Three kinds of gout, one shared failure: 12% at target

Design

population-based retrospective cohort with K-prototypes cluster analysis, SIDIAP primary care database, Catalonia, 2012-2023

Population

94,759 patients with incident gout

Primary outcome

clinical phenotype clusters, sustained serum urate control (<6 mg/dl for at least 80% of follow-up) and urate-lowering therapy adherence

Effect

three clusters (37.0%, 22.8%, 40.2%); only 12% achieved sustained urate control overall, worst in the most comorbid cluster despite higher prescribing

Using the SIDIAP primary care database in Catalonia, this study applied a clustering algorithm to 94,759 patients with incident gout diagnosed between 2012 and 2023. Three phenotypes emerged. Cluster 1 (37.0%) was younger, mean age 54, with few comorbidities. Cluster 2 (22.8%) was older, mean age 74, with type 2 diabetes in every member, obesity in 51.3%, and the heaviest cardiovascular burden. Cluster 3 (40.2%) was older with hypertension and dyslipidaemia but not diabetes.

The authors read these as three routes to hyperuricaemia — largely genetic, insulin-resistance-mediated, and renal-vascular — which is a reasonable interpretation but remains an inference from comorbidity patterns rather than a mechanistic finding.

The number that should stop a reader is not the clustering. Only 12% of the entire cohort achieved sustained serum urate below 6 mg/dl, defined as being at target for at least 80% of follow-up. And the pattern of failure is instructive: the most comorbid cluster received more urate-lowering prescriptions but achieved worse control than the youngest, healthiest group, which was prescribed less and adhered better. Prescribing is not the bottleneck. Titration to target, and the follow-up appointments that titration requires, are. Any service reading this and reaching for a formulary change is solving the wrong problem.

• Audit what proportion of your gout patients are at a serum urate below 6 mg/dl, not what proportion hold a prescription
• Book the urate recheck at the time of starting or changing urate-lowering therapy, not at the next flare
• Expect adherence, not prescription rates, to be the limiting step in the most comorbid patients
• Screen the metabolic cluster actively for diabetes and cardiovascular risk — it defines them
• A single normal urate does not equal control; the definition that mattered here was staying at target over time

The statistics, in plain English

Cluster analysis finds groups that differ on the variables fed into it; feeding in comorbidities guarantees clusters defined by comorbidities, so the three phenotypes describe the data rather than discover biology. The 12% figure is the robust finding here because it is a simple proportion with a prespecified definition, not a model output. Because this is routine primary care data, urate measurements happen when clinicians order them, so patients tested more often — usually the less well controlled — are represented differently from those rarely tested.

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.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.