Background Recent post-hoc analyses of patients participating in a trial of abatacept (ABA) vs. placebo, in combination with mycophenolate mofetil and corticosteroids for treatment of active lupus nephritis (IM101075) have provided an evidence-based rationale for choosing among alternative definitions of complete renal response (CRR) used in recent Phase 2/3 clinical trial in lupus nephritis.1
Objectives The objective of this analysis was to characterize the relationship between ABA exposure and CRR using these alternative definitions to guide the dose selection for a follow-on Phase 3 study.
Methods In IM101075, patients with biopsy-proven active, proliferative lupus nephritis were randomized to receive placebo (n=100) or one of two intravenous ABA regimens: 30 mg/kg on Days 1, 15, 29 and 57 followed by ~10 mg/kg every 4 weeks (30/10; n=99), or ~10 mg/kg for the entire 12 months (10/10; n=99). Pharmacokinetic (PK) data collected in the study were analyzed by population PK analysis, and were used to determine various summary measures of exposure (Cavg and Cmin over Month 1and 3, Cmin at Month 12 based on clearance observed at Month 1) in each subject. We assessed rates of CRR at 12 months according to 3 sets of criteria, from 1) an ongoing National Institutes of Health trial of ABA (Abatacept and Cyclophosphamide Combination: Efficacy and Safety Study [NCT00774852]), 2) the Aspreva Lupus Management Study trial of mycophenolate mofetil (NCT00377637), and 3) the Lupus Nephritis Assessment with Rituximab (LUNAR) trial of rituximab2. The probability of achieving a CRR at 12 months was described by a logistic regression exposure-response (E-R) model. Age, weight, sex and baseline urine protein creatinine ratio (UPCR) were assessed as covariates in the model. LUNAR criteria were selected for further model evaluation based on largest observed difference (CRR of ~ 6% in placebo group, compared to ~20-24% in ABA group). The final model using CRR (assessed by the LUNAR criteria) was applied to predict efficacy in a proposed Phase III trial.
Results The average model-predicted minimum concentration of ABA over the first month (Cmin (1 month)) was a significant predictor of CRR at Month 12. Higher baseline UPCR was also identified as a positive predictor of CRR at Month 12. There was a strong relationship between Cmin (0–1 month) and the probability of achieving CRR in nephrotic (UPCR >339 mg/mmol) and non-nephrotic (UPCR ≤339 mg/mmol) pts. An association was identified between proteinuria and exposure suggesting that ABA is cleared more rapidly in nephrotic subjects and that ABA exposure increased over time in some nephrotic subjects suggesting that reduction in proteinuria may reduce clearance over time. Model evaluation shows reasonable agreement between model-predicted and observed response rates predicting that there may be incremental benefit from induction with 30 mg/kg of ABA for 57 days. This regimen provides the most benefit compared to placebo over time.
Conclusions CRR based on LUNAR criteria shows a positive E-R relationship for ABA in the treatment of lupus nephritis and identifies proteinuria as a confounder in dose selection and a target of treatment effect. This supports further evaluation of ABA and dose selection for the treatment of lupus nephritis in an ongoing Phase III study.
Wofsy D, et al. Arthritis Rheum 2012;64:3660–65.
Rovin B, et al. Arthritis Rheum 2012;64:1215–26.
Disclosure of Interest: S. Suryawanshi Employee of: Bristol-Myers Squibb, M. Tagen Employee of: Bristol-Myers Squibb, B. Murthy Employee of: Bristol-Myers Squibb, J. Hillson Employee of: Bristol-Myers Squibb, A. Roy Employee of: Bristol-Myers Squibb