Rituximab, plasma exchange and immunoglobulins: an ineffective treatment for chronic active antibody-mediated rejection

We retrospectively reviewed our pathology database between 2006 and 2015, identified all patients with TG (cg ≥ 1 in the Banff histopathological classification) and re-evaluated them according to Banff 2017 classification criteria [3].

The inclusion criteria were the coexistence of c-aABMR TG with microvascular injury (MVI) ≥2 (g + ptc ≥ 2), with positive or negative C4d staining in peritubular capillaries and positive donor-specific antibodies (DSA). The patients with compatible histology, but negative DSA were included in the analysis as suspicious of c-aABMR. Also, TG plus positive C4d or TG with MVI =1 plus positive DSA were included. The presence of concomitant cellular rejection required at least g of 1.

The decisions to perform a renal biopsy and patient treatment were based on the clinical judgment at c-aABMR diagnosis.

Every patient who received treatment with RTX, IVIG, and PE after diagnosis of c-aABMR was included in the treatment group. Patients who did not receive RTX, IVIG neither PE, alone or in combination, were included in the control group.

PE was performed in Cobe Spectra or Spectra Optia separators (Terumo BCT, Lakewood, CO, USA) using 5% albumin (Albutein® 5%, Grífols, Spain) as a replacement solution. One plasma volume was exchanged in each session as previously reported [14].

The primary endpoint was graft survival. Secondary endpoints were the evolution of glomerular filtration rate and complications related to treatment.

The Institutional Ethics Committee approved the study. The registration number was 14,566/RG 24161 (Agencia Española del Medicamento y Producto Sanitario, AEMPS).

The clinical characteristics, immunosuppression, and treatment were analyzed at the time of c-aABMR diagnosis and in the follow-up period. We assessed Charlson comorbidity index (CCI) at diagnosis of c-aABMR [15]. In accordance with other groups, CCI was unadjusted for age, and the minimum score of our patients was 0 (2 points for renal insufficiency were not added) [16‐19].

We assessed renal function and proteinuria at c-aABMR diagnosis, 6 months before and 6, 12, 24 months after diagnosis and at the end of follow-up. Renal function was determined by serum creatinine and by estimated glomerular filtration rate (eGFR) using the Modification of Diet in Renal Disease (MDRD) formula [20]. Proteinuria was evaluated using the proteinuria/creatinine index [21].

Serum samples obtained at the moment of transplantation and rejection were screened for HLA class I and II antibodies using the Lifecodes LifeScreen Deluxe flow bead assay (Immucor, Stamford, CT, USA). Antibody specificities were determined using the Lifecodes Single Antigen bead assay (Immucor, Stamford, CT, USA) in patients with positive HLA antibodies.

Infections that required hospitalization at least 48 h during the first year after c-aABMR diagnosis were analyzed. All immediate adverse events (AE), after IVIG, PE and RTX infusion were registered.

Data were described as mean with standard deviation (SD) or standard error of the mean (SEM), in case of graphical presentation for quantitative variables and as absolute and relative (%) frequencies for qualitative variables. Group comparisons between patients with or without the intake of rituximab were made by Fisher’s exact, t-test or Mann Whitney U test for independent groups. Kaplan-Meier and comparison between both groups were made using Log-Rank test, considering non-related death as censure. We performed a crude estimation of the effect of treatment on the evolution of eGFR by generalized estimating equation (GEE) model using an AR(1) matrix to estimate the intra-subject correlation. These models included treatment effect, time of follow-up and their interaction with treatment. GEE models of treatment effect were adjusted for confounding factors including infectious disease. As a useful clinical evaluation of prediction values of change of eGFR (Δ of change) from 6 months before to rejection, we proposed a cutoff using the Likelihood Ratio (LR⁺) defined as ratio sensitivity/(1-specificity) from a ROC curve [22]. Estimation of risk of graft loss at 24 months due to treatment or a high delta of change of eGFR was performed calculating odds ratios (OR) and their confidence intervals at 95% (CI 95%) from logistic regression models. All statistical analyses were performed using IBM SPSS statistics V20.0 software (IBM Corp, Armonk, NY, USA). Two-sided P-values < 0.05 were used to indicate statistical significance.

In all treated patients RTX was initiated between one and 3 weeks after c-aABMR diagnosis. 82.6% of patients treated with RTX received two doses with a mean cumulative dose of 1008 ± 342 mg.

The mean number of PE sessions was 5.8 ± 0.38, and mean total processed volume was 24.2 ± 5.4 L. The dose of IVIG was 200 mg/kg, after every second PE.

In the control group, six patients presented concomitant acute cellular rejection and were treated with corticoids.

Graft survival censoring death was not different between both groups (Fig. 1), Log Rank P = 0.92. The proportion of graft loss at 12 and 24 months after c-aABMR diagnosis in treated and control groups was not statistically significant, 7 patients (30%) vs. 8 patients (34.7%) and 8 (34.7%) vs.13 (33.3%) respectively.

Four patients died in the treated group, two of sepsis (10 and 45 months after the initiation of treatment) and two of sudden death at home (3 and 64 months after c-aABMR treatment). None of the patients in the control group died.

The mean eGFR at diagnosis of c-aABMR and 6 months before was not different between groups (Table 1 and Fig. 2a). Also, proteinuria at diagnosis was similar in both groups (Table 1).

The mean eGFR at 6, 12 and 24 months in treated and control patients was not different (Fig. 2a). Even if we split according to graft outcome eGFR follow-up was similar between treated and control patient. (Figs 2b and Additional file 1: Figure S1).

An elevated ΔeGFR (eGFR 6 months before diagnosis – eGFR at diagnosis of cABMR) was related to graft loss during the first 24 months after diagnosis. A proposed cut of 13 ml/min in ΔeGFR was obtained from ROC analysis with LR⁺ = 3.34. A decrease of eGRF of 13 ml/min or more was an independent indicator of graft loss in the first 24 months with OR = 5 (95% CI = 1.5–16.9; P = 0.01). The impact of ΔeGFR was influenced neither in magnitude or statistical significance in a multivariate approach, adjusted by treatment (Table 2). Also proteinuria higher than 2.5 g/day (LR⁺ = 3.6), adjusted for treatment, was associated with loss of graft at 24 months in both groups (OR = 3; 95% CI = 1.22–7.37; P = 0.016).

We have evaluated the impact of treatment on the deterioration of renal function in a longitudinal model analysis (Fig. 3), showing that the impairment of eGFR is independent of the treatment in a crude estimation model. Also, we evaluated the impact of treatment adjusted by possible confounding factors such as age, Charlson’s index, graft loss, IFTA, microvascular inflammation (MVI), transplant glomerulopathy (TG), glomerulitis and peritubular capillaritis scores, proteinuria or presence of infections. However, the decrease in eGFR remains independent of the treatment (Fig. 3).

At diagnosis, histologically acute inflammatory and chronic lesions related to c-aABMR and TG were similar in both groups (Table 3). Transplant glomerulopathy and IFTA were similar in treated and control patients, 1.74 ± 0.83 vs. 1.83 ± 0.77 (P = 0.54) and 1.61 ± 0.78 vs. 1.83 ± 0.84 (P = 0.27) respectively. The presence of IFTA ≥ two was associated with graft loss at 24 months, (OR = 4.7; 1.19–18.5; P = 0.02).

Ten of the 23 treated patients had post-treatment follow-up biopsies within the first year after treatment with persisting active c-aABMR in all ten biopsies.

Infections were respiratory (8), urinary tract (6), cutaneous and mucosal (3), abdominal (4), disseminated zoster (1), and sepsis (1). The microbiological isolation was negative in 12 cases, 3 Pseudomonas aeruginosa, 2 Cytomegalovirus, 1 Klebsiella pneumoniae, 1 Escherichia coli, 1 Enterococcus faecalis, 1 Campylobacter jejuni and 1 Herpes Zoster virus.

A CCI of 3 was associated with more infectious complications in the control group (OR = 8.7; 95% CI = 1.15–65.9; P = 0.036), but not in the treated patients (P = 0.16). Related to adverse reactions in PE sessions or RTX infusion, only one patient developed tetany related to hypocalcemia.