RE: RESPONSE TO "VACCINATION AND THE MENSTRUAL CYCLE".
RE: RESPONSE TO "VACCINATION AND THE MENSTRUAL CYCLE".
复制标题
回复：对“疫苗接种和月经周期”的回应。
DOI：
10.1093/aje/kwad146
复制
发表时间：
2023
影响因子：
5
通讯作者：
Weinberg,ClariceR
中科院分区：
文献类型：
--
作者：
Weinberg,ClariceR
我感谢 Boniface 等人 (1) 对我的建议做出了回应，即长度偏差可能是他们发现 2019 年冠状病毒病 (COVID-19) 疫苗接种延长月经周期长度的基础 (2)。他们分析了 2 个大型在线数据库的前瞻性数据，发现接种 COVID-19 疫苗会延迟随后月经的到来（延迟不到 1 天）。该效果在后续周期中消失。我最初的评论忽略了最初的两篇论文（3、4）利用了来自同一时间和地点的未接种疫苗的对照组——这是他们方法的一个重要特征。然而，对照组的“名义”疫苗接种被指定为在每个未接种疫苗的女性的一系列周期中的第四个月经周期期间假发生的。与基于疫苗接种的抽样不同，他们对对照组采用的抽样是基于周期的，而不是基于天的。如果你随机抽取一天，那一天更有可能处于相对较长的周期，而不是相对较短的周期。相比之下，如果对随机周期进行采样，则该随机周期与其他周期之间不应存在系统性差异。因此，他们使用的“差异之差”方法无法控制我在评论中提到的长度偏差。
I thank Boniface et al.(1) for their response to my suggestion that length bias may underlie their finding that the coronavirus disease 2019 (COVID-19) vaccination extends the length of the menstrual cycle (2). They had analyzed prospective data from 2 large online databases and found that vaccination for COVID-19 delays the onset of the subsequent menstrual period (by less than 1 day). The effect is gone from subsequent cycles. My original commentary neglected to mention that the original 2 papers (3, 4) made use of an unvaccinated comparison group taken from the same general time and place—an important feature of their approach. However, the “notional” vaccination for the comparison group was assigned to have pseudo-occurred during the fourth menstrual cycle in a sequence of cycles from each unvaccinated woman. Unlike sampling based on vaccination, the sampling they employed for the comparison group was cycle-based and not day-based. If you sample a random day, that day is more likely to be in a relatively long cycle than in a relatively short cycle. By contrast, if you sample a random cycle, there should be no systematic difference between the random cycle and the other cycles. Thus the “difference of differences” method they used would not control for the length bias I wrote about in my commentary.