Blood Pressure

Blood pressure is the trait where genetics has to be honest about its place. It is highly heritable, it has been studied in over a million people, and none of that competes with the thing anyone can do in two minutes: put a cuff on an arm and read the number.

Two thousand positions, and a measurement that beats all of them

A 2018 study of more than a million people added 535 new blood-pressure loci in one paper. The count has since passed 2,000, and a 2025 review of the field says plainly that the causal genes and pathways behind most of them remain largely unknown.

Each of those positions moves blood pressure by a fraction of a millimetre of mercury. A reading tells you what your blood pressure actually is, today, including everything genetics cannot see — salt, weight, sleep, alcohol, stress, kidney function, the medicines you already take. That is why no guideline anywhere asks for a genotype before treating high blood pressure, and why this page cannot be used to decide anything.

These two came from an east Asian study

Most of the evidence on this site was gathered in people of European descent, and we say so on the pages where it matters. This page is the exception.

Both variants come from a meta-analysis in the AGEN-BP consortium: 19,608 people of east Asian ancestry, with new genotyping in 10,518 more and replication in a further 20,247.

• rs17249754 — near ATP2B1, a calcium pump in the membrane of the cells lining blood vessels. It was found first in European-ancestry studies and confirmed in this east Asian one, which is what a real association looks like.
• rs35444 — near TBX3, and this one was newly discovered in that study. TBX3 is a developmental transcription factor, better known for its role in heart and limb formation.

The one that was significant for hypertension itself

rs2681472 is also ATP2B1, and it comes from a different study with a distinction worth pulling out.

The CHARGE Consortium looked at 29,136 people and asked three questions rather than one: what moves systolic pressure, what moves diastolic pressure, and what predicts a diagnosis of hypertension. Those are not the same question — the first two are about a measurement, the third is about crossing a line.

Joined with a second consortium (34,433 more people), four loci reached genome-wide significance for systolic pressure and six for diastolic. For hypertension as a diagnosis, exactly one did: ATP2B1.

So two of the three positions on this page sit in the same gene, found by two different groups asking two different questions on two continents. That is about as much corroboration as a blood pressure locus gets.

What ATP2B1 actually does, and why it is a familiar idea

ATP2B1 encodes a calcium pump — the protein that pushes calcium out of a cell, including the smooth muscle cells that line arteries. How much calcium is inside those cells determines how tightly they squeeze, and how tightly they squeeze is blood pressure.

That mechanism will be familiar to anyone taking a calcium channel blocker: one of the oldest and most-used classes of blood pressure medicine works on the other side of the same traffic. The genetics did not lead to those drugs — they came first — but finding the strongest common signal on the same pathway is a decent check that the pathway matters.

It does not follow that carrying the variant predicts who responds to which drug. That has been looked for and is not established.

The part that connects to another page here

That same study found something striking: a very strong blood-pressure signal at chromosome 12q24.13, near ALDH2, with p-values of 7.9 x 10^-31 and 1.3 x 10^-35 — and it was ethnically specific, showing up in east Asians rather than in European-ancestry samples.

ALDH2 is the gene behind the alcohol flush reaction, and the variant that causes it is common in east Asia and almost absent elsewhere. A gene that governs how much alcohol a person can comfortably drink turning up as a blood-pressure locus in exactly the population where that gene varies is not a coincidence waiting to be explained — alcohol raises blood pressure, and this looks like that effect written into the genome.

It is also a good demonstration of why studying only one ancestry misses things. This signal is invisible in European-ancestry data because the variant is not there to be seen.

Twenty-nine more, from a study built to look for exactly this kind of difference

The International Genomics of Blood Pressure (iGEN-BP) Consortium asked a more pointed version of the same ancestry question: rather than replicating a European finding in an East Asian sample, it searched from the start for signals whose effect differs by ancestry. Across 289,038 people, principally of East Asian descent with a meta-analysis against European-ancestry data, it reported 19 new loci.

Twenty-nine of the positions on this page come from that study — nine for systolic pressure, twenty for diastolic. Three, including rs6590816 and rs6590811, sit near ARHGAP42, a smooth-muscle contraction gene already established in blood pressure; two more, rs9814480 and rs3774447, sit near CACNA1D, another calcium channel alongside ATP2B1; and rs138088171 and rs62525059 sit near CYP11B1 and CYP11B2, the genes for the last enzymatic step in making aldosterone, the hormone the kidney uses to hold onto sodium.

The study's main finding was methodological as much as it was genetic: at ten loci, the East Asian and European signals shared the same neighbourhood of the genome but not the same causal variant, and sometimes not even the same direction of effect. Across the whole genome, the underlying effect sizes still correlate strongly between ancestries — 0.898 for systolic pressure, 0.851 for diastolic — so most of blood pressure's genetics is shared. But the exceptions were not noise; the paper reports some carrying a signature of natural selection acting differently between the two populations, the same idea as the ALDH2 signal above, generalised into a method built to find it on purpose.

Thirteen more, specifically for pulse pressure

Pulse pressure — the gap between systolic and diastolic pressure, a rough marker of how stiff the large arteries have become — has its own genetics, and this page's remaining 13 loci come from four separate studies of it. The largest contribution, seven loci, comes from Kraja et al. 2017: 150,134 people in discovery and 228,245 in replication, cross-checked against whole-blood gene expression data from 12,607 people to implicate 48 genes across multiple independent resources, with three signals specific to kidney tissue — the kind of translational evidence aimed squarely at finding new drug targets rather than just new statistical hits.

rs62011052, near ADAMTS7, is one of them — the same gene already on this site's coronary artery disease page. Arterial stiffness and coronary disease share enough underlying vascular biology that this is a genuinely coherent overlap, not a coincidence of two studies happening to name the same gene.

A fourth of these loci come from Warren et al. 2017, an electronic-health-record study of 99,785 people with over 1.3 million individual blood pressure readings, later combined with two further consortia to 321,262 people total. One more comes from Liang et al. 2017, a study specifically of 31,968 people of African ancestry — its LLPH finding (this page's rs113866309) is one of several loci in that study whose associated allele is common only in African-ancestry populations, the same kind of ancestry-specific genetics this page discusses above for East Asian ancestry. The last comes from Surendran et al. 2018, which validated 107 independent UK Biobank loci and proposed using a genetic risk score to target early lifestyle intervention at the people it identifies as highest-risk.

Feitosa et al. 2018 took a different angle entirely: rather than just testing for a direct genetic effect, it tested whether a variant's effect on blood pressure changes depending on how much alcohol someone drinks. A discovery stage in about 131,000 people across several ancestries found 3,514 candidate SNVs, followed by independent replication in about 440,000 more people. The study confirmed 5 novel blood-pressure loci (380 SNVs in 21 genes) plus 49 previously reported loci in European- and multi-ancestry analyses, and a further correlated meta-analysis added 8 more novel loci. This page holds 35 of this study's variants, split across its pulse-pressure and mean-arterial-pressure interaction results, including an NPPB signal that appears in both — a gene already familiar to this page for its role in blood-pressure regulation.

The largest pulse-pressure study yet

A 2019 study built its discovery cohort trans-ethnically from the start rather than adding other ancestries onto a European-ancestry-first design: 776,078 people from the Million Veteran Program and collaborating studies. It found 208 novel common blood-pressure SNPs and 53 rare variants across systolic, diastolic and pulse pressure. Beyond the GWAS itself, a transcriptome-wide analysis linked 4,043 blood-pressure associations to the genetically predicted expression of 840 genes in 45 tissues, and single-cell RNA sequencing in mouse kidneys specifically flagged upregulated blood-pressure genes in renal tubule cells — evidence for the kidney's role in blood pressure that comes from a completely different method than a population GWAS alone. This page holds 56 of this study's pulse-pressure variants, plus 3 more from two smaller 2016 meta-analyses that tested rare and low-frequency variants alongside common ones, the same angle this larger study later scaled up trans-ethnically.

What to do about blood pressure

Everything that works is already known and none of it requires a test: knowing your numbers, salt, weight, alcohol, movement, sleep, and taking prescribed medicines as prescribed. High blood pressure is common, usually silent, and one of the most treatable causes of stroke and heart disease there is.

If it has been a while since anyone measured yours, that — not this page — is the useful next step.