Serum anion gap at admission as a predictor of mortality in the pediatric intensive care unit
Abstract An accurate method to predict the mortality in the intensive care unit (ICU) patients has been required, especially in children. The aim of this study is to evaluate the value of serum anion gap (AG) for predicting mortality in pediatric ICU (PICU). We reviewed a data of 461 pediatric patie...
Guardado en:
| Autores principales: | , , , , , , , , |
|---|---|
| Formato: | article |
| Lenguaje: | EN |
| Publicado: |
Nature Portfolio
2017
|
| Materias: | |
| Acceso en línea: | https://doaj.org/article/95c46f186ce342da98ff5c4f932ff30e |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: | Abstract An accurate method to predict the mortality in the intensive care unit (ICU) patients has been required, especially in children. The aim of this study is to evaluate the value of serum anion gap (AG) for predicting mortality in pediatric ICU (PICU). We reviewed a data of 461 pediatric patients were collected on PICU admission. Corrected anion gap (cAG), the AG compensated for abnormal albumin levels, was significantly lower in survivors compared with nonsurvivors (p < 0.001). Multivariable logistic regression analysis identified the following variables as independent predictors of mortality; cAG (OR 1.110, 95% CI 1.06–1.17; p < 0.001), PIM3 [OR 7.583, 95% CI 1.81–31.78; p = 0.006], and PRISM III [OR 1.076, 95% CI 1.02–1.14; p = 0.008]. Comparing AUCs for mortality prediction, there were no statistically significant differences between cAG and other mortality prediction models; cAG 0.728, PIM2 0.779, PIM3 0.822, and PRISM III 0.808. The corporation of cAG to pre-existing mortality prediction models was significantly more accurate at predicting mortality than using any of these models alone. We concluded that cAG at ICU admission may be used to predict mortality in children, regardless of underlying etiology. And the incorporation of cAG to pre-existing mortality prediction models might improve predictability. |
|---|