RESUMEN
AIMS: To describe real-life experience with sensor-augmented pump therapy with predictive low-glucose management (SAPT-PLGM), in terms of hypoglycemia and glycemic control after one year of follow-up in T1D patients with hypoglycemia as the main indication of therapy. METHODS: Retrospective cohort study under real life conditions. Baseline and one-year follow-up variables of glycemic control, hypoglycemia and glycemic variability were compared. RESULTS: Fifty patients were included, 31 on prior treatment with SAPT with low-glucose suspend (LGS) feature and 19 on multiple dose insulin injections (MDI). Mean HbA1c decreased in the MDI group (8.24%-7.08%; pâ¯=â¯0.0001). HbA1c change was not significant in the SAPT-LGS group. Area under the curve (AUC) below 70â¯mg/dl improved in both SAPT-LGS and MDI groups while AUC, %time and events below 54â¯mg/dl decreased in SAPT-LGS group. Glycemic variability improved in the MDI group. Less patients presented severe hypoglycemia with SAPT-PLGM in both groups, however the change was non-significant. CONCLUSIONS: Under real life conditions, SAPT-PLGM reduced metrics of hypoglycemia in patients previously treaded with MDI and SAPT-LGS without deteriorating glycemic control in SAPT-LGS patients, while improving it in patients treated with MDI.
Asunto(s)
Biomarcadores/análisis , Diabetes Mellitus Tipo 1/tratamiento farmacológico , Hiperglucemia/prevención & control , Hipoglucemia/prevención & control , Hipoglucemiantes/administración & dosificación , Sistemas de Infusión de Insulina/estadística & datos numéricos , Insulina/administración & dosificación , Adolescente , Adulto , Glucemia/análisis , Femenino , Estudios de Seguimiento , Hemoglobina Glucada/análisis , Humanos , Masculino , Pronóstico , Estudios Retrospectivos , Factores de Tiempo , Adulto JovenRESUMEN
This work presents the results of a new methodology for hybridizing metaheuristics. By first locating the active components (parts) of one algorithm and then inserting them into second one, we can build efficient and accurate optimization, search, and learning algorithms. This gives a concrete way of constructing new techniques that contrasts the spread ad hoc way of hybridizing. In this paper, the enhanced algorithm is a Cellular Genetic Algorithm (cGA) which has been successfully used in the past to find solutions to such hard optimization problems. In order to extend and corroborate the use of active components as an emerging hybridization methodology, we propose here the use of active components taken from Scatter Search (SS) to improve cGA. The results obtained over a varied set of benchmarks are highly satisfactory in efficacy and efficiency when compared with a standard cGA. Moreover, the proposed hybrid approach (i.e., cGA+SS) has shown encouraging results with regard to earlier applications of our methodology.