Refereed journal article or data article (A1)

Comprehensive biomarker profiling of hypertension in 36 985 Finnish individuals




List of Authors: Palmu Joonatan, Tikkanen Emmi, Havulinna Aki S, Vartiainen Erkki, Lundqvist Annamari, Ruuskanen Matti O, Perola Markus, Ala-Korpela Mika, Jousilahti Pekka, Würtz Peter, Salomaa Veikko, Lahti Leo, Niiranen Teemu

Publication year: 2021

Journal: Journal of Hypertension

Journal name in source: Journal of hypertension

Journal acronym: J Hypertens

ISSN: 0263-6352

eISSN: 1473-5598

DOI: http://dx.doi.org/10.1097/HJH.0000000000003051

Self-archived copy’s web address: https://research.utu.fi/converis/portal/detail/Publication/68196779


Abstract

Objective: Previous studies on the association between metabolic biomarkers and hypertension have been limited by small sample sizes, low number of studied biomarkers, and cross-sectional study design. In the largest study to date, we assess the cross-sectional and longitudinal associations between high-abundance serum biomarkers and blood pressure (BP).

Methods: We studied cross-sectional (N = 36 985; age 50.5 ± 14.2; 53.1% women) and longitudinal (N = 4197; age 49.4 ± 11.8, 55.3% women) population samples of Finnish individuals. We included 53 serum biomarkers and other detailed lipoprotein subclass measures in our analyses. We studied the associations between serum biomarkers and BP using both conventional statistical methods and a machine learning algorithm (gradient boosting) while adjusting for clinical risk factors.

Results: Fifty-one of 53 serum biomarkers were cross-sectionally related to BP (adjusted P < 0.05 for all). Conventional linear regression modeling demonstrated that LDL cholesterol, remnant cholesterol, apolipoprotein B, and acetate were positively, and HDL particle size was negatively, associated with SBP change over time (adjusted P < 0.05 for all). Adding serum biomarkers (cross-sectional root-mean-square error: 16.27 mmHg; longitudinal: 17.61 mmHg) in the model with clinical measures (cross-sectional: 16.70 mmHg; longitudinal 18.52 mmHg) improved the machine learning model fit. Glucose, albumin, triglycerides in LDL, glycerol, VLDL particle size, and acetoacetate had the highest importance scores in models related to current or future BP.

Conclusion: Our results suggest that serum lipids, and particularly LDL-derived and VLDL-derived cholesterol measures, and glucose metabolism abnormalities are associated with hypertension onset. Use of serum metabolite determination could improve identification of individuals at high risk of developing hypertension.


Downloadable publication

This is an electronic reprint of the original article.
This reprint may differ from the original in pagination and typographic detail. Please cite the original version.




Last updated on 2022-07-04 at 16:21