Predictors of mortality in patients with suspected propofol infusion syndrome.

OBJECTIVES To identify predictors of mortality in patients with suspected propofol infusion syndrome and to develop a simple scoring system to identify patients with suspected propofol infusion syndrome who are most at risk of death. DESIGN Retrospective, database analysis. SETTING MEDWATCH system. PARTICIPANTS Reports (1989-2005) where propofol was associated with > or = 1 of 24 published propofol infusion syndrome clinical manifestations. INTERVENTIONS None. MEASUREMENTS AND MAIN RESULTS After comparison of demographic and clinical manifestations between survivors and nonsurvivors, a multivariate logistic regression model was built through a stepwise selection process and then used to develop a simplified mortality scoring system. Of 1139 patients with suspected propofol infusion syndrome, 342 (30%) were fatal. Death was more likely if patients were < or = 18 yrs (odds ratio [95% confidence interval], 2.3 [1.7-3.2]), male (1.3 [1.1-1.7]), received a vasopressor (1.8 [1.3-2.5)]), or had the following clinical manifestations: cardiac (3.8 [2.88-4.91]), metabolic acidosis (3.7 [2.7-5.0]), renal failure (1.9 [1.4-2.6]), hypotension (1.8 [1.3-2.3]), rhabdomyolysis (1.8 [1.3-2.3]), or dyslipidemia (2.0 [1.2-3.4]). The multivariable modeling process found that cardiac symptoms, rhabdomyolosis, hypotension, metabolic acidosis, renal failure, and age each affected survival, although significant interactions existed between some of these factors. Based on the combination of the presence or absence of the six factors in the multivariate model, a propofol infusion syndrome mortality risk score of 0 to 4 resulted in a predicted %/observed % mortality for each score of 0 (10%/10%), 1 (24%/24%), 2 (47%/44%), 3 (72%/81%), and 4 (89%/83%). CONCLUSIONS A number of characteristics are independently associated with higher mortality in patients with suspected propofol infusion syndrome, only some of which are currently reflected in the package insert. Further research should focus on prospectively evaluating the mortality scoring system in patients with suspected propofol infusion syndrome.