The aim of this study was to create a model that forecasted the stay in the intensive care unit in post-liver transplantation.
Methods: Twenty-three consecutive patients who underwent liver transplantation provided samples for serum sodium, serum creatinine, total bilirubin, cholesterol, aspartate and alanine aminotransferase, alkaline phosphatase (ALP), albumin, and platelet count for correlation together with age at transplantation in a Pearson correlation model with intensive care unit stay. Multivariate analysis used a regression model to evaluate the relationship between the dependent variable "intensive care unit stay" and the predictor variables that were correlated by a Pearson correlation test. To test the acceptability and strength of the model, analyses of variance was performed and a multiple correlation coefficient R was calculated for the model.
Results: Pearson correlation test showed a strong correlation between intensive care unit stay and creatinine (correlation coefficient = 0.34, P = .03), serum sodium (correlation coefficient = -0.42, P < .01), and total bilirubin (correlation coefficient = -0.29, P = .06). Other variables showed no significant correlation, namely correlation coefficients < 0.24 (P > .1). The final model to evaluate the relationship between the dependent variable "intensive care unit stay" and laboratory parameters included ALP, serum creatinine, serum sodium, and total bilirubin as well as a correction for age.
Conclusions: The most significant parameters were total bilirubin, serum creatinine, and serum sodium. The proposal model significantly correlated with the variable "intensive care unit stay." Such data are particularly important since increased intensive care unit stay correlates with a significant reduction in 1-year survival rate.