The association between sedation practices and duration of mechanical ventilation in intensive care.

Choice of sedation agent may influence duration of mechanical ventilation in the intensive care unit (ICU). We conducted a retrospective observational analysis of 2102 consecutive mechanically ventilated ICU patients over an eight-year period at a Melbourne metropolitan hospital with a ten-bed general ICU to determine if propofol was associated with shorter duration of mechanical ventilation (MV) than midazolam. Data were extracted from the hospital administrative database, pharmacy supply order records and ICU database, to calculate rates of MV and tracheostomy, length-of-stay, propofol and midazolam infusion doses, illness severity and casemix and use of 'sedation scores' and 'sedation break' respectively. The primary end-points were duration of MV, tracheostomy rate and hospital outcome. Negative binomial regression and logistic regression were used to identify temporal trends. From 1 July 2002 to 30 June 2010 there were 5751 ICU admissions including 2102 (36.6%) with MV. Over this period there was a 70% decline in annual midazolam use and a greater than fivefold rise in propofol use. 'Sedation scoring' and 'sedation break' procedures were introduced from 2006. Over the eight-year observation period there were significant increases in the numbers of annual MV admissions and long-term (>96 hours) MV patients, but a decline in median duration of MV, tracheostomy rate, median ICU length-of-stay and median hospital length-of-stay. All temporal trends were significant (P <0.05). The temporal association with changes in sedation management practice, including primary sedative agent choice during MV, may explain these findings.