- Design
- Multicentre retrospective development and external validation study
- Population
- 443 patients with biopsy-confirmed lupus nephritis, five centres
- Primary outcome
- Identification of proliferative lupus nephritis
- Effect
- AUC 0.852 development, 0.895 external validation
A multicentre retrospective study of 443 patients with biopsy-confirmed lupus nephritis built a model to identify proliferative disease from routine data. It was developed in 337 patients from one centre and validated in 106 from four others.
The model used lymphocyte percentage, haematuria, leukocyturia, eGFR, uric acid, anti-dsDNA positivity and C3. Discrimination was good: AUC 0.85 in development and 0.90 in external validation, with acceptable calibration.
Proliferative lupus nephritis carries worse kidney outcomes and usually needs more intensive induction. Biopsy is the standard but is not always possible, which is a real constraint in parts of India.
The model helps prioritise urgent biopsy or guide treatment when biopsy cannot be done, but it does not replace histology. The validation cohort was small and from the same country.
- Haematuria, leukocyturia, low C3, positive anti-dsDNA and reduced eGFR together raise suspicion of proliferative disease.
- Use the model to prioritise biopsy, not to avoid it where biopsy is feasible.
- Where biopsy is contraindicated or unavailable, a high-risk estimate may support treating as proliferative.
- Validation was in a small cohort; check local performance before relying on it.
Why it matters
Where biopsy is delayed or unsafe, routine tests can still point to the patients who need intensive treatment.
The statistics, in plain English
An AUC of 0.90 is high, but in a validation group of 106 the estimate is imprecise, and performance often drops when a model is used elsewhere.
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