted into an artery to monitor these infants. Catheterization requires skill, is time consuming, prone to complications and is often painful. We report on the feasibility and accuracy of a non-invasive, wearable device (<15g) with an embedded thin capacitance sensor (~50μm), that continuously monitors BP from tiny neonates. The ultimate goal of this effort is to obviate the need for invasive arterial line placement for BP monitoring.
Subjects: Sixty (60) infants, ranging from preterm (~26 weeks gestation) to term (~40 weeks gestation) were included in the study. These infants were admitted to the ICU for monitoring for conditions such as respiratory distress to congenital cardiac malformations such as Tetralogy of Fallot, Coarctation of aorta, Hypoplastic left heart etc. These sensors were placed on the wrist and/or foot of preterm and term infants to acquire pulse waveform measurements. Using highly efficient artificial neural network (ANN) algorithms and a NN model built atop a historical arterial waveform library that has > 9000 hours of training data. We then inferred the systolic, diastolic and mean arterial pressure from the measured pulse waveform data. Next, the inferred non-invasive BP data were compared with corresponding invasive-umbilical arterial line data to determine accuracy.
Results: Across gestational age and pathologies, the systolic, diastolic and mean arterial pressures was as accurate as the BP from the invasive arterial line, per FDA-criteria. The device has obtained FDA breakthrough device designation. Additional data is being collected at Stanford for submitting to the FDA for clearance.