2045

HYDROXYCHLOROQUINE LEVELS DEFINE HIGH RATES OF NON ADHERENCE AND PREDICT WORSE RENAL OUTCOMES IN PATIENTS WITH LUPUS NEPHRITIS

Suceena Alexander1, Gary Chusney2, Vivienne D Chusney2, Janet Lee2, Tom Cairns1 and Liz Lightstone1.

1Imperial College Healthcare NHS Trust Lupus Centre and 2Leslie Brent Laboratory, HammersmithHospital, London, UK

Introduction: Non adherence (NA) to lupus nephritis (LN) treatment (Rx) is the most important factor for non-remission (NR) or renal relapses (RR). Hydroxychloroquine (HCQ) with a long terminal half-life of >40 days is ideally positioned to identify NA based on its blood levels. We report the utility of our newly developed HCQ assay in defining NA and poor outcomes in LN.

Aim: To identify non adherence defined as HCQBL ≤0.2mg/L and correlate with clinical outcomes in patients with ISN/RPSClass III, IV orV lupus nephritis, at baseline and at follow up.

Methods:154patients (pts) >18 years of age with ISN/RPS class III, IV or V LN and on HCQ for at least 3 months had HCQBLs measured during treatment. Complete remission(CR) was defined as serum creatinine (SCr) not more than 15%above baseline & urine protein:creatinine ratio (PCR) <50mg/mmol.Renal Relapse(RR) was defined as persistent increase of30% in proteinuria and/ or SCr requiring renal biopsy or increase/ change in immunosuppression. Persistent NA was defined as HCQBL ≤0.2mg/L on two separate patient clinic visits. The baseline of this study was the date of first HCQBL testing for each patient.

Results:Out of the 154 patients, 45.5% of patients were Indians and SE Asians. The baseline demographic and clinical variables were comparable between the non-adherent and adherent groups. The mean follow up period was 20.1±10.9 mths. The mean HCQBL dose was 5.4 ± 1.8mg/kg. 24.7% of patients were NA for HCQ and 17.3% (19/110) of them hadpersistent NA. CR was significantly greater in the adherent group than the NA group at baseline (65.5% vs. 36.8%, p <0.01) and at follow up (68.1%vs. 50%, p =0.04). HCQBL hadsignificant AUC in the ROC curve for detecting CR at baseline (AUC= 0.618±0.05, p =0.01). There was no significant association between HCQBLs and occurrence of renal relapses. Intra Class Correlation (ICC) / within-patient variability among adherent patients on HCQ 400mg/day (n = 96, total repeat measurements = 486) was 0.7. Spearman’s correlation with mycophenolic acid trough levels at baseline and follow up (n = 520) wassignificant (R2 =0.002, p <0.01).

Time to CR in all pts Time to CR in active ptsHCQBL at baseline

Conclusions:NA for HCQ was present in approximately25% of patients and persistent in 17.3%. NA is significantly associated with disease activity, fewer remissionsprolonged time to CR – the latter being a poor long term prognostic indicator with respect to renal outcomes. HCQBL had significant area under the curve to detect CR at testing. HCQBL had a significant but weak correlation with HCQ dose per kg. Key questions are whether NA with HCQ reflects more generalized NA with other medications and whether HCQ levels can be used prospectively to improve adherence and outcomes.