Design: Multicentre retrospective cohort with latent class analysis
Population: 168 patients with biopsy-proven lupus nephritis after renal response
Primary outcome: Renal flare and damage accrual
Effect: Combined serological-clinical activity: renal flare aHR 2.66 (1.34 to 5.30)

This multicentre retrospective cohort followed 168 patients with biopsy-proven lupus nephritis who had achieved renal response, for a median of four years. Latent class analysis grouped patterns of extra-renal activity.

Extra-renal activity appeared at least once in 92.3%, most often serological (anti-dsDNA or low complement), then mucocutaneous and musculoskeletal. Clinically significant extra-renal flares occurred in 37.5%. Higher non-renal SLEDAI-2K scores were associated with renal flare (aHR 1.18 per point, 95% CI 1.07 to 1.30) and damage (aHR 1.23, 1.13 to 1.35), and BILAG A extra-renal flares strongly with damage (aHR 10.56, 4.63 to 24.06). Of three activity patterns, only combined serological and clinical activity was associated with renal flare (aHR 2.66, 1.34 to 5.30) and damage (aHR 2.37, 1.07 to 5.25); isolated low complement was not.

The practical point is that renal response is not the end of the story. A patient whose kidneys have settled but who has a positive anti-dsDNA alongside rash or arthritis is the one to watch most closely, and perhaps not the one in whom to taper treatment.

Keep assessing extra-renal activity at every visit after renal response in lupus nephritis.
Monitor anti-dsDNA and complement regularly, not only urine protein.
Consider closer renal monitoring when serological activity coincides with clinical extra-renal flares.
Be cautious about tapering immunosuppression in patients with combined serological and clinical activity.
Record non-renal SLEDAI-2K scores to track risk over time.

Why it matters

Renal response can tempt treatment de-escalation, but extra-renal activity carries renal risk of its own.

The statistics, in plain English

An adjusted hazard ratio of 10.56 for damage after a BILAG A flare has a very wide interval (4.63 to 24.06), reflecting few events; the direction is clear but the size is uncertain. With 168 patients, the latent classes should be seen as exploratory.