Gout

Gout is arthritis caused by a crystal. When urate — the end product of purine breakdown — rises past what blood can hold, it can come out of solution as needle-shaped crystals in a joint. The immune system attacks the crystals, and the result is one of the most violently painful things a joint can do: usually the base of the big toe, usually overnight, red and hot and untouchable.

It has a reputation as a disease of excess, and diet does matter. It matters less than most people are told, and rather less than inheritance does.

The two genes that do the work

Urate is not made in an unusual way in people who get gout. It is removed in an unusual way. Both major genes here are transporters — proteins that carry urate across a membrane and out.

rs4475146 is in SLC2A9, which reabsorbs urate in the kidney: it decides how much of what is filtered out gets pulled back in. It has the largest effect on blood urate of any common variant known, in either direction.

rs1481012 is in ABCG2, which pumps urate out into the intestine. When it works less well, a route of disposal narrows.

The scale of these effects is unusual for common variants. In a 2008 study, the missense variant in SLC2A9 reached a p-value of 7 × 10⁻¹⁶⁸ for blood urate and was associated with gout at an odds ratio of 0.59 per copy of the protective allele; the missense variant in ABCG2 was associated with gout at OR 1.74. Those are not the fractional shifts common-variant genetics usually delivers.

How many positions, and how much they explain

A 2013 study of more than 140,000 people found 28 loci affecting blood urate — 18 of them new — and checked each one against gout itself and against how much urate the kidney excretes. Effects were of similar size in several non-European groups, which is not something that can be assumed.

The pathways the analysis pointed at were not only transport: glucose metabolism turned up too, which is part of why gout keeps company with obesity, diabetes and high blood pressure rather than arriving alone.

Sixteen more, from five further studies of urate itself

The two genes above are not the whole story. Sixteen more positions on this page come from five further studies of blood urate specifically, and SLC2A9 keeps turning out to be the strongest common-variant hit for the trait, study after study.

The discovery came first: rs6855911, in what was then called simply "GLUT9" — the same gene now known as SLC2A9 — was identified in 2007 in Sardinian and Tuscan cohorts, one of the earliest genome-wide scans to implicate this gene at all.

A 2010 CHARGE Consortium study of 28,283 people, replicated in 22,054 more, built an eight-locus "genetic urate score" — including three more of this page's variants, rs13129697 (also SLC2A9), rs675209 (RREB1) and rs1106766 (INHBC) — and used it as a natural experiment. The score was strongly associated with both urate and gout, but showed no association at all with blood pressure, glucose, kidney function or coronary heart disease. That is genetic evidence against, not for, the idea that urate itself damages the heart — gout and cardiovascular disease keep company for other reasons, not because one straightforwardly causes the other.

A larger 2021 UK Biobank study of 363,228 people, covering 35 different blood and urine biomarkers at once, contributed 56 of this page's variants in total. Using the same genetic-instrument logic as the CHARGE score above, at far greater scale, it confirmed a causal effect of urate on gout specifically — one of 51 such causal relationships the study established across all 35 biomarkers together.

The same study put that causal finding to practical use: it built a combined "multi-PRS" model, one polygenic risk score per biomarker used together rather than a single urate score alone, and tested it in an independent cohort of 135,500 people. For gout specifically, that combined model improved genetic risk stratification beyond what the urate score alone could manage — evidence that gout's genetic risk is not fully captured by urate genetics in isolation, even though urate is the proximate, causal driver described above.

The remaining loci come from a Japanese biochemical-trait GWAS (three variants, including further hits in ABCG2, SLC2A9 and SLC22A12) and a Scottish cohort study that used deeper genotype imputation to find two rarer variants, near FAM134B and ZNF160 — a reminder that even a trait mapped this thoroughly still has loci too rare for most studies to catch.

What this does not mean about your diet

High urate is necessary for gout and nowhere near sufficient: most people with raised urate never have an attack. And of the variation in blood urate between people, diet accounts for a strikingly small share compared with these transporters.

That is not permission, and it is not blame either. It is a correction to a story in which gout is something a person did to themselves.

One place a genotype does change a decision

Not these two. A different position, in HLA-B, does: the allele HLA-B*58:01 raises the risk of a rare but severe skin reaction to allopurinol, the most-prescribed urate-lowering drug. Prescribing guidance recommends testing before starting allopurinol in the populations where that allele is common — see the allopurinol page.

That is the reverse of the usual direction. It is not a genotype telling somebody to take something; it is a genotype telling a prescriber to look before they do.