Background Early identification of patients with lupus nephritis (LN) likely to achieve complete renal response (CRR) may expedite the evaluation of new therapies and guide clinical care. Prior analyses have shown that early improvement in proteinuria is associated with subsequent renal response.1,2 Several ongoing LN trials including NOBILITY, an assessment of the efficacy of the anti-CD20 monoclonal antibody obinutuzumab in combination with standard of care immunosuppression, will evaluate proteinuric response at 6 months as a key secondary endpoint.3 Whether short-term response accurately predicts future CRR, however, is uncertain.
Objectives To assess the predictive value of early measurements of the level of proteinuria and to identify proteinuria cutoffs that best identify patients who will achieve CRR at 18 months.
Methods LUNAR and BELONG were multicenter, double-blinded studies that in total randomized 522 patients with ISN/RPS class III or class IV LN to blinded investigational infusions or placebo in combination with standard of care immunosuppression.4,5 CRR was assessed at 18 months and defined for this analysis as achievement of urine protein to creatinine ratio (UPCR) <0.5 with normal serum creatinine that was not increased from baseline by >25%. Bootstrapping was used to generate nonparametric receiver operating characteristic (ROC) curves and estimate area under the curve (AUC). The Youden index was used to identify UPCR cutoff values that maximize sensitivity and specificity. Positive predictive value (PPV) and negative predictive value (NPV) were calculated.
Results ROC curves were constructed for proteinuria measurements at baseline and 3, 6, 9, and 12 months after randomization (Figure 1). AUC increased from baseline to month 3 (0.64 vs. 0.80, P <0.001) and from month 3 to month 6 (0.80 vs. 0.84, P <0.01) but did not increase beyond month 6 (P >0.05 for each pairwise comparison). Achievement of 6-month UPCR <1 was 83.8% sensitive and 71.0% specific for CRR at 18 months and had PPV and NPV of 64.9% and 87.2%, respectively. Evaluation of lower 6-month UPCR cutoff values yielded improvements in specificity and PPV but marked decreases in sensitivity and NPV. In multivariate analysis, the addition of 6-month serum creatinine and percent change in UPCR from baseline did not result in meaningful increases in AUC compared with 6-month proteinuria measurement alone.
Conclusions Level of proteinuria at 6 months alone was predictive of CRR at 18 months in aggregated data from two phase III LN clinical trials. After 6 months of treatment, UPCR <1 had high sensitivity and NPV for CRR at 18 months. This cutoff might be used to prospectively identify patients who are unlikely to achieve complete response within 18 months on the initial therapy for LN. The impact of these findings on guiding treatment decisions outside the setting of randomized clinical trials requires further investigation.
Tamirou Lupus Sci & Med 2015.
Dall'Era Arthritis Rheumatol 2015.
Schindler Ann Rheum Dis 2016.
Rovin B Arthritis Rheumatol 2012.
Mysler E Arthritis Rheumatol 2013.
Disclosure of Interest M. Cascino Employee of: Roche/Genentech, T. Schindler Employee of: Roche/Genentech, L. Gomez Mendez Grant/research support from: Roche/Genentech, P. Brunetta Employee of: Roche/Genentech, L. Dragone Employee of: Roche/Genentech, M. Dall'Era: None declared, J. Garg Employee of: Roche/Genentech