About

Marie-France Hivert is an Associate Professor in the Department of Population Medicine at Harvard Medical School and an Endocrinologist in the Diabetes Unit at Massachusetts General Hospital. Her primary research focus is on the etiology and primordial prevention of obesity and related co-morbidities, particularly type 2 diabetes and gestational diabetes. Her interests include fetal metabolic programming mechanisms and the integration of genetics, epigenetics, and environmental factors contributing to obesity and related disorders. She is the Principal Investigator of the Genetics of Glucose regulation In Gestation and Growth (Gen3G) and co-Principal Investigator of Project Viva, two independent prospective pre-birth cohorts investigating health determinants of mothers and children. Dr. Hivert is involved in numerous international consortia studying the genetic determinants of glycemic regulation during and outside of pregnancy, as well as epigenetic mechanisms linked to prenatal exposures and fetal programming. At Harvard Medical School, she directs the curricular theme 'Nutrition and Lifestyle Medicine' and co-directs the 'Metabolism, Nutrition, and Lifestyle Medicine' course, contributing to education in her field.

Research topics

• Medicine
• Political Science
• Internal medicine
• Endocrinology
• Genetics
• Biology
• Physiology
• Intensive care medicine
• Bioinformatics
• Family medicine
• Pathology
• Physical therapy
• Obstetrics
• Gerontology
• Andrology
• Demography
• Sociology
• Geography
• Environmental health
• Economics
• Evolutionary biology
• Economic growth

Selected publications

• Multi-Ancestry Epigenome-Wide Meta-Analysis Identifies Novel Bulk and Cell-Type-Specific Epigenetic Markers of Asthma with Severe Exacerbations

  medRxiv · 2026-04-18

  articleOpen access

  Background: Extreme-phenotype comparisons allowed the discovery of novel asthma genetic risk loci. However, this approach remains unexplored in epigenome-wide association studies (EWAS). We aimed to identify bulk and cell-specific methylation markers of asthma with severe exacerbations across diverse ancestry groups. Methods: We conducted a meta-EWAS of 739,543 CpGs in whole blood among 1,192 African American and Latino pediatric populations, comparing non-asthmatics and asthma exacerbators. Genome-wide CpGs were followed up for replication in a meta-analysis across 1,516 ethnically diverse participants and in a cross-tissue evaluation of 393 nasal samples. We conducted differentially methylated region (DMRs), cell-type-deconvoluted, and quantitative trait loci analyses (whole-genome sequencing n=1,668; RNA-seq n=1,209). We examined enrichment in traits, pathways, and druggable genes, and analyzed DNAm predictors of plasma proteins and aging. Results: , λ=1.05). We replicated 25 CpGs in blood cells, cross-validated 7 in nasal samples, and detected 42 cell-specific DNAm markers mainly driven by T cells. DNAm at 134 CpGs was associated with gene expression in whole blood, including 118 associations with T-cell receptor genes, and 446 CpGs were regulated by > 1 genetic variant. We found enrichment for previous associations with environmental exposures, immune disorders, immune and inflammatory pathways, and druggable genes by developmental drugs. 21 methylation-predicted plasma proteins, involved in host defense, and one lung aging clock were associated with asthma exacerbations. Conclusions: The first meta-EWAS of extreme asthma phenotypes identified hundreds of novel DNAm markers, suggesting novel methylation biomarkers and candidate drugs for asthma and supporting the role of T cells.

  PublisherOA PDFDOI

• Abstract 37: Associations of Prenatal Per- and Polyfluoroalkyl Substances (PFAS) Exposures and Childhood Blood Pressure in the Environmental influences on Child Health Outcomes (ECHO) Cohort

  Circulation · 2026-03-24

  article

  Background: PFAS are widespread, persistent chemicals that cross the placenta and may affect fetal development. Animal studies suggest prenatal PFAS exposure increases offspring blood pressure (BP); however, epidemiologic evidence is limited and inconsistent. Hypothesis: We hypothesized that prenatal PFAS exposure is associated with higher childhood BP. Methods: We included children aged 3-13 years in the U.S. ECHO Cohort with data on gestational plasma or serum concentrations of PFOA, PFOS, PFHxS, PFNA, PFDA, and MeFOSAA. Children’s BP was obtained from study visits or medical records. We derived age-, sex-, and height-specific BP percentiles according to the 2017 American Academy of Pediatrics Guidelines and defined elevated BP as systolic BP (SBP) or diastolic BP (DBP) ≥90 th percentile. We used multivariable linear and modified Poisson models with generalized estimating equations and repeated BP measures clustered by participant (adjustment variables shown in Figure 1 footnotes ) and addressed missing data using multiple imputation by chained equations. We conducted subgroup analyses by child sex, ethnicity, and race, and we examined trimester-specific associations. Results: Among 2,435 children (50% female, 21% Hispanic, 25% Black), 22% had elevated BP at their last visit at a median age of 7.3 years ( Table 1 ). In adjusted analyses ( Figure 1 ), each doubling of PFOA was associated with a 1.23 (95% CI: 0.02, 2.45) higher SBP percentile, a 0.95 (0.10, 1.79) higher DBP percentile, and 1.07 (1.00, 1.15) times the risk of elevated BP. Each doubling of PFOS and PFNA was associated with a 1.14 (0.01, 2.27) and 0.85 (0.09, 1.61) higher DBP percentile, respectively. The PFOA-SBP association was stronger in males, and associations of PFOA, PFOS, and PFNA with DBP were stronger in females and non-Hispanic children. In trimester-specific analyses ( Figure 2 ), first-trimester PFOA, PFOS, and PFDA were associated with elevated BP; second-trimester PFOA and PFNA were associated with higher SBP percentiles; and third-trimester PFOS and MeFOSAA were associated with higher DBP percentiles. Conclusions: Prenatal exposure to PFOA, PFOS, and PFNA was associated with higher childhood BP. The trimester-specific associations should be interpreted with caution as they may also reflect cohort differences in PFAS profiles. PFAS may contribute to early-life programming of elevated lifelong cardiovascular risk, reinforcing the importance of population-level prevention.

  PublisherDOI

• Abstract 71: Replacing Sedentary Time by Moderate-to-Vigorous Physical Activity or Sleep on a 24h Period in Early Adolescence Is Associated with Lower Insulin Resistance Five Years Later: Prospective Analysis Within the Project Viva Cohort

  Circulation · 2026-03-24

  articleSenior author

  Introduction: Physical activity, sedentary time and sleep are determinants of adolescent cardiometabolic health. However, few studies have examined the interrelatedness of these behaviors in a 24h period, which is more applicable to the real world and could better inform public health policies. Hypothesis: Reallocation of time from sedentary behaviors to physical activity over 24h in early adolescence is associated with the most cardiometabolic health benefits in late adolescence. Methods: We used data from the ongoing Project Viva prebirth cohort (Eastern Massachusetts, USA, births 1999-2002). In early adolescence (median age 12.9 years), participants wore a wrist accelerometer for 7-10 consecutive days and completed sleep logs. We determined the average minutes spent in sleep, sedentary time, light physical activity (LPA), and moderate-to-vigorous physical activity (MVPA) using validated cut points. In late adolescence (median age 17.5 years), we collected fasting blood for adiponectin, glucose, and insulin and calculated HOMA-IR. We analyzed how proportion of time spent in each of the four activity categories in early adolescence was associated with glucose, insulin, and adiponectin in late adolescence using compositional data analysis, adjusting for participants’ age, sex, and season of actigraphy in early adolescence and maternal education and household income at enrollment. Results: A total of 802 adolescents (n=394 with outcome data) had valid accelerometer data, with 51.9% female and 63.7% non-Hispanic white (Table). The sample average 24h time composition was 33% for sleep, 48% for sedentary activities, 17% for LPA and 2% for MVPA. Allocating time towards sleep or MVPA at the expense of sedentary time in early adolescence was associated with lower HOMA-IR in late adolescence (Figure). More specifically, reallocating 30 minutes of sedentary time to MVPA or to sleep was associated with a 14.8% (95%CI: -23.3; -5.1) and 4.7% (95%CI: -7.8; -0.6) reduction in HOMA-IR, respectively, while reallocating to LPA would yield no significant change in HOMA-IR (2.9%, 95%CI: -0.9; 7.4). Our results did not support associations of activity 24h composition with adiponectin or fasting glucose. Conclusions: Allocating more time in MVPA and sleep on a 24h period in early adolescence is associated with lower insulin resistance in late adolescence. Public health strategies should focus on promoting MVPA and sleep to preserve optimal cardiometabolic health in adolescents.

  PublisherDOI

• Cardiovascular health across childhood and adolescence and proteomic biomarkers in late adolescence

  European Journal of Preventive Cardiology · 2026-03-25

  article

  AIMS: The American Heart Association's Life's Essential 8 cardiovascular health (CVH) construct strongly predicts cardiovascular disease (CVD), yet biological processes associated with early-life CVH trajectories are unclear. We examined associations of CVH score and trajectory parameters across childhood with cardiovascular-related proteins. METHODS AND RESULTS: We derived CVH scores (0-100 points) in 424 children at median ages 3.2, 7.7, 13, and 17.5y. Segmented mixed-effect models estimated three CVH trajectory parameters: timing of inflection when CVH declines and slope before and after inflection. We assayed 92 proteins from plasma samples in late adolescence (17.5y). Linear regression models assessed cross-sectional (CVH in late adolescence) and longitudinal (CVH trajectory parameters) associations with proteins, with false discovery rate correction via Benjamini-Hochberg method. Mean (SD) late-adolescent CVH was 75.5 (10.5) points. Cardiovascular health slope before inflection was 0.6 (1.4) points/y, timing of inflection was 10.1y (0.7), and slope after inflection was -1.2 (1.2) points/y. Cross-sectionally, a 1-SD higher CVH was associated with 29 differentially abundant proteins (DAPs); five showed robust associations with log2-fold change ≥|0.2| (higher FGF-21, HAOX1, IL-1ra, LEP, and lower GH). Longitudinally, 1-SD increments in CVH trajectory parameters were associated with up to 13 DAPs; specifically, we observed robust associations of timing of CVH inflection with higher LEP and CVH slope after inflection with higher IL-1ra, LEP, and SERPINA12. These DAPs implicated lipid metabolism, inflammation, and glyoxylate/oxalate production pathways. CONCLUSION: We identified novel protein biomarkers reflecting adolescent CVH and cumulative impact of childhood CVH trajectories, which may inform preventive strategies early in life to alter CVD progression.

  PublisherDOI

• Impact of ambient temperature exposure on miRNA stability in human plasma

  BioTechniques · 2025-08-03

  articleOpen access

  MicroRNAs (miRNAs) are considered more stable than mRNA, but the impact of progressive thawing of biological samples after freezing as may happen during shipping delays has not been quantified. To address this, we utilized digital PCR to estimate the absolute concentrations of select miRNAs following progressive thawing of human plasma and maintenance at ambient temperature. Specifically, we quantified let-7b-3p, miR-144-5p, miR-150-5p, miR-517a-3p, miR-524-5p, and miR-1283, which have varying abundance in plasma. We observed a trend indicating a decline in miRNA concentration as plasma samples were progressively thawed. Notably, miR-150-5p and miR-517a-3p were the least stable and were degraded by 32% and 52% respectively after 24 hours of ambient temperature storage. We found that the variation in sensitivity to temperature was not due to the GC content of the miRNAs nor their initial abundance, suggesting that other factors, such as protein interactors and vesicles carrying these miRNAs, may impact sensitivity.

  PublisherOA PDFDOI

• Discovery of placental microRNAs associated with maternal insulin sensitivity during pregnancy

  Placenta · 2025-10-30

  articleOpen accessSenior author
  PublisherOA PDFDOI

• The Nasal Microbiome and Associations With Environmental Exposures and Respiratory Health

  Allergy · 2025-10-11 · 3 citations

  articleOpen access

  ABSTRACT Background The nasal microbiome is directly in contact with the external environment and may play a role in respiratory health. This study aimed to evaluate the association of the nasal microbiome with air pollutants, meteorological conditions, and respiratory health in adolescents. Methods We analyzed the nasal microbiome in 416 adolescents from the Project Viva cohort (mean age 13 years and 52% female). We tested for the association of alpha diversity, nasotypes, and bacterial genera abundance with environmental exposures from the past 2 days to the past year (PM 2.5 , NO 2 , O 3 , temperature, humidity, residential greenness) and respiratory outcomes (asthma, hay fever, wheezing, IgE, aeroallergen sensitization, FeNO, lung function) through regression models adjusted for confounders and corrected using a false discovery rate (FDR) < 5%. Results Bacterial diversity was positively associated with hay fever and short‐term exposure to NO 2 , while it was negatively correlated with temperature (FDR < 0.05). Adolescents whose nasal microbiome was dominated by Moraxella were exposed in the past week to lower O 3 levels (ORs: 0.73–0.76) and higher temperature and humidity (ORs: 1.19–1.26). Staphylococcus dominance was positively associated with aeroallergen sensitization compared to Propionibacterium dominance (OR: 4.48, FDR = 0.03). Thirteen and eight bacterial genera abundance were associated with short‐to‐medium‐term exposures (PM 2.5 , NO 2 , temperature) and respiratory outcomes (hay fever, wheezing, IgE, FeNO, lung function) (FDR < 0.05). Staphylococcus , Corynebacterium , Pelomonas , Lactococcus , Lachnospiraceae (unclassified) , and Faecalibacterium abundance were associated with both environmental exposures and respiratory traits. Conclusions Nasal microbiome diversity was associated with hay fever, NO 2 , and temperature exposure. Multiple short‐to‐medium‐term environmental exposures and respiratory outcomes were associated with nasotypes and bacterial genera abundance in adolescents.

  PublisherOA PDFDOI

• Adolescent “Lean PCOS” Is Characterized by Higher Insulin Resistance and Adverse Adipokine Profile

  The Journal of Clinical Endocrinology & Metabolism · 2025-12-03 · 3 citations

  articleOpen accessSenior author

  OBJECTIVE: Although polycystic ovary syndrome (PCOS) is associated with high body mass index (BMI), less is known about the cardiometabolic manifestations of PCOS without excess adiposity. Among female adolescents enrolled in the Project Viva longitudinal prebirth cohort, we characterized growth, adiposity, and cardiometabolic biomarkers among those with vs without PCOS, stratified by BMI category. METHODS: We defined PCOS at the mid-teen visit (mean age 17.7 years) as self-reported diagnosis or oligo-anovulation with clinical/biochemical hyperandrogenism. We obtained anthropometric and dual x-ray absorptiometry measurements. Within each BMI category (≥85th percentile vs < 85th percentile), we used unadjusted linear regression to compare growth trajectories, adiposity, and cardiometabolic biomarkers among those with vs without PCOS. We used mixed effects models to visually represent estimated BMI and linear growth trajectories. RESULTS: Among 358 females with data at the mid-teen visit, n = 51 (14%) participants met our criteria for PCOS. Among females with BMI <85th percentile, those with PCOS (n = 27) had earlier age at peak height velocity [β = -.57 years; 95% confidence interval (CI) -0.96, -0.18], higher Homeostatic Model Assessment of Insulin Resistance (β = .77, 95% CI 0.23, 1.30), and lower adiponectin-leptin ratio (β = -.35, 95% CI -0.65, -0.06) vs without PCOS. Females with BMI ≥85th percentile had similar biomarkers by PCOS status. Adiposity measures did not differ by PCOS status within either BMI category. CONCLUSION: Within this population-based cohort, adolescents with PCOS and BMI <85th percentile had greater insulin resistance and adipose tissue dysfunction vs without PCOS. PCOS-associated metabolic dysfunction exist even among adolescents with BMI <85th percentile.

  PublisherDOI

• Associations of maternal neighborhood and trauma-related stressors with mitochondrial DNA copy number and telomere length in maternal and cord blood

  Scientific Reports · 2025-08-09 · 1 citations

  articleOpen access

  Neighborhood and individual-level trauma-related stressors during pregnancy can increase oxidative stress, potentially altering cellular disease pathway biomarkers such as mitochondrial DNA copy number (mtDNAcn) and telomere length (TL). However, the biological mechanisms linking early-life stressors to long-term health outcomes remain understudied. In a subset of Project Viva participants (n = 415-917), we evaluated associations of neighborhood and individual-level stressors with mean relative mtDNAcn and TL measured in first trimester maternal blood and cord blood. Neighborhood stressors during pregnancy were assessed using the Child Opportunity Index (COI) and Social Vulnerability Index (SVI). Trauma-related stressors were measured using the Personal Safety Questionnaire (PSQ), administered mid-pregnancy, and maternal Adverse Childhood Experiences (ACEs), reported during a mid-life follow-up. In multivariable linear regression analysis, residence in a very high versus very low opportunity neighborhood was associated with lower maternal mtDNAcn ([Formula: see text]= - 0.09, 95% confidence interval (CI) - 0.17, - 0.02), while residence in a very high versus very low vulnerability area was associated with higher maternal mtDNAcn ([Formula: see text]= 0.06, 95% CI 0.01, 0.12). Additionally, residence in moderate versus very low opportunity neighborhoods was associated with longer cord blood TL ([Formula: see text]= 0.39, 95% CI 0.0002, 0.78), but associations were attenuated after cell-type adjustment. Our findings suggest that prenatal neighborhood stressors are associated with increased maternal mtDNAcn and neighborhood opportunity is associated with longer fetal TL, indicating possible links to biological pathways related to oxidative stress and cellular aging.

  PublisherOA PDFDOI

• Examining the impact of gestational diabetes genetic susceptibility variants on maternal glucose levels during and post-pregnancy

  medRxiv · 2025-06-30

  preprintOpen access

  Abstract Aim Gestational diabetes (GDM) has important environmental and genetic components. Genetic variants associated with GDM (n=14 SNPs) were recently classified into 2 groups: those with stronger effects on type 2 diabetes than GDM (Class-T, 3 SNPs) and those with stronger effects on GDM than on type 2 diabetes (Class-G, 8 SNPs), leaving 3 SNPs unclassified. It was suggested that the Class-G variants contribute to hyperglycaemia predominantly during gestation, but it is not known whether the effects of the two variant classes on maternal glucose levels vary with pregnancy status. We aimed to compare the effects of GDM-associated variants on glucose levels (fasting glucose and 2-hour post-OGTT) measured during vs. after pregnancy in longitudinal cohorts. Methods We calculated genetic scores (GS) by class (T_GS and G_GS) and overall (All_GS) in 10,225 pregnant women and 4,763 women post-pregnancy (mean 10.5 years post pregnancy) from 8 datasets representing 4 ancestrally-diverse cohorts: EFSOCH, Gen3G, HAPO and FinnGeDi. We used linear regression models adjusted for ancestry principal components to investigate associations between standardised GS and glucose levels during or after pregnancy. Analyses were performed separately in each dataset and then combined using inverse-variance weighted random-effects meta-analyses. Results In the meta-analysis, All_GS was associated with fasting glucose both during and after pregnancy (β[95%CI], in mmol/L per 1SD higher GS = 0.06 [0.04;0.08] during vs. 0.06 [0.04;0.07] post-pregnancy). All_GS was also associated with 2-hour post-OGTT glucose levels during pregnancy but not after (0.10 [0.04; 0.15] during vs. 0.01 [-0.04; 0.07] post-pregnancy). Both G_GS and T_GS showed consistent associations with fasting glucose during and post pregnancy (0.06 [0.04; 0.08] during and 0.05 [0.03; 0.07] post pregnancy for G_GS; 0.02 [0.01; 0.02] during and 0.02 [-0.001; 0.05] post pregnancy for T_GS). G_GS showed weak evidence of association with 2-hour glucose levels during pregnancy (0.06 [-0.002; 0.11]) and no association with 2-hour glucose levels post pregnancy (-0.03 [-0.08; 0.03]). However, T_GS was associated with 2-hour glucose during pregnancy and post pregnancy (0.10 [0.04; 0.16] and 0.06 [0.01; 0.12]). Conclusion Genetic scores for GDM have consistent associations with fasting glucose levels during and after pregnancy. This finding suggests that biological pathways underlying GDM genetic susceptibility to fasting hyperglycaemia are not pregnancy specific. However, the results for All_GS and 2-hour glucose provide evidence that some genetic associations with postprandial glucose may be stronger in pregnancy and should be followed up in larger samples.

  PublisherOA PDFDOI

Recent grants

• Prenatal environmental determinants of health in young adulthood: a lifecourse approach

  NIH · $21.9M · 1998–2028

• Gestational diabetes pathophysiology uncovered by placental transcriptomics

  NIH · $3.5M · 2018–2024

Frequent coauthors

• Emily Oken

  918 shared

• Luigi Bouchard

  Centre Intégré Universitaire de Santé et de Services Sociaux du Saguenay–Lac-Saint-Jean

  877 shared

• Sheryl L. Rifas–Shiman

  Harvard University

  814 shared

• Patrice Perron

  Centre Hospitalier Universitaire de Sherbrooke

  589 shared

• Andrés Cárdenas

  Stanford University

  366 shared

• Izzuddin M. Aris

  Harvard Pilgrim Health Care

  337 shared

• José C. Florez

  Harvard University

  295 shared

• Mariona Bustamante

  Barcelona Institute for Global Health

  290 shared