A feasibility study started in June 2016 and was followed by an unpublished pilot study starting in 2019. Available data from the pilot study will be incorporated into the current study. A longitudinal prospective cohort study will be carried out in the current format between August 2024 and October 2028. Participant recruitment will continue until October 2026. This study has two arms: (1) short follow-up—women with a current pre-eclamptic pregnancy (pre-eclamptic group) or an uncomplicated normotensive pregnancy (uncomplicated group); and (2) long follow-up—women from a historical cohort with a pre-eclamptic pregnancy 10-20 years ago.

Both arms will have identical investigations, detailed in online supplemental appendix 1, consisting of placental histopathology and a single postpartum CVD assessment. The CVD assessment includes medical history taking, blood tests, electrocardiography and echocardiography. Additionally, coronary CT angiography (CCTA) focusing on the presence of atherosclerotic plaques and calcium score will be carried out. Short follow-up women will have prospective collection of their placenta, and the single CVD assessment will be carried out 6–36 months postpartum. For the long follow-up, placental tissue is available for reanalysis from their pregnancy 10–20 years ago. The single CVD assessment will take place immediately and corresponds with 10–20 years postpartum. Figure 1 shows an overview of the participation timeline for the short and long follow-up.

Figure shows the timeline of study participation for both tracks and indicates when study inclusion and investigations are carried out. The lines indicate the histopathological analysis of the placenta. The upper bracket indicates the timeframe for the single postpartum CVD assessment for the short follow-up track whereas the lower bracket indicates the same for the long follow-up track. The left box indicates the contents of the single postpartum cardiovascular assessment. cIMT, carotid intima media thickness; CTA, CT angiography; FMD, flow-mediated dilatation; TTE, transthoracic echocardiography; USCOM, non-invasive ultrasonic cardiac output monitoring.

For the short follow-up, participants will be recruited during pregnancy in Maastricht University Medical Centre+ from the outpatient clinic or from the inpatient wards. Alternatively, participants can be recruited from external hospitals or midwifery practices if they express an interest. There is close collaboration with Erasmus MC (Rotterdam), University Medical Centre Groningen (Groningen), Amsterdam Medical Centre (Amsterdam) and Radboud University Medical Centre (Nijmegen).

Participant recruitment differs for the long follow-up. Women from a previously studied pre-eclamptic cohort in Radboud University Medical Centre, Nijmegen, The Netherlands will be invited to participate in this study.19–23 The earlier study included collection of various baseline characteristics and histopathological analyses of the placenta including data on acute atherosis. For participants who choose to take part, the baseline characteristics and placenta data will be made available for use in the current study.

All participants will give written informed consent prior to participating in this study.

To account for variation in pre-eclampsia definitions used in recent years and in different hospitals nationally, the traditional definition will be applied in this study. Pre-eclampsia will be defined as hypertension (systolic blood pressure (BP) ≥140 mm Hg and/or diastolic BP ≥90 mm Hg) developed after 20 weeks of pregnancy with de novo proteinuria (≥300 mg/24 hours or PCR >50 mg/mmol) at any gestational age.24 Uncomplicated pregnancies are pregnancies not complicated by hypertension, growth restriction, placental abruption or preterm birth.

For the long follow-up arm, women are included from a historical cohort with a pre-eclamptic pregnancy 10–20 years ago. The historic inclusion criteria for participating in the original cohort study were being admitted with a pre-eclamptic pregnancy to Radboud University Medical Centre, Nijmegen between 2004 and 2010 and and having histopathological analysis of the corresponding placenta. The historic exclusion criteria were multiple pregnancies, incomplete records or incomplete placental samples.19

For the current study, the following inclusion and exclusion criteria will be applied to participants in both the short and long follow-up arms. In the pre-eclamptic group, women are included with an index pregnancy complicated by pre-eclampsia, whereas in the uncomplicated group, women are included who have an uncomplicated normotensive index pregnancy. The index pregnancy is the pregnancy for which the placenta has been collected and analysed.

There are several exclusion criteria. In both the pre-eclamptic and uncomplicated groups, women will be excluded if they are under 18 or do not want their test results to be shared with their general practitioner, their specialist(s) or themselves. They will also be excluded if they have an allergy or contraindication to iodinated contrast media, beta-blockers or glyceryl trinitrate (GTN) spray. There are no further exclusion criteria for the pre-eclamptic group. In the uncomplicated group, women cannot have an index pregnancy complicated by hypertension, foetal growth restriction, placental abruption or preterm birth. Before their index pregnancy, women with a history of an uncomplicated pregnancy are excluded if diagnosed with chronic hypertension treated with antihypertensive medication, an autoimmune disorder including systemic lupus erythematosus, Crohn’s disease and rheumatoid arthritis or type 1 diabetes mellitus.

For both the pre-eclamptic and uncomplicated groups, women should not be pregnant at the time of the cardiovascular evaluation due to the iodinated contrast media injected for the CCTA. If they are pregnant, they can choose to postpone the postpartum CVD assessment. Breastfeeding is allowed and participants will be instructed to discard breastmilk for 24 hours after completing the investigations. Alternatively, women can also choose to postpone the postpartum assessment.

For the short follow-up, the placenta will be collected directly postpartum. It is either initially stored in a fridge before being transported to the pathology department the next day or transported to the pathology department directly for storage in 4% buffered formaldehyde. The standardised placenta sampling and reporting protocol of the Amsterdam Placental Workshop Group Consensus Statement will be used to guide sampling and analysis of the placentas.18 In addition to standardising placenta sampling and reporting, it also defines terminology for placental lesions and for placental patterns of disease. To account for sampling error of specific lesions in the placenta, a minimum of eleven biopsies will be taken. Two full-thickness sections will be taken from the umbilical cord: one from the maternal side and one from the foetal side. Two membrane rolls will be taken: one from the rupture edge to the placental margin and the other towards the insertion of the placental disk. Sampling of the thickest part of the membranes is preferred to increase the chance of sampling vasculature. Seven sections will be taken from the parenchyma: one full-thickness section at the umbilical cord insertion site, four central sections of the inner two-thirds of macroscopically normal parenchyma and two peripheral sections from both ends of the longest parenchymal slice. If there are any macroscopic lesions, additional sections will be taken.

The placenta will be analysed by an experienced pathologist. There will be particular focus in the analysis on the presence of acute atherosis. Acute atherosis will be defined by the presence of fibrinoid necrosis, lymphocytic infiltration and presence of lipid-laden macrophages (foam cells). Acute atherosis is a subtype of decidual vasculopathy (DV). DV has six subtypes, which can occur in isolation or in combination. Figure 2 shows the different DV subtypes including acute atherosis: A and B. Acute atherosis, C. Fibrinoid necrosis, D. Lymphocytic infiltration (chronic perivasculitis), E. Absence of spiral artery remodelling, F. Arterial thrombosis and G. Mural hypertrophy.18

Figure showing different subtypes of decidual vasculopathy in the maternal blood vessel of the placenta. (A) Acute atherosis consisting of fibrinoid necrosis (FN) and foamy macrophages (M) in the decidua parietalis (D). (B) Acute atherosis consisting of FN and foamy macrophages (M) shown in the placental membranes composed of amnion (A), chorion (C) and decidua parietalis (D). (C) FN with maternal blood vessels surrounded by lymphocytes (L) in the placental membranes composed of amnion (A), chorion (C) and decidua parietalis (D). (D) Chronic perivasculitis/lymphocytic infiltration around one maternal blood vessel which is cuff-like surrounded by lymphocytes (L) in the placental membranes composed of amnion (A), chorion (C) and decidua parietalis (D). (E) Incomplete remodelling of the maternal blood vessels in the decidua basalis characterised by a persistent muscular layer (mu) of the vessel wall. (F) Occluding thrombus (T) in a maternal blood vessel surrounded by normally remodelled blood vessels (nl) in the decidua basalis. (G) Mural hypertrophy (MH) characterised by a wall thickness >30% of diameter of the maternal blood vessels in the decidua basalis (D).

For the long follow-up group, the placentas were collected and analysed in the earlier study.19–23 Microscopic reanalysis of available stored placental blocks and slides will be carried out for the current study by an experienced pathologist following the Amsterdam Placental Workshop Group Consensus Statement.18

The single postpartum CVD assessment will be identical for both the short and long follow-up. Various investigations will be carried out to gain insight into different CVD parameters (Figure 1). All participants will be in a fasted state when they undergo the CVD assessment.

Participants will be asked about their medical history, drug history including intoxications, obstetric history and family history. They will complete a Short QUestionnaire to ASsess Health-enhancing physcial activity (SQUASH) lifestyle questionnaire.25 26 Basic measurements will be taken including weight, height, waist and hip circumference and BP. The median brachial artery BP will be recorded using a 30 minute measurement with 3 minute intervals using a semiautomatic oscillometric device (Dinamap Vital Signs Monitor 1846, Critikon, Tampa, Florida).27

Fasting blood and urine samples will be taken. Blood tests will include haemoglobin, mean corpuscular volume, renal function, liver function, thyroid function, glucose, insulin, HbA1c, cholesterol and NT-pro-BNP. Additional blood samples will be stored at −80°C. These samples will be used in future to determine biomarkers and perform DNA analyses based on new insights, which are anticipated in the coming years.

Cardiac function, structure and geometry will be investigated using ECG, transthoracic echocardiography (TTE) and non-invasive ultrasonic cardiac output monitoring (USCOM). TTE and USCOM measurements will be carried out by an experienced echocardiographer blinded to the patients’ history.

The procedure for TTE follows the most up to date guidelines whereby images will be obtained in a left lateral position.28 Images will be recorded to allow offline analysis. There will be a quantitative assessment of cardiac dimensions and left ventricular systolic function. Left ventricular systolic function will also be assessed according to current guidelines.29 Right ventricular function will be assessed in different ways including using the tricuspid annular plane systolic excursion and tissue Doppler imaging. Diastolic function will also be explored according to current guidelines, including using speckle tracking for strain analyses, and E/A and E/e’ ratios.29

For the USCOM measurement, the participant will be in the supine position. The probe will be placed in the suprasternal notch and directed across the aortic valve to acquire a spectral Doppler flow profile displayed as a time velocity plot. The optimal flow profile will be recognised by a triangular shape where the base covers the full systolic width, and the sides are straight and continuous before converging into a sharp peak. Once the optimal flow profile and highest velocity (minimum of 1 m/s) are obtained, the trace will be frozen. Automatic tracing of the flow profiles will allow calculation of the stroke volume (product of the velocity–time integral and the cross-sectional area of the aortic valve). The value for the cross-sectional area of the aortic valve will be automatically estimated using the height. Three serial measurements will be carried out to calculate cardiac output. The velocity–time integral is calculated for each individual measurement and will usually be calculated by recording all cardiac cycles during four respiratory cycles. The number of cardiac cycles recorded depends on the heart rate. USCOM will record more cardiac cycles than TTE, enabling more comprehensive analyses of the cardiac cycle accounting for both the baroreceptor-mediated fluctuation and the full respiratory cycle.30 Cardiac output will be calculated from the product of the heart rate and the stroke volume.

A CT scan of the heart will be performed using a third-generation dual-source CT scanner (Somatom Force, Siemens Healthineers, Forchheim Germany) or a photon counting CT scanner (NAEOTOM Alpha, Siemens, Healthineers, Forchheim Germany). ECG and heart rate will be monitored during CCTA. Prior to CCTA, a standardised non-contrast scan of the heart will be performed (eg, Agatston score) to determine the overall burden of coronary calcification. For this score, calcified plaques in each coronary artery are identified by voxels ≥1 mm2 with a minimum threshold of 130 Hounsfield units.31 Figure 3 shows examples of CCTA images and the presence of calcified plaque in the left anterior descending artery. The non-contrast scan of the heart is followed by a CCTA after administration of iodinated contrast media. In patients with a stable heart rate <65 bpm, a prospectively ECG-triggered ‘high pitch’ spiral protocol will be used. In patients with a stable heart rate between 60 bpm and 90 bpm, a prospectively triggered ‘adaptive sequence’ protocol will be used (prospective sequential data acquisition). Retrospective gating will be used in patients with an irregular heart rate or a heart rate >90 bpm. Oral or intravenous metoprolol will be given shortly before the scan, if deemed necessary. GTN will be given to all patients for vasodilation prior to the scan, in the absence of contraindications. Images will be reviewed in consensus by a dedicated cardiovascular radiologist and cardiologist according to the Society of Cardiovascular Computed Tomography guidelines.31 The presence and extent of CAD, and its subsequent plaque morphology and degree of stenosis will be described.

(A) Transverse CTA image of the heart depicting the calcified plaque in the proximal LAD (white arrow). (B) Non-contrast scan of the heart to determine the presence and extent of calcified plaques. White arrow indicates a calcified plaque in the LAD. (C) Curved MPR image of the LAD. The white arrows indicate the presence of calcified plaques. (D) 3D volume rendering image of the heart. Black arrow indicates the LAD. CTA, CT angiography; LAD, left anterior descending artery; MPR, multiplanar reformation.

The carotid arteries will be assessed bilaterally using B-mode ultrasound following an identical procedure for both carotid arteries. Initially, the systolic and diastolic flow is measured at the proximal common carotid, middle common carotid, distal common carotid, proximal internal carotid and the proximal external carotid. Afterwards, a straight section of the proximal common carotid, middle common carotid, distal common carotid and carotid bulb will be visualised. For each section, the distance between the border of the luminal intima and the border of the media adventitia will be measured to determine the carotid intima media thickness. Subsequently, the presence of plaque at the carotid bulb will be determined; if present, the length, thickness, circumference and area of the plaque will be measured.

In addition, the intima media thickness will also be measured in a different manner. A straight segment of the common carotid artery will be selected just before the carotid bulb or bifurcation and a recording is made of five to six cardiac cycles. This recording will be analysed postcapture using a vidArt program developed by professor A. Hoeks (Matlab R2013b, The Math-works Massachusetts). This program can identify the walls and borders of the artery by tracking these automatically. This can be adjusted manually if recording quality is poor or artefacts are present. Analysis using this program allows intima media thickness to be measured and also allows additional variables to be generated such as the SD of the intima media thickness and the homogeneity of the measurements.

Flow-mediated dilatation (FMD) measured will evaluate endothelial function using a brachial artery ultrasound. FMD is the endothelium-mediated dilatation response to a sudden increase in shear stress. The brachial artery diameter will be measured continuously over 9 min. A sphygmomanometer cuff will be placed on the dominant forearm of the subject. During the first 3 min baseline diameter and baseline blood flow, velocity of the brachial artery will be determined. Subsequently, the cuff on the forearm will be inflated up to 200 mm Hg and will remain at this pressure for 5 min to cause distal (forearm) hypoxia. The cuff will then be rapidly deflated and as a response to the hypoxia, vasodilatation of the vessels will occur. The arterial diameter and blood flow velocity will be measured for another 5 min. During the measurement, dual mode ultrasound images will be recorded. Afterwards, a new 3 min rest recording will be carried out followed by one spray of sublingual GTN which will be administered. GTN induces vasodilatation by converting into nitric oxide in the vascular wall. Recording of the vessel for 10 min after GTN administration enables discrimination of endothelial-dependent dilation and endothelial-independent dilation.