RESUMO
RESUMEN Introducción: El ArgenSCORE tiene una versión original (I) desarrollada en 1999 sobre una población con mortalidad del 8% y una versión II (recalibración del modelo en 2007) sobre una población con una mortalidad del 4%. Evaluamos en el registro CONAREC XVI la hipótesis de que el ArgenSCORE II podría estimar mejor el riesgo de mortalidad intrahospitalaria en los centros con baja mortalidad; en cambio, el ArgenSCORE I estimaría mejor la mortalidad en los centros con alta mortalidad. Material y métodos: Se analizaron 2548 pacientes de 44 centros del registro prospectivo y multicéntrico en cirugía cardíaca, CONAREC XVI. En cada centro se evaluó la mortalidad media observada (MO) y se calculó la mortalidad estimada media (ME) aplicando ambas versiones del ArgenSCORE. Se calculó la relación MO/ME de cada centro para los dos modelos y se evaluó si había diferencias significativas mediante el test Z. Resultados: La mortalidad intrahospitalaria del registro fue del 7,69%. El 75% de los centros (33/44) presentaban una mortalidad mayor del 6%. En centros con mortalidad menor del 6%, al aplicar el ArgenSCORE II, la relación MO/ME mostró valores cercanos a 1 y sin diferencias significativas. En centros con mortalidad mayor del 6%, el ArgenSCORE II subestima significativamente el riesgo. En cambio, cuando se aplica en estos centros el ArgenSCORE I, la relación MO/ME es cercana a 1 (sin diferencias significativas). Conclusiones: En centros con mortalidad menor del 6%, es recomendable utilizar el ArgenSCORE II-recalibrado; en centros con mortalidad mayor del 6%, tiene mejor desempeño el ArgenSCORE I-original.
ABSTRACT Background: The ArgenSCORE I was developed in 1999 on a population with 8% mortality. The ArgenSCORE II emerged after recalibrating the original model in 2007 on a validation population with 4% mortality. Using the CONAREC XVI registry, we evaluated the hypothesis that the ArgenSCORE II could better predict the risk of in-hospital mortality in centers with low mortality, whereas the ArgenSCORE I could better predict mortality in centers with high mortality. Methods: A total of 2548 patients from 44 centers of the prospective and multicenter cardiac surgery CONAREC XVI registry, were analyzed. Mean observed mortality (OM) and mean expected mortality (EM) were estimated applying both versions of the ArgenSCORE. The OM/EM ratio was calculated in each center for both models and the Z test was used to evaluate significant differences. Results: In-hospital mortality was 7.69% for the entire registry. In 75% of the centers (33/44) mortality was >6%. In centers with mortality <6%, the OM/EM ratio was close to 1 after applying the ArgenSCORE II, without significant differences. In centers with mortality >6%, the ArgenSCORE II significantly underestimated the risk. On the contrary, when the ArgenSCORE I was applied in these centers, the OM/EM ratio was close to 1, without significant differences. Conclusions: The recalibrated ArgenSCORE II is recommended in centers with mortality <6%, while in those with mortality >6% the original ArgenSCORE I has better performance.
RESUMO
This study aims to develop the first Latin-American risk model that can be used as a simple, pocket-card graphic score at bedside. The risk model was developed on 2903 patients who underwent cardiac surgery at the Spanish Hospital of Buenos Aires, Argentina, between June 1994 and December 1999. Internal validation was performed on 708 patients between January 2000 and June 2001 at the same center. External validation was performed on 1087 patients between February 2000 and January 2007 at three other centers in Argentina. In the development dataset the area under receiver operating characteristics (ROC) curve was 0.73 and the Hosmer-Lemeshow (HL) test was P=0.88. In the internal validation ROC curve was 0.77. In the external validation ROC curve was 0.81, but imperfect calibration was detected because the observed in-hospital mortality (3.96%) was significantly lower than the development dataset (8.20%) (P<0.0001). Recalibration was done in 2007, showing excellent level of agreement between the observed and predicted mortality rates on all patients (P=0.92). This is the first risk model for cardiac surgery developed in a population of Latin-America with both internal and external validation. A simple graphic pocket-card score allows an easy bedside application with acceptable statistic precision.