Association between modified body mass index and 30-day and 1-year mortality after intensive care unit admission: A retrospective cohort study

Abstract

Because conventional body mass index (cBMI) does not reflect fluid accumulation, modified BMI (mBMI, serum albumin multiplied by cBMI) is a more accurate measure of malnutrition status. This study aimed to determine whether mortality after intensive care unit (ICU) admission was associated with cBMI, mBMI, and/or serum albumin levels. The medical records of patients who were admitted to a tertiary hospital ICU between 1 January 2012 and 31 July 2016 were retrospectively reviewed. In total, 6169 ICU-admitted patients were included in the analyses. Multivariate Cox regression analyses revealed that low cBMI, mBMI and albumin level were significantly associated with 30-day and 1-year mortality after ICU admission (hazard ratio < 1.0, p < 0.05). The adjusted area under the curve (AUC) of mBMI for 1-year mortality was significantly higher than that of cBMI (p < 0.001), but not significantly different from that of albumin level (p = 0.098). Low values of mBMI, cBMI and albumin were independently associated with 30-day and 1-year mortality after ICU admission. Combining cBMI and albumin (mBMI) did not increase the validity of the AUC of albumin for 1-year mortality after ICU admission. Our study showed that serum albumin alone, rather than mBMI (combining cBMI), is recommended in predicting mortality among ICU patients.