<a href=”https://doi.org/10.1177/26335565261471615″>Associations of chronic disease and multimorbidity with healthcare utilization, mortality and adverse events in Estonia</a>
<em>Journal of Multimorbidity and Comorbidity</em>, 16 (2026)
DOI: 10.1177/26335565261471615

<strong>Authors
</strong>Angela Kannukene, Kaja Põlluste, Margus Lember, Carola Orrego, Jaak Vilo, Raivo Kolde, Sven Laur, Sulev Reisberg

<strong>Abstract</strong>
This population-based study examined chronic disease and multimorbidity in Estonia using linked administrative and clinical registry data from a 30% sample of the adult population. Chronic disease affected 60.3% of adults and multimorbidity, defined as two or more chronic conditions, affected 42.0%. Multimorbidity was strongly associated with increased mortality, healthcare use and healthcare-associated adverse events. Each additional chronic condition was linked to higher rates of inpatient, outpatient and emergency visits, as well as prescription use. The findings demonstrate the substantial population-level burden of multimorbidity and support the use of linked real-world health data for healthcare planning and patient-safety research.

<a href=”https://doi.org/10.1186/s12872-026-06447-8″>Heterogeneity in diagnosing heart failure – real world data from electronic health records</a>
<em>BMC Cardiovascular Disorders</em> (2026)
DOI: 10.1186/s12872-026-06447-8

<strong>Authors
</strong>Liis Roine, Teele Kasepalu, Helene Loorents, Märt Elmet, Alar Irs, Mai Blöndal, Marek Oja, Sulev Reisberg, Raivo Kolde, Jaak Vilo, Kerli Mooses

<strong>Abstract</strong>
This study examined diagnostic practices among newly diagnosed heart failure patients in Estonia using electronic health records and healthcare claims data. The analysis included a random 10% sample of the Estonian population between 2012 and 2019 and focused on echocardiography and natriuretic peptide testing. Within one year before and after diagnosis, 31.0% of patients received both recommended diagnostic tests, while 40.7% received neither. Patients diagnosed in emergency departments were most likely to undergo at least one heart-failure test. The findings show substantial variation and gaps in routine heart-failure diagnosis and highlight opportunities to improve diagnostic pathways and patient care.

<a href=”https://doi.org/10.1038/s41467-025-61661-w”>Resolving out of Africa event for Papua New Guinean population using neural network</a>
<em>Nature Communications</em>, 16, 6345 (2025)
DOI: 10.1038/s41467-025-61661-w

<strong>Authors
</strong>Mayukh Mondal, Mathilde André, Ajai K. Pathak, Nicolas Brucato, François-Xavier Ricaut, Mait Metspalu, Anders Eriksson

<strong>Abstract</strong>
This study re-examines the demographic history of the Papua New Guinean population, which has a distinctive history of early settlement, relative isolation and Denisovan ancestry. Using newly available samples and neural-network-based demographic modelling, the researchers found that patterns previously interpreted as evidence of admixture with an earlier out-of-Africa population are better explained by a strong population bottleneck and slower population growth. The analyses place Papua New Guinean populations as a sister group to other East Asian populations. The findings demonstrate the importance of accounting for population-specific demographic history when interpreting genetic patterns.

<a href=”https://doi.org/10.1038/s41562-026-02556-8″>Parental haplotype reconstruction in up to 440,209 UK and Estonian individuals reveals recent assortative mating dynamics</a>
<em>Nature Human Behaviour</em> (2026)
DOI: 10.1038/s41562-026-02556-8

<strong>Authors
</strong>Robin J. Hofmeister, Davide Marnetto, Théo Cavinato, Leona Knüsel, Giorgia Modenini, Erwan Pennarun, Estonian Biobank Research Team, Andres Metspalu, Lili Milani, Tõnu Esko, Mari Nelis, Georgi Hudjashov, Nicola Barban, Reedik Mägi, Tabea Schoeler, Luca Pagani, Zoltán Kutalik

<strong>Abstract</strong>
This study introduces a method for reconstructing parental haplotypes from close relatives and estimating recent genetic assortative mating without requiring genotyped couples. The approach was applied to 69 traits in 245,884 UK Biobank and 194,325 Estonian Biobank participants. It identified genetic assortative mating for traits including height, educational attainment, overall health and television viewing. The results indicate that assortative mating for height and education has accelerated in more recent generations. The method provides a way to study how partner choice contributes to the genetic architecture of human traits.

<a href=”https://doi.org/10.1093/eurheartj/ehag575″>Side effects in hypertension treatment: a pharmacogenomic analysis</a>
<em>European Heart Journal</em>, ehag575 (2026)
DOI: 10.1093/eurheartj/ehag575

<strong>Authors
</strong>Felix Vaura, Kristi Krebs, Tuomo Kiiskinen, Joel Rämö, Max Tamlander, Estonian Biobank research team, Simone Rubinacci, Lili Milani, Samuli Ripatti

<strong>Abstract</strong>
This pharmacogenomic study investigated genetic predictors of switching or discontinuing antihypertensive medications during the first year of treatment. Longitudinal medication data from more than 400,000 genotyped users in FinnGen, UK Biobank and the Estonian Biobank were analysed. The researchers identified 14 genome-wide significant loci associated with switching from angiotensin-converting enzyme inhibitors and dihydropyridine calcium channel blockers. The findings implicate the neurotensin-NTSR1 pathway in ACE inhibitor–induced cough and identify CYP3A4*22 as a potential predictor of calcium channel blocker switching. The results support the potential of genotype-guided prescribing for more personalised hypertension treatment.

<a href=”https://doi.org/10.2196/86183″>Extracting and Classifying Drug Discontinuations From Estonian Electronic Health Records: Development and Validation Study</a>
<em>Journal of Medical Internet Research</em>, 28, e86183 (2026)
DOI: 10.2196/86183

<strong>Authors
</strong>Hendrik Šuvalov, Nikita Umov, Maria Malk, Markus Haug, Sven Laur, Marek Oja, Sirli Tamm, Sulev Reisberg, Jaak Vilo, Raivo Kolde

<strong>Abstract</strong>
This study evaluated the use of large language models to extract and classify reasons for medication discontinuation from Estonian electronic health records. Prescription data and free-text clinical notes from a 10% sample of the Estonian population were used to analyse discontinuations of statins and antidiabetic medications. The models accurately extracted discontinuation-related phrases and reasons, with adverse reactions identified as the most frequent cause of discontinuation. The approach also distinguished whether treatment was stopped by the patient or clinician. The findings show that large language models can unlock clinically valuable information from unstructured health records, supporting scalable analyses of treatment patterns and adverse drug events.

<a href=”https://doi.org/10.2196/83465″>A Comprehensive Approach to Days’ Supply Estimation in a Real-World Prescription Database: Algorithm Development and Validation Study</a>
<em>Online Journal of Public Health Informatics</em>, 18, e83465 (2026)
DOI: 10.2196/83465

<strong>Authors
</strong>Maria Malk, Kerli Mooses, Marek Oja, Johannes Holm, Hanna Keidong, Nikita Umov, Sirli Tamm, Sulev Reisberg, Jaak Vilo, Raivo Kolde

<strong>Abstract</strong>
Accurate days’ supply information is essential for medication-use statistics and medication-adherence calculations, but it is often missing or incorrect in prescription data. This study developed and validated a workflow for cleaning prescription records, imputing daily doses and calculating days’ supply. It used prescription data from a random 10% sample of the Estonian population, comprising 150,824 patients and records from 2012 to 2019. By combining data-driven methods with medication-use guidelines, the approach determined days’ supply for 98.3% of dispensed prescriptions. The results show that a carefully designed imputation pipeline can improve prescription-data quality and yield more accurate, consistent medication-adherence measures.

<a href=”https://doi.org/10.1016/j.imu.2026.101806″>CohortContrast: Enrichment-based methodology for identification of clinically relevant concepts in real-world data</a>
<em>Informatics in Medicine Unlocked</em>, 65, 101806 (2026)
DOI: 10.1016/j.imu.2026.101806

<strong>Authors
</strong>Markus Haug, Neeme Ilves, Nikita Umov, Hendrik Šuvalov, Kerli Mooses, Sirli Tamm, Helene Loorents, Marek Oja, Sulev Reisberg, Jaak Vilo, Raivo Kolde

<strong>Abstract</strong>
CohortContrast is an open-source, OMOP-compatible methodology for identifying clinically relevant concepts in real-world health data. It compares target and control cohorts through statistical enrichment analysis, reduces temporal and semantic noise, and aggregates related concepts to support cohort characterisation and treatment-trajectory analysis. The approach was demonstrated using nationwide Estonian observational health data, including lung and prostate cancer cohorts. CohortContrast substantially reduced the number of concepts requiring manual review while retaining clinically interpretable treatment patterns. The method supports scalable, reproducible cohort analysis and real-world evidence research.

<a href=”https://doi.org/10.1093/ije/dyag113″>Data Resource Profile: EST-Health-30</a>
<em>International Journal of Epidemiology</em>, 55, dyag113 (2026)
DOI: 10.1093/ije/dyag113

<strong>Authors
</strong>Sulev Reisberg, Marek Oja, Kerli Mooses, Sirli Tamm, Ami Sild, Harry-Anton Talvik, Sven Laur, Raivo Kolde, Jaak Vilo

<strong>Abstract</strong>
EST-Health-30 is a large-scale, population-representative health-data resource based on a random 30% sample of the Estonian population. It contains linked longitudinal data for 509,858 individuals from 2012 onwards, integrating electronic health records, health-insurance claims, prescription data, cancer-registry data and causes-of-death records. The data are pseudonymised and harmonised to the OMOP Common Data Model using international standard vocabularies. The resource includes diagnoses, procedures, laboratory measurements, prescriptions, clinical notes, healthcare utilisation and costs. EST-Health-30 supports epidemiological, clinical and methodological research, including analyses of disease trajectories and healthcare outcomes, while enabling secure participation in international federated research networks.

<a href=”https://doi.org/10.1186/s12911-026-03790-w”>Syrona: an open-source visual analytics tool for pairwise comparison of health datasets in OMOP CDM</a>
<em>BMC Medical Informatics and Decision Making</em> (2026)
DOI: 10.1186/s12911-026-03790-w

<strong>Authors
</strong>Maarja Pajusalu, Marek Oja, Kerli Mooses, Silver Heinsar, Triin Laisk, G. Grents, Taavi Tillmann, Sven Laur, Sulev Reisberg, Jaak Vilo, Raivo Kolde

<strong>Abstract</strong>
Syrona is an open-source visual analytics tool for systematic pairwise comparison of health datasets and subcohorts represented in the OMOP Common Data Model. The tool extracts annual prevalence data for conditions, procedures and drugs, calculates prevalence ratios across demographic groups, and synthesises results using multilevel meta-analysis. Its interactive dashboard combines distributional overviews, stratified heatmaps, forest plots and absolute-prevalence comparisons. Three case studies using Estonian national health data from 495,000 people, collected between 2012 and 2024, demonstrate its use for assessing dataset representativeness, comparing institutional practices and identifying coding artefacts. Syrona can be applied to any OMOP CDM database and supports systematic assessment of datasets in distributed observational research.

<a href=”https://doi.org/10.1093/eurheartj/ehag400″>Predicting lifetime cardiovascular risk and benefits of preventive treatment in patients with established atherosclerotic cardiovascular disease: the SMART-REACH2 model</a>
<em>European Heart Journal</em>, 47, 5840–5856 (2026)
DOI: 10.1093/eurheartj/ehag400

<strong>Authors
</strong>Joris Holtrop, Mari N. Gynnild, Salil V. Deo, Carl-Emil Lim, Stelios Boulitsakis Logothetis, Kristi Läll, Minna Oksanen, Håvard Dalen, John Munkhaugen, Peter Ueda, Tomas Jernberg, José Tuñón, Ignacio Mahíllo Fernández, Ingvild Saltvedt, Hanne Ellekjær, Martin Teraa, Manon G. van der Meer, Ynte M. Ruigrok, Urmo Võsa, Małgorzata Chlabicz, Karol Kaminski, Ph. Gabriel Steg, Ivana Burazor, Reedik Mägi, Taavi Tillmann, David A. Wood, Kausik K. Ray, Dirk de Bacquer, Janne Martikainen, John William McEvoy, Stephen Kaptoge, Frank L. J. Visseren, Emanuele Di Angelantonio, Deepak L. Bhatt, Angela Wood, Steven H. J. Hageman, Jannick A. N. Dorresteijn

<strong>Abstract</strong>
SMART-REACH2 is an updated risk-prediction model for people with established atherosclerotic cardiovascular disease. It estimates 10-year and lifetime risk of recurrent cardiovascular events, cardiovascular disease-free life expectancy, and the potential benefit of preventive treatment. The model accounts for sex-specific and geographical variation and was externally validated in more than 2 million patients from 54 countries. It supports personalised preventive care and shared treatment decisions.

The Estonian Biobank’s Journey from Biobanking to Personalized Medicine.
Nature Communications, 16, 3270 (2025)
DOI: 10.1038/s41467-025-58465-3

Authors
Lili Milani, Maris Alver, Sven Laur, Sulev Reisberg, Toomas Haller, Oliver Aasmets, Erik Abner, Helene Alavere, Annely Allik, Tarmo Annilo, Krista Fischer, Robin Hofmeister, Georgi Hudjashov, Maarja Jõeloo, Mart Kals, Liis Karo-Astover, Silva Kasela, Anastassia Kolde, Kristi Krebs, Kertu Liis Krigul, Jaanika Kronberg, Karoliina Kruusmaa, Viktorija Kukuškina, Kadri Kõiv, Kelli Lehto, Liis Leitsalu, Sirje Lind, Laura Birgit Luitva, Kristi Läll, Kreete Lüll, Kristjan Metsalu, Mait Metspalu, René Mõttus, Mari Nelis, Tiit Nikopensius, Miriam Nurm, Margit Nõukas, Marek Oja, Elin Org, Marili Palover, Priit Palta, Vasili Pankratov, Kateryna Pantiukh, Natalia Pervjakova, Natàlia Pujol-Gualdo, Anu Reigo, Ene Reimann, Steven Smit, Diana Rogozina, Dage Särg, Nele Taba, Harry-Anton Talvik, Maris Teder-Laving, Neeme Tõnisson, Mariliis Vaht, Uku Vainik, Urmo Võsa, Burak Yelmen, Tõnu Esko, Raivo Kolde, Reedik Mägi, Jaak Vilo, Triin Laisk & Andres Metspalu

Abstract
Large biobanks have set a new standard for research and innovation in human genomics and implementation of personalized medicine. The Estonian Biobank was founded a quarter of a century ago, and its biological specimens, clinical, health, omics, and lifestyle data have been included in over 800 publications to date. What makes the biobank unique internationally is its translational focus, with active efforts to conduct clinical studies based on genetic findings, and to explore the effects of return of results on participants. In this review, we provide an overview of the Estonian Biobank, highlight its strengths for studying the effects of genetic variation and quantitative phenotypes on health-related traits, development of methods and frameworks for bringing genomics into the clinic, and its role as a driving force for implementing personalized medicine on a national level and beyond.

Pathway level metabolomics analysis identifies carbon metabolism as a key factor of incident hypertension in the Estonian Biobank
Scientific Reports 15, 8470 (2025)
DOI: 10.1038/s41598-025-92840-w

Authors
Liis Hiie, Anastassia Kolde, Natalia Pervjakova, Anu Reigo, Estonian Biobank Research Team, Erik Abner, Urmo Võsa, Tõnu Esko, Krista Fischer, Priit Palta & Jaanika Kronberg

Abstract
The purpose of this study was to find metabolic changes associated with incident hypertension in the volunteer-based Estonian Biobank. We used a subcohort of the Estonian Biobank where metabolite levels had been measured by mass-spectrometry (LC-MS, Metabolon platform). We divided annotated metabolites of 989 individuals into KEGG pathways, followed by principal component analysis of metabolites in each pathway, resulting in a dataset of 91 pathway components. Next, we defined incident hypertension cases and controls based on electronic health records, resulting in a dataset of 101 incident hypertension cases and 450 controls. We used Cox proportional hazards models and replicated the results in a separate cohort of the Estonian Biobank, assayed with LC-MS dataset of the Broad platform and including 582 individuals. Our results show that body mass index and a component of the carbon metabolism KEGG pathway are associated with incident hypertension in both discovery and replication cohorts. We demonstrate that a high-dimensional dataset can be meaningfully reduced into informative pathway components that can subsequently be analysed in an interpretable way, and replicated in a metabolomics dataset from a different platform.

Exploiting uniqueness: seed-chain-extend alignment on elastic founder graphs
Bioinformatics, 41 (2025)
DOI: 10.1093/bioinformatics/btaf225

Authors
Nicola Rizzo, Manuel Cáceres & Veli Mäkinen

Abstract
Sequence-to-graph alignment is a central challenge of computational pangenomics. To overcome the theoretical hardness of the problem, state-of-the-art tools use seed-and-extend or seed-chain-extend heuristics to alignment. We implement a complete seed-chain-extend alignment workflow based on indexable elastic founder graphs (iEFGs) that support linear-time exact searches unlike general graphs. We show how to construct iEFGs, find high-quality seeds, chain, and extend them at the scale of a telomere-to-telomere assembled human chromosome.

Development and validation of a risk prediction algorithm for high-risk populations combining genetic and conventional risk factors of cardiovascular disease

PLOS One, 20 (10) (2025)
DOI: 10.1371/journal.pone.0335064

Authors

Tuuli Puusepp, Ave Põld, Lili Milani, Aet Elken, Estonian Biobank Research Team, Mikk Jürisson & Krista Fischer

Abstract

Aim: To develop a model for cardiovascular disease (CVD) risk, combining polygenic risk score (PRS) with traditional risk factors while assessing the added value of PRS in two cohorts of biobank participants.

Methods: Data of 128 209 participants from the Estonian Biobank recruited between 2002–2017 and 2018–2022 without prevalent cardiovascular disease, was included. Hazard ratios (HR) for polygenic risk versus conventional risk factors were estimated with Cox proportional hazards models, cumulative incidence was assessed with Aalen-Johansen curves. Predictive performance was tested using a split-sample approach and competing risk modelling. Age at CVD event served as the outcome, and the impact of the PRS was evaluated by age group (25–59 vs. 60+), sex, and recruitment period, using HRs, Harrell’s C-index, and net reclassification indices (NRI).

Results: The estimated HR per one standard deviation (SD) of PRS ranged from 1.1, 95% CI 1.06–1.15 (age 60 + , earlier cohort) to 1.36, 95% CI 1.24–1.49 (men 25–59, later cohort). Adding PRS to the conventional risk factors in the age group 25–59 increased the C-statistic by 0.028 (p < 0.0001) for men. In the age group 60 + , the increase was 0.016 (p = 0.0002) across all. In the independent validation set, the continuous NRI was 19.1% (95% CI 13.3%–24.9%) in the 25–59 group and 13.9% (95% CI 8.1%–19.6%) in the 60 + group.

Conclusions: In a high-risk population, PRS is a strong independent risk factor for CVD and should be considered in routine risk assessment, starting at a relatively young age.

Analysis of follow-up data in large biobank cohorts: a review of methodology
Frontiers in Genetics, 16, 1534726 (2025)
DOI: 10.3389/fgene.2025.1534726

Authors
Anastassia Kolde, Merli Koitmäe, Meelis Käärik, Märt Möls & Krista Fischer

Abstract
This study focuses on key methodological challenges in genome-wide association studies (GWAS) of biobank data with time-to-event outcomes, analyzed using the Cox proportional hazards (CPH) model. We address four primary issues: left-truncation of the data, computational inefficiency of standard model-fitting algorithms, relatedness among individuals, and model misspecification. To manage left-truncation, the common practice is to use age as the timescale, with individuals entering the risk set at their age of recruitment. We assess how this choice of timescale influences bias and statistical power, under realistic GWAS conditions of varying effect sizes and censoring rates. In addition, to alleviate the computational burden typical in large-scale data, we propose and evaluate a two-step martingale residual (MR) approach for high-dimensional CPH modeling. Our results show that the timescale choice has minimal effect on accuracy for small hazard ratios, though using time since birth as the timescale – ignoring recruitment age – yields the highest power for association detection. We find that relatedness, when ignored, does not substantially bias effect size estimates, while omitting key covariates introduces significant bias. The two-step MR approach proves to be computationally efficient, retaining power for detecting small effect sizes, making it suitable for large-scale association studies. However, when precise effect size estimates are critical, particularly for moderate or larger effect sizes, we recommend recalculating these estimates using the conventional CPH model, with careful attention to left-truncation and relatedness. These conclusions are drawn from simulations and illustrated with data from the Estonian Biobank cohort.

Characterization of prevalent genetic variants in the Estonian Biobank body-mass index GWAS
Nature Communications, 16, 8956 (2025)
DOI: 10.1038/s41467-025-64006-9

Authors
Erik Abner, Kanwal Batool, Nele Taba, Tiit Nikopensius, Kristi Läll, Anastasiia Alekseienko, Anders Eriksson, Joel Rämö, Hele Haapaniemi, Hanna Maria Kariis, Liis Haljasmägi, Urmo Võsa, Taavi Tillmann, Uku Vainik, Kelli Lehto, Hanna M. Ollila, Kai Kisand, Estonian Biobank Research Team & Tõnu Esko

Abstract
Population-specific genome-wide association studies can reveal high-impact genomic variants that influence traits like body-mass index (BMI). Using the
Estonian Biobank BMI dataset (n = 204,747 participants) we identified 214 genome-wide significant loci. Among those hits, we identified a common noncoding variant within the newly associated ADGRL3 gene (−0.18 kg/m²; P = 3.21 × 10⁻⁹). Moreover, the missense rare variant PTPRT:p.Arg1384H is
associated with lower BMI (−0.44 kg/m²; P = 2.51 × 10⁻¹⁰), while the proteintruncating variant POMC:p.Glu206* was associated with considerably higher
BMI (+ 0.81 kg/m²; P = 1.48 × 10−12), both likely affecting the functioning of the leptin-melanocortin pathway. POMC:p.Glu206* was observed in different North-European populations, suggesting a broader, yet elusive, distribution of this damaging variant. These observations indicate the previously unrecognized roles of the ADGRL3 and PTPRT genes in body weight regulation and suggest an increased prevalence of the POMC:p.Glu206* variant in European populations, offering avenues for developing interventions in obesity management.

Diverse landscape of genomic research within the Estonian Biobank
Human Molecular Genetics, 1–6 (2025)
DOI: 10.1093/hmg/ddaf026

Authors
Reedik Mägi

Abstract
The Estonian Biobank (EstBB) is a national biobank hosted by the Institute of Genomics at the University of Tartu. Established in 2000, it is one of the largest population-based biobanks in the world, with the sample size exceeding 212 000 individuals
(including more than 74 000 parent–child pairs and 38 000 sibpairs) and thus representing more than 20% of the adult population in Estonia. The biobank collects and stores health and lifestyle data, biological samples (DNA, blood plasma, buffy coat etc.), and medical information from national health databases on a large segment of the Estonian population to facilitate scientific research and the development of personalized medicine. This extensive coverage provides a population-representative sample of the Estonian population and additionally enables various analyses, including genome-wide association studies and studies involving related individuals or families.

Genetic effects on migration behavior contribute to increasing spatial differentiation at trait-associated loci in Estonia
iScience, 28, 12 (2025)
DOI: 10.1016/j.isci.2025.114013

Authors
Ivan A. Kuznetsov, Estonian Biobank Research Team (Andres Metspalu, Lili Milani, Tõnu Esko, Reedik Mägi, Mari Nelis, Georgi Hudjashov), Mait Metspalu, Uku Vainik, Luca Pagani, Francesco Montinaro & Vasili Pankratov

Abstract
Emerging evidence suggests that migration behavior can be selective with respect to individuals’ genotypes, producing genotype-environment correlations that standard methods used in genetic association studies cannot correct. We investigate this phenomenon by examining the spatial dynamics of polygenic scores (PGSs) in Estonia. Our analyses show that contemporary migrations intensify inter-regional differences in PGSs for multiple traits, with educational attainment (EA) PGS showing the strongest effect and largely explaining the inter-regional variation of other PGSs. This differentiation is mainly driven by individuals with higher EA PGS migrating to Estonia’s two largest cities from the rest of the country. Importantly, this pattern replicates within families: individuals migrating to the major cities have, on average, higher EA PGS than their siblings staying elsewhere. This trend has persisted since the mid-20th century, despite significant societal changes. These findings illustrate how migration shapes genetic differentiation within a population and highlight direct genetic effects influencing this process.

Pharmacokinetic recall study of Estonian Biobank participants with novel genetic variants in CYP2C19 and CYP2D6
npj Genomic Medicine (2026)
DOI: 10.1038/s41525-025-00549-6

Authors
Kristi Krebs, Laura Birgit Luitva, Anette Caroline Kõre, Raul Kokasaar, Maarja Jõeloo, Georgi Hudjashov, Kadri Maal, Elisabet Størset, Birgit Malene Wollmann, Liis Karo-Astover, Krista Fischer, Estonian Biobank Research Team, Volker M. Lauschke, Magnus Ingelman-Sundberg, Espen Molden, Alar Irs, Kersti Oselin, Jana Lass & Lili Milani

Abstract
CYP2C19 and CYP2D6 are involved in the hepatic metabolism of approximately 35–40% of clinically used drugs. We conducted an in vivo phenotyping study encompassing 114 Estonian Biobank participants to evaluate the functional impact of rare or novel single-nucleotide and structural variants in the CYP2C19 and CYP2D6 genes using omeprazole and metoprolol as respective probe drugs. Plasma concentrations of these drugs and their metabolites were measured at 10 time points, and parent drug-to-metabolite ratios were calculated to determine enzymatic activity. Long-read sequencing enabled high-resolution star allele calling. Our results provide the first in vivo confirmation that partial gene and intragenic deletions in CYP2C19 (CYP2C19*37 and CYP2C19*42), enriched in Estonians and Finns, are associated with poor metaboliser phenotypes (P < 1.2 × 10⁻⁷). Additionally, we offer in vivo evidence of reduced metabolic activity of the CYP2D6*124 allele and a novel missense variant (c.940C>A) in exon 6 of CYP2D6. Furthermore, we observed that inhibitor exposure was significantly associated with higher metabolic ratios for both CYP2C19 (P = 3.0 × 10⁻⁶) and CYP2D6 (P = 0.02). Our findings emphasise the importance of identifying genetic variants in CYP2C19 and CYP2D6 beyond commonly assessed star alleles and that profiling for drug interactions can provide more precise assignments of metabolic phenotypes and improve personalised treatment.

Systematic evaluation of medication adherence determinants across 137 active substances on population-level real-world health data 
Communications Medicine 6, 237 (2026)
DOI: 10.1038/s43856-026-01515-8

Authors
Kerli Mooses, Marek Oja, Maria Malk, Helene Loorents, Maarja Pajusalu, Nikita Umov, Sirli Tamm, Johannes Holm, Hanna Keidong, Taavi Tillmann, Sulev Reisberg, Jaak Vilo & Raivo Kolde

Abstract
The current knowledge about medication adherence is based on studies focusing only on few health conditions and little is known about how strongly adherence is shaped by person-specific behaviour. The aim of the cohort study is to 1) evaluate the effect of multiple factors affecting medication adherence in a consistent manner across 137 active substances, and 2) calculate individual medication adherence score (IMAS), evaluate its predictive power, stability over time, and impact on health outcomes. In essence, IMAS describes persons’ medication-taking “baseline”.

Methods

We utilised a representative dataset with electronic health records, claims, and dispensed medications across 137 active substances and applied continuous multiple interval measures of medication availability (CMA). To assess the effect of various demographic, health, and medication-related variables on CMA, we employed linear mixed models.

Results

Here we show that the medication adherence ranged from 0.423 (albuterol, 95% CI 0.414–0.432) to 0.922 (warfarin, 95% CI 0.917–0.926). The demographic, health- and medication-related factors explained 11.6% and IMAS 22.0% of the variation in adherence. IMAS predicted adherence across medication classes, reduced the risk of overall hospitalisation (hazard ratio = 0.76, 95% CI 0.60–0.97, p < 0.05) and cause-specific incidence for 17 conditions.

Conclusions

Thus, IMAS represents a person-level metric that captures baseline medication-taking behaviour across therapeutic classes and predicts both medication adherence as well as health outcomes. Our analysis suggests that medication-taking behaviour represents a broader patient-level phenomenon manifesting consistently across medications, suggesting its potential for personalised interventions in clinical practice and more efficient public health strategies and policies.

Genetic analysis of circulating metabolic traits in 619,372 individuals
Nature (2026)
DOI: 10.1038/s41586-026-10532-5

Authors
Ralf Tambets, Mihkel Jesse, Jaanika Kronberg, Adriaan van der Graaf, Erik Abner, Urmo Võsa, Ida Rahu, Nele Taba, Anastassia Kolde, Dzvenymyra Yarish, Sariyya Abdullayeva, Anastasiia Alekseienko, Andres Veidenberg, Estonian Biobank Research Team, Krista Fischer, Zoltán Kutalik, Tõnu Esko, Kaur Alasoo & Priit Palta

Abstract
Interpreting the association of genetic variants with complex traits can be improved by gaining a greater understanding of the molecular consequences of these variants. Although genome-wide association studies (GWAS) for complex diseases routinely profile over one million individuals1,2,3,4,5, studies of molecular traits have lagged behind. Here we performed a GWAS meta-analysis for 249 circulating metabolic traits in the Estonian Biobank and the UK Biobank in up to 619,372 individuals. We identified 88,127 common and low-frequency locus–trait associations from 8,398 loci that converged on shared genes and pathways. Using statistical fine mapping, systematic phenome-wide colocalization and cis-Mendelian randomization, we explored putative causal links between metabolic traits and disease outcomes. We predict that although plasma branched-chain amino acids (BCAAs) have been associated with type 2 diabetes in observational studies6,7, lowering BCAA levels by targeting the BCAA catabolism pathway is unlikely to reduce type 2 diabetes risk. Leveraging our large sample size and high-quality genotype imputation, we found that 19.4% of the confidently fine-mapped variants had minor allele frequencies between 0.1 and 1%, and these variants were twofold enriched for predicted missense and splice-altering variants. Our results highlight the value of integrating low-frequency variants into genetic association studies.

Pragmatic trial assessing polygenic risk driven statin therapy for cardiovascular disease prevention: study protocol for the EE-PRS trial
BMJ Open 2026;16:e120048
DOI: 10.1136/bmjopen-2026-120048

Authors
Ave Voit, Aet Elken, Margus Viigimaa, Ruth Kalda, Krista Fischer, Karolin Toompere, Alar Irs, Lili Milani, Helene Alavere, Sander Pajusalu, Elin Org, Janika Alloja, Katrin Kaarna, Mikk Jürisson

Abstract

Introduction: Atherosclerotic cardiovascular disease (ASCVD) continues to be a leading cause of preventable death globally. Polygenic risk scores (PRS) offer a way to detect individuals at higher relative risk of developing ASCVD, but they have not yet been incorporated into routine clinical practice. Pragmatic trials offer a way to evaluate the integration of PRS into cardiovascular disease prevention in primary care, allowing for a more accurate assessment of their effectiveness in everyday clinical practice.

Methods and analysis: This trial evaluates the effectiveness of preventive statin treatment on reducing the incidence of cardiovascular events and death over 5 years in women (aged 55–80) and men (aged 45–80) with a high coronary artery disease PRS. This is a pragmatic, multicentre, open-label, parallel group, randomised clinical trial with a 1:1 allocation ratio to intervention (n=1350), receiving preventive rosuvastatin 20 mg treatment or control (n=1350), receiving current standard of care. Following the intention-to-treat principle, all randomised participants are analysed according to their allocated group, with the primary outcome defined as time to first major adverse cardiovascular event, comprising ischaemic heart disease, stroke or transient ischaemia, peripheral vascular occlusion and stenosis, revascularisation or cardiovascular death.

Ethics and dissemination: This trial has received ethical approval from the Estonian Committee on Bioethics and Human Research (13.06.2024 nr 1.1-12/1531) and the Estonian Ethics Committee for Medicinal Products (19.12.2024 nr RKU-4/92).

Polygenic risk scores reveal shared polygenic risk across five common cardiovascular diseases
Atherosclerosis 417 (2026) 120763
DOI: 10.1016/j.atherosclerosis.2026.120763

Authors
Tigist D. Adane, Maryam Kavousi, Bahar Sedaghati–Khayat, Fernando Rivadeneira, Lianne Trap, Asrat B. Dagnaw, Joyce B.J. van Meurs, André G. Uitterlinden, Jeroen G.J. van Rooij

Abstract

Background and aim: The five cardiovascular diseases (CVDs): Coronary artery disease (CAD), stroke, heart failure (HF), peripheral arterial disease (PAD), and atrial fibrillation(Afib) often co-occur. Most polygenic risk scores (PRSs) are developed for individual CVDs, limiting their clinical utility across multiple CVDs. We investigated associations between five PRSs and (1) their CVDs, (2) cross-disease associations for pleiotropy, and (3) across multimorbidity clusters.

Methods: We calculated restricted PRSs using genome-wide significant variants from published genome-wide association studies and validated them in 484,154 UK Biobank participants. For the time-to-onset and sequence analysis, we included 34,394 individuals with a CAD diagnosis. We used logistic regression, Cox regression, and accelerated failure time analysis; p ≤ 0.05 was considered significant.

Results: All PRSs were significantly associated with their respective CVDs (ORs 1.08–1.60 per 1-SD). The odds of CAD increased by 55% per 1-SD increase in PRSCAD (OR 1.55, 95% CI: 1.53–1.57). In cross-disease associations, all 25 PRS-CVD pairs were significant. For some CVDs, PRS from other CVDs showed a stronger association than their own. For instance, HF was more strongly associated with PRSAfib than PRSHF (OR 1.21 vs 1.17). In CAD patients, those in the top 10% of PRSCAD had a 2.66-fold higher PAD risk (HR 2.66, 95% CI: 1.06–6.71) and developed PAD 5.6 years earlier than those in the bottom 10% (time ratio 0.40, 95% CI: 0.17–0.93).

Conclusion: PRSs reveal shared polygenic risk across CVDs and provide insight into CVD multimorbidity trajectories, such as timing and sequence of events.

Association and risk prediction of 19 complex diseases with polygenic scores and socioeconomic status
Communications Medicine (2026)
DOI: 10.1038/s43856-026-01716-1

Authors
Fiona A. Hagenbeek, Anne Richmond, Max Tamlander, Kira Detrois, Zhiyu Yang, Tuomo Hartonen, Daniel L. McCartney, FinnGen, Riccardo E. Marioni, Pekka Martikainen, Nina Mars, Andrea Ganna & Samuli Ripatti

Abstract

Background: Differences in disease risk are linked to inherited genetic variation and social circumstances, but how these factors jointly relate to multiple diseases is not fully understood. This study investigates the independent and combined associations of genetic predisposition and socioeconomic status with common diseases.

Methods: We analyzed data from 743,194 adults from three biobank studies in Finland and the United Kingdom, including European and non-European ancestry individuals. Genetic predisposition was measured using polygenic scores for 19 common diseases. Socioeconomic status was assessed through education and occupation. We evaluated associations of genetic and socioeconomic factors with disease risk and whether genetic associations differed across education levels. Models including both genetic and educational factors were compared with models including only one factor and models including their interaction.

Results: Here, we show that both genetic predisposition and socioeconomic status are associated with disease risk across most conditions. For seven diseases, genetic associations are stronger in highly educated groups, while occupation shows smaller differences. Patterns are generally consistent across ancestry groups, though education associations are less stable in non-European individuals. While both education and polygenic scores independently improved prediction, combining both provided additional modest improvement to prediction for 16 and 12 out of 19 diseases in FinnGen and the UK Biobank, respectively.

Conclusion: Genetic variation and social circumstances are linked to differences in disease risk both independently, and in combination with one another. Considering both factors modestly improves identification of higher-risk groups and highlights the value of accounting for social context alongside inherited variation in population-level assessments.

Predicting lifetime cardiovascular risk and benefits of preventive treatment in patients with established atherosclerotic cardiovascular disease: the SMART-REACH2 model
European Heart Journal, ehag400
DOI: 10.1093/eurheartj/ehag400

Authors
Joris Holtrop, Mari N Gynnild, Salil V Deo, Carl-Emil Lim, Stelios Boulitsakis Logothetis, Kristi Läll, Minna Oksanen, Håvard Dalen, John Munkhaugen, Peter Ueda, Tomas Jernberg, José Tuñón, Ignacio Mahíllo Fernández, Ingvild Saltvedt, Hanne Ellekjær, Martin Teraa, Manon G van der Meer, Ynte M Ruigrok, Urmo Võsa, Małgorzata Chlabicz, Karol Kaminski, Ph Gabriel Steg, Ivana Burazor, Reedik Mägi, Taavi Tillmann, David A Wood, Kausik K Ray, Dirk de Bacquer, Janne Martikainen, John William McEvoy, Stephen Kaptoge, Frank L J Visseren, Emanuele Di Angelantonio, Deepak L Bhatt, Angela Wood, Steven H J Hageman, Jannick A N Dorresteijn,

Abstract

<a href=”https://doi.org/10.1038/s41598-026-35027-1″>Evaluating the impact of communicating high polygenic risk for coronary artery disease on the health behaviour of young overweight adults: a pragmatic randomised controlled trial</a>
<em>Scientific Reports</em>, 16, 4907 (2026)
DOI: 10.1038/s41598-026-35027-1

<strong>Authors
</strong>Ave Põld, Karolin Toompere, Aet Elken, Margus Viigimaa, Alar Irs, Lembi Caroline Nohrin, Krista Fischer, Lili Milani, Helene Alavere, Janika Alloja, Kersti Pärna, Tanel Traks, Ruth Kalda, Elisabeth Widen, Samuli Ripatti, Mikk Jürisson

<strong>Abstract</strong>
This pragmatic randomised controlled trial evaluated whether communicating high coronary artery disease polygenic risk could improve health behaviour and cardiovascular risk factors among young overweight adults. A total of 1,020 participants were randomised to intervention and control groups and followed for 12 months. The intervention did not result in statistically significant differences in body mass index, waist circumference, blood pressure, lipid profile, blood glucose, cardiovascular risk, smoking, physical activity, or preventive medication use. The findings indicate that communicating high polygenic risk alone is insufficient to produce meaningful short-term behavioural change and highlight the need to combine genetic-risk communication with lifestyle counselling and preventive treatment.

Listen to Vikerraadio’s “Reedene intervjuu” with Merilin Pärli and Lili Milani, Head of the Estonian Biobank and Professor of Pharmacogenomics at the University of Tartu.

https://eestiraadio.err.ee/vikerraadio/1609906742/reedene-intervjuu-merilin-parli-ja-lili-milani/saatest

The interview features Lili Milani and discusses topics related to genetics, health data and personalised medicine in Estonia.

It offers an accessible introduction to how biobank research and the responsible use of genetic information can support prevention and more precise healthcare.

Listen to Vikerraadio’s “Reedene intervjuu” with Merilin Pärli and Lili Milani, Head of the Estonian Biobank and Professor of Pharmacogenomics at the University of Tartu.

https://eestiraadio.err.ee/vikerraadio/1609906742/reedene-intervjuu-merilin-parli-ja-lili-milani/saatest

The interview features Lili Milani and discusses topics related to genetics, health data and personalised medicine in Estonia.

It offers an accessible introduction to how biobank research and the responsible use of genetic information can support prevention and more precise healthcare.

Listen to podcast “Caring with data in a healthy digital society” with Lili Milani, Head of the Estonian Biobank and Professor of Pharmacogenomics at the University of Tartu. e-Governance Academy, 29.10.2025.

https://ega.ee/caring-with-data-healthy-digital-society/

In this episode, Lili Milani discusses how the Estonian Biobank has developed from a genetics research initiative into a trusted national resource that supports more personalised and preventive healthcare.

The conversation explores the responsible use of health and genetic data, the importance of participant trust and consent, and how digital tools such as the MyGenome portal can help people understand and manage their health risks.

Watch “Pealtnägija” report: “Around one hundred women at high cancer risk in Estonia have had their breasts removed” (Eestis on sadakond vähiriskiga naist lasknud rinnad eemaldada). ERR, 30.04.2025.

https://www.err.ee/1609680422/pealtnagija-eestis-on-sadakond-vahiriskiga-naist-lasknud-rinnad-eemaldada

The report explores how genetic testing can identify people at high inherited risk of breast and ovarian cancer and support earlier, more personalised prevention. It includes perspectives from women with BRCA mutations and from medical geneticist Sander Pajusalu.

The story also explains how genetic risk information can guide screening and prevention decisions. It highlights both the potential of risk-based healthcare and the personal considerations involved in preventive treatment choices.

Read article: “The best investment in healthcare is prevention” (Parim investeering tervishoidu on ennetus). Postimees, 06.03.2026.
https://www.akadeemia.ee/diary/7-marts-2026/

Preventing disease is less costly than treating it, and maintaining people’s ability to work is less costly than restoring it after illness. Urmas Varblane and Andres Võrk explain why prevention should be viewed as a long-term investment in both health and the economy.

The article highlights the importance of evidence-based prevention and health-policy decisions that support a healthier population and a more sustainable healthcare system.

Read discussion: “Artificial intelligence and the future of Estonia’s knowledge infrastructure” (Tehisintellekt ja Eesti teadmustaristu tulevik). Vikerkaar, 6/2025.
https://www.vikerkaar.ee/tehisintellekt-ja-eesti-teadmustaristu-tulevik/

Researchers Aro Velmet, Kadri Aavik, Krista Fischer and Oliver Laas discuss how artificial intelligence is reshaping education, research and Estonia’s knowledge infrastructure.

The discussion considers the benefits and limitations of AI, including data bias, privacy, environmental impact and the need to preserve critical thinking. It argues for a careful, evidence-based public debate on how AI should be used in society.

Read article: “Estonian health data can save lives if we put it to work” (Eesti terviseandmed võivad päästa elusid, kui me need tööle paneme). Riigikogu Toimetised, 2026.
https://rito.riigikogu.ee/nr-53/eesti-terviseandmed-voivad-paasta-elusid-kui-me-need-toole-paneme/

Estonia has extensive health data, but its potential is limited when information remains fragmented across systems and cannot be used securely for research, clinical improvement and health-policy decisions. Sulev Reisberg explains why this data should be put to practical use.

The article discusses how common data standards, secure research environments and clear rules for responsible data use can help transform routinely collected health information into better prevention, treatment and patient care.

Read article: “Can artificial intelligence save Estonian healthcare – or simply make it more humane?” (Kas tehisaru päästab Eesti tervishoiu – või teeb selle lihtsalt inimlikumaks?). Postimees, 27.01.2026.
https://arvamus.postimees.ee/8402264/jaak-vilo-kas-tehisaru-paastab-eesti-tervishoiu-voi-teeb-selle-lihtsalt-inimlikumaks

Artificial intelligence offers opportunities to use health data more effectively, support clinical decisions and improve how healthcare services are organised. Jaak Vilo considers what this could mean for Estonia’s healthcare system.

The article discusses how technology should help healthcare professionals understand the patient’s full care pathway and make services more responsive to people’s needs, while keeping human judgement at the centre of care.

Read discussion: “Artificial intelligence and the future of Estonia’s knowledge infrastructure” (Tehisintellekt ja Eesti teadmustaristu tulevik). Vikerkaar, 6/2025.
https://www.vikerkaar.ee/tehisintellekt-ja-eesti-teadmustaristu-tulevik/

Researchers Aro Velmet, Kadri Aavik, Krista Fischer and Oliver Laas discuss how artificial intelligence is reshaping education, research and Estonia’s knowledge infrastructure.

The discussion considers the benefits and limitations of AI, including data bias, privacy, environmental impact and the need to preserve critical thinking. It argues for a careful, evidence-based public debate on how AI should be used in society.

Read article: “Genetic risk is not destiny” (Geneetiline risk ei ole saatus). Edasi, 2026.

https://edasi.org/271760/lili-milani-geneetiline-risk-ei-ole-saatus/

Genetic risk information can help identify people who could benefit most from preventive healthcare, but genes do not determine a person’s future. Lili Milani explains how the Estonian Biobank can support more precise and preventive medicine.

The article discusses the importance of evidence-based implementation, cost-effective planning and public trust when bringing genetic information into everyday healthcare.

Read article: “Data loop: how health data enables a learning healthcare system” (Andmering: kui terviseandmed muutuvad õppimisvõimeks. OMOP-andmekihil põhinev ravikvaliteedi tagasisideraamistik südameveresoonkonnahaiguste näitel). 23.08.2026.

https://ojs.utlib.ee/index.php/EA/article/view/29714

Estonia’s healthcare system generates extensive data, but it is not yet routinely used to provide systematic feedback on treatment quality. This article introduces Andmering, a framework for turning routinely collected health data into a recurring learning process for healthcare.

The framework uses the OMOP data layer to bring information from different sources into one analytical format. Its first application focuses on cardiovascular diseases, including acute myocardial infarction, atrial fibrillation and heart failure.

Seminar “How health data help to prevent and treat diseases?” (Kuidas terviseandmed aitavad haigusi ennetada ja ravida?) in co-operation with the Estonian Academy of Sciences (EAS). 24.03.2025

Large volumes of diverse health data are generated in healthcare systems on a daily basis. The best modern medical care relies on extensive individual patient health data, which are highly valuable for analysing how disease prevention can be improved, treatments made more effective, and informed healthcare policy decisions taken.

At present, only a small proportion of the available data is being used. Although data-processing capacity has increased dramatically over the past decade, continued attention to data quality remains essential. Presenters discussed how Estonian researchers have recently advanced the management of health data and what would be a sensible way forward.

Click on the presentation title to watch the recording (in Estonian).

Introduction (Sissejuhatus teemasse). Sander Pajusalu, MD, Associate Professor of Clinical Genetics, University of Tartu; Head of the Genetics and Personalised Medicine Clinic, Tartu University Hospital
Health informatics – how data are generated and how they are used (Terviseinformaatika – kuidas andmed tekivad ja kuidas neid kasutatakse). Jaak Vilo, Professor of Bioinformatics, University of Tartu
From data to medicines (Andmetest ravimiteni). Kadi Liis Saar, Researcher, University of Cambridge
The black box of colorectal cancer (Soolevähi must kast). Indrek Seire, MD, Junior Research Fellow, Institute of Clinical Medicine, University of Tartu
Beautiful health data (Ilusad terviseandmed). Maarja Pajusalu, Junior Research Fellow, Institute of Computer Science, University of Tartu
Health data in joint replacement (Endoproteesimise terviseandmed). Kaspar Tootsi, MD, Researcher in Orthopaedics, University of Tartu

Listen to episode “Genetic study of cardiovascular health“ (Südametervise geeniuuring). 17.11.2025

What questions should a five-year genetic study on cardiovascular health answer? The study includes 2,700 participants and forms part of the TeamPerMed project, which focuses on advancing personalised medicine. The topic is discussed by Mikk Jürisson, Associate Professor of Public Health at University of Tartu.

Listen to episode “Personalised medicine” (Personaalmeditsiin). 13.09.2025

Which genetic data could general practitioners use? What do studies show about people’s health behaviour, and what do researchers recommend?
These questions are addressed by Mikk Jürisson, Associate Professor of Public Health at University of Tartu.

Listen to episode “Genetic research and personalised medicine” (Geeniuuringud ja personaalmeditsiin). 02.04.2025

Estonian Biobank is sending invitations to selected gene donors to participate in a large-scale cardiovascular health research project. Based on the data collected, researchers aim to determine whether cardiovascular disease can be prevented more effectively by prescribing cholesterol-lowering treatment according to an individual’s genetic risk.

The opportunities offered by genetic research are discussed in more detail by medical geneticist and Head of the Genetics and Personalised Medicine Clinic at Tartu University Hospital, Dr Sander Pajusalu.

Listen to episode “Margus Viigimaa, MD: My heart rejoices over the new generation of cardiologists” (Margus Viigimaa: mu süda rõõmustab kardioloogide uue põlvkonna üle). 19.04.2025

As a practising cardiologist, Dr Margus Viigimaa consults up to twenty patients on some days at the Heart Centre of the Regional Hospital, while also conducting scientific research at Estonian universities.

Dr Viigimaa talks about cutting-edge cardiac genetic studies recently launched with his involvement and leadership, as well as new biological therapies, wireless pacemakers and other innovations that have only just been introduced into cardiology.