In this single center retrospective observational cohort study, we evaluated the outcome and clinical relevance of intimal arteritis in the early posttransplant biopsies performed within 12 months from January 2014 to December 2017. For comparison, we retrospectively reviewed medical records of 881 patients who had undergone kidney transplantation at the same time to identify those with histologic features of ABMR in early biopsies and enrolled them for further investigation. Data collection was finalized in January 2021 when all the study subjects reached a 3-year follow-up. Demographics of the study cohort are given in Table 2 and the enrollment is described in Figure 1. All patients signed the informed consent with case or surveillance biopsies and medical data assessment, and the local Ethical Board approved the study under No.: 15-265191A.

Kidney allograft biopsies were performed using a percutaneous ultrasound-guided 16G biopsy needle. Diagnoses of acute rejection were established at a median time 70 days after transplantation by for-cause biopsies in 120 (69%) of patients. Fifty-two (31%) rejections were found in protocol biopsies at the third month after transplantation.

Biopsy samples were assessed for Banff scored lesions (2) as glomerulitis (g), peritubular capillaritis (ptc), transplant glomerulopathy (cg), intimal arteritis (v), interstitial inflammation (i), tubulitis (t), mesangial matrix increase (mm), vascular intimal fibrosis (cv), arteriolar hyaline thickening (ah), interstitial fibrosis (ci), or tubular atrophy (ct). The microvascular inflammation (MVI) score was defined as the sum of glomerular (g) and peritubular capillary inflammation (ptc). Immunofluorescence detection of C4d was performed in all cases.

Patients with glomerulitis, peritubular capillaritis, intimal arteritis, and/or TG were divided into five groups (TCMRV N = 26, ABMRV/DSA– N = 29, ABMRV/DSA+ N = 17, ABMR/DSA– N = 47, ABMR/DSA+ N = 55) according to the histopathological finding and presence or absence of DSA (Table 1).

T cell-mediated vascular rejection was characterized by intimal arteritis and tubulointerstitial inflammation (TI) in the absence of glomerulitis of rejection origin, C4d, and DSA. ABMRV included features of MVI and C4d (positive or negative) in addition to v-lesion. Patients were further divided according to their DSA status as ABMRV/DSA– when DSA were negative and ABMRV/DSA+ when DSA were positive. ABMR, defined by the presence of MVI and C4d (positive or negative), were further divided according to DSA status as described above (ABMR/DSA–, ABMR/DSA+). Four phenotypes fulfilling criterion 1 (histologic evidence of acute tissue injury) and 2 (evidence of recent antibody interaction with endothelium) of Banff ABMR definition were called histologic ABMR groups.

Identification of circulating donor-specific anti-HLA antibodies (DSA) was performed by Luminex bead-based assay (One Lambda Inc., Canoga Park, CA, USA). Class I A, B, and class II DR antibodies were evaluated for specificity in all recipients before transplantation. Class II DP and DQ antibodies were evaluated for specificity after transplantation because donors' HLA DP or DQ typing was not available at the time of transplantation. DSA positivity at the time of transplantation was revealed in 14 (82%) of the patients with ABMRV/DSA+ and 47 (86%) of them with ABMR/DSA+. In the rest of the patients from these phenotype categories, DSA was detected at the time of diagnostic biopsy.

Patients received maintenance immunosuppression based on tacrolimus, mycophenolate mofetil, and corticosteroids. All patients received induction treatment. Patients at low risk (N = 73) received basiliximab whereas those at high risks, such as retransplants and patients with anti-HLA antibodies or DSA, received rabbit anti-thymocyte globulin (rATG) (N = 49). Patients with DSA positivity (<5,000 MFI) at transplant had undergone plasma exchange (PE) prior to transplantation and intravenous immunoglobulin (IVIG) as additional treatment (N = 52). The later described multivariable Cox regression model was adjusted based on induction immunosuppression to eliminate confounding variables.

Patients with rejection were treated by steroid pulses [18 (69%) TCMVR; 8 (28%) ABMRV/DSA–; 1 (6%) ABMRV/DSA+; 37 (79%) ABMR/DSA-; 17 (31%) ABMR/DSA+] or rATG [7 (27%) TCMRV; 15 (52%) ABMRV/DSA–; 3 (18%) ABMRV/DSA+; 2 (4%) ABMR/DSA–, none of ABMR/DSA+] and/or plasmapheresis/IVIG [1 (4%) TCMRV; 6 (21%) ABMRV/DSA-; 13 (77%) ABMRV/DSA+; 5 (11%) ABMR/DSA–; 33 (60%) ABMR/DSA+]. Altogether, eight patients were not treated mostly due to concomitant infectious complications.

Altogether, 124 patients from the whole cohort underwent at least one biopsy after the diagnostic examination. All patients who experienced intimal arteritis in early biopsies were eligible for surveillance biopsy at 3 months after the first biopsy. After obtaining written informed consent, 60 out of the 72 patients (83%) who experienced intimal arteritis underwent the surveillance biopsy. Patients from ABMR/DSA– and ABMR/DSA+ groups underwent subsequent biopsies either at the time of center protocol biopsy at the third month or when clinically indicated due to graft function worsening or proteinuria.

Continuous variables were expressed as median and range. Categorical variables were expressed as n and percentage of the total. Categorical Banff histologic scores were expressed as count per category. The Chi-square, ANOVA, Kruskal–Wallis, and Wilcoxon tests were used for hypothesis testing when appropriate. p < 0.05 were considered statistically significant. Survival analyses were performed with the Kaplan–Meier method using the log-rank test. To identify factors associated with death-censored graft failure and the development of TG, univariable Cox regression was created. For the multivariable model, all variables with p < 0.1 were selected, and only variables without missing data were included. Data analyses were performed using IBM SPSS 22 (SPSS, Inc. Chicago, IL), R Studio 4.0.3. (2020-10-10), and GraphPad Prism5 (Graph Pad, San Diego, CA).

In 174 patients, five different rejection phenotypes were identified: 26 TCMRV, 29 ABMRV/DSA–, 17 ABMRV/DSA+, 47 ABMR/DSA–, and 55 ABMR/DSA+ (Table 1). Diabetes, ischemic nephropathy, and glomerulonephritis were the most common original diseases. DSA– phenotypes were more common after the first transplantation. DSA positive phenotypes were more frequent in retransplantation. Peak PRA was significantly higher in DSA positive groups. Cohort demographics are given in Table 2.

Surveillance biopsies at 3 months after vascular rejection diagnosis were performed in 21 out of 26 patients (81%) with TCMRV, 24 out of 29 (83%) patients with DSA negative ABMRV, 15 out of 17 (88%) patients with DSA positive ABMRV. At case biopsy, DSA negative patients with ABMRV did not differ from DSA positive ABMRV patients in any of the histologic scores, indicating acute inflammation (g, ptc, i, t, v). The only significant difference was the intensity of C4d (p = 0.003). Surveillance biopsies revealed the resolution of rejection in 19 (90%) TCMRV cases, 14 (58%) ABMRV/DSA- cases, and only four (27%) ABMRV/DSA+ cases (p = 0.006). When comparing case and surveillance biopsies, patients with TCMRV had significantly improved i, t, and v scores, and DSA– patients with ABMRV had significantly improved in g, ptc, i, t, and v scores; however, in DSA+ ABMRV cases, only the v-score had improved (Figure 2, Table 3).

Significant differences between ABMR rejection phenotypes were found when we studied the development of the TG. For cause and surveillance, biopsies at the 3-year follow-up were included in the analysis. Estimate of the cumulative proportion of patients without TG (TG-free survival) at 36 months was 100% in TCMRV, 85% in ABMR/DSA-, 65% in ABMRV/DSA–, 54% in ABMR/DSA+, and 31% ABMRV/DSA+ (log rank p < 0.001). Pairwise comparisons showed a significantly longer time to TG in the TCMRV group than in other groups (log rank p < 0.05). Interestingly, DSA– ABMRV patients had a significantly shorter time to TG development than DSA– ABMR patients (mean for survival time without TG was 26 and 32 months, respectively) (log rank p = 0.035) (Figure 3).

Cox regression models were performed to identify risk factors for the development of TG in patients with histologic ABMR. Risk factors identified in the univariable analysis were included in the multivariable model (retransplantation status, DSA positivity, rATG therapy, rATG + PE/IVIG therapy, intimal arteritis at case biopsy, and C4d positivity). Multivariable model adjusted for the variables mentioned above found intimal arteritis to be associated with TG development (HR 2.1, p = 0.01) (Table 4).

The graft survival was significantly impaired in DSA+ patients with intimal arteritis as a part of ABMR compared with others (Figure 4). Death-censored kidney graft survival at 36 months was 96% in TCMRV, 98% in ABMR/DSA–, 86% in ABMRV/DSA–, 79% in ABMR/DSA+, and 64% in the ABMRV/DSA+ group (log rank p = 0.001). Interestingly, DSA– patients with intimal arteritis as part of ABMR experienced similar graft survival as DSA+ patients with ABMR (p = 0.507).

Cox regression assessing risk factors for graft failure in ABMR revealed PRA > 20% (HR 4.4, p = 0.001), DSA positivity (HR 3.9, p = 0.007), retransplantation (HR 4.0, p = 0.002), induction with ATG and ATG with additional PE, IVIG (HR 4.1, p = 0.021), and C4d positivity (HR 2.5, p = 0.041) in the univariable analysis (Table 5). None of these variables were found to be significant in the multivariable model.