- Design
- two-stage case-control epigenetic study with independent validation cohort and penalised regression classifiers
- Population
- 124 with gout, 121 with asymptomatic hyperuricaemia, 124 normouricaemic controls; discovery 83, validation 286
- Primary outcome
- discrimination between disease states by DNA methylation signature
- Effect
- gout vs asymptomatic hyperuricaemia area under the curve 0.67 clinical, 0.87 combined (DeLong P = 0.018)
Only a minority of people with hyperuricaemia ever develop gout, and nothing in routine assessment explains which. This two-stage study profiled peripheral blood DNA methylation in 124 patients with gout, 121 with asymptomatic hyperuricaemia and 124 normouricaemic controls, with a discovery cohort of 83 and independent testing in 286.
Differentially methylated regions appeared across all three pairwise comparisons, mapping to 84 genes that overlapped known gout and serum urate risk loci — including HLA-DRB1, TAP2 and GNAS — and were enriched at transcription factor binding sites. One hundred and thirteen validated differentially methylated cytosines were used to build classifiers.
Performance tracked the difficulty of the question. Separating gout from normouricaemic controls was easy (test-set area under the curve 0.96) and not clinically interesting. Separating gout from non-gout controls, methylation outperformed clinical variables (0.91 against 0.78). The distinction that actually matters — gout against asymptomatic hyperuricaemia — was the hardest, and adding methylation to clinical variables took it from 0.67 to 0.87.
That is a substantial improvement on a question clinical assessment answers poorly, and the convergence with genetic risk loci gives it biological plausibility rather than leaving it as a pattern-matching exercise. It is not a test. The authors say prospective validation is needed before clinical deployment, and a case-control design cannot tell you whether the signature predicts who will develop gout or merely reflects having had it — which is the question a clinician seeing a raised urate would want answered.
- Continue to manage asymptomatic hyperuricaemia on current principles; no methylation test is available or validated
- Note that clinical variables alone separated gout from asymptomatic hyperuricaemia poorly (area under the curve 0.67)
- Do not read the overlap with genetic risk loci as evidence that the signature predicts future gout
- Expect stratification research in this area to use case-control designs until prospective cohorts report
- Keep documenting whether a raised urate was ever symptomatic; that clinical history remains the only available discriminator
Why it matters
The question of which hyperuricaemic patient will develop gout has had no better answer than waiting to see.
Don't overread it
A case-control study cannot show that the signature predicts future gout, only that it differs between groups already defined by their diagnosis.
The statistics, in plain English
An area under the curve of 0.87 against 0.67 is a large gain on a question where clinical variables barely beat guessing — 0.5 would be chance. The classifier was built with a penalised regression method designed to avoid overfitting, and tested in an independent cohort, which is the right approach and makes the numbers more credible than a single-cohort model. The design remains case-control, so the signature distinguishes people who have gout from people who do not; whether it would have identified them before the first attack is untested.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for rheumatology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free