///2019 Abstract Details
2019 Abstract Details2019-07-13T07:45:15-05:00

Photoplethysmography for Prediction of Preeclampsia

Abstract Number: F3C-455
Abstract Type: Original Research

Tammy Y Euliano MD1 ; Kostas Michalopoulos PhD2; Neil Euliano PhD3


Preeclampsia (PE) is a major cause of maternal and neonatal death, particularly in low-resource settings. When treated properly, including management of blood pressure, seizure prophylaxis and often preterm delivery of the fetus, the maternal mortality rate is low. Advance prediction of PE would facilitate increased surveillance, early recognition and treatment of these patients, reducing complications. Biomarkers (e.g. S-Flt and PlGF) and uterine artery doppler have been proposed for this purpose and several groups have reported promising results.(1) The primary focus of this project is to develop an inexpensive, non-invasive, point-of-care test that can improve birth outcomes.


After written, informed consent, women receiving prenatal care at our medical center were enrolled and monitored with pulse oximetry and ECG at prenatal visits after 13 weeks. Upon delivery, subjects were labeled normotensive, HTN, or preE (including super-imposed). We analyzed the following data as indirect measures of arterial compliance during pregnancy: 1) Time lag between ECG and photoplethysmograph (PPG); 2) Heart rate variability; and 3) PPG waveform features. We used the Least Absolute Shrinkage and Selection Operator (LASSO) procedure to select the most informative features regarding the state of the subject. The classifier is a Linear Discriminant Analysis classifier trained with five-fold cross validation. To ensure the classifier appropriately models the data, it was trained and tested with 500 different trials.


PE vs Normotensive: Using second trimester collections (24 PE, 37 controls), the system obtained an AUC of 0.83 (sensitivity 0.76, specificity 0.73). Using third trimester collections (23 PE, 84 controls), the system had an AUC of 0.90 (sensitivity 0.80, specificity 0.83).

PE vs HTN: Using second trimester collections (24 PE, 22 HTN), the system obtained an AUC of 0.91 (sensitivity 0.80, specificity 0.80). Using third trimester collections (23 PE, 41 HTN), the system had an AUC of 0.87 (sensitivity 0.84 specificity 0.79).


Vascular reactivity changes pre-date the onset of preeclamptic. Doppler assessment of uterine artery flow and measurement of circulating angiogenic proteins has shown promise in the prediction of preeclampsia.(1) These tests, however, incur significant expense and are not practical for application in the developing world. Arterial stiffness and wave reflection also predict placental-mediated disease.(2) This study shows that inexpensive, non-invasive and reusable ECG and PPG technology may substitute for more expensive applanation tonometry.

This may allow prediction of preeclampsia and enhance prenatal management and preventative therapy research.


(1) Akolekar et al. Fetal Diagnosis and Therapy 33(1):8-15, 2013.

(2) Osman et al. Journal of Hypertension 36(5):1005-14, 2018.

SOAP 2019