Abstract
BACKGROUND: Regional ventilation/perfusion (V/Q) mismatch in intrapulmonary shunt in dependent regions has always been considered a hallmark of ARDS. However, little is known about the spatial distribution of shunt, and a clear definition has been lacking. The aim of the study was to propose two phenotypes for the spatial distribution of intrapulmonary shunt using electrical impedance tomography (EIT) and to investigate the clinical characteristics and outcomes in the two preset phenotypes. METHODS: A total of 76 ARDS patients who received EIT saline contrast examination were included in this retrospective study. Deadspace(%), Shunt(%), and V/Qmismatch(%) were calculated based on the lung V/Q matching map. EIT maps were divided into two horizontal anterior-to-posterior regions of interest, ranging from gravity-independent regions to gravity-dependent regions. The dosal shunt proportion (Shunt(dosal)/Shunt(global)%) was defined as the percentage of shunt in gravity-dependent regions. Based on Shunt(dosal)/Shunt(global)%, the patients were divided into a dependent-shunt group (D-shunt, Shunt(dosal)/Shunt(global)% > 50%) and a nondependent-shunt group (ND-shunt, Shunt(dosal)/Shunt(global)% ≤ 50%). RESULTS: The D-shunt group (n = 46) had lower dorsal ventilation, lower dorsal deadspace, and a higher Shunt(dosal)/Shunt(global)% than the ND-shunt group (n = 30). Multivariable Cox regression analysis showed that Shunt(dosal)/Shunt(global)% was an independent predictive factor for 28-day mortality (HR = 0.06; 95% CI, 0.01-0.36; P = 0.002). There was no significant difference in regional perfusion distribution, global shunt, global deadspace and global V/Q mismatch between the two groups. Moreover, a higher BMI (25.4 [22.9, 29.2] vs. 22.9 [20.8, 26.4], P = 0.04) and more extrapulmonary ARDS patients [65% (30/46) vs. 33% (10/30), P = 0.01] were found in the D-shunt group. A similar PaO(2)/FiO(2) ratio was found between the two groups on Day 0, but the D-shunt group had a higher PaO(2)/FiO(2) ratio on Day 4. A higher 28-day mortality (40% vs. 17%, P = 0.03) and fewer ventilation-free days (VFDs) on day 28 (11.0 [0, 21.8] vs. 20.5 [4.8, 24.0], P = 0.04) were found in the ND-shunt group. CONCLUSION: Two phenotypes of regional shunt gravitational distribution can be revealed by EIT. Patients exhibiting a predominance of dependent shunt were characterized by a higher BMI and extrapulmonary ARDS and may experience faster improvement in oxygenation as well as better clinical outcomes. Further research is necessary to evaluate shunt distribution patterns to guide the individualized treatment of ARDS patients.