In 1935 Leon Chesley¹ initiated a follow-up of a small cohort of women who experienced eclampsia to try to find the answer to the question of whether eclampsia causes vascular damage, eventually culminating in CVD in women. Close to a century later, we are still struggling to understand the link between hypertensive disorders of pregnancy (HDP) and later cardiovascular health.

In this special issue of Paediatric and Perinatal Epidemiology, Johnston and colleagues² propose a set of guidelines for studies using administrative data that aim to explore the risks of cardiovascular disease (CVD) risk in women with a history of HDP. They document key, important features to be aware of when planning studies to explore the associations between HDP and CVD. Over a decade ago, the association between pregnancy complications and CVD risk was judged to be remarkably consistent across studies.³ As with all scientific discoveries, the research has evolved from acknowledging the link to trying to better understand and characterise it.

The paper by Johnston et al.² includes a detailed illustration of the hypothesised pathological mechanisms underlying these associations. The authors highlight that the nature of the associations observed is still unknown. Does experiencing HDP cause CVD? This is yet to be answered. The paper describes methodological and conceptual challenges, not only useful for the specific question, but also relevant to other life course issues that we confront in reproductive epidemiology. The authors state that future research is needed to ‘monitor HDP burden and associated CVD risk’. Within this lies an aim to prevent later CVD in women exposed to HDP. How can we achieve this?

Johnston et al.² point to the remaining questions on the heterogeneity of HDP and CVD phenotypes. There is an important design aspect to explore the association between HDP and later CVD: Studying exposure to HDP using longitudinal data within a woman to capture their full reproductive history. Johnston et al. highlight the importance of correct classification on HDP exposure (Table 1 in their manuscript) where they list five categories of challenges, noting ‘… consider participants' lifetime exposure to HDP’. Why is this important?

The inclusion of a woman's complete reproductive history provides several insights: Understanding recurrent risks, an accurate description of whether she was previously exposed to HDP, the order in which the complications occurred, the total number of pregnancies a woman achieved, and if some or all of her pregnancies were complicated. All of this information is vital to moving us towards an understanding of which women are actually at high risk after experiencing HDP.

Cross-sectional studies or studies focusing on first pregnancies, as also discussed by Johnston et al.,² do not take this aspect into account. Women who stop reproduction after a first pregnancy complicated by preterm preeclampsia have a 9-fold higher risk of CVD mortality compared to women with two or more pregnancies without preeclampsia.⁴ Furthermore, the order in which preeclampsia occurs in a woman's reproductive history matters.⁵ Pregnancy complications in one pregnancy may be linked to other complications in a subsequent pregnancy, even complications at term. By using a woman's complete pregnancy history, it is possible to evaluate the association between preeclampsia and long-term CVD mortality by the pattern of occurrence of HDP,⁶ as well as other complications, in later pregnancies. For example, women who experience HDP in their first pregnancy and remain normotensive in subsequent pregnancies may have a different CVD risk profile than women with other patterns of complications and might be less in need of follow-up for CVD prevention.^{1, 7}

Johnston et al. highlight the limitations of using administrative data when evaluating the link between HDP and CVD, with reimbursement systems possibly playing a role in coding. Birth registries, such as those in Nordic countries are ideal data sources for surveillance purposes, as a result of their longitudinal capture of births and availability of historical data on a woman's reproductive history. Also, other data sources have been used to study subsequent births to a mother.^{6, 8} Furthermore, as Johnston et al. point out, sufficient follow-up time remains vital for studying the link between HDP and later cardiovascular events. However, a limitation of registries is the lack of granular information on clinical variables and historical data on lifestyle factors, such as smoking and body mass index. Future research should prioritise data sources with sufficient follow-up time to enable researchers to effectively examine the link between HDP and CVD.

Johnston et al. do not emphasise the influence of diabetes in relation to preeclampsia. Rich-Edwards and colleagues³ in their review state that ‘… metabolic dysregulation in pregnancy defines gestational diabetes (GDM) and is a strong risk factor for preeclampsia. There is a considerable overlap of the 2 conditions with twice the rate of preeclampsia in diabetic versus non-diabetic pregnancies’. Johnston et al. mention the need to consider the co-occurrence of other complications, as Bane et al show,⁸ outcomes are related. In Table 1, we provide a matrix that highlights the complex picture of co-occurring complications between 1st and 2nd pregnancies to be aware of when studying HDP and CVD. As expected, there is a high recurrent risk of the same complication, shown on the diagonal of the matrix, but there are also other associations that are positively and negatively associated with each other.

Though the underlying pathways for HDP and later CVD are not fully understood, pregnancy history might be included in clinical care as a means of identifying women with pregnancy complications that should be followed up more closely to prevent premature CVD mortality. Several studies have found no added benefit of including HDPs as predictors in CVD risk prediction tools.^{9, 10} However, if the goal is to develop a prediction tool that accurately identifies women at high risk of CVD, we need to understand which patterns of pregnancy complications are associated with an increased risk of CVD. If these patterns are not fully understood, there is no hope that a prediction tool will identify a subgroup of high-risk women destined to develop CVD.

We agree with Johnston and colleagues that there is a need for new methods to understand the relationship between pregnancy complications and CVD. These goals will not be reached without adequate data, including a woman's complete reproductive history.

1935 年，Leon Chesley ¹对一小群经历过子痫的女性进行了随访，试图找到子痫是否会导致血管损伤并最终导致女性 CVD 问题的答案。近一个世纪后，我们仍在努力了解妊娠期高血压疾病 (HDP) 与后来的心血管健康之间的联系。

在本期儿科和围产期流行病学特刊中，Johnston 及其同事²提出了一套使用管理数据进行研究的指南，旨在探讨有 HDP 病史的女性患心血管疾病 (CVD) 的风险。他们记录了在规划研究探索 HDP 和 CVD 之间关联时需要注意的关键、重要特征。十多年前，妊娠并发症与心血管疾病风险之间的关联被认为在各项研究中非常一致。³与所有科学发现一样，研究已经从承认这种联系发展到试图更好地理解和表征它。

约翰斯顿等人的论文。^图2详细说明了这些关联背后的假设病理机制。作者强调，观察到的关联的性质仍然未知。经历 HDP 会导致 CVD 吗？这还有待回答。本文描述了方法论和概念上的挑战，不仅对特定问题有用，而且与我们在生殖流行病学中面临的其他生命历程问题相关。作者指出，未来的研究需要“监测 HDP 负担和相关的 CVD 风险”。其目的是预防暴露于 HDP 的女性日后发生 CVD。我们怎样才能做到这一点？

约翰斯顿等人。²指出有关 HDP 和 CVD 表型异质性的剩余问题。探索 HDP 与后来的 CVD 之间的关联有一个重要的设计方面：使用女性体内的纵向数据研究 HDP 暴露，以捕获其完整的生殖史。约翰斯顿等人。强调了对 HDP 暴露进行正确分类的重要性（其手稿中的表 1），其中列出了五类挑战，并指出“……考虑参与者终生暴露于 HDP”。为什么这很重要？

纳入女性完整的生育史可以提供一些见解：了解复发风险、准确描述她之前是否接触过 HDP、并发症发生的顺序、女性怀孕的总数以及是否部分或全部怀孕。她的怀孕经历很复杂。所有这些信息对于帮助我们了解哪些女性在经历 HDP 后实际上面临高风险至关重要。

正如 Johnston 等人所讨论的，横断面研究或关注首次怀孕的研究²没有考虑到这一方面。与两次或两次以上妊娠但没有先兆子痫的女性相比，第一次怀孕后因患有早产先兆子痫而停止生育的女性的 CVD 死亡风险高出 9 倍。⁴此外，先兆子痫在女性生育史中发生的顺序也很重要。⁵一次妊娠的妊娠并发症可能与随后妊娠的其他并发症有关，甚至是足月并发症。通过使用女性的完整妊娠史，可以通过妊娠后期 HDP ⁶以及其他并发症的发生模式来评估先兆子痫与长期 CVD 死亡率之间的关联。例如，在第一次怀孕时经历 HDP 并在随后的怀孕中保持正常血压的女性可能与有其他并发症模式的女性有不同的 CVD 风险状况，并且可能不太需要 CVD 预防的随访。^{1, 7}

约翰斯顿等人。强调在评估 HDP 和 CVD 之间的联系时使用管理数据的局限性，报销系统可能在编码中发挥作用。出生登记处（例如北欧国家的出生登记处）是用于监测目的的理想数据源，因为它们可以纵向捕获出生情况并提供有关妇女生育史的历史数据。此外，其他数据源也被用来研究母亲随后的生育情况。^{6, 8}此外，正如约翰斯顿等人。指出，足够的随访时间对于研究 HDP 与后来的心血管事件之间的联系仍然至关重要。然而，登记的局限性是缺乏临床变量的详细信息和生活方式因素（例如吸烟和体重指数）的历史数据。未来的研究应优先考虑有足够后续时间的数据源，以使研究人员能够有效地检查 HDP 和 CVD 之间的联系。

约翰斯顿等人。不要强调糖尿病与先兆子痫的影响。 Rich-Edwards 及其同事³在他们的评论中指出，“……妊娠期代谢失调定义了妊娠期糖尿病 (GDM)，并且是先兆子痫的一个重要危险因素。这两种情况有相当大的重叠，糖尿病妊娠的先兆子痫发生率是非糖尿病妊娠的两倍。约翰斯顿等人。提到需要考虑其他并发症的同时发生，正如 Bane 等人所表明的，^{8 个}结果是相关的。在表 1 中，我们提供了一个矩阵，突出显示了在研究 HDP 和 CVD 时需要注意的第一次和第二次妊娠之间同时发生的并发症的复杂情况。正如预期的那样，相同并发症的复发风险很高，如矩阵对角线所示，但也存在其他相互正相关和负相关的关联。

尽管 HDP 和后来的 CVD 的潜在途径尚不完全清楚，但妊娠史可能会被纳入临床护理中，作为识别患有妊娠并发症的女性的一种手段，应更密切地随访这些并发症，以防止过早 CVD 死亡。几项研究发现，将 HDP 作为 CVD 风险预测工具中的预测因素并没有带来额外的好处。^{9, 10}然而，如果目标是开发一种预测工具来准确识别 CVD 高风险女性，我们需要了解哪些妊娠并发症模式与 CVD 风险增加相关。如果没有完全理解这些模式，预测工具就不可能识别出注定会患上 CVD 的高危女性亚群。

我们同意 Johnston 及其同事的观点，即需要新的方法来了解妊娠并发症与 CVD 之间的关系。如果没有足够的数据，包括女性完整的生育史，这些目标就无法实现。