EUVIMEDEUVIMED entdecken
EditorialKommentar

Modelling relations between blood pressure, cardiovascular phenotype, and clinical factors using large scale imaging data

Turkay Kart, Mohanad Alkhodari, Winok Lapidaire, Paul Leeson iD

European Heart Journal - Cardiovascular Imaging · 2023

DOI: 10.1093/ehjci/jead161 ↗

Kurz gefasst

Editorial in der Zeitschrift European Heart Journal - Cardiovascular Imaging, 2023. Verfasst von Turkay Kart und 3 weiteren Autor:innen. Beteiligte Einrichtung: Cardiovascular Clinical Research Facility, Radcliffe Department of Medicine, John Radcliffe Hospital, University of Oxford, Oxford OX3 9DU, UK. Hauptthemen laut MEDLINE: Cardiovascular System, Hypertension. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 13 Literaturangaben. Gefördert von: Medical Research Council, Oxford NIHR Biomedical Research Centre.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This editorial refers to ‘Characterizing the hypertensive cardiovascular phenotype in the UK Biobank’, by H. Elghazaly et al., https://doi.org/10.1093/ehjci/jead123. Hypertension remains one of the commonest cardiovascular risks.1 However, both provision and uptake of screening as well as subsequent availability of advice about how to control elevated blood pressure remain poor.2 This can exacerbate health inequalities for certain groups including women, who often first present very early in life with hypertension during pregnancy,3 and ethnic groups, who appear to suffer higher rates of hypertension-related complications.2 Into this clinical paradigm, Elghazaly et al.4 have now published an extensive exploration of cardiac phenotypes associated with hypertension in the UK Biobank imaging study taking into account both sex and ethnicity. Cardiovascular risk in those with hypertension is augmented if there is also evidence of cardiac or vascular remodelling.2 Failure to account for progression of underlying end organ disease may be leading to under treatment of disease state while missing an opportunity to provide more personalized care to groups at risk.3,5 UK Biobank is a unique data resource, openly available to researchers, providing large-scale data on health and illness, within the UK.6 The sample comprises of 500 000 participants, over the age of 45 years at first involvement, of which 100 000 are currently undergoing multi-organ imaging.6 Around 40 000 imaging data sets are currently available for research including cardiac and vascular magnetic resonance7 from which Elghazaly et al.4 have compiled an impressive set of quality-controlled metrics. A combination of analysis techniques has been deployed including deep-learning–based automated image analysis that covers chamber sizes, functional measures, including strain, and myocardial characteristics such as T1 mapping.8–10 Furthermore, through careful use of linkage between UK Biobank data and primary care records, an in-depth clinical picture of hypertension management has been developed.6 With approximately 40 000 participants, this is the one of the largest studies in the literature addressing hypertensive cardiac phenotypes. The data structure, including a significant heterogeneity in demographic and lifestyle characteristics, as well as a high prevalence of conditions, has enabled a detailed associative analysis. Of course, the data are not perfect. Ninety-seven per cent of participants were of a White ethnicity, with an average age at imaging of 64 years, which limits generalizability across age groups and populations. However, a strength of the manuscript is the simplicity of the statistical approach, which means that the findings are robust and, perhaps most important, straightforward for the reader to understand. A range of multi-variable linear regression models, including analyses stratified for variations in demographics, time since hypertension diagnosis, and blood pressure control, add depth to the information provided. The majority of associations follow expectations such as the associations between hypertension and measures of left ventricular and atrial volumes and wall thickness.4 Even associations between hypertension and lower left atrial ejection fraction and aortic distensibility are well known. An unexpected finding was lower myocardial T1 associated with hypertension and this has led to speculation about possible technical reasons11 and the conclusion ‘further work is needed’. Arguably the most interesting findings relate to variation in phenotype with time from clinical diagnosis and treatment with both linear associations with ventricular function and mass12 as well as non-linear associations with left atrial ejection fraction and volume.13 A limitation of this analysis is the batch sampling of ‘time since diagnosis’ that only allows conclusions to be made across two or three time points. Greater fidelity will be required to fully explore potential non-linearity of associations.12 Reassuringly, good blood pressure control prevented or reversed these changes.2 Or did it? This is an associative study and direction of causality is not clear. A valid alternative explanation is that blood pressure control is better, or simpler, in those who are not predisposed to develop cardiac remodelling, so, ‘further work is needed’. Investigations such as those undertaken by Elghazaly et al.4 both consolidate our understanding of the morphological cardiac features associated with hypertension and confidently demonstrate that particular patterns of cardiac remodelling are not inevitable for all patients with hypertension. Large databases such as the UK Biobank provide the opportunity to investigate differences that may go unnoticed, or be missed, in smaller studies. Further use of large-scale imaging data sets over the next few years, particularly when combined with additional features, such as clinical, genomic, and proteomic information, should generate fascinating insights into patterns and trajectories of cardiac remodelling.5 From these discoveries, the way may open for personalized and new approaches to blood pressure management, which take into account underlying organ disease state. This work was supported by the Medical Research Council (grant number MR/W003686/1). P.L. acknowledges support of the NIHR Oxford Biomedical Research Centre. M.A. is funded by the Healthcare Engineering Innovation Center (HEIC) at Khalifa University, United Arab Emirates (award number: 8474000132). No new data were generated or analysed in support of this research.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Cardiovascular SystemHypertensionHumansBlood PressureBiological Specimen BanksPhenotypeUnited Kingdom

Autor:innen und Institutionen

  • Cardiovascular Clinical Research Facility, Radcliffe Department of Medicine, John Radcliffe Hospital, University of Oxford, Oxford OX3 9DU, UK.
  • Healthcare Engineering Innovation Center (HEIC), Department of Biomedical Engineering, Khalifa University, Shakhbout Bin Sultan St., Hadbat Al Za'faranah - Zone 1, P.O. Box 127788, Abu Dhabi, UAE.

ORCID-Profile: Paul Leeson

Förderung

  • Medical Research Council
  • Oxford NIHR Biomedical Research Centre

Zitieren

APA

Kart, T., Alkhodari, M., Lapidaire, W., & Leeson, P. (2023). Modelling relations between blood pressure, cardiovascular phenotype, and clinical factors using large scale imaging data. European Heart Journal - Cardiovascular Imaging. https://doi.org/10.1093/ehjci/jead161

Vancouver

Kart T, Alkhodari M, Lapidaire W, Leeson P. Modelling relations between blood pressure, cardiovascular phenotype, and clinical factors using large scale imaging data. European Heart Journal - Cardiovascular Imaging. 2023. doi:10.1093/ehjci/jead161

BibTeX
@article{Kart2023,
  author = {Kart, Turkay and Alkhodari, Mohanad and Lapidaire, Winok and Leeson, Paul},
  title = {Modelling relations between blood pressure, cardiovascular phenotype, and clinical factors using large scale imaging data},
  journal = {European Heart Journal - Cardiovascular Imaging},
  year = {2023},
  doi = {10.1093/ehjci/jead161}
}
Kostenlos starten →