A prospective observational cohort study highlights kidney biopsy findings of lupus nephritis patients in remission who flare following withdrawal of maintenance therapy

One of the most difficult management issues in lupus nephritis (LN) is the optimal duration of maintenance immunosuppression after patients are in clinical remission. Most patients receive immunosuppression for years, based mainly on expert opinion. Prospective data are unavailable. Complicating this issue are data that patients in clinical remission can still have histologically active LN; however, the implications of this are unknown. To study this, the Lupus Flares and Histological Renal Activity at the end of Treatment study (ClinicalTrial.gov, NCT02313974) was designed to examine whether residual histologic activity predisposes to LN flares in class III and IV LN. Patients in complete clinical remission for at least 12 months who had received at least 36 months of immunosuppression were eligible. Patients consented to a second kidney biopsy, were tapered off maintenance immunosuppression and were then followed prospectively for LN flares over 24 months. Forty-four patients were enrolled, and 36 completed the study. LN flares occurred in 11 patients, and ten of these had residual histologic activity on the second biopsy. All patients with an NIH activity index over two flared. The activity index and duration of systemic lupus erythematosus at the second biopsy were independent predictors of flare. A predictive equation based on these variables discriminated between flare and no flare with a sensitivity of 100%, specificity of 88%, and a misclassification rate of 8.3%. Thus, a repeat kidney biopsy may be useful in managing maintenance immunosuppression in LN, and patients in histologic remission may be candidates for withdrawal of therapy. One of the most difficult management issues in lupus nephritis (LN) is the optimal duration of maintenance immunosuppression after patients are in clinical remission. Most patients receive immunosuppression for years, based mainly on expert opinion. Prospective data are unavailable. Complicating this issue are data that patients in clinical remission can still have histologically active LN; however, the implications of this are unknown. To study this, the Lupus Flares and Histological Renal Activity at the end of Treatment study (ClinicalTrial.gov, NCT02313974) was designed to examine whether residual histologic activity predisposes to LN flares in class III and IV LN. Patients in complete clinical remission for at least 12 months who had received at least 36 months of immunosuppression were eligible. Patients consented to a second kidney biopsy, were tapered off maintenance immunosuppression and were then followed prospectively for LN flares over 24 months. Forty-four patients were enrolled, and 36 completed the study. LN flares occurred in 11 patients, and ten of these had residual histologic activity on the second biopsy. All patients with an NIH activity index over two flared. The activity index and duration of systemic lupus erythematosus at the second biopsy were independent predictors of flare. A predictive equation based on these variables discriminated between flare and no flare with a sensitivity of 100%, specificity of 88%, and a misclassification rate of 8.3%. Thus, a repeat kidney biopsy may be useful in managing maintenance immunosuppression in LN, and patients in histologic remission may be candidates for withdrawal of therapy. Systemic lupus erythematosus (SLE) has a relapsing and remitting course, with patients experiencing episodic disease activity (flares) over time. Kidney biopsy plays an important role in the initial diagnosis and staging of lupus nephritis (LN). It also guides the appropriate selection of treatment, especially for high-risk patients.1Weening J.J. D'Agati V.D. Schwartz M.M. et al.The classification of glomerulonephritis in systemic lupus erythematosus revisited.Kidney Int. 2004; 65: 521-530Abstract Full Text Full Text PDF PubMed Scopus (1113) Google Scholar, 2Kidney Disease: Improving Global Outcomes (KDIGO) Glomerulonephritis Work GroupKDIGO clinical practice guideline for glomerulonephritis.Kidney Int Suppl. 2012; 2: 221-232Google Scholar The duration of maintenance treatment for International Society of Nephrology/Renal Pathology Society (ISN/RPS) class III and IV LN is currently based on the clinical evolution of the LN flare. Many guidelines recommend at least 3 years of total treatment and 1 year of complete remission before withdrawing treatment. These recommendations are based mainly on expert opinion, as there are few data to develop an evidence-based guideline. Protocol repeat biopsy studies after complete remission show continuing histologic activity in a significant number of patients.3Alvarado A.S. Malvar A. Lococo B. et al.The value of repeat kidney biopsy in quiescent argentinian lupus nephritis patients.Lupus. 2014; 23: 840-847Crossref PubMed Scopus (62) Google Scholar, 4Malvar A. Pirruccio P. Alberton V. et al.Histologic versus clinical remission in proliferative lupus nephritis.Nephrol Dial Transplant. 2017; 32: 1338-1344Crossref PubMed Scopus (104) Google Scholar, 5Alsuwaida A. Husain S. Alghonaim M. et al.Strategy for second kidney biopsy in patients with lupus nephritis.Nephrol Dial Transplant. 2012; 27: 1472-1478Crossref PubMed Scopus (53) Google Scholar, 6Alsuwaida A. The clinical significance of serial kidney biopsies in lupus nephritis.Mod Rheumatol. 2014; 24: 453-456Crossref PubMed Scopus (13) Google Scholar, 7Zickert A. Sundelin B. Svenungsson E. et al.Role of early repeated renal biopsies in lupus nephritis.Lupus Sci Med. 2014; 1: e000018Crossref PubMed Scopus (76) Google Scholar Stopping maintenance immunosuppression in such patients may theoretically put them at risk of renal flare. Management of maintenance therapy is even more uncertain in patients with stable partial renal remission and no extrarenal lupus activity. Such patients often have ongoing proteinuria, and thus continue to receive immunosuppression indefinitely. Repeat biopsies in such patients3Alvarado A.S. Malvar A. Lococo B. et al.The value of repeat kidney biopsy in quiescent argentinian lupus nephritis patients.Lupus. 2014; 23: 840-847Crossref PubMed Scopus (62) Google Scholar, 4Malvar A. Pirruccio P. Alberton V. et al.Histologic versus clinical remission in proliferative lupus nephritis.Nephrol Dial Transplant. 2017; 32: 1338-1344Crossref PubMed Scopus (104) Google Scholar, 8Condon M.B. Ashby D. Pepper R.J. et al.Prospective observational single-centre cohort study to evaluate the effectiveness of treating lupus nephritis with rituximab and mycophenolate mofetil but no oral steroids.Ann Rheum Dis. 2013; 72: 1280-1286Crossref PubMed Scopus (307) Google Scholar have shown that many have no histologic activity and are in histologic remission. Persistent proteinuria may be from past injury and scarring, so continuing treatment may put such patients at risk for infectious morbidity with little benefit for the LN.2Kidney Disease: Improving Global Outcomes (KDIGO) Glomerulonephritis Work GroupKDIGO clinical practice guideline for glomerulonephritis.Kidney Int Suppl. 2012; 2: 221-232Google Scholar, 9Pagni F. Galimberti S. Goffredo P. et al.The value of repeat biopsy in the management of lupus nephritis: an international multicentre study in a large cohort of patients.Nephrol Dial Transplant. 2013; 28: 3014-3023Crossref PubMed Scopus (43) Google Scholar We suggest that a repeat biopsy in LN patients on long-term maintenance therapy who have been in complete renal remission for at least 1 year may help guide the withdrawal of maintenance immunosuppression. We postulated that patients with remaining histologic activity will have a greater tendency toward LN flares than those with no remaining activity, and tested this hypothesis prospectively in the Lupus Flares and Histological Renal Activity at the End of the Treatment (LuFla) study (ClinicalTrial.gov identifier: NCT02313974). The overall study design and patient flow of LuFla is shown in Figure 1. LuFla recruited 44 patients, and 36 completed the study. All patients were Hispanic and white, 25 (83%) were female, and the average age of the cohort at biopsy 1 was 31.6 ± 11.3 years. Table 1 presents the demographic, clinical, and histologic characteristics of the patients at biopsy 1 segregated by renal flare status after stopping immunosuppression.10Levey A.S. Stevens L.A. Schmid C.H. et al.A new equation to estimate glomerular filtration rate.Ann Intern Med. 2009; 150: 604-612Crossref PubMed Scopus (16096) Google Scholar Patients who experienced flare were comparable to patients who did not in all respects except for having a longer overall duration of SLE. There were fewer males in the flare group, but this did not reach significance. LN histologic classes were distributed as follows: IV A (n = 7), IV A/C (n = 16), III A (n = 7), III A/C (n = 6), and 4 of the class III A patients had concomitant class V LN.Table 1Clinical and histologic findings at biopsy 1VariableEntire cohort (n = 36)Flare group (n = 11)No-flare group (n = 25)P valueaCalculated between the flare and nonflare groups by the nonparametric Mann-Whitney test for continuous data and Fisher’s exact test for proportions.Age (yr)31.6 ± 11.330.0 ± 6.532.0 ± 130.92% male16.79.1200.64Duration of SLE (mo)54 (1-240)120 (12–240)48 (1–240)0.03% with prior history of renal flare33.355240.12Proteinuria (g/d)2.1 (0.2–20)2.1 (0.17–4.6)2.0 (0.3–20)0.66SCr (mg/dl)30.78 (0.46–2.80)0.74 (0.55–1.00)0.80 (0.46–2.80)0.51eGFR (ml/min per 1.73 m2)bCalculated by the CKD-EPI formula (Levey et al.10).103 (20–143)109 (86–122)99 (20–143)0.39C3 (mg/dl)81 (25–140)89 (40–118)68 (25–140)0.36C4 (mg/dl)12 (0–32)12 (3–23)11 (0–32)0.78% low C352.827.364.00.07% low C469.472.768.01.00% anti-dsDNA–positive80.681.880.01.00Activity index8 (3–16)11 (4–16)7 (3–13)0.23Chronicity index3 (0–6)2 (0–4)3 (0–6)0.24C3, complement component 3; C4, complement component 4; dsDNA, double-stranded DNA autoantibody; eGFR, estimated glomerular filtration rate; SCr, serum creatinine concentration; SLE, systemic lupus erythematosus.Data are presented as mean ± SD, proportion (%) of patients, or median (range).a Calculated between the flare and nonflare groups by the nonparametric Mann-Whitney test for continuous data and Fisher’s exact test for proportions.b Calculated by the CKD-EPI formula (Levey et al.10Levey A.S. Stevens L.A. Schmid C.H. et al.A new equation to estimate glomerular filtration rate.Ann Intern Med. 2009; 150: 604-612Crossref PubMed Scopus (16096) Google Scholar). Open table in a new tab C3, complement component 3; C4, complement component 4; dsDNA, double-stranded DNA autoantibody; eGFR, estimated glomerular filtration rate; SCr, serum creatinine concentration; SLE, systemic lupus erythematosus. Data are presented as mean ± SD, proportion (%) of patients, or median (range). After a minimum of 36 months of immunosuppression and at least 12 months of clinical renal remission, a repeat kidney biopsy (biopsy 2) was performed. The clinical and histologic findings at biopsy 2 are summarized in Table 2. Overall, 20 patients (55.6%) achieved complete histologic remission with an activity index (AI) of 0. Nine patients (25%) had an AI of 1 or 2. The remaining 7 patients (19.4%) had an AI between 3 and 5. Despite complete clinical renal remission, persistent histologic activity was present in 16 patients (44.4%). The histologic components of the AI that were found in these patients were endocapillary proliferation in 13 (81%), subendothelial deposits in 14 (88%), and interstitial inflammation in 4 (25%). No patient had persistent glomerular crescents or necrosis.Table 2Clinical and histologic findings at biopsy 2VariableEntire cohort (n = 36)Flare group (n = 11)No-flare group (n = 25)P valueaCalculated between the flare and nonflare groups by the nonparametric Mann-Whitney test for continuous data and Fisher’s exact test for proportions.Duration of treatment (mo)38 (36–54)38 (36–48)38 (36–54)0.61Time to remission (mo)24 (12–40)24 (16–36)24 (12–40)0.75Duration of remission (mo)12 (12–30)12 (12–20)13 (12–30)0.43Proteinuria (g/d)0.11 (0.03–0.48)0.16 (0.06–0.48)0.07 (0.03–0.48)0.06SCr (mg/dl)bCalculated by the CKD-EPI formula.0.70 (0.50–1.12)0.66 (0.60–0.90)0.70 (0.50–1.12)0.70eGFR (ml/min per 1.73 m2)114 (81–135)114 (95–127)114 (81–135)0.85C3 (mg/dl)112 (55–188)100 (55–170)116 (64–188)0.19C4 (mg/dl)19 (3–51)15 (3–28)20 (6–51)0.20% low C313.927.38.00.15% low C436.145.532.00.47ΔC3cThe decline in complement C3 and C4 from 6 months before biopsy 2 to biopsy 2 (value at biopsy 2 minus value 6 months before biopsy 2).1 (–36 to 77)–7 (–30 to 26)10 (–36 to 77)0.07ΔC4cThe decline in complement C3 and C4 from 6 months before biopsy 2 to biopsy 2 (value at biopsy 2 minus value 6 months before biopsy 2).0 (–15 to 15)–3 (–13 to 11)0 (–15 to 15)0.42% anti-dsDNA–positive22.236.316.00.21Activity index0 (0–5)3 (0–5)0 (0–2)<0.0001% endocapillary proliferationdHistologic components of the activity index expressed as percentage of patients positive for the finding at biopsy 2.30.690.94<0.0001% subendothelial depositsdHistologic components of the activity index expressed as percentage of patients positive for the finding at biopsy 2.38.990.916<0.0001% glomerular leukoctyesdHistologic components of the activity index expressed as percentage of patients positive for the finding at biopsy 2.2545.5160.075Chronicity index3 (0–5)3 (0–4)2 (0–5)0.13C3, complement component 3; C4, complement component 4; dsDNA, double-stranded DNA autoantibody; eGFR, estimated glomerular filtration rate; SCr, serum creatinine concentration.Data are expressed as proportion (%) of patients, or median (range).a Calculated between the flare and nonflare groups by the nonparametric Mann-Whitney test for continuous data and Fisher’s exact test for proportions.b Calculated by the CKD-EPI formula.c The decline in complement C3 and C4 from 6 months before biopsy 2 to biopsy 2 (value at biopsy 2 minus value 6 months before biopsy 2).d Histologic components of the activity index expressed as percentage of patients positive for the finding at biopsy 2. Open table in a new tab C3, complement component 3; C4, complement component 4; dsDNA, double-stranded DNA autoantibody; eGFR, estimated glomerular filtration rate; SCr, serum creatinine concentration. Data are expressed as proportion (%) of patients, or median (range). After maintenance therapy was tapered and discontinued, LN flared in 11 patients (30.5%). The clinical findings at flare for these patients are shown in Table 3. All but 1 flare (91%) occurred in patients who had active histology at biopsy 2, and everyone with an AI > 2 experienced flare (Figure 2). Among the no-flare group, 6 patients (24%) had an AI of 1 to 2 on biopsy 2. In the entire cohort, the incidence of renal flare in patients who had an AI ≤ 2 at biopsy 2 was 13.8%.Table 3Clinical findings of patients who experienced flarePatientTime to renal flare (mo)aTime from stopping maintenance immunosuppression.Proteinuria (g/d)SCr (mg/dl)eGFRbCalculated by the CKD-EPI formula. (ml/min per 1.73 m2)C3 (mg/dl)C4 (mg/dl)Anti-dsDNA antibodies (present/absent)Urinalysis (RBC/hpf)cAll samples contained acanthocytes.Biopsy 2FlareBiopsy 2FlareBiopsy 2FlareBiopsy 2FlareBiopsy 2FlareBiopsy 2FlareFlare1180.431.400.660.501171291701462417AA4–8260.121.120.800.8597908246142PP12–14360.101.020.720.7411611110781147PP4–84150.070.650.600.80127102666756AP8–125120.061.060.901.001189495781912AP10–146180.482.000.840.60959513473115AA2–47120.4260.540.600.551101141241102412AA14–208210.160.330.610.681191141001071512PP30–509210.082.590.771.40107528326185.7PP70–8010210.301.470.600.671191141221282834AA8–101160.402.060.600.60114113555533AA10–20dsDNA, double-stranded DNA autoantibody; eGFR, estimated glomerular filtration rate; hpf, high-power field; RBC, red blood cells; SCr, serum creatinine concentration.a Time from stopping maintenance immunosuppression.b Calculated by the CKD-EPI formula.c All samples contained acanthocytes. Open table in a new tab dsDNA, double-stranded DNA autoantibody; eGFR, estimated glomerular filtration rate; hpf, high-power field; RBC, red blood cells; SCr, serum creatinine concentration. Clinical, serologic, and histologic findings at biopsy 2 are provided in Table 2. Although proteinuria decreased below 500 mg/d in all patients, the flare patients showed a trend (P = 0.06) toward more proteinuria than the no-flare patients in remission. Additionally, while not statistically significant, more patients who experienced flare were positive for anti–double-stranded DNA antibodies, had low C3 and C4 levels at biopsy 2, and showed a decline in C3 in the 6 months preceding biopsy 2 (P = 0.06). Proteinuria, change in C4, and anti–double-stranded DNA antibody status at biopsy 2 did not correlate with the presence or absence of persistent histologic activity in biopsy 2 (AI = 0 vs. AI > 0). The decline in C3 in the 6 months preceding biopsy 2 showed a tendency to associate with the AI (P = 0.073 by logistic regression), but its correlation was not strong (Spearman r = –0.20; P = 0.23). Chronicity at biopsy 2, measured by the chronicity index (CI), did not correlate with proteinuria at biopsy 2 (Spearman r = 0.003; P = 0.99). This lack of association was also observed when proteinuria was correlated with chronicity in the glomerular compartment (glomerulosclerosis plus fibrous crescents) or chronicity in the tubulointerstitial compartment (tubular atrophy plus interstitial fibrosis) (data not shown). By logistic regression, the AI at biopsy 2 was significantly associated with the odds of an LN flare within 2 years of stopping maintenance immunosuppression (P < 0.0001). If the AI was at least 1, the odds ratio for LN flare was 31.7 (95% confidence interval 3.3–300). For flare prediction an AI cutoff of 1 had a sensitivity of 91%, a specificity of 76%, and a misclassification rate of 16.7%. Similarly, when other variables were tested as potential predictors, only the duration of SLE (P = 0.0297) and the decline in C3 in the 6 months preceding biopsy 2 (P = 0.0386) were significant. Univariate predictors that were not significant are provided in Supplementary Table S2. Multivariable logistic regression models were developed for predicting future LN flare using the predictors found to be significant on univariate analysis (Table 4). The table contains the intercepts and slope coefficients for the presented models. Of these, model 4 is parsimonious and has the lowest misclassification rate and the highest area under the receiver operating characteristic curve. Using Model 4, flare is predicted if Y is greater than cutoff value C, where Y = 3.48 x (AI biopsy 2) + 4.58 x natural log (duration of SLE in months) – 24.83. A C value of 0 gave the maximum sum of sensitivity (100%) and specificity (88%) to discriminate between flare and no flare. The odds ratios for LN flare per unit change of AI and log (duration SLE) were 32.5 (95% confidence interval 4.4–1912) and 97 (95% confidence interval 3.8–5049), respectively. There was no relationship between AI and duration of SLE (Spearman correlation r = 0.023, P = 0.89), indicating that they can be viewed as independent predictors.Table 4Performance characteristics of logistic regression models to predict future LN flareModelPredictoraPredictors are defined as follows: 1 = AI at biopsy 2, 2 = log (duration of SLE), 3 = decline in complement C3 from 6 months before biopsy 2 to biopsy 2, 4 = endocapillary proliferation, 5 = subendothelial deposits, and 6 = glomerular leukocytes.Model P valuebBased on the likelihood ratio test.Misclassification ratecUsing predicted probability of 0.5 as the cutoff.AUCInterceptPredictor coefficient (P value)AI at biopsy 2Duration of SLEdNatural log duration (mo).ΔC3eDecline in complement C3 from 6 months before biopsy 2 to biopsy 2 (value at biopsy 2 minus value 6 months before biopsy 2).Endocap at biopsy 2SubendoGlom leuk at biopsy 211<0.00010.170.91–3.121.66 (<0.0001)220.0300.220.72–6.841.28 (0.030)330.0390.330.69–0.74–0.035 (0.039)41, 2<0.00010.0830.98–26.503.48 (<0.0001)4.58 (0.0014)51, 3<0.00010.0830.92–3.221.74 (<0.0001)–0.053 (0.078)62, 30.0060.250.81–8.251.60 (0.015)–0.043 (0.019)71, 2, 3<0.00010.0830.99–43.915.07 (<0.0001)8.01 (0.0013)–0.081 (0.07)84<0.00010.0560.94–3.185.26 (<0.0001)95<0.00010.140.90–3.093.68 (<0.0001)1060.0670.280.65–1.251.48 (0.067)112,4<0.00010.0560.99–29.075.07 (0.031)9.64 (<0.0001)AI, activity index; AUC, area under the receiver operating characteristic curve; Endocap, endocapillary proliferation on biopsy 2; Glom leuk, glomerular leukocyte infiltration at biopsy 2; SLE, systemic lupus erythematosus; Subendo, subendothelial deposits on biopsy 2.a Predictors are defined as follows: 1 = AI at biopsy 2, 2 = log (duration of SLE), 3 = decline in complement C3 from 6 months before biopsy 2 to biopsy 2, 4 = endocapillary proliferation, 5 = subendothelial deposits, and 6 = glomerular leukocytes.b Based on the likelihood ratio test.c Using predicted probability of 0.5 as the cutoff.d Natural log duration (mo).e Decline in complement C3 from 6 months before biopsy 2 to biopsy 2 (value at biopsy 2 minus value 6 months before biopsy 2). Open table in a new tab AI, activity index; AUC, area under the receiver operating characteristic curve; Endocap, endocapillary proliferation on biopsy 2; Glom leuk, glomerular leukocyte infiltration at biopsy 2; SLE, systemic lupus erythematosus; Subendo, subendothelial deposits on biopsy 2. Individual components of the AI listed in Table 2 were examined as possible independent predictors of LN flare. In univariate analysis, endocapillary proliferation and subendothelial deposits were both significantly associated with the odds of an LN flare, but endocapillary proliferation was more robust (Table 4). Combining endocapillary proliferation and duration of SLE led to a flare prediction model (model 11, Table 4) that reduced the misclassification rate and increased the AUC compared with model 4. Using model 11, flare is predicted if Y is greater than the cutoff value C, where Y = 9.64 x (endocapillary proliferation at biopsy 2) + 5.07 x natural log (duration of SLE in months) – 27.53. A C value of 0 gave the maximum sum of sensitivity (100%) and specificity (92%) to discriminate between flare and no flare. These data demonstrate that despite extensive and long-term immunosuppression, patients with LN who enter a complete clinical renal remission still have a high rate of relapse following withdrawal of maintenance immunosuppression. Patients prone to relapse cannot be identified a priori by clinical or demographic variables that are commonly collected during standard office visits. However, examination of kidney histology during treatment and after clinical remission provides information that can be used to predict who is likely to relapse and who is likely to remain in remission after immunosuppression is stopped. Almost every patient who developed an LN flare had persistent histologic LN activity, and every patient in this cohort with an activity index above 2 had an LN flare. These data suggest that a kidney biopsy should be considered when withdrawal of maintenance therapy is contemplated. The activity index of this biopsy, and more specifically the extent of endocapillary proliferation present in this biopsy, combined with the duration of a patient’s SLE may be used to predict who is or is not likely to experience a flare. Patients predicted not to experience a flare may be candidates for stopping treatment. Persistent histologic activity was found in 44% of this cohort. Previous studies have also shown that about 30% to 60% of LN patients have residual evidence of active inflammation on protocol biopsies performed during maintenance therapy after complete clinical remission.5Alsuwaida A. Husain S. Alghonaim M. et al.Strategy for second kidney biopsy in patients with lupus nephritis.Nephrol Dial Transplant. 2012; 27: 1472-1478Crossref PubMed Scopus (53) Google Scholar, 11Esdaile J.M. Joseph L. Mackenzie T. et al.The pathogenesis and prognosis of lupus nephritis: information from repeat renal biopsy.Semin Arthritis Rheum. 1993; 23: 135-148Abstract Full Text PDF PubMed Scopus (80) Google Scholar, 12Grootscholten C. Bajema I.M. Florquin S. et al.Treatment with cyclophosphamide delays the progression of chronic lesions more effectively than does treatment with azathioprine plus methylprednisolone in patients with proliferative lupus nephritis.Arthritis Rheum. 2007; 56: 924-937Crossref PubMed Scopus (97) Google Scholar, 13Moroni G. Pasquali S. Quaglini S. et al.Clinical and prognostic value of serial renal biopsies in lupus nephritis.Am J Kidney Dis. 1999; 34: 530-539Abstract Full Text Full Text PDF PubMed Scopus (114) Google Scholar The predominant persisting lesions found in our patients’ biopsies were endocapillary proliferation and subendothelial immune complexes. Cellular crescents and glomerular necrosis were not found, similar to in other repeat biopsy cohorts.4Malvar A. Pirruccio P. Alberton V. et al.Histologic versus clinical remission in proliferative lupus nephritis.Nephrol Dial Transplant. 2017; 32: 1338-1344Crossref PubMed Scopus (104) Google Scholar The endocapillary proliferation component of the AI predicted flare better than did the combined AI. At biopsy 1 the flare and no-flare patients were similar clinically and histologically, except the no-flare group had more patients with a low C3 level. Interestingly, 2 of the 3 flare patients who had low C3 at biopsy 1 continued to have a low C3 at biopsy 2, while all but 2 of the 16 no flare patients with low C3 at biopsy 1 had normal levels by biopsy 2. At biopsy 2 the flare group had somewhat more proteinuria than the no-flare group, and more patients in the flare group had a positive double-stranded DNA titer, low C3, and low C4. Additionally, in the 6 months preceding biopsy 2, the patients who experienced flare tended to show a decrease in C3, and the patients who did not tended to show an increase in C3. Taken together, and despite individual measurements not reaching statistical significance, these findings suggest that a low level of clinical activity persisted in the flare group. Despite this, no clinical measurement obtained around biopsy 2 was significantly associated with, or could robustly predict, the AI of biopsy 2, but this needs to be examined in a larger cohort. In our cohort, anyone with an AI of 1 or more on biopsy 2 had a high risk of future LN flare, and everyone with an AI of >2 experienced flare. Interestingly, a Middle Eastern cohort of LN patients who had protocol biopsies during maintenance therapy was found to have a poor (44%) 10-year kidney survival rate if the AI was >2 at re-biopsy, fair (80%) survival if the AI was 1 or 2, and 100% survival if the AI was 0.5Alsuwaida A. Husain S. Alghonaim M. et al.Strategy for second kidney biopsy in patients with lupus nephritis.Nephrol Dial Transplant. 2012; 27: 1472-1478Crossref PubMed Scopus (53) Google Scholar These findings suggest that complete histologic remission may be a target treatment goal for LN. This study had several limitations. The cohort was a relatively small and ethnically homogeneous population, and results may differ in an ethnically and racially diverse LN population. No sample size calculations were done, but the small P values for the slopes associated with AI and log (duration of SLE) and the lower confidence limit for the odds ratios for predicting future flare being away from 1 suggest the data are robust. The finding that duration of SLE was associated with future relapse may be due to the fact that patients with persistent disease activity have a longer follow-up duration. However, even among those with long disease duration, all but 1 patient who relapsed had persistent histologic activity. The study was under-powered to draw firm conclusions regarding proteinuria and serologic markers at biopsy 2. The duration of follow-up was limited. Patients with no histologic activity may flare, but it may take longer than 2 years. The principal investigator was not blinded to the results of the second biopsy, and this could have influenced flare diagnosis; however, for LN flares objective data were required, mitigating bias. The most important limitation was that LuFla did not demonstrate that continuation of maintenance therapy in patients with persistent histologic activity and complete clinical remission would have prevented future LN flares. This is a critical question and will need to be examined in a follow-up study, powered by the LuFla outcomes. In summary, the LuFla investigation supports the use of a kidney biopsy to help manage the duration of maintenance therapy for LN patients who have been in complete renal remission for at least 1 year and have had at least 3 years of immunosuppression. Persistent histologic activity on this biopsy, especially if sufficient to be labeled as a National Institutes of Health AI of >2, is associated with LN relapse. If patients have an activity index of 0, indicating complete histologic remission, withdrawal of immunosuppression may be reasonable. LuFla was a prospective observational cohort study of adult patients (ages 18–70) with class III/IV±V LN to assess the relationship of post-therapy LN flares to kidney histology found on biopsies performed immediately before maintenance therapy was tapered off. Written informed consent was obtained from all participants. The study was approved by the University Hospital Ethics Committee. The investigators and hospital ethics committee judged that clinical equipoise existed for LuFla because the consequences of stopping immunosuppression for patients lacking clinical evidence of LN