Hypertensive disorders of pregnancy are not consistently associated with long-term subclinical hypertension-mediated organ damage across cerebral, cardiac, vascular, and renal systems in middle-aged parous females, according to a matched cohort investigation utilizing UK Biobank data. Led by researchers at McGill University and collaborating institutions, the study examined whether parous women with a validated history of pregnancy-related hypertension demonstrate measurable target-organ injury decades after index delivery compared to matched controls. Evaluating standardized multi-organ biomarkers collected during follow-up, the investigators found that after multivariate adjustment for baseline cardiovascular risk factors, prior pregnancy hypertension was not significantly associated with elevated white matter hyperintensity volume on brain magnetic resonance imaging, reduced left ventricular ejection fraction on cardiac imaging, heightened pulse wave arterial stiffness, or elevated urine albumin-to-creatinine ratio. These findings reveal that while hypertensive pregnancies carry established epidemiological associations with future cardiovascular morbidity, subclinical multisystem damage is not uniformly detectable using single-timepoint cross-sectional metrics during midlife.
Hypertensive pregnancy conditions, encompassing gestational hypertension, preeclampsia, and eclampsia, occur in an estimated five to ten percent of all pregnancies globally and remain primary drivers of maternal and fetal morbidity and mortality. Clinically defined by systolic blood pressure reaching at least one hundred forty millimeters of mercury or diastolic pressure reaching at least ninety millimeters of mercury developing after twenty weeks of gestation, these disorders involve complex maternal-fetal pathophysiology characterized by impaired placentation, immune dysregulation, oxidative stress, and profound angiogenic imbalance. A key feature of this antiangiogenic state is the elevation of soluble fms-like tyrosine kinase-1, an antiangiogenic factor that precipitates systemic maternal endothelial dysfunction. Because sustained endothelial injury, persistent microvascular dysfunction, and shared metabolic traits such as dyslipidemia, higher body mass index, insulin resistance, and chronic vascular inflammation are common to both hypertensive gestation and systemic small-vessel pathology, investigators hypothesized that affected individuals would exhibit heightened measurable hypertension-mediated organ damage across multiple vascular beds later in life.
To evaluate these long-term multisystem trajectories, the research team analyzed the extensive prospective dataset of the UK Biobank, which originally enrolled over five hundred thousand participants across twenty-two assessment centers throughout the United Kingdom. Commencing with an initial sampling frame of two hundred seventy-three thousand two hundred twenty-seven female participants, rigorous methodological exclusion criteria were systematically applied to isolate genuine gestational exposures and prevent confounding from pre-existing chronic conditions. The team excluded fifty-one thousand seventeen individuals reporting no prior live births, seven hundred ninety-six with missing parity documentation, four hundred thirty-seven lacking documentation of the year of first live birth, and fifty-eight who experienced deliveries during or following the baseline recruitment period. An additional two hundred nine thousand one hundred forty-four participants lacking complete measurements across all four specified target-organ biomarkers were removed, along with sixteen identified statistical outliers. Subsequent filtering excluded individuals with missing dates of hypertension onset, pre-existing hypertension predating pregnancy, or hypertension diagnoses preceding first delivery, ultimately establishing a refined analytic cohort of one hundred seventy-four females with documented pregnancy hypertension matched one to four with six hundred ninety-four unexposed parous controls.
Matching was conducted with precision based on participant age at first live birth and total parity, ensuring balanced reproductive chronologies between the exposed and unexposed cohorts. At initial baseline evaluation, the overall cohort demonstrated a mean age of fifty years with a standard deviation of seven years, showing comparable chronological ages between exposed individuals at fifty point two years and controls at fifty point eight years. The average latency interval spanning from first delivery to baseline study evaluation was twenty-one point one years among exposed females and twenty-one point five years among unexposed controls, which expanded to twenty-nine point seven years and thirty point zero years, respectively, by the time of subsequent follow-up imaging examinations. Evaluation of metabolic and lifestyle characteristics demonstrated identical mean low-density lipoprotein cholesterol levels of three point five millimoles per liter across both groups. However, females with past pregnancy hypertension exhibited statistically significant elevations in baseline waist-to-hip ratio averaging zero point eight compared to controls, as well as a heightened prevalence of gestational diabetes mellitus at three point five percent versus zero point four percent. Conversely, control participants exhibited a significantly higher documented history of depression at ten point five percent compared to three point five percent among the exposed group, while daily alcohol consumption, smoking history, and atypical sleep duration showed no significant differences.
Assessment of long-term organ damage incorporated standardized multi-organ protocols executed across UK Biobank assessment centers, capturing continuous and categorical representations of structural and functional physiological integrity. Cerebral small-vessel pathology was quantified through total white matter hyperintensity volume derived from sagittal three-dimensional T1-weighted and T2-fluid-attenuated inversion recovery sequences acquired on high-resolution Siemens Skyra three-tesla magnetic resonance imaging scanners, with automated volumetric segmentation executed via the verified UK Biobank imaging pipeline. Cardiac systolic performance was evaluated as left ventricular ejection fraction using comprehensive cardiac magnetic resonance imaging on Siemens MAGNETOM Aera one point five-tesla scanners utilizing electro-cardiogram-gated long-axis and short-axis cine acquisitions combined with automated Cardiac Dot Engine analyses. Large-artery structural stiffness was captured via pulse wave arterial stiffness index derived from ten to fifteen-second digital finger photoplethysmography using PulseTrace PCA2 CareFusion instruments adjusted for individual height-dependent peak-to-peak transit intervals. Renal glomerular and microvascular injury was quantified from baseline spot urine collections through urine albumin-to-creatinine ratio expressed in milligrams per millimole, applying automated imputation for undetectable urinary albumin concentrations alongside standardized creatinine quantification.
Initial categorical examinations demonstrated that parous females with prior pregnancy-induced hypertension exhibited a significantly greater raw proportion of elevated white matter hyperintensity burden, defined as falling within the highest tertile of volumetric distribution, compared to their unexposed counterparts, registering thirty-nine point seven percent versus thirty-one point six percent, which reached statistical significance with a p-value of zero point zero four four. However, categorical evaluations of the remaining three target-organ biomarkers revealed no statistically significant differences between the exposure cohorts. Impaired left ventricular ejection fraction, defined as resting systolic function below fifty percent, was identified in eight point one percent of exposed women compared to seven point five percent of controls, yielding a nonsignificant p-value of zero point seven five one. Similarly, pulse wave arterial stiffness index values exceeding ten meters per second occurred in thirty-one point six percent of exposed subjects compared to thirty-four point one percent of unexposed subjects, while microalbuminuria, designated as a urine albumin-to-creatinine ratio of three milligrams per millimole or greater, appeared in nine point eight percent versus nine point one percent, both demonstrating no statistical divergence.
To rigorously eliminate residual confounding from comorbid cardiometabolic traits, the researchers performed multivariate linear regression modeling adjusting for recruitment age, low-density lipoprotein cholesterol, waist-to-hip ratio, personal history of diabetes, recorded history of depression, daily alcohol consumption, cigarette smoking history, sleep duration, and subsequent incident hypertension. In the fully adjusted primary models, a documented history of pregnancy hypertension was not significantly associated with continuous white matter hyperintensity volume, yielding an adjusted beta coefficient of negative zero point three one with a ninety-five percent confidence interval of negative one point zero four to positive zero point four one and a p-value of zero point four zero one. Similarly, log-transformed cerebral lesion volumes displayed no association, demonstrating a beta of positive zero point zero one and a p-value of zero point seven nine five. Cardiac systolic performance likewise showed no significant association, with left ventricular ejection fraction yielding an adjusted beta of positive zero point seven six, a confidence interval of negative zero point three eight to positive one point nine zero, and a p-value of zero point one nine three. Furthermore, pulse wave arterial stiffness exhibited an adjusted beta of positive zero point one four, while urine albumin-to-creatinine ratio demonstrated an adjusted beta of negative zero point one two, neither demonstrating meaningful statistical significance.
To explore whether latent organ vulnerability is unmasked within distinct clinical phenotypes, the investigators stratified the study population across five clinically meaningful subgroups: subsequent development of chronic hypertension, menopausal status, age fifty years or older, active antihypertensive medication use, and elapsed time since index pregnancy hypertension of ten years or longer. Among participants who went on to develop chronic hypertension, multivariate regression identified an isolated, statistically significant inverse association with log-transformed white matter hyperintensity volume, presenting an adjusted beta coefficient of negative zero point four one with a ninety-five percent confidence interval of negative zero point seven nine to negative zero point zero three and a p-value of zero point zero three three. However, this finding was not mirrored in untransformed cerebral volumes, which yielded a nonsignificant beta of negative four point six one. Among postmenopausal participants, an initial unadjusted association with heightened log-transformed cerebral lesion volume was attenuated to nonsignificance following multivariate adjustment, yielding a beta of zero point zero nine. Across all remaining stratified subsets—including individuals aged fifty or older, those prescribed antihypertensive medications, and those evaluated beyond a ten-year post-index latency—adjusted regression analyses yielded uniformly nonsignificant associations across all four organ biomarkers.
In contextualizing these null and isolated associations within the broader literature of post-pregnancy cardiovascular risk, the study authors emphasize that long-term pathological trajectories may be far more heterogeneous and time-dependent than captured by static single-timepoint assessments. Epidemiological evidence indicates that relative risks for cardiovascular and cerebrovascular complications are highest in the initial years immediately following an affected pregnancy, with renal risks demonstrating pronounced early peaks: previous nationwide cohorts have demonstrated hazard ratios for chronic kidney disease falling from six point one one within five years of pregnancy to two point zero six after five or more years. Because participants in this study averaged fifty years of age and were evaluated approximately three decades after their first pregnancy, early transient vascular abnormalities may have stabilized, or alternatively, age-related degenerative processes may have begun to obscure pregnancy-specific contributions. Furthermore, specific physiological biomarkers may fail to capture subtle microvascular phenotypes: myocardial endothelial injury often manifests as coronary microvascular dysfunction or structural concentric remodeling rather than depressed global ejection fraction, while ongoing glomerular podocyte shedding can persist without manifesting as macroscopic albuminuria detectable via conventional spot ratios.
The study provides notable strengths through its concurrent, standardized multi-organ imaging and biochemical evaluation of cerebral, cardiac, arterial, and renal parameters within an exceptionally well-characterized matched cohort. Nonetheless, several methodological limitations must be considered when interpreting these findings. The cross-sectional acquisition of imaging biomarkers at a single follow-up timepoint precluded evaluation of longitudinal progression rates and restricted the establishment of clear temporality. The final sample size of one hundred seventy-four exposed individuals, while rigorously matched, limited statistical power to detect modest effect sizes or perform stratified analyses across distinct hypertensive classifications, such as separating mild gestational hypertension from severe preeclampsia with features of hemolysis, elevated liver enzymes, and low platelets. Furthermore, participation in the UK Biobank imaging substudy is subject to documented healthy-volunteer selection bias, as attending individuals generally exhibit more favorable cardiovascular risk profiles than the general population. Given that the cohort was predominantly composed of individuals of European ancestry and relied partly on self-reported reproductive histories, the authors conclude that larger prospective longitudinal studies incorporating repeated multi-organ assessments are necessary to elucidate true lifetime organ damage trajectories.
Subject of Research: Long-term multisystem organ damage following hypertensive disorders of pregnancy
Article Title: Long-term multisystem organ damage following hypertensive disorders of pregnancy across clinical risk subgroups: a UK Biobank study
Article References: Hmimas, M., Angarita-Fonseca, A., Sharma, A., Berry, C., Kaur, A., Zhou, Y., Mavrakanas, T., Rajah, M. N., & Pilote, L. (2026). Long-term multisystem organ damage following hypertensive disorders of pregnancy across clinical risk subgroups: a UK Biobank study. Frontiers in Cardiovascular Medicine, 13, Article 1996732. https://doi.org/10.3389/fcvm.2026.1996732
Image Credits: AI Generated
DOI: 10.3389/fcvm.2026.1996732
Keywords: Hypertensive disorders of pregnancy, Hypertension-mediated organ damage, Preeclampsia, UK Biobank, White matter hyperintensities, Left ventricular ejection fraction, Pulse wave arterial stiffness index, Urine albumin-to-creatinine ratio, Cerebral small-vessel injury, Cardiovascular risk, Endothelial dysfunction, Brain MRI, Cardiac magnetic resonance, Arterial stiffness, Microvascular disease, Gestational hypertension, Renal injury, Matched cohort study
News Source: Harold Sullivan. (November 14, 2026). Hypertensive Pregnancy Disorders Show No Consistent Long-Term Multisystem Organ Damage. Scienmag.



