Background Few studies have addressed glucocorticoid (GC) withdrawal in lupus nephritis (LN). Yet, this remains a pivotal issue due to GC-induced side effects on the one hand, and to the risk of relapse on the other hand.
Objectives We reviewed the data of the Louvain Lupus Nephritis Inception Cohort (LOULUNIC) i): to determine the percentage of patients able to permanently or transiently stop GC; ii): to compare their baseline and follow-up characteristics to patients who never stopped; and iii): to assess the consequences of GC withdrawal.
Methods Ninety patients with new-onset biopsy-proven LN were included. All were under the care of the same senior physician (FAH) during follow-up. Clinical, pathological and biological data were extracted from our data base. The SLICC/ACR-DI was assessed at last visit. Unpaired t-tests, Mann-Whitney tests and ANOVA were used, as appropriate.
Results Out of 90 patients with incident LN, 43 (48%) ever stopped GC (group E), of which 32 permanently (group P). Median time to stop GC was 37 months. 47 patients (52%) never stopped GC (group N).
At baseline, serum creatinine, uP/C ratio, ISN/RPS classes, activity and chronicity indices did not differ between groups, nor did the mean initial dose of methylprednisolone (MP) (N: 28 mg/d; E: 32 mg/d; P: 31 mg/d), the use of IV MP pulses (82 and 77% in N and E groups, respectively) and of IV cyclophosphamide (81 and 77%, respectively).
During the first year, mean (SD) uP/C decreased statistically more in group E compared to group N (p=0.028 by ANOVA), with striking differences at month 3 (N: 1.73±1.87; E: 0.96±1.34; p=0.038 by unpaired t-test). This difference at month 3 was also noticed for group P patients (0.85±0.76; p=0.02 by unpaired t-test). Interestingly, the mean MP dose at month 3 was statistically higher in group E (19±8) and P (20±9) compared to group N (15±6) (p=0.005 by unpaired t-test).
At last follow-up, serum creatinine was statistically lower in E and P patients compared to N patients. Eight of the 11 patients from the T group suffered form a renal relapse, justifying restart of GC, after a median time of 30 months. Importantly, SLICC/ACR-DI was significantly lower in E and P patients, compared to N patients (p=0.0068 and 0.0027, respectively).
Conclusions In half of LN patients, complete GC withdrawal is achievable and in one third it can be maintained long term. As expected, patients able to stop GC display less damage at last followup. Patients who were able to stop GC decreased their proteinuria much more promptly during the first year of treatment. Interestingly, they received more GC within the first 3 months of therapy, thereby suggesting that a higher dose of GC during the first 3 months of treatment might be associated with a higher probability of later GC withdrawal.
Disclosure of Interest None declared