#3704 Plasma cell infiltration predicts elevated steroid demand during short-term management of severe ANCA-associated renal vasculitides

Abstract Background and Aims Steroid-sparing regimen are key to improved management of the potentially life-threatening autoimmune conditions ANCA-associated renal vasculitides (AArV). However, treatment reality is hallmarked by undeniable steroid dependency necessitating indulgent weighting of steroid dosing and tapering due to steroid's toxicities and sequelae. Though outcome guidance heavily depends on renal histopathology, to date, impacting factors predicting steroid demand remain elusive. Therefore, we aimed to identify histopathological denominators of an increased steroid demand during the initial hospitalization phase in patients with AArV. Method A retrospective single-center study enrolling all cases of biopsy-affirmed AArV within the observational period of 2015 until 2021 was performed. After ethics approval, medical records were utilized to assess patient characteristics (age, sex, comorbidities, medication), date of admission and discharge from hospital, laboratory serum parameters at baseline (creatinine, eGFR, BUN, potassium, white blood cells, platelets, C-reactive protein, C3/C4 levels), and conducted treatment measures to induce remission (rituximab, cyclophosphamide, plasma exchange). Cumulative prednisolone demand (CPD) and ICU scores were calculated (APACHE II, SAPS II, SAPS III). Blinded histopathological analysis was independently performed by two nephropathologists evaluating the extent of tubulointerstitial immune cell infiltration, especially plasma cell (PC) richness (PC+) vs. PC scarcity (PC-). ROC curve analysis and maximized Youden index (MYI) calculation was conducted. Survival curve analyses were performed by means of Kaplan-Meier method and log rank testing. The event was defined as achieving a CPD level above the total cohort median of 2395 mg, exitus was censored. Results The total cohort (TC) contained n = 50 patients (22/50 being female) with a median age of 63 y (53–74), an in-hospital stay of 14 d (7–26) and a mean CPD of 2582 ± 1683 mg. CPD independently correlated with APACHE II score (β = 0.51, P < 0.001) implying the state of critical illness to be crucial for CPD elevation. We then identified the cut-off of 15.5 points (APACHE II) by means of MYI calculation in line with ROC curve analysis (AUC: 0.8, 95% CI: 0.7–0.9, P = 0.0003) (Fig. 1A), which enabled cohort subgrouping (ICU vs. non-ICU). Subgroup baseline characteristics were compared: sig. increase of CPD levels (4.5 g ± 1.1 g vs. 2.1 g ± 1.4 g, P < 0.0001) and sig. elevation of serum creatinine levels, 5.6 (4.4–7.2) mg/dL vs. 2.0 (1.1–3.6) mg/dL, characterized the ICU subgroup. CPD levels independently correlated with serum creatinine elevation (TC: β = ß.53, P = 0.019; ICU: β = 0.65, P < 0.029) (Fig. 1B) and PC infiltration (TC: β = 0.328, P = 0.03; ICU: β = 0.71, P = 0.033) (Fig. 1B), which was not observable in the non-ICU group. Finally, we performed a survival curve analysis applying the ICU-subgroup median of PC infiltration (0.3%) as a cut-off defining two groups: PC+ vs. PC−. PC infiltration of >0.3% sig. predicted increased CPD levels (logrank: P = 0.0036, HR 2.8, 95% CI 1.1–6.7) (Fig. 1C). Conclusion We here show that prognosis-affecting steroid demand during short-term management is determinable depending on distinct histopathological lesions and the state of critical illness, which might be implemented in future models enabling individualized outcome prediction for patients with ANCA-associated renal vasculitides.