Predicting early prone position ventilation responsiveness in patients with acute respiratory distress syndrome based...

Studies have demonstrated that over 20% of patients with moderate to severe acute respiratory distress syndrome (ARDS) do not exhibit significant improvement in oxygenation following prone positioning ventilation (PPV). It is vital to investigate the modifiable characteristics associated with PPV, which would facilitate targeted interventions and minimize the adverse effects of PPV. This study aimed to investigate the physiological effects of PPV by using electrical impedance tomography (EIT), and to explore the predictors of response to PPV in patients with moderate to severe ARDS during the early phase.

This is a single-center, prospective, observational study. Ninety-four adult patients who were diagnosed with moderate-to-severe ARDS according to the Berlin definition (PaO₂/FiO₂ < 150mmHg) and underwent protocolized PPV within 48 h post-intubation were enrolled. EIT data and clinical variables were collected at two time points: before PPV (in the supine position) and 16 h after PPV (PPV 16 h, in the prone position). EIT data were analyzed for the global, separated ventral/dorsal regions, and separated four gravity-dependent regions of interest (ROI 1-4).

Of the 94 enrolled patients, 78 (83%) were PPV responders and 16 (17%) were nonresponders. Compared to the nonresponders, the respiratory system compliance (Crs) in responders was significantly higher (35.6 ± 7.5 cmH₂O/mL vs. 31.4 ± 5.8 cmH₂O/mL, p = 0.037) before PPV. The center of ventilation (COV) index was higher in responders compared with nonresponders (median [Interquartile range] 45.1 [41.7-48.1] vs. 40.5 [38.0-45.7], p = 0.035) before PPV. Compared to the responders, the V/Q mismatch% in ROI 3 was higher in the nonresponders (16.5% [12.8%-20.5%] vs. 9.6% [6.6%-13.7%], p < 0.001) before PPV. The shunt% in ROI 3 was also higher in the nonresponders (14.3% [10.4%-19.6%] vs. 4.9% [1.6%-8.5%], p < 0.001) before PPV. The multivariate logistic regression analysis showed that the shunt% in ROI 3 before PPV was an independent predictor of response to PPV (OR 1.190, 95% CI [1.055-1.342], p = 0.005). The receiver operating characteristic (ROC) curve analysis indicated that the shunt% in ROI 3 before PPV predicted response to PPV with an AUC of 0.89 (95% CI [0.82-0.95], p < 0.001), and the cut-off value less than 8.4% before PPV could predict the response to PPV with a sensitivity of 94% and specificity of 76%. When using the shunt% in ROI 3 combined with Crs before PPV as a diagnostic system to predict PPV responsiveness, the AUC of the ROC curve was 0.92, the sensitivity was 0.94, and the specificity was 0.89.

For moderate to severe ARDS patients, the shunt% in ROI 3 before PPV could help to predict the response of PPV during the early phase.