Breathing-phase selective filtering of respiratory data improves analysis of dynamic respiratory mechanics

Abstract

Background:

The analysis of non-linear respiratory system mechanics under the dynamic conditions of controlled mechanical ventilation is affected by systemic disturbances of the respiratory signals. Cardio-pulmonary coupling induces cardiogenic oscillations to the respiratory signals, which appear prominently in the second half of expiration.

Objective:

We hypothesized that breathing phase-selective filtering of expiratory data improves the analysis of respiratory system mechanics.

Methods:

We retrospectively analyzed data from a multicenter-study (28 patients with injured lungs, under volume-controlled ventilation) and from two additional studies (3 lung healthy patients and 3 with injured lungs, under pressure-controlled ventilation). Data streams were recorded at different levels of positive end-expiratory pressure. Using the gliding-SLICE method, intratidal dynamic respiratory mechanics were analyzed with and without low-pass filtering of expiratory or inspiratory data separately. The quality of data analysis was derived from the coefficient of determination (R²).

Results:

Without filtering, R² lay below 0.995 for 87 of 280 investigated data streams. In 68 cases expiration-selective low-pass filtering improved the quality of analysis to R² ⩾ 0.995. In contrast, inspiration-selective filtering did not improve R².

Conclusions:

The selective filtering of expiration data eliminates negative side-effects of cardiogenic oscillations thus leading to a significant improvement of the analysis of dynamic respiratory system mechanics.