Assessment of a volume-dependent dynamic respiratory system compliance in ALI/ARDS by pooling breathing cycles

New methods were developed to calculate the volume-dependent dynamic respiratory system compliance (C(rs)) in mechanically ventilated patients. Due to noise in respiratory signals and different characteristics of the methods, their results can considerably differ. The aim of the study was to establish a practical procedure to validate the estimation of intratidal dynamic C(rs). A total of 28 patients from intensive care units of eight German university hospitals with acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) were studied retrospectively. Dynamic volume-dependent C(rs) was determined during ongoing mechanical ventilation with the SLICE method, dynostatic algorithm and adaptive slice method. Conventional two-point compliance C(2P) was calculated for comparison. A number of consecutive breathing cycles were pooled to reduce noise in the respiratory signals. C(rs)-volume curves produced with different methods converged when the number of pooling cycles increased (n ≥ 7). The mean volume-dependent C(rs) of 20 breaths was highly correlated with mean C(2P) (C(2P,mean) = 0.945 × C(rs,mean) - 0.053, r(2) = 0.968, p < 0.0001). The Bland-Altman analysis indicated that C(2P,mean) was lower than C(rs,mean) (-2.4 ± 6.4 ml cm(-1) H(2)O, mean bias ± 2 SD), but not significant according to the paired t-test (p > 0.05). Methods for analyzing dynamic respiratory mechanics are sensitive to noise and will converge to a unique solution when the number of pooled cycles increases. Under steady-state conditions, assessment of the volume-dependent C(rs) in ALI/ARDS patients can be validated by pooling respiratory data of consecutive breaths regardless of which method is applied. Confidence in dynamic C(rs) determination may be increased with the proposed pooling.