Medical records of 55 patients (age <18 years) of Indian ethnicity treated in our nephrology clinic with a histological diagnosis of lupus nephritis from Jan-2000 to Dec-2007 were reviewed [5]. Patients were excluded if they had Class-I renal histology (n=1) and did not have minimum 5-year follow-up (n=20), leaving behind 34 patients available for analysis.

The incidence rate of flares was calculated as the number of flares per person-year. Univariate logistic regression was used to evaluate risk factors for being refractory to treatment. Kaplan Meier estimates and proportional hazards models were used to assess the effect of flares and refractory illness on renal outcome.

The median (IQR) follow-up period was 7 (5.8, 8.0) years. Induction therapy resulted in complete response in 24 (70.6%) and partial response in 8 (23.5%) patients. Of 32 responders, 14 (42.4%) had at least one episode of renal flare; >1 flares were present in 10 (31.2%) (Table I). The median duration to onset of first flare was 29 months. There were a total of 36 renal flares (36% nephritic), with an incidence rate of 0.16 flares/person/year. Male gender (P=0.002), presence of tubulointerstitial fibrosis (P=0.014) and nephrotic syndrome at presentation (P<0.001) were associated with higher incidence of flares, while hydroxychloroquine therapy (P=0.039) was associated with reduced incidence of flares (Table II).

TABLE I	Characteristics at Presentation with Lupus Nephritis in Patients with or without Renal Flares

TABLE II	Incidence Rate of Flares in Relation to Patient Characteristics at Presentation

Eight (23.5%) patients (two at end of induction phase, six after a flare) had refractory illness. The median time to diagnosis of refractory illness was 3.9 years. The presence of tubulointerstitial fibrosis was associated with increased risk of refractory illness (OR=13.22 [1.39-124.91]; P=0.02). The presence of Class IV nephritis, type of maintenance therapy- MMF/AZA and proteinuria at onset failed to show significant association. Eight patients (23.5%) developed an adverse renal outcome. Five-year patient and renal survival were 93% and 79%, respectively. Refractory illness (hazard ratio, HR=11.52 [2.30-57.56], P=0.003) and occurrence of multiple flares (HR=11.18 [1.29-96.97]; P=0.03) were associated with adverse renal outcome (Fig. 1), while occurrence of single renal flare was not.

Fig. 1 Renal survival of patients with lupus nephritis in relation to presence of (a) refractory illness and (b) multiple flares.

About two-fifths of the patients of lupus nephritis in our series experienced at least one episode of renal flare and about quarter of the patients were considered refractory to conventional management. Patients with multiple flares and refractory illness were at increased risk of adverse renal outcome.

This was a retrospective study and has all the limitations of such design. Although we evaluated the associations between various clinical characteristics and different outcomes, the study could not account for potential confounding factors using multivariate models due to the small sample size. As the study included patients over a long period of time, variations in therapeutic practices (hydroxychloroquine treatment becoming a standard practice) and diagnostic criteria over time might have affected the results of the study.

The incidence of renal flares reported in this study is similar to studies in Caucasian children [9,10]. However, this is lower than the 63% reported in a study by Srivastava, et al. [11] in Indian children. But the incidence rate of 0.09 flares/patient/year in their study was lower than that reported in the present study suggesting that the incidence of multiple flares might have been less in their cohort. Flares were associated with male gender, presence of tubulointerstitial fibrosis and nephrotic range proteinuria which have been linked to severe disease and/or adverse outcomes [12,13]. Other pediatric studies report non-response rates of 10-29% that are comparable to the 23% reported in the current study [4,13]. Our finding that tubulointerstitial fibrosis relates to therapy refractoriness suggests irreversible kidney injury might play an important role in prognosis as reported in other studies [14]. Similar to our study, Gibson, et al. [4] showed that non-response to standard treatment predicted end-stage renal disease. Our findings that multiple renal flares are associated with adverse outcome concur with those of Parikh, et al. [15], who showed that poor renal outcomes are associated with the number of flares per year and the time spent in flare.

The present study adds to the information on the disease course and medium term outcomes in pediatric lupus nephritis, and their determinants. Our findings underscore the role of multiple disease relapses and subsequent non-response to therapy in predisposing to advanced kidney disease, and emphasize the importance of vigilance for disease relapses and the role of effective immunosuppressive strategies in maintaining disease remission.

Contributors: All authors have contributed to study concept and design, drafting of manuscript, and its critical revision; JG, KPS, AS, AB: acquisition, analysis, or interpretation of data; JG, KPS, AS: statistical analysis;　 AS, PH, AKD, AB: Administrative, technical, or material support; AS, AB: study supervision.

Funding: None; Competing Interest: None stated.