About

Rosemarie de la Rosa is an Assistant Professor of Environmental Health Sciences at the School of Public Health at the University of California, Berkeley. Her research focuses on understanding how the social environment becomes biologically embedded over the life course and influences susceptibility to pollutant exposures. She is a laboratory scientist working at the intersection of environmental toxicology, epidemiology, and molecular biology, with a particular emphasis on health equity and environmental justice. Her primary research goal is to understand how the social environment and context during childhood 'gets under the skin' and affects vulnerability to the toxic effects of environmental pollutants throughout life. Current projects include collaborations with investigators at UCSF examining the relationship between adverse childhood experiences, resilience, and biomarkers of toxic stress using an allostatic load framework. She also evaluates the joint effect of psychosocial stress and air pollution on health outcomes and stress-related biomarkers in children, adolescents, and caregivers from diverse communities.

Research topics

• Sociology
• Medicine
• Emergency medicine
• Internal medicine
• Psychiatry
• Dermatology
• Pediatrics
• Gender studies
• Demography
• Immunology

Selected publications

• Arsenic exposure at birth, socioeconomic status, and epigenetic aging among adults in northern Chile

  Environmental Research · 2026-03-29

  articleOpen access

  Arsenic exposure remains a major global health concern, and early-life exposure has been linked to cancer, cardiovascular disease, and diabetes. Epigenetic biomarkers of aging may capture long-term effects of arsenic, yet whether exposure during sensitive developmental windows leaves detectable epigenetic signatures decades later remain unclear. Socioeconomic status (SES) may modify these relationships, yet its role as a modifier has not been examined. We leveraged a natural experiment in northern Chile, where municipal drinking-water arsenic concentrations were extremely high from 1958-1972. Decades later, leukocyte DNA methylation was measured among 358 adults (mean age = 65.7 years) using the Illumina EPIC v2 array. Arsenic concentration at birth (0-860μg/L) was assigned from historical water records. Current exposure was measured in urine (2-646μg/g creatinine). We evaluated associations of birth and current arsenic exposure with epigenome-wide methylation and epigenetic aging, incorporating multiplicative interaction and stratification by SES. No CpGs were linked to birth arsenic exposure, whereas nine CpGs passed the Bonferroni threshold for current urinary arsenic ( p Bonferroni <0.05). Each doubling of arsenic at birth was associated with older epigenetic age (Hannum: b = 0.11, 95% CI = 0.0027, 0.23; Zhang: 0.03, 0.00070, 0.06). SES modified associations for phenotypic age (PC-PhenoAge) and pace of aging (DunedinPACE) ( p interaction <0.05), especially among lower-SES individuals (PC-PhenoAge: b = 0.16, 95% CI = 0.02, 0.29; DunedinPACE: b = 0.0035, 95% CI = 0.00016, 0.0069). Early-life arsenic exposure was associated with accelerated epigenetic aging in adulthood, whereas current exposure was associated with CpG-specific methylation changes, particularly among socioeconomically disadvantaged individuals. • Higher arsenic at birth was linked to accelerated epigenetic aging decades later. • Epigenetic aging linked to arsenic at birth was evident in lower socioeconomic status. • Current arsenic exposure was associated with widespread CpG methylation changes. • Social disadvantage shaped epigenetic responses to arsenic across the life course.

  PublisherDOI

• Estimating the effect of reducing average annual ambient PM2.5 concentrations on allostatic load in children

  Environmental Research · 2025-12-23

  articleSenior author
  PublisherDOI

• Author response for "Unmet social needs and greater symptom burden among children with eosinophilic asthma"

  2025-01-10

  peer-review1st authorCorresponding
  PublisherDOI

• Estimating the Effect of Reducing Average Annual Ambient Pm2.5 Concentrations on Allostatic Load in Children

  SSRN Electronic Journal · 2025-01-01

  preprintOpen accessSenior author
  PublisherDOI

• Unmet social needs and greater symptom burden among children with eosinophilic asthma

  Pediatric Allergy and Immunology · 2025-02-01

  letter1st authorCorresponding

  Asthma, one of the most common diseases among US children, disproportionately affects Black, Latinx, and Indigenous children facing higher rates of prevalence, deaths, and hospitalizations than White children.1 There is robust evidence demonstrating that structural and social drivers of health, like economic and housing instability, are linked to both the development of asthma and worse outcomes, including increased symptom burden and exacerbations,2 Furthermore, interventions addressing social needs have been associated with decreased asthma-related ED visits and hospitalizations among children.3 Therefore, addressing inequities in social needs is necessary to reduce asthma disparities and improve health outcomes in historically marginalized communities. Asthma is also a heterogenous disease comprising multiple endotypes with distinct etiology, severity, and treatment approaches. The eosinophilic asthma endotype is associated with worse asthma outcomes, including increased risk for exacerbations, asthma-related hospital readmissions, and poor asthma control.4 However, less is known about the relationships between social and environmental factors, the prevalence of asthma endotypes, and asthma symptom severity. In this study, we investigated whether greater exposure to unmet social needs was associated with more severe asthma symptoms and increased risk of emergency department visits. Additionally, we explored whether these associations differed between children with and without eosinophilia. We conducted a cross-sectional analysis of 191 children with asthma from the Pediatric ACEs Screening and Resiliency Study (PEARLS) recruited during well-child visits at UCSF Benioff Children's Hospital Oakland Primary Care Clinic (UCSF BCH Oak) between March 2017 and October 2018. Study details are described in Thakur et al.5 Our study included participants who had a caregiver-reported asthma diagnosis confirmed by chart review and also blood eosinophil counts (BECs) measured as part of the study. There were 46 children (24%) with eosinophilic asthma, characterized by blood eosinophil counts (BECs) of ≥300 cells/μL, a cutoff is associated with more severe asthma.4 We also conducted a sensitivity analysis using a BEC threshold of 150 cells/μL that has been associated with greater hospitalization risk among patients with asthma.6 The number of unmet social needs in the child's household was assessed using a questionnaire completed by the caregiver that has been previously used to screen for social needs in a pediatric clinical care setting at UCSF BCH Oak.7-9 The survey covered the following social domains: food insecurity, problems paying utility bills, problems finding employment, housing instability, living in an unhealthy environment, other housing concerns, problems paying medical bills, lack of health insurance, being cut off or denied access to programs that provide income support, lacking a primary care physician, disability impairing ability to work, lack of access to mental health care for someone in the household, problems with a current or former job, and concerns about pregnancy-related work benefits. Each unmet need was coded as 0 if not experienced in the last 12 months or 1 if disclosed. Disclosures were summed to create a score (Possible Range: 0–30). Caregivers were asked to complete the International Study of Asthma and Allergies in Childhood (ISAAC) questionnaire about their child's asthma symptoms. Electronic health records were used to obtain emergency department (ED) visits and hospitalizations in the 12 months prior to recruitment (any visit vs. none). Study demographics are presented in Table 1. The mean age of participants was 7.7 years; 60.7% were male, 59.7% identified as non-Hispanic Black, 85.3% had caregivers with a high school education, and 20.4% were exposed to tobacco smoke while in utero. One third of caregivers reported that their child used either an inhaled corticosteroid (ICS) or ICS-long-acting beta-agonist (LABA) inhaler. The average number of unmet needs was 3.6, with no statistically significant differences by demographic factors (data not shown). However, we observed that median unmet needs scores were 1.6 times higher among participants with reported ICS usage compared to those without (median scores 4 vs. 2.5; p = .067). Children with eosinophilic asthma were older, had a greater proportion of reported ICS usage, and included fewer non-Hispanic Black and Hispanic participants compared to children with non-eosinophilic asthma. We then analyzed associations between unmet social needs and asthma symptom burden (Table 2). Having higher unmet social needs was associated with greater odds of reporting wheeze during exercise. We did not observe statistically significant associations with any other asthma symptom or ED visits. Additionally, we examined whether the association between unmet social needs and asthma symptom burden differed between children with eosinophilic and non-eosinophilic asthma. In analyses stratified by endotype, associations between unmet social needs and adverse asthma outcomes were generally stronger among children with eosinophilic asthma compared to those without eosinophilia (Table 2). We detected a statistically significant interaction between unmet social needs score and eosinophilic asthma status on ED visits (p = .054). Participants with eosinophilic asthma had 1.47 times the odds of having an ED visit in the past 12 months for each additional unmet social need, while no association was observed among children with non-eosinophilic asthma. Effect modification by eosinophilic asthma status on the association between unmet social needs score and ED visits remained statistically significant when using a BEC cutoff of ≥150 cells/μL (p = .068), but the magnitude of the association was attenuated (aOR: 1.15; 95% CI: 0.97, 1.36). Results from our analyses indicate that children with asthma and high BECs (at both ≥150 and 300 cells/μL) are more likely to have ED visits when they have unmet social needs, compared to children with asthma and lower BECs. Our study found that higher reports of unmet needs were associated with an increased risk of exercise-induced wheeze among children with asthma. However, we did not observe associations with any other asthma symptom or ED visits. Our findings align with a similar study conducted among asthmatic children from an urban population, where having 3+ unmet social needs was associated with 59% greater odds of persistent severe symptoms.10 Having a greater number of unmet social needs was associated with worse symptom burden and ED visits among children with eosinophilia compared to those with non-eosinophilic asthma. This endotype is characterized by inflammation which might be exacerbated by the activation of stress-response pathways due to unmet social needs, leading to more severe asthma symptoms. Therefore, children with eosinophilic asthma may be more sensitive to social stressors. Interestingly, only 24% of children in our study had eosinophilic asthma, compared to the prevalence of 57% reported among children in NHANES.11 The higher proportion of children with non-eosinophilic asthma in this study population may reflect structural inequities that increase exposure to factors linked to nonatopic asthma, such as household mold, environmental pollutants, and psychosocial stress.12, 13 Future research is needed to identify factors that drive the prevalence of asthma endotypes in this population. Furthermore, practitioners should adhere to the American Academy of Pediatrics' recommendation to screen for unmet social needs to optimize care and reduce symptom burden, especially for children with allergic asthma. A limitation of our study is that we used BECs alone to characterize the Th2-high asthma phenotype since associated cytokines (IL-4, IL-5, and IL-13) were not measured in this study population. IgE concentrations was also not quantified, limiting our ability to identify children with atopic asthma. However, a nationally representative study found that 81% of children with eosinophilic asthma (BEC ≥300 cells/μL) also had atopic asthma, suggesting substantial overlap between these phenotypes.11 Therefore, our findings may be generalizable to other asthma phenotypes and warrant further exploration. A key strength of our study was the recruitment from a Federally Qualified Health Center, where the majority of participants were Black and Latinx, with 40% reporting asthma. However, this recruitment approach may have affected the generalizability of our results. The number of social need in this population was comparable to averages reported in other studies conducted at BCH Oak and at two safety net hospitals in San Francisco.8 Lastly, the cross-sectional study design limits our ability to establish a causal relationship between unmet needs and asthma symptoms. Collectively, these findings highlight that unmet social needs may be an important social determinant of acute asthma symptoms among children and suggest potential susceptibility to social stress based on asthma endotype. Rosemarie de la Rosa: Conceptualization; writing – original draft; writing – review and editing; formal analysis; visualization; investigation. Adali Martinez: Writing – review and editing; methodology. Morgan Ye: Writing – review and editing; data curation. Danielle Hessler: Writing – review and editing; data curation; methodology; funding acquisition. Kadiatou Koita: Writing – review and editing; methodology. Monica Bucci: Writing – review and editing; methodology. Dayna Long: Writing – review and editing; methodology; resources; funding acquisition. Neeta Thakur: Conceptualization; writing – original draft; methodology; resources; funding acquisition. We would like to thank our families for sharing and trusting us with their personal experiences and sensitive information. We would also like to acknowledge the contributions of our study coordinators—Nitasha Sharma, Cherri Harris, Roberto Mok, and Nai Pharn—without whom this study would not be possible. This work was supported in part by the TARA Health Foundation, Genentech Corporate Giving, and the California Initiative to Advance Precision Medicine; P0551412. RD was supported by the University of California President's Postdoctoral Fellowship. The authors have no conflicts to declare. The study was approved by the UCSF Benioff Oakland institutional review board (Protocol# 21-33697). Informed consent was obtained from legally authorized representatives and/or guardians of all participants. The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/pai.70034.

  PublisherDOI

• Long-Term Effect of Early-Life Arsenic Exposure on Morning Plasma Cortisol in Adults from Antofagasta, Chile

  medRxiv · 2025-09-02

  preprintOpen access1st authorCorresponding

  Over 100,000 people were exposed to arsenic-contaminated drinking water in Antofagasta, Chile from 1958-1970. Individuals born during this high exposure period have elevated rates of cancer, lung and cardiovascular disease, and hypertension. However, the mechanisms of long-term arsenic toxicity remain unclear. We investigated whether early-life arsenic exposure was associated with altered glucocorticoid levels in adulthood. This study included 114 individuals born in Antofagasta during the high exposure period and 118 individuals born elsewhere. Arsenic exposure metrics were constructed based on residential histories and included: concentration at birth, peak and highest 5-year average between ages 0-10 years, and highest lifetime 5-year average, and lifetime cumulative exposure. Morning plasma cortisol concentrations were measured using a cell-based bioassay. Individuals in the highest quartile of highest lifetime 5-year average of arsenic exposure had approximately 11% lower mean log cortisol levels than those in the lowest quartile of exposure (β = -0.116; 95% CI: -0.229, -0.003). In sex-stratified analyses, associations were stronger among females. For example, females in the highest quartile of cumulative exposure had 22.0% lower cortisol levels compared to those in the lowest quartile (β = -0.248; 95% CI: -0.444, -0.053) and the test for interaction by sex was statistically significant (p = 0.036). This study is the first to show that early-life arsenic exposure may have lasting effects on cortisol. These findings highlight endocrine disruption as a mechanism contributing to long-term health effects of early arsenic exposure.

  PublisherOA PDFDOI

• Adverse Childhood Experiences and Related Events are Associated with Asthma Symptoms in Children

  Academic Pediatrics · 2024-01-19 · 8 citations

  article
  PublisherDOI

• Associations Between Early-Life Adversity, Ambient Air Pollution, and Telomere Length in Children

  Psychosomatic Medicine · 2024-04-08 · 4 citations

  articleOpen access1st authorCorresponding

  OBJECTIVE: Examine the independent associations and interaction between early-life adversity and residential ambient air pollution exposure on relative buccal telomere length (rBTL). METHODS: Experiences of abuse, neglect, household challenges, and related life events were identified in a cross-sectional sample of children aged 1 to 11 years ( n = 197) using the 17-item Pediatric ACEs and Related Life Event Screener (PEARLS) tool. The PEARLS tool was analyzed both as a total score and across established domains (Maltreatment, Household Challenges, and Social Context). Ground-level fine particulate matter (PM 2.5 ) concentrations were matched to residential locations for the 1 and 12 months before biospecimen collection. We used multivariable linear regression models to examine for independent associations between continuous PM 2.5 exposure and PEARLS score/domains with rBTL. In addition, effect modification by PEARLS scores and domains on associations between PM 2.5 exposure and rBTL was examined. RESULTS: Study participants were 47% girls, with mean (standard deviation) age of 5.9 (3.4) years, median reported PEARLS score of 2 (interquartile range [IQR], 4), median 12-month prior PM 2.5 concentrations of 11.8 μg/m 3 (IQR, 2.7 μg/m 3 ), median 1-month prior PM 2.5 concentrations of 10.9 μg/m 3 (IQR, 5.8 μg/m 3 ), and rBTL of 0.1 (IQR, 0.03). Mean 12-month prior PM 2.5 exposure was inversely associated with rBTL ( β = -0.02, 95% confidence interval = -0.04 to -0.01). Although reported PEARLS scores and domains were not independently associated with rBTL, we observed a greater decrement in rBTL with increment of average annual PM 2.5 as reported Social Context domain items increased ( p -interaction < .05). CONCLUSIONS: Our results suggest that adverse Social Context factors may accelerate the association between chronic PM 2.5 exposure on telomere shortening during childhood.

  PublisherOA PDFDOI

• Biological Burden of Adverse Childhood Experiences in Children

  Psychosomatic Medicine · 2023 · 16 citations

  1st authorCorresponding
  • Medicine
  • Demography
  • Pediatrics

  OBJECTIVE: This study aimed to examine relationships between adverse childhood experiences (ACEs) and related life events and allostatic load (AL)-"wear and tear" from chronic stress-in a pediatric population. METHODS: Children were screened with the PEdiatric ACEs and Related Life Event Screener (PEARLS) tool, a 17-item questionnaire capturing experiences of abuse, neglect, household challenges, and related life events. Biological data were available for 207 participants, and AL was operationalized using clinical or empirical cutoff points across 4 physiological systems (i.e., cardiac, metabolic, inflammatory, neurologic). Covariate-adjusted multivariable regression models were used to examine associations between AL with adversity and health. RESULTS: Children (mean age = 6.5 years, range = 1-11 years) had an average AL score of 1.9 (standard deviation = 1.7), and a U-shaped relationship was observed with child's age. Continuous PEARLS and original ACE scores were not associated with AL. However, children with a reported PEARLS score of 1 to 2 or original ACEs score of 1 to 3 had 1.5 (incidence rate ratio [IRR] = 1.50, 95% confidence interval [CI] = 1.09-2.08) and 1.4 (IRR = 1.41, 95% CI = 1.08-1.84) times greater AL, respectively, compared with participants with none reported. In secondary analyses, caregiver mental illness was associated with higher child AL (adjusted IRR = 1.27, 95% CI = 1.01-1.58). AL was also associated with poorer perceived child general health (adjusted β = -0.87, 95% CI = -1.58 to -0.15) and greater odds of child obesity (adjusted odds ratio = 1.51, 95% CI = 1.23-1.89). CONCLUSIONS: Measuring AL in a pediatric population requires careful consideration of age. Higher AL was associated with a greater number of reported adversities and worse child health.

  PublisherOA PDFDOI

• Abstract P171: Associations Between Metals and Allostatic Load Among Non-Hispanic Asians and Non-Hispanic Whites From the National Health and Nutrition Examination Survey: 2015-2020

  Circulation · 2023-02-28

  articleSenior author

  Introduction: Environmental metal exposure has been linked with multi-system toxicity and elevated stress hormones, and high metal concentrations have been observed among Asian Americans. Metal exposure may also increase Allostatic Load (AL), or physiologic “wear-and-tear” on the body. Hypothesis: Metal biomarker concentration is associated with high AL among non-Hispanic Asian (NHA) and non-Hispanic White (NHW) adults. Methods: AL scores were calculated based on 9 biomarkers representing cardiovascular, metabolic, and immune system function using data from the National Health and Nutrition Examination Survey (2015-2020). High AL was defined as having ≥3 high-risk biomarkers based on empirically defined thresholds. Lead, cadmium, and mercury were quantified in blood while arsenic, monomethylarsonic acid (MMA), and dimethylarsinic acid (DMA) were measured in urine. Metal concentrations were categorized into empirically-defined sample tertiles. Survey-weighted logistic regression analyses controlled for demographic, socioeconomic, and health behavior variables. Analyses were also stratified by NHA and NHW. Results: The study sample consisted of 5,012 adults [mean age: 50.4, SE: 0.5] (21.6% foreign-born NHA, 2.7% US-born NHA, and 75.7% NHW). Prevalence of high AL was 38.2% among NHW, 32.4% among foreign-born NHA, and 19.3% among US-born NHA. Metal exposure demonstrated heterogenous relationships with AL ( Table ). Individuals in the highest tertile of cadmium exposure had 38% greater odds of having high AL than those in the lowest tertile, however this association was attenuated after controlling for socioeconomic status and health behaviors. Lead, mercury, arsenic, DMA, and MMA were associated with lower odds of high AL. Relationships were similar when stratified by NHW and NHA. Conclusion: The association between metals and AL differs by metal type in our study population. Future studies should explore the mechanistic pathways between high metal exposure and AL.

  PublisherDOI

Frequent coauthors

• Neeta Thakur

  University of California, San Francisco

  11 shared

• Kadiatou Koita

  Université des Sciences, des Techniques et des Technologies de Bamako

  9 shared

• Morgan Ye

  University of California, San Francisco

  9 shared

• Dayna Long

  University of California, San Francisco

  9 shared

• Monica Bucci

  Youth Development

  9 shared

• Danielle Hessler

  King's College London

  9 shared

• Nicole R. Bush

  Cohort (United Kingdom)

  8 shared

• Austin Le

  Illinois College

  6 shared