Association of Ultrastructural Changes in Renal Allograft Biopsies With Diagnostic Elements of Antibody-Mediated Rejection and Graft Outcomes

Background. The role of electron microscopy (EM) in the diagnosis of kidney allograft pathologies, particularly immunologic injury has not been well studied. Methods. In this retrospective, single-center cohort study, we examined EM features in 796 biopsies from 623 patients at high risk for antibody-mediated rejection, with glomerular abnormalities in light microscopy, presence of donor-specific antibody (DSA), or any degree of albuminuria/proteinuria. Results. Glomerular endothelial cell enlargement (GECE) > 50% was present in 29.1%, subendothelial expansion/basement membrane duplication in 24.5%, and peritubular basement membrane multilamellation > 4 (PTCML) in 18.5%. There was an incremental odds of worsening GECE from no DSA to class I DSA (odds ratio [OR], 2.75, P < 0.001; 95% confidence interval [CI], 1.7–4.5), class II DSA (OR, 3.44, P < 0.001, 95% CI, 2.5–4.7) and both classes (OR, 6.3, P < 0.001; 95% CI, 4.1–9.8). Moreover, the increase in number of antibodies was predictive of higher likelihood of worsening GECE (OR, 2.81, P < 0.001; 95% CI, 2.1–3.8 for 1 DSA; OR, 5.29, P < 0.001; 95% CI, 3.5–7.9 for 2–3; and OR, 8.45, P < 0.001; 95% CI, 4.7–15.3 for ≥4). Similar association was observed with PTCML. In multivariate analysis including DSA, subendothelial expansion/basement membrane duplication, and GECE >50%, but not PTCML were independently predictive of graft failure over mean follow-up of 63 mo (hazard ratio [HR], 1.6, P = 0.006, 95% CI, 1.2–2.3; HR, 2.0, P < 0.001; 95% CI, 1.4–2.9, respectively). Among a cohort with g, ptc, cg, and C4d scores 0, GECE >50% was independently associated with graft failure (HR, 2.58, P < 0.001, 95% CI, 1.6–4.3). Conclusions. These observations support the wider use of EM in kidney transplant biopsies to help with earlier diagnosis of antibody-mediated rejection and to risk stratify the graft outcome.