Objectives The study aimed to determine if serum albumin at 12 months predicts long-term renal outcome at 48 months. Data from the NYU SAMPLE (Specimen and Matched Phenotype Linked Evaluation) Lupus Registry were used to compare the performance of albumin, anti-double-stranded DNA, C3/C4, proteinuria and haematuria.

Methods 82 patients with SLE with data at time of renal biopsy, at 12 months and at a second visit, and up to 48 months were included. The significance of each biomarker as a predictor of an adverse renal outcome (ARO), defined as doubling of serum creatinine, as creatinine >4 mg/dL if initial >2.5 mg/dL or ESRD, was evaluated in univariate and exploratory multivariable Cox proportional hazards models. Hazard ratios (HRs) for ARO with 95% CIs were generated. The receiver operating characteristic (ROC) curves at 48 months were used to identify the optimal cut-off point for albumin and proteinuria to predict ARO. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated for albumin and proteinuria.

Results Serum albumin and proteinuria had statistically significant HRs for ARO (0.140 and 1.459, respectively). The model with both albumin and proteinuria indicated no additional independent contribution of proteinuria to albumin alone. The ROC curves identified cut-offs of 3.7 g/dL for albumin and 0.964 urine protein to creatinine ratio for proteinuria. Albumin had a sensitivity of 94%, specificity of 87%, PPV of 64% and NPV of 98%.

Conclusions This study demonstrates serum albumin >3.7 g/dL is a predictor of a favourable long-term renal outcome. These results support the inclusion of albumin as an outcome in lupus nephritis trials and treat-to-target guidelines.