Base excess contains useful information about net non-lactate metabolic acid/base status for mortality prediction

Abstract

Aims & Objectives: Lactate on admission to PICU may predict mortality better than the absolute value of Base Excess (BE). Using only Lactate discards other information that contributes to the overall BE value. We investigated non-Lactate components of Plasma Base Excess (PBE, the gold standard) and BE, and whether they might be considered in mortality prediction. Methods A single-centre dataset of 6212 PICU admissions with measurement on admission of pH, Bicarbonate, BE, Lactate, Chloride, Albumin, Phosphate, Sodium, Potassium, Calcium and Magnesium (all mEq/L) was analysed to calculate PBE and its components. Non-Lactate BE (i.e. BE+Lactate, since Lactate contributes negatively to BE) was compared with non-Lactate PBE, calculated accurately from its components. Results BE and PBE include Lactate and are widely spread and skewed negatively (range-34 to +38, means-1.0 and-1.8 respectively. Non-Lactate PBE is symmetrical and unskewed around mean 0.19, median 0.12, SD 4.61. Non-Lactate BE behaves similarly, mean 0.38, median 0.37, SD 4.94 (approximately Normal, Figure 1). The difference between them on a Bland-Altman plot (Figure 2) is symmetrical about mean-0.2, median-0.34, SD 1.77. The mean is statistically significantly different from zero p<0.001, but not clinically significant. Lactate does not correlate with non-Lactate PBE or any of it components (maximum R2=0.03). Locally weighted regression shows an association between positive and negative values of non-Lactate BE and mortality (Figure 3). Conclusions BE+Lactate can contribute valuable information about the net non-Lactate components, without collinearity with Lactate, and should be considered in multivariate analysis of PICU mortality. (Figure prsented).L6238161182018-09-13 <br />