Mechanistic Evaluation of Treatments for Acute Antibody-Mediated Rejection of the Kidney Transplant

Design: Prospective observational study of patients enrolled in TAR:GET-1

Objectives: TAR-GET-1 (starting in 2019) is a multi-centre UK clinical trial that will compare standard of care (SOC) treatment of Acute Antibody-Mediated Rejection of the Kidney Transplant (AAMR) to standard of care plus rituximab (SOCR). This trial presents a rare opportunity to investigate the mechanisms of action of treatments (including rituximab) in AAMR. The rationale for using rituximab is the depletion of CD20-positive precursors to antibody-producing cells. However, rituximab has effects on other B cell subtypes and not all studies have shown antibody level reductions. Furthermore, AAMR is a heterogeneous disease, as defined by variability in mechanistic markers of AAMR in blood (B cell and antibody characteristics) and in transplant biopsies (histological features and gene expression).

The objectives of this study are:

1. To investigate how treatment with rituximab affects these mechanistic markers, by comparing SOC to SOCR
2. To predict treatment benefit early, by analysing the effect of treatment on the mechanistic markers

Methodology: In patients enrolled in TAR:GET-1 (n=170), mechanistic markers will be analysed with established proof of concept (B cell and antibody characteristics, histological and transcriptional markers in the biopsy) at baseline and on follow-up, making use of clinical data, surplus diagnostic material from TAR:GET-1, and material from optional extra blood samples for research taken as part of TAR:GET-1, at 3,6 and 12 months. An optional protocol biopsy at 6 months post treatment will also be offered.

TARGET-1 is assessing DSA MFI and target and B lymphocyte numbers at baseline and 3 months post-treatment. This study will in addition analyse:

1. Serum DSA MFI and target at 6 months; complement-activating ability at baseline, 3 months and 6 months;
2. Biopsy Banff histological scores at baseline and in post-treatment indication biopsies and an optional protocol biopsy;
3. Biopsy AAMR-related transcript signature, assessed using Nanostring nCounter technology, at baseline and in post-treatment biopsies indication biopsies and an optional protocol biopsy;
4. B lymphocyte numbers at 6 months.

Differences pre- and post-treatment will be analysed comparing SOC and SOCR. Mediation analysis models will be used to decompose total effect of treatment effect on graft loss at 4 years (primary end point) into an indirect effect, which measures how much of the effect acts through the intermediate variable(s), and a residual direct effect.

Number of Subjects and Study populations: 170 subjects enrolled in TAR:GET-1

Outcome measures: same as primary outcome of TAR:GET-1 (graft loss at 4 years after initiation of treatment)

Inclusion Criteria: All patients enrolled in TAR:GET-1 ; Signed informed consent prior to any study specific procedures.

Exclusion Criteria: None

Duration of study: 8 years (duration is dictated by the need for primary outcome of TAR-GET-1 to be reached)

Statistical Methods: Data from this research project will be merged and linked with data from the TAR:GET-1 clinical trial by the statistician team.

Descriptive data: This will include description of all the biological characteristics analysed in 2 populations (rituximab + SOC and SOC) at baseline and at all follow-up time points. Summary statistics (mean, median, SD, range for continuous variables, and frequency tables for categorical variables) will be reported broken down by treatment, with exploratory figures (box-and-whisker plots for continuous, and stacked bar-charts for categorical).

Outcome analysis: Appropriate measures of differences between treatment groups will be estimated with point estimates, 95% confidence intervals, and p-values. Depending on the type of variable (continuous, categorical, time-to-event), regression techniques (linear, logistic, cox) and transformation of the endpoints will be considered.

Differences in characteristics comparing pre- to post-treatment samples at sequential time points will be analysed comparing the 2 groups of patients (SOCR versus SOC) using r mixed effect models repeat measurement techniques.

Mediation analysis, a special case of regression analysis, will be used to document that the treatment effects a change in the biomarker. Mediation analysis models will be used to decompose total effect of treatment effect on graft loss into an indirect effect, which measures how much of the effect acts through the intermediate variable(s), and a residual direct effect. All the mediation analyses will be based on the Intention-to-Treat principle. The same form of outcome models will be fitted as for the efficacy analyses, including the mediator as a covariate. The mediation package within R will be used which was developed specifically to make valid causal inference in explanatory analyses of the mechanisms of treatment-induced change in clinical outcomes in randomised clinical trials. All estimates of treatment differences will be summarised using effect size estimates and 95% confidence intervals (Dunn 2013). Collaborator for mediation analysis: Prof Ian White, Professor of Statistical Methods for Medicine, UCL.