Rev. cir. (Impr.); 74 (1), 2022
Publication year: 2022
Resumen Introducción:
Existen distintos scores pronósticos para determinar probabilidad de muerte en pacientes quemados. El Índice de Garcés (IG) utilizado en Chile, no es aceptado mundialmente. Objetivo:
Decidimos evaluar la correlación entre IG y Abbreviated Burn Severity Index (ABSI), de uso generalizado. Materiales y Método Estudio descriptivo analítico transversal de pacientes quemados subidos a la plataforma online "Registro Nacional de Quemados" y luego ingresados a nuestro centro entre julio de 2017 y julio de 2018. Se calculó coeficiente de correlación de Pearson entre IG y ABSI. Se normalizan ambas variables para hacerlas comparables y se calculó concordancia entre estas con coeficiente de correlación de Lin. Con análisis de regresión logística se calculó probabilidad de fallecer asociada para ambas variables y se compararon áreas de las curvas ROC de ambos scores. Resultados:
De 141 pacientes, 15 fallecieron durante hospitalización, con medias de IG de 104 ± 49 puntos y ABSI 7 ± 2. El coeficiente de correlación de Pearson entre IG y ABSI demostró ser muy alto con un valor de 0,928. El coeficiente de correlación de Lin entregó concordancia moderada con un valor de 0,93. Al comprobar cuál de las dos puntuaciones predice mejor la probabilidad de muerte mediante análisis de sensibilidad y especificidad, la curva ROC cubre más área en IG (0,82 vs. 0,83), sin ser estadísticamente significativo. Conclusiones:
Ambos índices, IG y ABSI, son muy similares para predecir la probabilidad de muerte, por lo cual son equiparables a la hora de compartir resultados de estudios científicos.
Introduction:
There are different prognostic scores to determine the probability of death in burned patients. The Garces Index (GI) used in Chile is not accepted worldwide. Aim:
We decided to evaluate the correlation between IG and Abbreviated Burn Severity Index (ABSI), which is widely used. Materials and Method:
Cross-sectional analytical and descriptive study of burned patients uploaded to the online platform "National Burns Registry" and then admitted to our center between July 2017 and July 2018. Pearson's correlation coefficient between IG and ABSI was calculated. Both variables were normalized to be able to make them comparable and the agreement between them was calculated with Lin's correlation coefficient. With logistic regression analysis, the associated probability of death was calculated for both variables and areas of the ROC curves of both scores were compared. Results:
Of 141 patients, 15 died during hospitalization, with a mean GI 104 ± 49 and ABSI 7 ± 2. Pearson's correlation coefficient between GI and ABSI showed a very high correlation with a value of 0.928. Lin's correlation coefficient gave moderate agreement with a value of 0.93. When checking which of the two scores best predicts the probability of death through sensitivity and specificity analysis, the ROC curve covers more area in IG (0.82 vs 0.83) without being statistically significant. Conclusions:
Both scores, IG and ABSI, are very similar when it comes to predicting the probability of death, which is why they are comparable at the time of sharing the results in scientific studies.