More evidence is needed for the association between serum myasthenia gravis and adverse pregnancy, delivery, and neonatal outcomes

作者：Yi‐Ping Song, Liang Ma · 发表于：European Journal of Neurology · 年份：2023 · DOI：10.1111/ene.16161 · 研究领域：Myasthenia Gravis and Thymoma、Pregnancy and Medication Impact、Pharmacological Effects and Toxicity Studies

We read with the great interest the article titled "Pregnancy Outcomes for Women With Myasthenia Gravis and Their Newborns: A Nationwide Register-Based Cohort Study" by Laura O'Connor et al. [1]. The authors found that pregnancy complications did not show an increased risk in patients with myasthenia gravis (MG). The authors have made valuable contributions to the existing literature, but there are several crucial aspects that require further deliberation. First, the study failed to address important confounders. It has been suggested that factors such as marital status, infant sex, and monthly household income may serve as potential confounders [2], particularly considering the association of anemia with adverse pregnancy outcomes [3]. However, the authors did not incorporate them as confounding variables in the logistic regression model, which could potentially impact the adjusted odds ratio of MG and outcomes, thereby influencing the determination of any association between these factors. Furthermore, there are certain limitations associated with the logistic regression approach employed in this study. (i) The criteria employed by the authors for variable selection remain undisclosed. In the case of multiple regression models, there typically exist two approaches to variable selection. First, variables were chosen based on the statistical outcomes of univariate analysis. Mickey and Greenland proposed employing a significance level of 0.25 as a threshold for selecting indep...