The SHIPE Index: A Tertile-Based Approach to Predict End-Stage Kidney Disease in ANCA-Associated Vasculitis Authors: Hyunsue Do, Lucy Eunju Lee, Jang Woo Ha, Jason Jungsik Song, Yong-Beom Park, Sang-Won Lee Published: 2026 ABSTRACT AbstractBackgroundThis study aimed to develop a simple, robust, and ethnoregion-specific index for predicting progression to end-stage kidney disease (ESKD) in Korean patients with antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV).MethodsBaseline data from 239 Korean patients with AAV enrolled in a tertiary hospital cohort were analyzed to identify independent predictors of ESKD using univariable and multivariable Cox proportional hazards models. The SHIPE (Summation of the Highest tertiles of Independent Predictors of ESKD) index was constructed by summing the highest tertiles of the identified predictors: Five-Factor Score (≥ 2) and serum creatinine (≥ 1.3 mg/dL), with 1 point assigned to each criterion met. Cumulative ESKD-free survival rates across SHIPE index scores were compared using Kaplan-Meier analysis.ResultsA SHIPE index score of 2 was significantly associated with lower cumulative ESKD-free survival rate compared to scores of 0 or 1 (P < 0.001). By integrating two independent predictors, the SHIPE index provides a simple yet effective tool for risk stratification, demonstrating greater robustness against variability in single predictors such as serum creatinine. This index demonstrated clinical utility, providing a straightforward method for identifying high-risk patients in real-world settings.ConclusionThe SHIPE index is a novel, robust, and convenient tool for predicting ESKD progression in Korean patients with AAV. This study introduces a practical framework for developing tailored prediction indices that can be adapted for diverse ethnic and regional populations. Its simplicity and practical applicability make it well-suited for routine clinical use. Future prospective studies are needed to validate the SHIPE index in larger, multi-ethnic cohorts and ensure its broader applicability.Keywords: ANCA-Associated Vasculitis, End-Stage Kidney Disease, Prognosis, Biomarkers, Risk Assessment INTRODUCTION Antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) is a systemic vasculitis that primarily affects small vessels, including capillaries, arterioles, and venules, with occasional involvement of medium-sized arteries. AAV comprises three subtypes: microscopic polyangiitis (MPA), granulomatosis with polyangiitis (GPA), and eosinophilic GPA (EGPA).1 Among major internal organs, the kidney and lungs have the largest capillary surface area, rendering them particularly susceptible to AAV-related damage.2 Renal involvement occurs in 90–100% of MPA patients, 50–80% of GPA patients, and 4–50% of EGPA patients.3 AAV-induced renal involvement can lead to progressive kidney function decline, with some patients ultimately advancing to end-stage kidney disease (ESKD). This progression not only increases all-cause mortality but also exacerbates systemic complications of AAV.4 Directly preventing renal involvement or halting its progression at the time of AAV diagnosis is often challenging, if not impossible. However, the ability to predict ESKD progression at an early stage could significantly improve patient outcomes. Early prediction would facilitate more frequent follow-ups, timely assessment of renal involvement using Birmingham vasculitis activity score (BVAS), and optimization of treatment regimens, including dose adjustments and tailored therapeutic combinations.5 Several predictors of ESKD, such as sex, age, body mass index, serum creatinine, and estimated glomerular filtration rate, have been proposed. However, relying on a single predictor has proven inadequate for achieving accurate and reliable predictions.6,7 Furthermore, the development of predictive tools must address two critical challenges: accounting for ethnic and geographical variations and ensuring clinical practicality. Therefore, this study aimed to develop a novel, robust, ethnoregion-specific, and clinically practical index that incorporates multiple independent predictors of ESKD in Korean patients with AAV. METHODS Patients We retrospectively reviewed the medical records of 239 patients with AAV who were enrolled in the Severance Hospital ANCA-associated VasculitidEs (SHAVE) cohort, a prospective observational AAV cohort study conducted between November 2016 to December 2022. The inclusion criteria for the SHAVE cohort were as follows: i) Patients newly diagnosed with MPA, GPA, or EGPA at this tertial university hospital by specialized rheumatologists; i) Patients newly diagnosed with MPA, GPA, or EGPA at this tertial university hospital by specialized rheumatologists; ii) Patients classified and validated (or reclassified) as having AAV based on the 2007 European Medicine Agency algorithm, the revised 2012 Chapel Hill Conference Consensus nomenclature, and the 2022 American College of Rheumatology and European Alliance of Associations for Rheumatology criteria1,2,8,9,10,11; ii) Patients classified and validated (or reclassified) as having AAV based on the 2007 European Medicine Agency algorithm, the revised 2012 Chapel Hill Conference Consensus nomenclature, and the 2022 American College of Rheumatology and European Alliance of Associations for Rheumatology criteria1,2,8,9,10,11; iii) Patients with sufficient medical records to collect clinical, laboratory, histological, and radiological data at the time of AAV diagnosis and during follow-up; iii) Patients with sufficient medical records to collect clinical, laboratory, histological, and radiological data at the time of AAV diagnosis and during follow-up; iv) Patients with recorded AAV-specific indices, such as BVAS and Five-Factor Score (FFS), or with sufficient information to assess these indices at diagnosis12,13; iv) Patients with recorded AAV-specific indices, such as BVAS and Five-Factor Score (FFS), or with sufficient information to assess these indices at diagnosis12,13; v) Patients with ANCA test results available at diagnosis; v) Patients with ANCA test results available at diagnosis; vi) Patients followed for at least six months after diagnosis; vi) Patients followed for at least six months after diagnosis; vii) Patients without serious concomitant medical conditions, such as malignancies, severe infections requiring hospitalization, or autoimmune diseases mimicking AAV manifestations8,9,10; vii) Patients without serious concomitant medical conditions, such as malignancies, severe infections requiring hospitalization, or autoimmune diseases mimicking AAV manifestations8,9,10; viii) Patients who had not been exposed to immunosuppressive drugs for AAV treatment within four weeks prior to diagnosis. viii) Patients who had not been exposed to immunosuppressive drugs for AAV treatment within four weeks prior to diagnosis. Clinical data at diagnosis Demographic data, including age, and sex, were collected at the time of diagnosis. Additional variables included AAV subtypes, ANCA positivity, AAV-specific indices (BVAS and FFS), the results of acute-phase reactants (erythrocyte sedimentation rate [ESR] and C-reactive protein [CRP]). Routine laboratory test results were also recorded. Myeloperoxidase (MPO)-ANCA and proteinase 3 (PR3)-ANCA levels were measured using immunoassays, while perinuclear (P)-ANCA and cytoplasmic (C)-ANCA were detected via indirect immunofluorescence assays.14 Both testing methods were accepted based on the 2022 newly proposed criteria for AAV.8,9,10 Comorbidities diagnosed prior to AAV diagnosis, including type 2 diabetes mellitus, hypertension, and hyperlipidemia, were also documented. Clinical data during follow-up ESKD was identified as a poor outcome, with only cases occurring after AAV diagnosis included in this study. The follow-up period for ESKD was defined as the time from AAV diagnosis to the initiation of renal replacement therapy in patients with ESKD and from diagnosis to the last clinical visit in those without ESKD.15 Additionally, the number of patients receiving glucocorticoids and other immunosuppressive drugs after diagnosis was documented. SHIPE index: Summation of the Highest tertiles of Independent Predictors of ESKD The SHIPE index (Summation of the Highest tertiles of Independent Predictors of ESKD) was developed through a systematic three-step process: i) Identification of independent predictors: Independent predictors of ESKD were identified using multivariable Cox proportional hazards regression analysis. i) Identification of independent predictors: Independent predictors of ESKD were identified using multivariable Cox proportional hazards regression analysis. ii) Defining significant abnormal values: The highest tertile of each independent predictor was designated as a significant abnormal value, based on an arbitrary yet clinically meaningful threshold. ii) Defining significant abnormal values: The highest tertile of each independent predictor was designated as a significant abnormal value, based on an arbitrary yet clinically meaningful threshold. iii) Calculation of the SHIPE index: The SHIPE index was calculated as the sum of these significant abnormal values derived from the highest tertiles. iii) Calculation of the SHIPE index: The SHIPE index was calculated as the sum of these significant abnormal values derived from the highest tertiles. To accommodate cohort-specific variations, such as differences in ethnicity or geography, the tertile thresholds can be adjusted (e.g., to the highest quartile or upper half) as needed. Adjustment of FFS scale for SHIPE index development For the development of the SHIPE index, the FFS was utilized on a 0 to 5 scale rather than the conventional 0 to 2 scale. This adjustment provided greater granularity for defining tertile values of the FFS variable, aligning with the SHIPE index’s methodology of using tertile thresholds for independent predictors. By adopting this approach, the scoring system remained compatible with the tertile-based framework while preserving the resolution of clinical risk factors in the analysis. Statistical analyses All statistical analyses were performed using SPSS Statistics for Windows, version 28 (IBM Corp., Armonk, NY, USA) or R software, version 4.3.2 (R Foundation for Statistical Computing, Vienna, Austria). Continuous variables were expressed as medians with interquartile ranges, while categorical variables were expressed as numbers (percentages). Multivariable Cox proportional hazards analysis was performed on variables with statistical significance in univariable analysis to estimate hazard ratios (HRs) over the follow-up period. Multicollinearity was evaluated using the variance inflation factor (VIF) and the condition index. The Kaplan-Meier survival analysis, combined with the log-rank test, was used to compare cumulative ESKD-free survival rates between groups. Relative risks (RRs) for all-cause mortality cut-off values were calculated using contingency tables and χ2 tests. A P value of < 0.05 was considered statistically significant. Ethics statement The study was approved by the Institutional Review Board (IRB) of Severance Hospital (Seoul, Korea; IRB No. 4-2016-0901), and was conducted in accordance with the principles of the Declaration of Helsinki. Given its retrospective design and the use of anonymized patient data, the requirement for written informed consent was waived by the IRB. RESULTS Characteristics of patients with AAV The median age of the 239 patients was 59.9 years, with 82 (34.3%) being male. Among the patients, 128 (53.6%) were classified as MPA, 59 (24.7%) as GPA, and 52 (21.8%) as EGPA. ANCA was detected in 190 patients (79.5%), with 160 (66.9%) testing positive for MPO-ANCA (or P-ANCA) and 40 (16.7%) for PR3-ANCA (or C-ANCA). The median values of AAV-specific indices and acute-phase reactants at diagnosis were as follows: BVAS, 12.0; FFS, 1.0; ESR, 60.0 mm/hr; and CRP, 13.8 mg/L. Additional laboratory results are presented in Table 1. Comorbidities included hypertension in 98 patients (41.0%), type 2 diabetes mellitus in 60 patients (25.1%), and dyslipidemia in 44 patients (18.4%). Table 1Characteristics of 239 AAV patientsVariablesValuesVariables at diagnosisDemographic dataAge, yr59.0 (49.0–69.0)Male sex82 (34.3)AAV subtypesMPA128 (53.6)GPA59 (24.7)EGPA52 (21.8)ANCA positivityMPO-ANCA (or P-ANCA) positive160 (66.9)PR3-ANCA (or C-ANCA) positive40 (16.7)Both ANCA positive10 (4.2)ANCA positivity190 (79.5)AAV-specific indicesBVAS12.0 (7.0–18.0)FFS1.0 (1.0–2.0)Acute phase reactantsESR, mm/hr59.0 (22.0–96.0)CRP, mg/L13.8 (1.6–71.0)Laboratory resultsWhite blood cell count, /mm39,040.0 (6,400.0–12,660.0)Haemoglobin, g/dL11.4 (9.5–13.3)Platelet count, × 1,000/mm3301.0 (228.0–393.0)Fasting glucose, mg/dL102.0 (91.0–124.0)Blood urea nitrogen, mg/dL17.6 (12.3–31.0)Serum creatinine, mg/dL0.9 (0.7–1.8)Total protein, g/dL6.7 (6.1–7.2)Serum albumin, g/dL3.7 (3.1–4.2)ComorbiditiesType 2 diabetes mellitus60 (25.1)Hypertension98 (41.0)Dyslipidaemia44 (18.4)During follow-upESKD41 (17.2)Follow-up duration based on ESKD, mon33.7 (10.2–74.0)MedicationsGlucocorticoids224 (93.7)Cyclophosphamide128 (53.6)Rituximab41 (17.2)Mycophenolate mofetil37 (15.5)Azathioprine131 (54.8)Tacrolimus22 (9.2)Methotrexate23 (9.6)Plasma exchange21 (8.8)Values are expressed as a median (interquartile range) or number (%).AAV = ANCA-associated vasculitis, MPA = microscopic polyangiitis, GPA = granulomatosis with polyangiitis, EGPA = eosinophilic granulomatosis with polyangiitis, ANCA = antineutrophil cytoplasmic antibody, MPO = myeloperoxidase, P = perinuclear, PR3 = proteinase 3, C = cytoplasmic, BVAS = Birmingham vasculitis activity score, FFS =Five-Factor Score, ESR = erythrocyte sedimentation rate, CRP = C-reactive protein, ESKD = end-stage kidney disease. Characteristics of 239 AAV patients Values are expressed as a median (interquartile range) or number (%). AAV = ANCA-associated vasculitis, MPA = microscopic polyangiitis, GPA = granulomatosis with polyangiitis, EGPA = eosinophilic granulomatosis with polyangiitis, ANCA = antineutrophil cytoplasmic antibody, MPO = myeloperoxidase, P = perinuclear, PR3 = proteinase 3, C = cytoplasmic, BVAS = Birmingham vasculitis activity score, FFS =Five-Factor Score, ESR = erythrocyte sedimentation rate, CRP = C-reactive protein, ESKD = end-stage kidney disease. During a median follow-up duration of 33.7 months, 41 patients (17.2%) progressed to ESKD. Glucocorticoids were administered to 224 patients (93.7%). Cyclophosphamide was the most commonly used induction therapy (53.6%), while azathioprine was the most frequently prescribed maintenance therapy (54.8%). Detailed treatment information is summarized in Table 1. Independent predictors of ESKD Univariable and multivariable Cox proportional hazard analyses were performed to identify baseline independent predictors of ESKD. In the univariable analysis, MPO-ANCA (or P-ANCA) positivity, BVAS, FFS, hemoglobin, blood urea nitrogen, serum creatinine, serum albumin, and hypertension at diagnosis were significantly associated with progression to ESKD during follow-up in patients with AAV. In the multivariable analysis, both FFS (HR, 1.732; 95% confidence interval [CI], 1.170–2.563; P = 0.006), and serum creatinine (HR, 1.376; 95% CI, 1.204–1.571; P < 0.001) were identified as independent predictors of ESKD (Table 2). Table 2Cox hazards model analysis of variables at diagnosis for end-stage kidney disease during follow-up in 239 AAV patientsVariablesUnivariable analysisMultivariable analysisHR95% CIP valueHR95% CIP valueDemographic data at diagnosisAge1.0140.991–1.0370.239Male sex0.9380.476–1.8480.853AAV subtypes at diagnosisMPA versus GPA or EGPA2.0781.075–4.0160.030* 0.7710.335–1.7720.540ANCA positivity at diagnosisMPO-ANCA (or P-ANCA) positive2.4241.117–5.2640.025* 0.9370.363–2.4180.892PR3-ANCA (or C-ANCA) positive0.7450.313–1.7750.507AAV-specific indices at diagnosisBVAS1.0851.037–1.135< 0.001*** 1.0300.973–1.0910.311FFS2.5211.895–3.354< 0.001*** 1.7321.170–2.5630.006** Acute phase reactants at diagnosisESR1.0071.000–1.0150.070CRP1.0040.999–1.0090.102Laboratory results at diagnosisWhite blood cell count1.0001.000–1.0000.851Haemoglobin0.6910.592–0.806< 0.001*** 0.9610.798–1.1560.673Platelet count0.9990.996–1.0010.256Fasting glucose1.0000.993–1.0080.908Blood urea nitrogen1.0311.024–1.038< 0.001*** 1.0070.994–1.0210.287Serum creatinine1.4901.374–1.615< 0.001*** 1.3761.204–1.571< 0.001*** Total protein1.0200.974–1.0690.391Serum albumin0.5950.399–0.8870.011* 0.8250.458–1.4850.521Comorbidities at diagnosisDiabetes mellitus0.8290.396–1.7370.619Hypertension2.1311.144–3.9690.017* 1.4210.727–2.7760.304Dyslipidaemia1.0610.490–2.2990.881AAV = ANCA-associated vasculitis, HR = hazard ratio, CI = confidence interval, MPA = microscopic polyangiitis, GPA = granulomatosis with polyangiitis, EGPA = eosinophilic granulomatosis with polyangiitis, ANCA = antineutrophil cytoplasmic antibody, MPO = myeloperoxidase, P = perinuclear, PR3 = proteinase 3, C = cytoplasmic, BVAS = Birmingham vasculitis activity score, FFS = Five-Factor Score, ESR = erythrocyte sedimentation rate, CRP = C-reactive protein.*P < 0.05, **P < 0.01, ***P < 0.001. Cox hazards model analysis of variables at diagnosis for end-stage kidney disease during follow-up in 239 AAV patients AAV = ANCA-associated vasculitis, HR = hazard ratio, CI = confidence interval, MPA = microscopic polyangiitis, GPA = granulomatosis with polyangiitis, EGPA = eosinophilic granulomatosis with polyangiitis, ANCA = antineutrophil cytoplasmic antibody, MPO = myeloperoxidase, P = perinuclear, PR3 = proteinase 3, C = cytoplasmic, BVAS = Birmingham vasculitis activity score, FFS = Five-Factor Score, ESR = erythrocyte sedimentation rate, CRP = C-reactive protein. *P < 0.05, **P < 0.01, ***P < 0.001. Given that FFS reflects renal function, we investigated whether FFS and serum creatinine were truly independent predictors of ESKD in the multivariable analysis. Multicollinearity was assessed using the VIF and the conditional index. The VIF between these two variables was 1.29, and the highest condition index was 2.24, indicating no significant multicollinearity. Thus, FFS and serum creatinine at diagnosis were confirmed to be baseline independent predictors of ESKD in Korean patients with AAV enrolled in the SHAVE cohort. Highest tertiles of predictors and cumulative ESKD-free survival rates To enhance convenience—one of the objectives of this study—cut-offs for predicting future progression to ESKD were defined as the lower limit of the highest tertile for each baseline independent predictor. These cut-offs were determined using frequency analysis rather than ROC analysis. For FFS and serum creatinine, the cut-offs were set at 2 and 1.3 mg/dL, respectively. Patients with FFS ≥ 2 at diagnosis had significantly lower cumulative ESKD-free survival rates compared to those with FFS < 2. Similarly, patients with serum creatinine ≥ 1.3 mg/dL at diagnosis exhibited significantly reduced cumulative ESKD-free survival rates compared to those with serum creatinine < 1.3 mg/dL (Fig. 1). Comparison of cumulative ESKD-free survival rates by the highest tertiles of independent predictors of ESKD. (A, B) Patients with FFS ≥ 2 and serum creatinine ≥ 1.3 mg/dL at diagnosis had significantly lower cumulative ESKD-free survival rates compared to those with FFS < 2 and serum creatinine < 1.3 mg/dL, respectively. FFS = Five-Factor Score, ESKD = end-stage kidney disease, Cr = creatinine. ***P < 0.001. Development of an index for predicting progression to ESKD To create a simple and practical tool for predicting progression to ESKD, we assigned 1 point to the highest tertile of each independent predictor (FFS and serum creatinine) and 0 point to the mid and lowest tertiles. The scores for each predictor were then summed to develop the index. The formula for the SHIPE index was established as follows: Both FFS and serum creatinine values were measured at diagnosis. Based on the SHIPE index, AAV patients were categorized into three groups: SHIPE index = 0 (118 patients), SHIPE index = 1 (58 patients), and SHIPE index = 2 (63 patients) (Fig. 2A). Patients with a SHIPE index of 2 exhibited significantly lower cumulative ESKD-free survival rates compared to those with SHIPE indices of 0 and 1 (P < 0.001) (Fig. 2B). The SHIPE index for AAV patients. (A) The SHIPE index equation and the distribution of patients across group. (B) Comparison of cumulative ESKD-free survival rates based on the SHIPE indices. Combined analysis of cumulative ESKD-free survival rates for groups with SHIPE indices of 0 and 1 compared to a group with a SHIPE index of 2. SHIPE = Summation of the Highest tertiles of Independent Predictors of ESKD, FFS = Five-Factor Score, Cr = creatinine, AAV = antineutrophil cytoplasmic antibody-associated vasculitis, ESKD = end-stage kidney disease. ***P < 0.001. DISCUSSION In this study, we aimed to develop a novel baseline index that is ethnoregional, robust, and convenient for predicting future progression to ESKD in Korean patients with AAV. Several key findings emerged from our investigation. First, we prioritized creating an ESKD-predicting index with ethnoregional specificity. To achieve this, independent predictors of ESKD were identified using univariable and multivariable Cox proportional hazard analyses of baseline variables in Korean patients with AAV. FFS and serum creatinine at diagnosis were successfully validated as baseline predictors tailored to Korean population. We then developed the SHIPE index, a novel equation based on the highest tertiles of FFS and serum creatinine, prioritizing simplicity and practical applicability over conventional ROC-derived cut-offs. Importantly, a SHIPE index of 2 was demonstrated to be a significant Korean-specific predictive marker for ESKD progression. This study provides not an absolute formula but a methodological framework adaptable to diverse ethnic and regional cohorts. The second objective was to ensure the robustness of the SHIPE index. Comparative analysis of cumulative ESKD-free survival rates demonstrated that the SHIPE index ≥ 2 offered superior predictive utility compared to FFS ≥ 2 or serum creatinine ≥ 1.3 mg/dL alone. Although serum creatinine ≥ 1.3 mg/dL at diagnosis exhibited the highest RR (17.710; 95% CI, 6.875–46.492), the SHIPE index ≥ 2 (RR, 9.933; 95% CI, 5.118–19.471) provided greater robustness by incorporating two predictors. This is particularly critical in clinical scenarios where serum creatinine levels may fluctuate due to reversible conditions such as dehydration or acute illness.16,17 By integrating FFS, the SHIPE index compensates for these variations, ensuring reliability as a predictive tool. Additionally, the SHIPE index demonstrated significant predictive capacity for ESKD progression in real-world settings, offering a more reliable and practical approach compared to single-variable models like serum creatinine alone. Fig. 3 further illustrates the distribution of SHIPE index groups among non-ESKD and ESKD patients classified by other predictors, FFS and serum creatinine. Non-ESKD patients in these predictors predominantly included those with lower SHIPE index scores (0 or 1), while ESKD patients were mostly composed of individuals with a SHIPE index of 2. This observation highlights that the SHIPE index consolidates and enhances the predictive power of these individual variables, providing superior robustness and reliability in clinical applications. Comparison of three predictors—FFS, serum creatinine, and SHIPE index—for ESKD. Comparison of predictive performance among (A) FFS, (B) serum creatinine, and (C) SHIPE index. FFS = Five Factor Score, ESKD = end-stage kidney disease, Cr = creatinine, SHIPE = Summation of the Highest tertiles of Independent Predictors of ESKD. The FFS, a prognostic index specific to AAV, is composed of five items: age > 65 years, renal insufficiency, cardiac involvement, gastrointestinal involvement, and absence of ENT manifestations. It is primarily used to predict severe outcomes such as mortality, with 5-year mortality rates of 9%, 21%, and 40%, for total FFS scores of 0, 1, and 2, respectively.13 In this study, we emphasized that our intention was not to validate FFS itself but to pragmatically incorporate an accessible, composite prognostic score to improve risk stratification in our Korean cohort.13 The mechanisms by which baseline FFS predicts progression to ESKD are both direct and indirect pathways. Directly, baseline FFS reflects renal function decline at diagnosis, as one of its components, renal insufficiency, is defined by serum creatinine ≥ 1.7 mg/dL.13 Indirectly, baseline FFS reflects systemic involvement and inflammation in AAV, which increase ESKD risk over time. Persistent systemic inflammation may cause primary renal damage via inflammatory mediators and secondary renal damage due to organ dysfunction.18,19 These mechanisms highlight the broader prognostic role of FFS in predicting ESKD beyond its renal-specific components. Because a parameter of serum creatinine levels of more than 1.7 mg/dL is already included in FFS, there might be a concern about redundancy when deriving a formula using FFS (including a parameter of serum creatinine levels) and another parameter of serum creatinine levels of more than 1.3 mg/dL. However, we cautiously suggest that this concern could be overcome for three reasons: first, no multicollinearity issues were found between the two variables; second, independent predictive validity was ensured in multivariable Cox analysis; and third, after verifying independence as continuous variables in multivariable Cox analysis, the cut-off of each variable was determined using a new concept of the highest tertile. Another objective of this study was to prioritize the convenience of the ESKD-predicting index. To achieve this, the SHIPE index was specifically designed as a simple and practical tool for routine clinical use. The simplicity of the SHIPE index lies in its reliance on easily obtainable baseline variables—FFS and serum creatinine at diagnosis—and its use of categorical thresholds rather than requiring complex calculations or advanced statistical methods. Regarding the choice of cutoff values, we adopted a tertile-based approach rather than ROC-derived thresholds for several reasons. To begin with, tertile classification offers a straightforward and reproducible method that mitigates the risk of overfitting to a single-center cohort. Moreover, it provides flexibility for adaptation across populations with differing baseline characteristics, including geographic and ethnic variations in serum creatinine levels or disease severity. In addition, tertile cutoffs are less susceptible to sampling variability in relatively small datasets and have been employed in previous scoring systems that prioritized ease of clinical use and external applicability. For example, cardiovascular risk stratification models such as the Framingham Risk Score have commonly categorized continuous variables into tertiles or quartiles to facilitate group-based comparisons of risk (D’Agostino et al.’s study20). Similarly, a risk prediction model for chronic kidney disease (CKD) among patients with type 2 diabetes in Taiwan (DCMP cohort) used a Framingham-style risk point system and stratified scores into risk groups—with inherent grouping analogous to tertiles—to compare CKD incidence across strata (Lin et al.’s study21). Thus, while the thresholds of FFS ≥ 2 and serum creatinine ≥ 1.3 mg/dL were derived from our dataset, they are grounded in both clinical rationale and a practical framework that supports external adaptability. Unlike approaches that necessitate determining slopes or deriving cut-offs using multivariable regression and ROC analysis, the SHIPE index minimizes computational complexity. This focus on simplicity enhances its practicality, ensuring it can be readily implemented in diverse clinical settings without specialized tools or expertise. The most significant strength of this study was the development of the SHIPE index—an ethnicity- and region-tailored, robust, and convenient tool that incorporates two independent predictors of ESKD progression in AAV patients. This study not only demonstrates the clinical significance of the SHIPE index but also provides a framework for developing similar indices. Tailored to specific cohorts. The inclusion of a substantial sample size further enhances the reliability and generalizability of the findings supporting immediate application of the SHIPE index in clinical practice. However, this study has several limitations. First, its retrospective design may lead to missing data, preventing the incorporation of additional variables. Second, while the sample size in this study was reasonably large for AAV research, selection bias cannot be ruled out given the study population consisted of severely ill patients referred to a tertiary hospital. Third, a direct comparison between the SHIPE index and established tools such as the renal risk score was not conducted in this study, as these models require additional information (e.g., histopathology) that were not available in our cohort. Future studies using standardized clinical data will be necessary to determine whether the SHIPE index provides additional prognostic value for renal outcomes compared to existing models. Although, the SHIPE index offers simplicity and clinical feasibility, we acknowledge that dichotomizing continuous predictors may reduce the granularity of risk estimation. This limitation is particularly relevant for patients with borderline characteristics, who may not be adequately represented by binary classification. Future refinements may involve the use of continuous or tiered scoring systems to improve predictive accuracy while preserving clinical usability. Finally, independent predictors were based solely on baseline data, limiting the ability to assess renal function changes over time. Future multi-center prospective studies with serial monitoring of kidney function will be essential for providing more reliable, dynamic, and generalizable data to develop tailored indices for predicting ESKD progression in diverse AAV cohorts. In this study, we developed the SHIPE index, a novel and practical prediction tool for ESKD progression that integrates FFS and serum creatinine at diagnosis. The SHIPE index demonstrated significant clinical utility as a robust, simple-to-use, and ethnoregion-specific approach for identifying high-risk patients with AAV. By prioritizing convenience and practicality, the SHIPE index provides a user-friendly method suitable for routine clinical application. Additionally, this study offers a methodological framework for developing tailored prediction indices that account for specific ethnic and regional characteristics. Further multi-center and prospective studies are needed to validate the SHIPE index’s generalizability and predictive performance across diverse populations, ensuring its broader applicability in global clinical practice. Footnotes Funding: This study was funded by Eisai Korea Inc., Seoul, Republic of Korea (4-2024-0700), Yuhan Corporation, Seoul, Republic of Korea, as part of their “2024 Investigator Initiated Translation Research Program” (4-2025-0044), and Samsung Bioepis Co., Ltd., Incheon, Republic of Korea (4-2025-1066). The funder was not involved in the study design, collection, analysis, interpretation of data, the writing of this article, or the decision to submit it for publication. Therefore, we disclose that these analyses were pre-specified and conducted independently. Disclosure: The authors have no potential conflicts of interest to disclose. Author Contributions: Conceptualization: Do H, Lee LL, Lee SW.Data curation: Do H, Lee LL, Lee SW.Formal analysis: Do H, Lee LL.Funding acquisition: Lee SW.Investigation: Lee LL.Methodology: Lee LL.Project administration: Do H, Lee LL, Lee SW.Resources: Lee SW.Software: Lee LL.Supervision: Song JJ, Park YB, Ha JW, Lee SW.Validation: Lee SW.Visualization: Do H.Writing - original draft: Do H, Lee LL, Lee SW.Writing - review & editing: Song JJ, Park YB, Ha JW, Lee SW. Conceptualization: Do H, Lee LL, Lee SW. Conceptualization: Do H, Lee LL, Lee SW. Data curation: Do H, Lee LL, Lee SW. Data curation: Do H, Lee LL, Lee SW. Formal analysis: Do H, Lee LL. Formal analysis: Do H, Lee LL. Funding acquisition: Lee SW. Funding acquisition: Lee SW. Investigation: Lee LL. Investigation: Lee LL. Methodology: Lee LL. Methodology: Lee LL. Project administration: Do H, Lee LL, Lee SW. Project administration: Do H, Lee LL, Lee SW. Resources: Lee SW. Resources: Lee SW. Software: Lee LL. Software: Lee LL. Supervision: Song JJ, Park YB, Ha JW, Lee SW. Supervision: Song JJ, Park YB, Ha JW, Lee SW. Validation: Lee SW. Validation: Lee SW. Visualization: Do H. Visualization: Do H. Writing - original draft: Do H, Lee LL, Lee SW. Writing - original draft: Do H, Lee LL, Lee SW. Writing - review & editing: Song JJ, Park YB, Ha JW, Lee SW. Writing - review & editing: Song JJ, Park YB, Ha JW, Lee SW. References