THE DIAGNOSTIC ACCURACY OF PERITUBULAR CAPILLARY C4D STAINING FOR ACUTE ANTIBODY MEDIATED REJECTION IN KIDNEY TRANSPLANT RECIPIENTS: A SYSTEMATIC REVIEW
R. Sapir-Pichhadze, S.P. Curran, A.C. Tricco, E. Uleryk, R. John, K. Tinckam, J. Beyene, A. Laupacis, A. Logan, S.J. Kim
University of Toronto, Toronto, Ontario, Canada
Introduction & objectives: Past limitations in histopathology and donor-specific antibody (DSA) assays necessitated the development of a construct reference standard (histopathology, DSA, and peritubular capillary C4d) for the diagnosis of acute antibody-mediated rejection (AAMR). Based on original reports showing favorable performance, C4d has become the cornerstone of AAMR diagnosis. Over the past two decades, definitions of histopathological changes have been refined and the performance of solid phase DSA assays has improved. In light of these advances, we conducted a systematic review to assess the performance of C4d compared withcurrent histopathological features of AAMR and DSA assays.
Methods:MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials and ProQuest Dissertations & Theses Database were searched until December 2012 with an information specialist. Cross-sectional and cohort studies discussing kidney transplant recipients who underwentallograft biopsies to rule out rejection in the presence of acute allograft dysfunction and allowing the derivation of a diagnostic table for the performance of C4d by immunofluorescence (IF) or immunohistochemistry (IHC) compared with histopathological changes of AAMR or DSA assays, were eligible for inclusion. Histopathological changes of AAMR included microcirculatory inflammation (peritubular capillaritis or glomerulitis) and vascular changes (capillary fibrinoid necrosis or thrombotic microangiopathy). DSA assaysincluded solid phase or cell-based cytotoxic antibody screening. Modified QUADAS-2 criteria were used to assess study quality. We evaluated between test agreement and computed sensitivity, specificity, and likelihood ratios for positive and negative tests for each study. Reference screening, data abstraction and quality appraisal were conducted by 2 reviewers.
Results: Of 3,027 unique abstracts, thirteen eligible studies including 2,245 allograft biopsies were identified. Risk of bias was primarily related to participant selection and lengthy intervals between C4d and comparator tests. In the highest quality studies, focal or diffuse C4d by IHC-paraffin exhibited fair agreement with peritubular capillaritis (Kappa: 0.3), glomerulitis (Kappa range: 0.1-0.4), andsolid-phase DSA assays (Kappa range: 0.2-0.3). Low sensitivity and moderate specificityfor AAMR detection, compared with peritubular capillaritis (sensitivity range: 0.19-0.49; specificity range: 0.83-0.96), glomerulitis (sensitivity range: 0.23-0.72; specificity range: 0.79-0.96), andsolid-phase DSA assays (sensitivity range: 0.31-0.68; specificity range: 0.63-1.00) was also identified.Specificity of IHC-paraffin C4d was lower in studies using thresholds of <10%. Two studies assessing IF-frozen C4d provided markedly different diagnostic characteristics as a consequence of heterogeneity in sample selection, comparator test definitions, and C4d thresholds. The positive likelihood ratio of C4d by IF-frozen and IHC-paraffin was typically ≥ 2, and the negative likelihood ratio was approximately 1.
Conclusions: Acknowledging limitations related to heterogeneity between studies and partial reporting of study details preventing complete assessment of quality items, our findings suggest that absence of C4d does not rule out AAMR and support the inclusion of “C4d negative AAMR” in the Banff classification. Whether DSA or microcirculatory inflammation in isolation, justify treatment for AAMR, warrants further study.