Abstract
BACKGROUND: We aim to develop a non-invasive, bed-side method for supporting personalised ventilation of neonates with congenital diaphragmatic hernia (CDH). Currently, there are no CDH severity measures to do it. As ventilation inhomogeneity (VI) resulting from lung hypoplasia is highly variable in CDH patients, mechanical ventilation is a real challenge and the risk of lung injury is high. METHODS: We conducted 250 simulations of conventional ventilation of CDH cases using the infant hybrid (numerical-physical) respiratory simulator and a ventilator. Utilising simulation results, we searched for a regression model describing patient ventilation parameters as a function of the respiratory system parameters, ventilator settings and two new CDH severity measures: VI-degree defined as a ratio of time constants ratio of the contralateral and ipsilateral lung (T(1)/T(2)) and chest-wall-to-lung compliance ratio (C(W)/C(L)). The regression model aimed to find the T(1)/T(2) and C(W)/C(L) values for real CDH cases and estimate optimal, matched to VI-degree, peak inspiratory and mean airway pressure (PIP, MAP). RESULTS: The developed regression models (R(2) = 0.78 ÷ 0.98; P < 0.001) enabled to find clinically hard-to-measure values of T(1)/T(2) and C(W)/C(L) ratios for three patients, respectively: 9 and 6.52 (P(1)), 3.5 and 4.96 (P(2)), and 4 and 5.02 (P(3)). The T(1)/T(2) and C(W)/C(L) correlated with defect size (gamma coefficient: 1; P < 0.05), duration of mechanical ventilation and hospitalization (Spearmen's coefficient: 0.99; P < 0.01). The clinical and estimated PIP and MAP didn't differ statistically. CONCLUSION: The T(1)/T(2) and C(W)/C(L) indices can help to personalize CDH infants' ventilation and might be used for prognostication.