Objective: To develop a measure of renal activity in systemic lupus erythematosus and use it to develop a renal response index.
Methods: Abstracted data from the medical records of 215 patients with lupus nephritis were sent to 8 nephrologists and 29 rheumatologists for rating. Seven nephrologists and 22 rheumatologists completed the ratings. Each physician rated each patient visit with respect to renal disease activity (none, mild, moderate, or severe). Using the most commonly selected rating for each patient as the gold standard, stepwise regression modeling was performed to identify the variables most related to renal disease activity, and these variables were then used to create an activity score. This activity score could then be applied to 2 consecutive visits to define a renal response index.
Results: The renal activity score was computed as follows: proteinuria 0.5-1 gm/day (3 points), proteinuria 0.5-1 gm/day = 3 points, proteinuria >1-3 gm/day = 5 points, proteinuria >3 gm/day = 11 points, [corrected] urine red blood cell count > = 5/hpf = 3 points, [corrected] urine white blood cell count > or = 5/hpf = 1 point. [corrected] The chance-adjusted agreement between the renal response index derived from the activity score applied to the paired visits and the plurality physician response rating was 0.69 (95% confidence interval 0.59-0.79).
Conclusion: Ratings derived from this index for rating of renal response showed reasonable agreement with physician ratings in a pilot study. The index will require further refinement, testing, and validation. A data-driven approach to create renal activity and renal response indices will be useful in both clinical care and research settings.