(370) Application of a Multiscale Model of Heart Growth in Pregnancy to Patient-Specific Echocardiographic Data

Objective: Computational models are a useful tool in predicting changes during pregnancy, as real-time monitoring can be invasive or not practically feasible. Here, we apply a multiscale model of left ventricular (LV) growth throughout pregnancy. The model was scaled to patient masses and mean arterial pressures (MAP) to assess how the model represents individual pregnant patients.

Study Design: The MATLAB code for the previously developed model in the rat is publicly available and allometric scaling was used to adapt the rat model’s hemodynamics to human physiology. Hormone levels (estrogen, progesterone, and angiotensin II) were adapted to human concentrations reported by the literature. Then, using an obstetric outcomes database, we identified 663 pregnant patients without comorbidities and at least one normal echocardiogram during pregnancy. The model was executed for 18 patients, which was scaled to the patient’s pre-pregnancy mass and matched to MAP. Hormones were considered the same for all patients.

Results: The model demonstrates an increase in LV mass by approximately 60-70% through pregnancy. The allometric scaling between patients did not significantly affect the model’s growth, as each patient’s LV mass growth curve was nearly the same (Figure 1). The model overpredicted LV mass in patients with a pre-pregnancy mass of < 60 kg, but was nearly all within a 20% error range of the patient’s LV mass in patients > 60 kg (Table 1). The model’s predictions for end-diastolic diameter (EDD) and cardiac output
(CO) were overestimated in the patients weighing > 60 kg prior to pregnancy (Table 1).

Conclusion: A multiscale model demonstrating heart growth during pregnancy has been scaled from rats to humans. It is a novel use of this model to compare to human echocardiographic data that makes reasonable predictions, especially in normal-weight patients. Future directions include comparison of the model’s accuracy to longitudinal data for specific patients to determine which factors are most instrumental in predictions and how to make determinations about those at risk for pregnancy-associated cardiac morbidity.