Examinations were performed at Skåne University Hospital, Lund, Sweden, between 2014 and 2019. The Regional Ethical Review board in Lund, Sweden, approved the study (Dnr 2013/244). All participants, and their guardians when appropriate, provided written informed consent before participation. Participants underwent 24-hour ABPM and MRI, and blood were sampled for measures of kidney function. Weight and height were measured in conjunction with 24-hour ABPM or MRI. Body surface area (BSA) was calculated using the Mosteller formula.20

As part of a prospective cohort study, the current study included adolescents born very preterm with early onset FGR actively delivered due to fetal blood flow velocity abnormality between the years 1998 and 2004. Subjects had a birth weight <2 standard deviations,21 absent or reversed end-diastolic blood flow in the umbilical artery as determined by a standardized protocol using Doppler velocimetry, and were delivered with cesarean section between 24–29 gestational weeks (preterm FGR). Two control groups with birth weight appropriate for gestational age (AGA) born in the same time-period (1998–2004) were identified and a total of 102 participants were included prospectively and divided into three groups forming 34 matching triplets (Fig. 1). The first control group was a subset (n = 34) of all children (n = 371) who were born very preterm and who were admitted to the neonatal intensive care unit at Skåne University Hospital, Lund, Sweden. The causes of very preterm birth in this group included preterm premature rupture of the membranes, ablatio placentae, chorioamnionitis and being delivered as the healthy matched twin to a same sex individual in the preterm FGR group.9 This group was selected to match the preterm FGR group for sex, gestational age at delivery, and year of birth and thus consisted of those born very preterm without FGR (preterm AGA). Twins formed a matched pair if the following criteria were met; one twin was born with FGR, the other twin was born AGA, was of same sex, had normal blood flow in the umbilical artery, and twin-to-twin transfusion had been excluded.9 The second control group consisted of children born at term after normal pregnancy and were matched to both groups for sex and year of birth, forming the third triplet (term AGA). Detailed peri- and neonatal data and follow-up studies of cardiovascular, neuro-cognitive and pulmonary outcomes during childhood have been reported previously.9,22,23,24

Upper panel shows initial inclusion of the original cohort. Middle panel shows the current follow-up in adolescence. Lower panel shows the number of individuals who underwent 24-hour ABPM and MRI. FGR fetal growth restriction, AGA appropriate for gestational age, MRI magnetic resonance imaging, 24-hour ABPM 24-hour ambulatory blood pressure measurements.

All individuals from the original cohort were contacted via mail and by phone and asked to participate in this prospective cohort study in adolescence. Potential differences in the presence of peri- and neonatal confounders, e.g., maternal smoking, preeclampsia and bronchopulmonary dysplasia in those who participated versus those who opted out from follow-up were evaluated.

Preeclampsia was defined as diastolic blood pressure >90 mmHg on two or more consecutive occasions >4 h apart, arising after 20 weeks of gestation, and proteinuria >300 mg/L in two random clean-catch midstream urine specimens collected ≥4 h apart. Bronchopulmonary dysplasia was defined as need for supplemental oxygen (FiO2 >0.30) at an age corresponding to 36 gestational weeks.

24-hour ABPM was performed according to clinical routine. In short, systolic, diastolic, and mean arterial blood pressure were measured every 20 min over 24 hours, and activity and potential symptoms were reported. Daytime and nighttime periods were reviewed both separately and combined (SpaceLabs Medical ABP-monitor model 90207 or Ultralite TM 90217A, Issaquah).

Quality assessment and evaluation of the activity diary together with blood pressure readings was performed by a physician (PS) with 10 years of experience. Generally, >80% of successful readings was deemed sufficient for inclusion. Evaluation was in accordance with European Society of Cardiology (ESC) guidelines, using reference values for children and adolescents based on sex, age, and height.16,25 Mean systolic and diastolic blood pressures were graded as normal (<90th percentile), prehypertension (≥90th to <95th percentile) or hypertension (≥95th percentile) for daytime and nighttime separately. A nocturnal decrease in mean arterial blood pressure <10% was graded as pathological and in addition to reference values for children,25 systolic or diastolic blood pressures above adult reference values were graded as hypertension. Normal values for adults were defined for 24 hours as <130/80 mmHg, for daytime <135/85 mmHg, and for nighttime <120/70 mmHg.16

Participants were imaged in supine position using a 1.5T MR scanner (Philips Achieva, Best, the Netherlands; or Magnetom Aera, Siemens Healthineers, Erlangen, Germany). Flow data were acquired in the ascending aorta and descending aorta at diaphragm level using a 2D phase-contrast gradient recalled echo sequence with retrospective ECG gating. Typical parameters were 35 timeframes per cardiac cycle, temporal resolution 23 ms, TR/TE = 9/6 ms, flip angle = 15°, 1.2 × 1.2 × 6 mm (Philips), and TR/TE = 10/3 ms, flip angle = 20°, 1.5 × 1.5 × 5 mm (Siemens). Velocity encoding was 150–250 cm/s to optimize for individual velocity resolution and to avoid aliasing. Respiratory-gated 3D angiography of the thoracic aorta was acquired using a T2-prepared balanced steady-state free precession sequence with isotropic resolution 0.88 mm (Philips) or 0.55 mm (Siemens).

Images were analyzed in Segment 3 (http://segment.heiberg.se, Medviso AB, Lund, Sweden).26 Two observers with 5 (JL) and 6 years (DR) of MRI experience performed analyses.

The ascending aorta and descending aorta at diaphragm level were delineated in magnitude images throughout the cardiac cycle using automatic segmentation with manual correction as needed, and guided by phase-contrast images where appropriate. Linear background phase correction was performed.27 Quantitative flow, flow curves, and cross-sectional area change throughout the cardiac cycle were analyzed.

Aortic PWV was assessed as previously described,28 using the time-to-foot transit time method and blood pulse traveling distance by manual centerline 3D angiography between flow planes (Fig. 2), with flow-curve baseline correction.19

a Shows non-contrast-enhanced 3D angiography of the thoracic aorta with flow measurement planes (solid lines) perpendicular to the ascending aorta and descending aorta at diaphragm level, and the aortic centerline distance (Δd; dashed line) between flow measurement planes. b Shows delineations of the ascending aorta in a magnitude image (b1) and phase-contrast image (b2). c Shows corresponding delineations of the descending aorta at diaphragm level (c1 and c2). d Shows flow curves for the ascending aorta (solid line) and descending aorta (dashed line) used to assess pulse wave velocity using the time-to-foot method. Pulse wave traveling time (Δt) was calculated as the time between upslope tangents intersecting the baseline. Pulse wave velocity was calculated by dividing the aortic centerline distance (Δd) with the time difference (Δt).

Thoracic aortic distensibility in the ascending aorta and descending aorta at diaphragm level was calculated as \(\frac \,-\, A_}} \,\cdot\, \Delta P}}\), where Amax and Amin are the maximum and minimum cross-sectional areas during the cardiac cycle, and \(\Delta P\) equals the brachial blood pressure difference between systole and diastole.29 Oscillometric brachial blood pressure was acquired immediately after the respective flow acquisition.

Blood samples were collected in EDTA test tubes and directly centrifuged (Thermo Scientific Megafuge 8, Thermo Fisher Scientific, Waltham) at 1500 G for ten minutes, pipetted (Eppendorf Research plus 100–1000 μL, Hamburg, Germany) into cryotubes, and stored at −80 °C (Panasonic Ultra-low Temperature freezer, MDF-DU702VH-PE, Panasonic Healthcare Co., Ltd, Tokyo, Japan). Cystatin C was measured at Skåne University Hospital Clinical Chemistry laboratory using validated routine clinical assays. Estimated glomerular filtration rate (eGFR) was calculated using the Caucasian, Asian, Pediatric and Adult (CAPA) equation.30

Statistical analyses were performed using SPSS 26.0 (IBM Corp, Armonk, New York) and GraphPad Prism 9 (GraphPad Software, La Jolla, California). Artwork was performed using the free and open-source vector graphics editor Inkscape (https://inkscape.org/). For 24-hour ABPM data, boys and girls were analyzed separately as per clinical routine whereas MRI data were analyzed both with both sexes combined and separately. Data are expressed as median (range). Kruskal-Wallis with Bonferroni’s multiple comparison test assessed group differences. The Jonckheere-Terpsta and Kendall’s Tau-b tests assessed trends between groups. Pearson’s chi-squared test or Fisher’s Exact Test assessed categorical variables. P values <0.05 were considered to show statistically significant differences.