1995 OPEN FORUM Abstracts
PREDICTING THE WEANING TIME IN POST-OPERATIVE CARDIOTHORACIC SURGERY PATIENTS
Dulsie Pilman CRTT, Melissa Batchelor CRTT, and William Burke RRT, PhD; Veterans Administration Medical Center, and The Respiratory Therapy Program, School of Allied Health Sciences, Indiana University School of Medicine, Indiana University, Indianapolis, IN 46202.
BACRGROUND Early detection of respiratory or hemodynamic problems in the post-operative recovery period is paramount in minimizing the time spent in the ICU by patients recuperating from open heart surgery. However, it is not known how commonly measured respiratory or hemodynamic variables change during the recovery period. We examined the possibility that routine measurements made early in the recovery period could be used to predict the patients weaning time. METHODS We collected pulmonary mechanics and hemodynamics data every 2-to-3 hours from ICU admission to extubation in 13 consecutive patients scheduled to undergo cardiothoracic surgery at the local VA Medical Center. Data were analyzed by dividing the total time of weaning into 5 equal time brackets, each bracket containing 20% of the patients weaning time (bracket 1 contained the first 20% of time, bracket 5 contained the last 20%). Routine pulmonary mechanics and hemodynamics variables were averaged within each time bracket and analyzed for significance. In addition, we defined a SHORT WEAN to be a wean that consisted of progressive decreases in set ventilator rate followed by extubation. Any wean not following this definition was defined as a LONG WEAN. We used ANOVA to determine how routine mechanics or hemodynamics variables changed across the time brackets or affected WEAN TIME. Once the variables affected by the time brackets or WEAN TIME were determined, we used multiple regression to determine what variable combinations could best predict weaning time. RESULTS The mean wean times were 21.7±9.2 and 64.0±27.5 hours for the SHORT WEAN and LONG WEAN groups, respectively. Many variables were affected by both the time brackets and the WEAN TIME including cardiac output, heart rate, peak alveolar pressure, PaO_2, and stroke volume. A regression model of the following form could predict the hours needed for weaning and explained 73% of the variance in weaning time, 83.87+0.61[cdot]HR-4.25[cdot]PkPalv, where HR and PkPalv are the patients heart rate and peak alveolar pressure from time bracket 1. CONCLUSION During time bracket 1, patients who weaned faster tended to have; a lower HR, CVP, mean PAP, and spontaneous resp. rate; and a higher PkPalv, PaO_2, and stroke volume.
SEE ORIGINAL GRAPH