Pilot study comparing the childhood arthritis and rheumatology research alliance consensus treatment plans for induction therapy of juvenile proliferative lupus nephritis

Details about these plans have been previously published [4]. In brief, for induction therapy of proliferative LN, CARRA CTPs recommend either intravenous (IV) cyclophosphamide (CYC) 500–1000 mg/m2 (max 1500 mg) every 4 weeks × 6 months (6–7 doses) or mycophenolate mofetil (MMF) 600 mg/m2/dose BID (ma× 3000 mg/day) in addition to one of three high-dose corticosteroid (steroid) CTPs. Steroid CTP options include primarily oral, primarily IV, or mixed oral/IV regimens. High-dose pulse IV methylprednisolone 30 mg/kg (max 1000 mg/dose) × 3 doses is recommended at the start of therapy in the primarily IV and mixed oral/IV steroid CTPs and optional in the primarily oral steroid CTP. Tapering schedules for prednisone or prednisolone are outlined for each steroid regimen. Use of mesna, anti-emetics, gonadotropin releasing hormone agonists, and antimicrobials for Pneumocystis jiroveci prophylaxis are at the provider’s discretion. Maintenance immunosuppression CTP options include MMF, azathioprine (AZA), or quarterly IV CYC in addition to low-dose prednisone or prednisolone with a goal to taper to ≤10 mg/day by 12 months and to ≤5 mg/day by 24 months from the start of induction therapy.

A multicenter prospective observational cohort study was conducted from May 2012 through October 2015. Patients at participating sites were enrolled in the CARRA registry and treated per the induction CTPs at the discretion of the pediatric rheumatology provider. Patients with complete or partial renal response at the 6-month visit were treated according to one of the three maintenance CTPs. Main study entry criteria included new diagnosis of biopsy-proven active proliferative LN (ISN-RPS class III or IV) with or without concurrent class V disease, fulfillment of ≥4 of 11 American College of Rheumatology revised classification criteria for SLE or presence of 3 criteria provided one is histological evidence of LN [5], age at diagnosis with SLE ≤ 16 years, and age at study enrollment ≤20 years. Exclusion criteria were: severe infection, pregnancy or lactation, presence of another chronic or genetic disease or organ involvement that significantly influenced treatment of LN, and treatment with MMF or CYC not indicated per provider.

Study visits occurred at baseline and 3, 6, 9, 12, 18, and 24 months from the start of induction therapy. Standard-of-care clinical and laboratory data were captured at each visit. Data was collected using standardized case report forms through the InForm electronic data capture system managed by the Duke Clinical Research Institute. Patients or guardians were consented for data collection through the Legacy CARRA registry. The Legacy CARRA registry general protocol and consent was approved by Duke University institutional review board (IRB) and all participating site IRBs. Because the CTP study is not interventional and patients receive standard-of-care treatment at the discretion of their provider, only consent for data collection as a participant in the CARRA registry was required.

The provider’s reasons for CTP selection were assessed using standardized responses (Table 1) with the ability to select multiple reasons. Reasons for induction immunosuppression and steroid CTP selection were assessed separately at baseline. Reasons for maintenance immunosuppression CTP selection were assessed in responders at the 6-month visit.

Adherence to induction immunosuppression and steroid CTP regimens was assessed by medication log and provider-report. Medications used during the study period were recorded at every visit. Overall adherence to the induction CTPs was evaluated by providers at the 3- and 6-month visits by asking whether the CTP had been followed as intended. Reasons for not following a CTP were assessed by multiple choice with ability to select multiple reasons: patient non-adherence, patient-reported intolerance, physician drug adjustment due to intolerance, adverse event, disease flare, lack of response, laboratory abnormality, pregnancy, and other.

Renal response was assessed by providers at the 6-month visit. Responder criteria were established as part of the previously published CARRA LN CTP and adapted from the 2006 ACR response criteria for proliferative and membranous renal disease in SLE clinical trials [4, 6]. Complete renal response (CR) was defined as normalization of estimated glomerular filtration rate (GFR), inactive urine sediment (< 5 WBC/hpf, < 5 RBC/hpf, and no cellular casts) and spot urine protein-to-creatinine ratio (UPCR) < 0.2 mg/mg. Partial renal response (PR) was defined as at least 50% improvement in two core renal parameters (GFR, urinary sediment, proteinuria), maximum UPCR of < 1.0 mg/mg, and no clinically relevant worsening of the remaining renal core parameters. Laboratory measures of renal function were collected at every visit. GFR was estimated using the modified Schwartz formula [7].

The Systemic Lupus Erythematosus Disease Activity Index 2000 (SLEDAI-2 K) score was reported at baseline and at each follow-up visit [8]. Providers assessed whether the patient had experienced a disease flare since the previous visit and whether the flare was renal or non-renal. Specific flare criteria were not provided.

Adverse Events (AEs) were graded using the Common Terminology Criteria for Adverse Events (CTCAE v4.0) [9]. AEs of grade two and higher and serious adverse events were recorded at each study visit. Serious AEs were defined as death, life-threatening, hospitalization, disability or permanent damage, congenital anomaly or birth defect, or event that does not fit the defined outcomes but may require intervention to prevent one of the defined outcomes.

This was not a randomized study and comparisons of baseline characteristics between CTP groups were performed using Chi-square test, Fisher’s exact test, and Wilcoxon rank sum tests to evaluate for possible biases impacting CTP selection. To quantify deviation from the oral steroid tapers during induction therapy, the difference between the expected dose recommended per the chosen CTP and the reported daily dose was calculated for each patient. An average percent daily difference for each week of induction therapy was generated. Deviations from the IV pulse component of the primarily IV and mixed steroid CTPs were calculated similarly by taking the difference between the expected (per CTP) and recorded number of pulses in the medication log.

Exploratory analyses on clinical outcomes were performed using multivariate logistic regression and mixed effect models for repeated measures. The impact of induction immunosuppression treatment (CTP) on the renal response at the 6-month visit was evaluated by multivariate logistic regression analyses with adjustment of baseline characteristics including age (years), proteinuria (mg/dL), class of proliferative LN (III, IV), steroid CTP regimen (primarily IV, mixed IV/oral, primarily oral). Colinearity of continuous covariates were examined. Differences in longitudinal outcomes of GFR, proteinuria, and SLEDAI-2 K between induction immunosuppression CTPs were assessed using mixed models with repeated measures with adjustment for baseline characteristics including age, gender, time of scheduled visits, steroid CTP regimen, and baseline values of these outcomes. Study treatment (CTP) was considered a fixed effect and subjects were considered random effects. Missing data points were considered missing at random. Multiple variance structures were explored such as unstructured and spatial power. If convergence was reached with multiple covariance structures, standard goodness-of-fit measures were used to select the model with the best fit. Statistical analyses were conducted using STATA® 14.0 (StataCorp LLC) and SAS® 9.3 (SAS Institute Inc.). All tests are two-sided. P-values were not adjusted for multiple comparisons. Tests with p-values of < 0.05 were considered statistically significant.

Eighty-five patients were screened with 41 patients ultimately enrolled at 10 CARRA sites. The most common reasons for not participating were failure to meet inclusion criteria (66%) and provider decision not to use a CTP to guide treatment (20%). Baseline demographic and clinical characteristics are shown in Table 2. Significantly more patients in the CYC group had class IV LN (79% vs. 35%, p = 0.005) and hematuria (96% vs. 47%, p = 0.001) compared to the MMF group.

All patients completed at least 6 months of follow-up. Retention declined over time with 35 (85%) and 18 (44%) patients completing the 12- and 24-month visits respectively. Overall, 60% of visits occurred within four weeks before or after the target visit date.

CYC was selected for 24 (59%) patients and MMF for 17 (41%) patients (Fig. 1). Most sites used both regimens (Fig. 2). The most common reasons for selecting CYC were “This is what I or my group always does” (54%) and “I think this treatment works best” (54%). Concern for patient non-adherence was the rationale for initiating CYC for 8 patients. The most common reason for MMF selection was “This is what I or my group always does” (41%).

Of the three steroid CTPs, the mixed regimen was the most commonly used (n = 17, 41%), followed by primarily IV (n = 15, 37%), and primarily oral (n = 22%). Several sites used only one regimen (Fig. 2). The most common reasons for CTP selection were: “I always select this regimen” (47%) and “This steroid regimen works best” (47%) for the mixed group, “This steroid regimen works best” (80%) for the primarily IV group, and “My patient prefers oral medications” (33%) for the primarily oral group. IV-based steroid CTPs (primarily IV and mixed) were more frequently used in conjunction with CYC (p = 0.002).

Per medication logs, adherence to the immunosuppression CTPs was acceptable. In the MMF group, 84% and 86% of patients were at the target dose of ≥600 mg/m2 BID by the 3- and 6-month visits respectively, In the CYC group, 63% received the expected number of 6 or 7 infusions; the median number of infusions was 6 (IQR 5–6). The median cumulative CYC dose was 6290 mg (IQR 5040-8700). The median number of infusions was 6 (IQR 5–6) with a median monthly dose of 1100 mg/m2 (IQR 849–1256).

Providers reported the immunosuppression CTPs were followed as intended in 76% of patients at the 3-month visit and 64% at the 6-month visit. The most common reason for not following a CTP as intended was patient non-adherence (17%). Although many providers reported not following a CTP, the vast majority (95%) of patients stayed on the initially selected treatment during the first 6 months (Fig. 3). Two patients switched therapy; one switched to MMF after the first CYC infusion due to an allergic reaction and another switched to CYC from MMF due to patient non-adherence. Two patients were treated with additional immunosuppression during the induction period. Concurrent medications are described in Additional file 1.

Oral steroid and IV pulse exposure through week 24 was highly variable, indicating poor adherence to the steroid CTPs (Table 3). For the primarily IV and mixed groups, there was a tendency to prescribe fewer IV pulses than outlined in the CTPs however, a substantial number of patients (n = 22) had incomplete IV records and were excluded from the IV analysis.

Providers reported adhering to the steroid CTPs in 68% of patients at 3 months and just 37% of patients at 6 months. Reasons for not following a steroid CTP were similar across the regimens; the most common reasons were patient non-adherence (22%) and other (17%). Review of free-text responses revealed a theme of tapering steroids more quickly than recommended.

The median prednisone or prednisolone dose at 24 weeks for responders was 12 mg/day (IQR 10–20) or 0.2 mg/kg (IQR 0.2–0.3). Of patients with complete tapering data at 12 months, 74% were in alignment with the CTP tapering goal of ≤10 mg/day (median 7.7 mg/day or 0.2 mg/kg/day, IQR 0.1–0.2). By 24 months, 78% were on a dose of ≤5 mg/day (median 3.4 mg/day or 0.1 mg/kg/day, IQR 0–0.1).

Provider assessment of renal response (CR, PR) was confirmed by laboratory values in 24 of 41 (59%) patients. However, we were unable to corroborate the provider’s assessment in 17 patients due to: missing laboratory data (n = 9) and inconsistency between the laboratory values and the reported response (n = 8). To conservatively estimate the proportion of patients achieving renal response (CR or PR) at the 6-month visit using only the reported laboratory data, we counted the nine patients with missing data as non-responders, resulting in a CR rate just above 40% for both CYC (10/24, 42%) and MMF (7/17, 41%) groups. The total proportion of responders (CR or PR) in the CYC group was 63% (15/24) and 53% (9/17) in the MMF group, p = 0.54. Courses of non-responders are summarized in Additional file 2.

Median GFR, proteinuria, and SLEDAI-2 K scores over the course of the study are shown in Fig. 4. Exploratory analyses evaluating the effect of induction immunosuppression CTP (CYC vs. MMF) on outcomes of proteinuria, GFR, and SLEDAI-2 K over the study period were conducted using mixed effects models. No significant differences were found between CYC and MMF groups and GFR, proteinuria, or SLEDAI-2 K over time.

Of the 30 patients with a CR or PR at month 6, four patients experienced disease flare (2 renal flares) by month 24; all four patients were on MMF at the time of flare.

AEs are summarized in Table 4. Two serious adverse events were reported during the 6-month induction period; one patient was hospitalized for depression and suicidal ideation and one patient developed an opportunistic infection. Study is available in Additional file 3.