Abstract
Objective To evaluate the ability of the Renal Activity Index for Lupus (RAIL) score, a urine biomarker-derived score, to capture and predict the course of active lupus nephritis (LN) in adult patients.
Methods Available serial urine samples collected up to week 52 from a subset of adults with active biopsy-proven proliferative LN participating in the double-blind randomized ALLURE trial of abatacept (ClinicalTrials.gov: NCT01714817) were used to calculate RAIL scores from creatinine-adjusted urine biomarkers (neutrophil gelatinase-associated lipocalin [NGAL], kidney injury molecule 1 [KIM-1], monocyte chemotactic protein 1 [MCP-1], adiponectin, hemopexin, ceruloplasmin). Discriminative performance of RAIL scores alone over time were compared with urine protein/creatinine ratio (UPCR), kidney function (estimated glomerular filtration rate [eGFR]), and mixed model analysis of RAIL score adjusted for baseline UPCR, eGFR, age, weight, sex, and race, with comparisons by renal response states including complete renal response (CRR), partial renal response but not CRR (PRR-only), and nonresponse (NR).
Results The analysis included 240 patients who contributed 599 samples. At weeks 12/24/52, there were 44/22/15 patients with PRR-only, 27/33/18 with CRR, and 127/61/15 NR. RAIL scores, eGFR, and UPCR improved over time irrespective of abatacept use, but were significantly lower with CRR compared to NR. The eGFR alone had poor accuracy (area under the receiver-operating characteristic curve [AUC] < 0.51) to discriminate renal response. Only after correction of baseline UPCR and eGFR, the RAIL score had excellent accuracy to reflect CRR from other renal response states at the current (AUC = 0.83-0.84) and next visit (AUC = 0.84-0.85) and performed better than UPCR; without correction, UPCR and RAIL score had similarly good accuracy.
Conclusion RAIL scores identify active LN and longitudinally predict the course of adult LN. (ClinicalTrials.gov: NCT01714817)
As a complex, multisystem, chronic autoimmune disease with diverse phenotypes, systemic lupus erythematosus (SLE) can be challenging to diagnose and predict its course. Lupus nephritis (LN) is a common, severe organ manifestation of SLE and a leading cause of morbidity and mortality.1,2 In adult cohorts, LN affects up to 40% of patients with SLE, with a higher prevalence observed in women, individuals of lower socioeconomic status, and with African American, African Caribbean, Asian race, and Hispanic ethnicity.1,3-6
Challenges in treating LN include a lack of noninvasive, accurate biomarkers to assess disease activity reliably, evaluate treatment response, and predict worsening in patients with LN.7,8 Baseline kidney biopsies are the gold standard for LN diagnosis and are used to guide initial treatment decisions, based on the International Society of Nephrology/Renal Pathology Society (ISN/RPS) classification and the extent of kidney damage.1,9,10 Disease course assessment with repeat kidney biopsies is less established, as biopsies are associated with patient risk given the invasiveness of the procedure, and they are costly, decreasing feasibility for routine surveillance in clinical practice, and especially in clinical trials.1,11,12 Repeat kidney biopsies are best to characterize a treatment’s effect on the kidneys, but there is no consensus on timing for postbaseline biopsies, and there remain discrepancies between the degree of structural injury demonstrated on kidney biopsy vs in functional markers.1,13-15
Currently, the presence and degree of proteinuria are among the key variables used to diagnose LN clinically, assess the degree of kidney inflammation, and monitor treatment response.1,16 However, proteinuria can also occur with preexisting kidney disease, hypertension, and LN damage, although some patients with marked kidney inflammation only have low levels of proteinuria.1,16 Another LN laboratory measure is kidney function (estimated glomerular filtration rate [eGFR]), which can vary with hydration status, concurrent tubular injury, and kidney damage.17 Further, conventional SLE biomarkers such as complement proteins and anti-dsDNA cannot differentiate active kidney disease from damage.11,18 Given that immune effectors mediating LN pathogenesis, including autoantibodies, complement proteins, inflammatory cytokines, immune mediators, and leukocytes, have been detected in urine samples from patients with LN,12 assessment of biomarker concentrations in the urine may provide a noninvasive tool for estimating the degree of LN activity and the associated kidney pathology in patients with LN.
Histological findings in kidney biopsies with LN vary in degree and type of mesangial, tubular, glomerular, and interstitial pathology, making it unlikely that a single biomarker measured in the urine suffices to estimate LN activity. Thus, the Renal Activity Index for Lupus (RAIL) was developed as a scoring system informed by the urine concentrations of 6 urine biomarkers with strong biological plausibility.19,20 RAIL biomarkers are neutrophil gelatinase-associated lipocalin (NGAL), monocyte chemotactic protein 1 (MCP-1), ceruloplasmin, adiponectin, hemopexin, and kidney injury molecule 1 (KIM-1).19 Higher RAIL scores reflect higher active inflammation in LN. Additionally, changes in pediatric patients’ RAIL scores correlate with response to treatment and LN flare.19,20 Two different RAIL algorithms have been developed: one for adults and one for pediatric patients with LN.21 When used in pediatric patients, the pediatric RAIL scores were able to predict US National Institutes of Health Activity Index (NIH-AI) scores with > 92% accuracy and Tubulointerstitial Activity Index scores with > 80% accuracy.19,21 When prospectively validated in an adult LN cohort, the RAIL has been found to accurately discriminate the level of LN activity as assessed by NIH-AI scores.21 Further, improvement of RAIL scores in pediatric patients with LN captures and even precedes clinical response of LN by at least 12 weeks, with > 90% accuracy.22 Although there is initial evidence that the RAIL also captures response to LN in adults, the ability of the RAIL biomarkers to monitor therapeutic response over time or anticipate the course of LN has not yet been well investigated in adults.23
Abatacept, a fusion protein of the Fc region of IgG1 and the extracellular domain of cytotoxic T lymphocyte-associated protein 4 (CTLA-4), prevents T cell activation. This medication has been approved for several rheumatic diseases, including rheumatoid arthritis and juvenile idiopathic arthritis. The potential benefits of abatacept for the treatment of LN were studied in 2 phase II trials (ClinicalTrials.gov: NCT00430677 and NCT00774852).24-26 Findings of these trials prompted the phase III ALLURE trial (NCT01714817) to further evaluate the benefits of abatacept when used for the treatment of LN.27-29
In this exploratory biomarker study, we used data and available samples collected during the ALLURE trial to investigate the ability of RAIL to differentiate LN treatment responders from nonresponders (NR) in adult patients, as well as to predict future kidney disease flare, under consideration of abatacept exposure.
METHODS
Study design and patients. The ALLURE trial (NCT01714817) was a 104-week, phase III randomized, double-blind, placebo-controlled study to evaluate the efficacy and safety of abatacept in patients with active class III or IV LN, with or without class V.27-29
Upon enrollment, eligible patients were randomized 1:1 to receive monthly infusions of either placebo or abatacept together with stable background treatment that included mycophenolate mofetil (MMF) and corticosteroids, which could be tapered during the study. The abatacept group received 30 mg/kg (based on body weight at baseline) of intravenous abatacept on days 1, 15, 29, and 57. This was followed, starting at week 12, by a weight-tiered monthly dose of abatacept at 500 mg, 750 mg, or 1000 mg for patients weighing < 60 kg, 60-100 kg, and > 100 kg, respectively, throughout the remainder of the study. The primary endpoint was complete renal response (CRR) measured at week 52, and secondary endpoints included the achievement of partial renal response (PRR) at week 52.
Eligible patients were > 18 years of age, had a kidney biopsy within 12 months of study enrollment screening that indicated the presence of active proliferative LN as per the 2003 ISN/RPS classification,30 had evidence of active LN within 3 months of screening, and had a baseline urine protein/creatinine ratio (UPCR) > 1 mg/mg (113.17 mg/mmol) and serum creatinine < 3 mg/dL.
Full study design details, including patient inclusion and exclusion criteria for the ALLURE trial, and results have been previously presented.27-29 In brief, 405 patients were enrolled into the ALLURE trial, with 316 patients remaining in the study for at least 52 weeks. Similar proportions of patients in the placebo (33.5%) and the abatacept groups (33.5%) achieved CRR in intention-to-treat analysis at week 52; hence, the primary endpoint was not achieved.27,28 The last collected samples used in our study were at 52 weeks, and many patients were discontinued even earlier. The ALLURE trial was conducted in accordance with the Declaration of Helsinki, and the International Conference on Harmonization Good Clinical Practice Guideline. Independent ethics committee or independent institutional review board approvals were obtained, and all patients provided written informed consent, including for the secondary use of urine samples for research.24
Definition of renal outcomes. CRR was defined as meeting all of the following: (1) eGFR (based on the Chronic Kidney Disease Epidemiology Collaboration [CKD-EPI] equations) normal (≥ 60 mL/min/1.73 m2) or at least 85% of the eGFR at baseline; (2) UPCR < 0.5; and (3) inactive urinary sediment, which we defined as urine red blood cell (RBC) casts ≤ 5 and urine RBC ≤ 10. Accordingly, PRR was defined by meeting all of the following 3 criteria: (1) eGFR ≥ 85% of baseline eGFR; (2) UPCR < 0.5, or UPCR decreased as ≥ 50% to value of < 1, for baseline UPCR < 3, or ≥ 50% reduced from baseline to UPCR < 3 if baseline UPCR ≥ 3; and (3) urine RBC ≤ 10 and urine RBC casts ≤ 5. Patients not achieving at least PRR were considered as having NR.
Urine biomarker measurement. Urine samples were frozen within 24 hours from collection prior to testing in individual RAIL protein assays. RAIL urine biomarkers were quantified using singleplex assays per the manufacturers’ instructions. ELISAs were used to measure the remaining biomarkers; all sample testing was done in singlicate. Human urinary KIM-1 (R&D Systems; DKM100) had a lower limit of quantitation (LLOQ) of 0.009 ng/mL. Human MCP-1 (R&D Systems; DCP00) had a lower limit of detection (LLOD) of 1.7 pg/mL. Human adiponectin (R&D Systems; DRP300) had a LLOD of 0.246 ng/mL. Human ceruloplasmin (Assaypro; EC4201-1) had a LLOD of 0.085 ng/mL. Human hemopexin (Assaypro; EH2001-1) had a LLOD of 4.2 ng/mL. KIM-1 and MCP-1 used a 4-parameter logistic curve to fit the standard curve, and adiponectin, ceruloplasmin, and hemopexin instead used a log/log curve. Analyte concentrations that were above or below the limits of detection were imputed by 50% of the level of LLOD and 50% over the upper limit of detection, respectively. Natural log-transformed quantities of NGAL, adiponectin, hemopexin, and ceruloplasmin are reported in ng/mL, whereas those of KIM-1 and MCP-1 are reported in pg/mL. A Roche Cobas C311 clinical chemistry analyzer, using a commercially available assay (BioPorto, Catalog KIT ST001RA for NGAL; Roche Diagnostics, Reference 03263991 190 for creatinine), was used to measure both human NGAL and urine creatinine. NGAL had a LLOQ of 9.8 ng/mL, and creatinine had a LLOQ of 1.1 mg/dL.
RAIL score calculation. We have shown in the past that the RAIL biomarkers are produced locally in the kidney with active LN31 and that standardization of biomarker quantities by urine creatinine but not unspecific proteinuria results in the best reflection of histologic activity of LN as per kidney biopsies (NIH-AI score), in both children and adults with LN.19,21,32 When using RAIL scores, the age of the patient must be considered. RAIL scores were developed in pediatric patients with SLE and then validated in adult patients previously and in our study.19,21 NGAL, KIM-1, and MCP-1 urine concentrations have been shown to increase with age, whereas adiponectin concentrations decrease with age.11,31 Significantly higher concentrations of NGAL, hemopexin, and ceruloplasmin have been found in urine samples from female patients than in those from male patients.11,31 As previously published, the RAIL score for adults was calculated from the urine quantities of the RAIL biomarkers, after natural logarithmic (ln) transformation of values after normalization by urine creatinine concentrations as follows21:
Statistical analysis. Summary statistics of baseline sample characteristics were reported for treatment arms by frequency and corresponding percentage for categorical variables; median and IQR were reported for numerical variables. LN responses were classified into CRR, PRR-only (excluding CRR responders), and NR, for all patients at each study visit (12, 24, and 52 weeks). The RAIL scores were compared between groups of responders (PRR-only or CRR) and NR at weeks 12, 24, and 52. To understand the minimum clinically important differences, relative (or percentage) changes in RAIL scores were calculated between 2 adjacent visits (0-12, 12-24, and 24-52 weeks) expressed as the percentage of the previous values of RAIL score. The mean difference and its corresponding SD between responders (CRR, PRR-only) and NR were reported for all and by treatment arm at each follow-up visit. Standardized mean difference was calculated to express the group mean differences relative to the SD of each group; t tests were used for group comparisons.
To evaluate the ability of the RAIL score to discriminate between the responder and NR groups, generalized estimating equation (GEE) regressions were used to model the effect of the RAIL score on CRR, compared against the performance of traditional kidney function markers (eGFR based on CKD-EPI) or UPCR at baseline with and without adjustment for baseline UPCR and eGFR, as well as patient age, weight, sex, and race. The area under the receiver-operating characteristic (ROC) curves (AUC) were generated for each marker (RAIL score, eGFR, UPCR). Comparisons considered were CRR plus PRR-only vs NR, CRR vs PRR-only plus NR, and finally CRR vs NR in all patients and by treatment arm. In general, values of AUC are considered outstanding, excellent, good, fair, or poor in terms of accuracy for AUC of 0.90-1.00, 0.81-0.90, 0.71-0.80, 0.61-0.70, or 0.51-0.60, respectively. Two sets of analyses were performed: one modeled concurrent RAIL score, and the other modeled the RAIL score from the previous visit. The same GEE analyses were performed in a sensitivity analysis excluding the week 52 data.
RESULTS
Patients and samples. This analysis included 240 patients enrolled in the ALLURE trial for whom 599 urine samples are available (consented) for future research at baseline. Urine samples at weeks 12, 24, and 52 were available for 198, 116, and 48 patients (96, 57, and 18 remaining in the abatacept group; and 102, 59, and 30 in the placebo group), respectively. Demographics and baseline disease characteristics are presented in Table 1, with 121 patients randomized to intravenous abatacept and 119 to placebo, respectively. Kidney biopsy results (class only) were available for a total of 197 patients. At baseline, there were no statistically significant differences (all P ≥ 0.52) in kidney measures (UPCR, eGFR, urine RBCs) between treatment groups. However, the proportion of White patients and patients with class IV (or IV/V) LN was higher in the abatacept group than the placebo group, whereas RAIL scores were comparable between groups (P = 0.75).
Patient characteristics by treatment arm at baseline (week 0).
Kidney measures at weeks 12, 24, and 52 by treatment group. Traditional kidney measures showed trends of improvement from baseline by week 12 and continued to further improve at weeks 24 and 52. Only at week 12, but not at weeks 24 or 52, was the UPCR significantly lower in the abatacept group than the placebo group [median (IQR); abatacept vs placebo: 1.13 (0.55-2.29) vs 1.55 (0.72-3.22); P = 0.04], whereas other conventional kidney measures (eGFR, UPCR) were similar throughout the study (Supplementary Table S1, available with the online version of this article). Prednisone use was similar in the abatacept group and the placebo group.
Renal response status and RAIL scores and their changes over time by renal response status. Of the patients who contributed urine samples used in this study, there were 13.6% and 22.2% of patients with CRR and PRR-only at week 12; then 28.5% CRR and 19% PRR-only at week 24. At week 52, only 48 urine samples were available from the study, of which 37.5% and 31.3% showed CRR and PRR-only, respectively. RAIL scores decreased over time, generally without important differences between treatment groups. Compared to NR, mean (SD) RAIL scores were lower in patients achieving CRR or PRR-only (Table 2); these differences were only statistically significant for patients who achieved CRR at weeks 12, 24, and 52, and for those who achieved PRR-only at week 52. These absolute RAIL scores are shown graphically in Figure 1. Differences in RAIL score based per treatment group showed similar trends (Supplementary Table S2, available with the online version of this article). We also assessed relative changes in RAIL scores since baseline and the preceding visit. Generally, changes lacked statistical significance, irrespective of whether changes from baseline or since the prior visit were considered (Supplementary Table S3-S4).
Absolute RAIL scores by renal response status at each follow-up timepoint.
RAIL scores with renal response states during the study. Box (SD) and whisker (95% CI) plots illustrating cross-sectional comparison of the change in RAIL score from baseline to weeks 12, 24, and 52 for (A) all the patients together, (B) the placebo group alone, and (C) the abatacept group alone. Absolute changes in RAIL scores are depicted for patients considered at these timepoints to have one of the evaluated renal outcomes: CRR, PR-only, or NR. CRR: complete renal response; NR: nonresponder; PRR-only: partial renal response but not CRR; RAIL: Renal Activity Index for Lupus.
Accuracy of the RAIL score, UPCR, and eGFR to discriminate LN response status. As summarized in Table 3, the eGFR was a poor predictor of renal response (AUC all ≤ 0.51, P ≥ 0.68), irrespective of the comparison made (CRR vs PRR-only plus NR; CRR plus PRR-only vs NR; CRR vs NR). Without correction of baseline differences of renal response groups, the RAIL score had fair accuracy in differentiating response groups, irrespective of the comparison made (P < 0.005), and remained significant after adjustment for the UPCR and eGFR (P range = 0.01-0.04). The ROC analyses suggested that, after adjusting for baseline UPCR, eGFR, age, weight, sex, and race, the RAIL score outperformed the UPCR and had excellent accuracy for differentiating CRR from other renal response states (AUC ≥ 0.83) and still had good accuracy (AUC = 0.77) to discriminate patients who achieved at least PRR from NR (Figures 2A,C). Sensitivity analyses revealed the same pattern.
Accuracy of the RAIL score, UPCR, and eGFR to discriminate LN response status concurrently or at the next visit.
Ability of biomarkers and their combination to reflect current renal response state. ROC curves of RAIL score and other biomarkers alone or in combination to differentiate renal response status. Comparisons include (A) CRR vs PRR-only + NR, (B) CRR + PRR-only vs NR, and (C) CRR vs NR. Horizontal dotted lines depict the sensitivity at 90%, 80% and 70%. CRR: complete renal response; eGFR: estimated glomerular filtration rate; NR: nonresponder; PRR: partial renal response, RAIL: Renal Activity Index for Lupus; ROC: receiver-operating characteristic; RAIL-adjBL: RAIL score adjusted for baseline UPCR, eGFR, age, weight, sex, and race; UPCR: urine protein/creatinine ratio.
Using RAIL scores to anticipate future renal response based on the current RAIL score. GEE analyses revealed a strong predictive role of the RAIL score, with and without adjustment for the UPCR and eGFR. Without adjustment for baseline differences of patients with different response states (CRR, PRR-only, NR), the RAIL score had fair accuracy (AUC = 0.61-0.63) to discriminate CRR from other response states (Table 3). As shown in Figure 3, after adjusting for baseline eGFR, UPCR, age, weight, sex, and race, the accuracy of the RAIL score to anticipate CRR at the next visit was excellent (AUC = 0.84-0.85; Figures 3A,C) and had good ability to differentiate NR from other response states (AUC = 0.77, Figure 3B). Again, the eGFR was not useful to predict response (AUC ≤ 0.51) and was inferior to the UPCR (Table 3). No important differences were observed in the performance of the RAIL score in sensitivity analysis.
Ability of biomarkers and their combination to reflect next renal response state. ROC curves illustrating the sensitivity and specificity of each variable in predicting renal response for the next visit, with the AUC labeled for each. Comparisons include CRR vs PRR-only + NR (A), CRR + PRR-only vs NR (B), and CRR vs NR (C). Horizontal dotted lines depict the sensitivity at 90%, 80% and 70%. AUC: area under the curve; CRR: complete renal response; eGFR: estimated glomerular filtration rate; NR: nonresponder; PRR: partial renal response, RAIL: Renal Activity Index for Lupus; RAIL-adjBL: RAIL score adjusted for baseline UPCR, eGFR, age, weight, sex, and race; ROC: receiver-operating characteristic; UPCR: urine protein/creatinine ratio.
DISCUSSION
Despite therapeutic advances in SLE management, LN remains a common severe disease manifestation, conferring significant morbidity and mortality.1,3-5 The RAIL score, based on a set of urine biomarkers, was developed as a noninvasive tool to help diagnose LN and monitor the course of pediatric LN.19,21 Here, we newly demonstrate the ability of the RAIL biomarkers from adult patients with LN to predict future renal response state. Further, we confirm that absolute change in RAIL scores since baseline differentiates renal response states (PRR, CRR, NR). Collectively, these results suggest that the RAIL score is an effective, noninvasive tool for assessing kidney activity and treatment response over time in adult patients with LN.
In the past, we showed that active LN can be discriminated from inactive or absent LN for scores ≥ 4.25 with 80% accuracy.23 Consistent with prior observations,19,21 patients with CRR could be discriminated from NR for scores ≤ 4.18.
In contrast to our prior studies in adults with LN,23 this is the first to show, to our knowledge, that high RAIL scores at a given visit are predictors of future nonresponse of adults with LN. This finding is again consistent with our prior observations in pediatric patients with LN where we reported that changes in RAIL score anticipated the clinically observed course at least 3 months earlier.22 Together, our findings might be exploited in clinical care for the surveillance of LN, especially in patients with known proteinuria from kidney damage in whom treatment resistance to the current treatment regimen is suspected. Indeed, RAIL scores that remain 6.5 or higher may indicate the need to adjust immunosuppressive therapy or explore nonadherence.
The shortcoming of proteinuria to gauge LN activity has long been recognized.33 For example, proteinuria can reflect kidney damage or even comorbid disease such as hypertension and diabetes mellitus.34,35 Current definitions of renal response used in the ALLURE trial are reliant on changes in proteinuria, eGFR, and abnormal urine sediment.26 The latter 2 components were either very rarely abnormal (ie, urine sediment) or did not contribute to gauging renal response (as was the case for the eGFR). In the ALLURE trial, no participant who reached the 1-year endpoint had active sediment. In our present study, the UPCR performed similarly well to the RAIL score to predict or identify patients who experienced PRR or CRR as opposed to NR. The observed result might be expected, as the definition of CRR and PRR was based on a composite measure that included the UPCR, eGFR, and RBC or RBC casts, with the latter occurring only rarely in the study population. Altogether, we consider that the RAIL score provides additional value to the assessment of LN, as we have shown in the past that the RAIL biomarkers do not reflect kidney damage,20,21 different from the UPCR. Notably, a direct head-to-head comparison of these biomarkers has not been done.
The ALLURE trial in LN was stopped after the assessment of the primary endpoint.27-29 The study population included in this current study resembled the overall study population.27-29 Herein we provide supportive evidence for this decision, given that the RAIL scores in the 2 treatment arms (abatacept, placebo) remained almost identical over time through week 24 and at baseline. We note, however, a decrease in RAIL scores at week 52 among patients with CRR and PRR in the abatacept group, a trend that was not observed in the placebo group. Given that the trial was stopped as the primary endpoint was not met, the number of patients in the study by week 52 and the limited number of urine samples available limit the generalizability of this observation. However, sustained CRR rates were initially higher with abatacept, an analysis we were unable to duplicate considering the RAIL biomarkers, given the available urine samples.
Although we consider the availability of kidney biopsy samples to confirm active LN as a strength of this study,24 we were unable to confirm response to therapy by serial kidney biopsies since the study protocol did not mandate repeat biopsies. However, requiring serial kidney biopsies without evidence of ongoing activity (eg, significant proteinuria) for the sole purpose of research or a clinical trial may be ethically unacceptable. Further, additional validation studies are needed to expand on the interpretation of RAIL scores over time in adults, best with available serial kidney biopsies to estimate response to therapy.
Besides the RAIL biomarkers, there are other LN biomarkers that have been shown to reflect LN and its course. None of the proposed LN biomarkers, including the RAIL biomarkers, are specific to LN, but they capture aspects of pathological changes with LN. Such novel biomarkers include CD163 and activated leukocyte adhesion molecule (ALCAM), among many others.36 The assessment of other biomarkers was outside the scope of this research, but we cannot exclude that any of these biomarkers alone or in combination with the RAIL biomarkers could be superior to the RAIL score in reflecting the course of LN activity in adults.
In conclusion, these findings support the clinical utility of the RAIL biomarkers and the RAIL score to noninvasively assess kidney inflammation and response to therapy in adult LN, to guide early treatment decisions, including identifying patients who may need therapy intensification, and thus allow for a personalized monitoring approach.
ACKNOWLEDGMENT
The authors would like to thank the investigators, research staff, healthcare providers, and patients who contributed to this study and the clinical trials on which its analyses are based. A special thank you to Ms. Megan Quinlan-Waters and Angela Merritt for the management of the biosamples and the support of the research regulatory activities, respectively. We thank Bristol Myers Squibb for the donation of the urine samples and clinical data.
Footnotes
CONTRIBUTIONS
JR performed the testing of the urine samples. All authors contributed to the development of the manuscript, including interpretation of results, substantive review of drafts, and approval of the final draft for submission.
FUNDING
The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health (NIH). Main funding and study medication were both provided by Pfizer Inc. Additionally, the study was supported by the Center for Clinical and Translational Science and Training at the University of Cincinnati and is funded by the NIH Clinical and Translational Science Award (CTSA) program, grant UL1TR001425. The CTSA program is led by the NIH National Center for Advancing Translational Sciences. Further, the project described was supported by the National Institutes of Arthritis and Musculoskeletal Skin Diseases under Awards Number P30AR076316 and P30AR070549, as well as a Transformational Research Award from the Marianne and Ralph Falk Medical Research award. SKO was supported by NIH T32AR069512.
COMPETING INTERESTS
HIB has received speaking fees from Novartis, Pfizer, and GSK; consultancies/honoraria from AbbVie, AstraZeneca-Medimmune, Boehringer Ingelheim, BMS, Eli Lilly, EMD Serono, Janssen, GSK, F. Hoffmann-La Roche, Merck, and Novartis. The Cincinnati Children’s Hospital, where HIB works as a full-time public employee, has received contributions from the following industries in the past 3 years: BMS, F. Hoffmann-La Roche, Janssen, Novartis, and Pfizer. This funding has been reinvested for the research activities of the hospital in a fully independent manner, without any commitment to third parties. PD has received grant/research support from NIH, and licensing agreements on submitted patents related to NGAL as a biomarker of kidney disease from BioPorto. JL and MAM are employees and shareholders of BMS. The remaining authors declare no conflicts of interest relevant to this article.
ETHICS AND PATIENT CONSENT
The ALLURE trial was conducted in accordance with the Declaration of Helsinki, and the International Conference on Harmonization Good Clinical Practice Guideline. Independent ethics committee or independent institutional review board (IRB) approvals were obtained, and all patients provided written informed consent, including for the secondary use of urine samples for research. This study was approved initially on November 28, 2022, by the Cincinnati Children’s Hospital IRB (IRB #: 2022-0859; Federal Wide Assurance #00002988).
DATA AVAILABILITY
Data are available on reasonable request.
- Accepted for publication October 30, 2025.
- Copyright © 2026 by the Journal of Rheumatology
This is an Open Access article, which permits use, distribution, and reproduction, without modification, provided the original article is correctly cited and is not used for commercial purposes.
REFERENCES
SUPPLEMENTARY DATA
Supplementary material accompanies the online version of this article.
