Publication: Estimación del riesgo de aborto y muerte neonatal en embarazos de gemelos con dos enfoques diferentes: longitudinal y tiempo hasta el evento
Loading...
Official URL
Full text at PDC
Publication Date
2020
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Facultad de Estudios Estadísticos
Citations
Exportar
Abstract
Los embarazos gemelares son embarazos con un alto riesgo de muerte fetal (aborto), por lo que es frecuente inducir el parto antes de que éste llegue a término. Sin embargo, el parto prematuro se asocia con un mayor riesgo de muerte neonatal. Por lo tanto, se necesitan estimaciones sólidas para el riesgo de muerte fetal y de muerte neonatal, para poder decidir el tiempo óptimo para el parto.
En el artículo de Cheong-See F. et al. se aborda este problema mediante una revisión sistemática. El análisis se realiza con un enfoque de datos longitudinales y analizando la edad gestacional en semanas, como variable categórica. Tomando este artículo como punto de partida, el riesgo de aborto y muerte neonatal se han estimado mediante dos enfoques y técnicas estadísticas. Primero se han estimado los riesgos con un enfoque de modelos longitudinales, analizando la semana gestacional, no solo como variable cualitativa, sino también como variable cuantitativa. Los resultados obtenidos son similares a los publicados en el artículo de referencia; en el que se concluyó que la diferencia de riesgos (DR) es cero en la semana 37.
Dado que la ocurrencia de aborto o muerte neonatal ocurren solo una vez en cada embarazo; la incidencia de evento se ha estimado aplicando técnicas de análisis de supervivencia con riesgos competitivos para cada evento, y la técnica bootstrap para estimar la diferencia de incidencias, por edad gestacional (en semanas). En este análisis se concluye que la incidencia de aborto se incrementa de forma considerable a partir de la semana 37, en línea con los resultados previos. Sin embargo, el enfoque de análisis de supervivencia tiene limitaciones importantes cuando se trata de analizar la incidencia de muerte neonatal y estos resultados deben ser interpretados con cautela.
Twin pregnancies are high risk, with a high risk of stillbirth, which is why it is common to induce labor before term. However, premature deliveries are associated with a higher risk of neonatal death. Therefore, robust estimates of stillbirth from continuing the pregnancy and the neonatal risk from early delivery to decide on the optimal timing of delivery are needed [1]. In the article by Cheong-See F. et al. [1] this problem is addressed through a systematic review. The analysis uses an approach based on longitudinal data and meta-analysis, with the gestational age analyzed as a categorical variable. Taking this article [1] as a starting point, stillbirth and neonatal death risk are estimated following two different approaches. First risks are estimated with a longitudinal approach, analyzing the gestational not only as a qualitative variable but also as a quantitative variable. The obtained results are similar to the ones published in the reference article; in which it is concluded that risk difference is zero at week 37. Due to the fact the occurrence of stillbirth or neonatal death happens only once in each pregnancy; the incidence of the events has been estimated by applying survival analysis techniques with competitive risks for each event, and bootstrap techniques to estimate the difference of incidences; by gestational age (in weeks). In this analysis, it is concluded that the risk of stillbirth increases considerably from week 37, in line with previous results. However, the survival analysis approach has important limitations when it comes to analyzing the incidence of neonatal death and these results should be interpreted with caution.
Twin pregnancies are high risk, with a high risk of stillbirth, which is why it is common to induce labor before term. However, premature deliveries are associated with a higher risk of neonatal death. Therefore, robust estimates of stillbirth from continuing the pregnancy and the neonatal risk from early delivery to decide on the optimal timing of delivery are needed [1]. In the article by Cheong-See F. et al. [1] this problem is addressed through a systematic review. The analysis uses an approach based on longitudinal data and meta-analysis, with the gestational age analyzed as a categorical variable. Taking this article [1] as a starting point, stillbirth and neonatal death risk are estimated following two different approaches. First risks are estimated with a longitudinal approach, analyzing the gestational not only as a qualitative variable but also as a quantitative variable. The obtained results are similar to the ones published in the reference article; in which it is concluded that risk difference is zero at week 37. Due to the fact the occurrence of stillbirth or neonatal death happens only once in each pregnancy; the incidence of the events has been estimated by applying survival analysis techniques with competitive risks for each event, and bootstrap techniques to estimate the difference of incidences; by gestational age (in weeks). In this analysis, it is concluded that the risk of stillbirth increases considerably from week 37, in line with previous results. However, the survival analysis approach has important limitations when it comes to analyzing the incidence of neonatal death and these results should be interpreted with caution.