1996 OPEN FORUM Abstracts
THE SPONTANEOUS PATTERN OF BREATHING DURING SIMV VENTILATION IS MORE CHAOTIC IN PATIENTS WHO FAIL EXTUBATION.
Mohamad EL-Khatib. Div. Pediatric Pharmacology & Critical Care. Rainbow Babies and Childrens Hospital, Cleveland, OH.
Introduction: Many studies have established criteria to predict outcome of extubation from mechanical ventilation in the ICU settings. Most of these studies have combined simple respiratory variables (e.g., the number of breaths per minute, tidal volume, blood gas analysis, etc) and combined them in integrative indexes that are supposed to be superior than traditional parameters in predicting outcome of mechanical ventilation. However a value for an integrative index can be achieved with infinite combinations of values for the individual parameters constituting this index and such without a qualitative evaluation of the respiratory pattern, these indexes might be misleading.
Objective: Geometric and numerical techniques from non-linear dynamics were used to characterize the spontaneous respiratory pattern during SIMV ventilation in an ICU setting for patients who passed and failed extubation.
Methods: Fourty-one children (0-15 yrs) mechanically ventilated in the PICU, clinically stable and considered extubatable by their attending physicians were studied. A respiratory monitoring system (VENTRAK, Novametrix, CT) was used to measure the spontaneous volume and flow signals. Patients were then extubated. Failing extubation was defined as reintubation within 24 hours of extubation. Observations of the spontaneous breath trajectories within the two dimensional (ie, volume and flow) respiratory phase portrait were used to qualitatively assess the dynamical breathing behaviors of the two patients groups: successful vs. unsuccessful extubations. The Kolmogorov entropy, or the net exponential rate at which adjacent spontaneous respiratory trajectories diverge in the volume-flow space, which is a dynamic property related to chaotic behavior, was derived. The divergence constantly produces new information (ie, bits) over time. Also the dimension of the pattern, or the number of clusters of nearby trajectories of spontaneous breaths, was derived. High values for both entropy and dimension reflect a more chaotic breathing pattern. Results: Of the 41 patients included, 33 extubations (80%) were successful and 8 (20%) were not.
Successful (n=33) Unsuccessful (n=8)
Entropy (bits/cycle) 0.05±0.03 0.32±0.05 (p < 0.05)
Dimension 1.06±0.08 3.03±0.35 (p < 0.05)
Both the Kolmogorov entropy and the dimension of the breathing pattern were significantly smaller in the successfully extubated group. Conclusion: The spontaneous breathing pattern during SIMV ventilation is more chaotic in patients who failed extubation whereas it is stable and more reproduciable in patients who failed extubation. Thus characterizing the spontaneous breathing pattern might be a useful tool in differentiating between extubation success and failure.