Background. Long-term kidney graft monitoring still relies on delayed clinical indicators such as eGFR and proteinuria, which change only after substantial injury. To investigate biological processes not captured by standard metrics, we hypothesized that plasma proteomics could identify biomarkers and reveal the pathways underlying the pathological process, thereby providing independent prognostic value for late graft failure. Methods. In this retrospective matched case-control study, we analyzed pre-event plasma samples from 29 kidney transplant recipients who later experienced late, death-censored graft failure (DCGF; occurring ≥1 y posttransplant) and 49 matched recipients without DCGF from the TransplantLines Biobank. We quantified 800 proteins by label-free liquid chromatography with tandem spectrometry, followed by differential expression analysis and survival-related feature selection. Independent survival markers were evaluated using random survival forests. Biological context was analyzed via STRING network and enrichment analyses. Results. The cohort (n = 78) had a baseline age of 49.8 ± 14.1 y and estimated glomerular filtration rate of 36.9 ± 13.2 mL/min/1.73 m 2 . Median time between transplantation and baseline was 2.7 (0.7–7.8) y with 4.6 (2.1–5.8) y of follow-up. Seventeen proteins differed at baseline between cases and controls, including HLA-A and lipopolysaccharide-binding protein. Feature selection identified a 113-protein signature enriched in lipid metabolism, coagulation, and immunity. This signature provided significant prognostic value independent of clinical parameters such as estimated glomerular filtration rate and proteinuria (C-index 0.878 versus 0.851, P < 0.001). After accounting for confounders, apolipoprotein L1 and C1 were the most prominent. Conclusions. Proteomic profiling may provide novel biological insights into the pathogenesis of DCGF and may complement existing clinical metrics.
Shehab et al. (Fri,) studied this question.