Background Oxidative stress is one of the most important components of the ischemia-reperfusion process after kidney transplantation (KTx) and increases with graft dysfunction. Methods This prospective study was conducted on 40 consecutive KTx recipients to evaluate time-dependent changes in oxidative stress-related parameters within the first week after KTx and to assess their performance in predicting delayed graft function (DGF=dialysis requirement during initial posttransplant week) and graft function at 1 year. Blood samples were collected before (day 0) and after KTx (days 1, 2, 4, and 7). Total antioxidant capacity, plasma levels of malondialdehyde (MDA), and activities of glutathione peroxidase, glutathione reductase and superoxide dismutase were measured. Multivariable linear mixed and linear regression models, receiver-operating characteristic (ROC), and areas under ROC curves (AUC-ROC) were used. Results At all time points after KTx, mean MDA levels were significantly higher in patients developing DGF (n=18). Shortly after KTx (8–12 hr), MDA values were higher in DGF recipients (on average, +0.16 μmol/L) and increased further on following day, contrasting with prompt functioning recipients. Day 1 MDA levels accurately predicted DGF (AUC-ROC=0.90), with a performance higher than SCr (AUC-ROC=0.73) and similar to cystatin C (AUC-ROC=0.91). Multivariable analysis revealed that MDA levels on day 7 represented an independent predictor of 1-year graft function. Antioxidant enzyme activities were not significantly changed during the study period and were not predictors of 1-year graft function. Conclusions Increased MDA levels on day 1 after KTx might be an early prognostic indicator of DGF, and levels on day 7 might represent a useful predictor of 1-year graft function.

Detrimental impact of preformed donor-specific antibodies (DSAs) against human leucocyte antigens on outcomes after kidney transplantation are well documented, however, the value of their capacity to bind complement for predicting antibody-mediated rejection (AMR) and graft survival still needs to be confirmed. We aimed to study DSA characteristics (strength and C1q binding) that might distinguish harmful DSA from clinically irrelevant ones. We retrospectively studied 60 kidney-transplanted patients with preformed DSA detected by single antigen bead (SAB) assays (IgG and C1q kits), from a cohort of 517 kidney graft recipients (124 with detectable anti-HLA antibodies). Patients were divided into DSA strength (MFI < vs. ≥ 15 000) and C1q-binding ability. AMR frequency was high (30%) and it increased with DSA strength (P = 0.002) and C1q+ DSA (P < 0.001). The performance of DSA C1q-binding ability as a predictor of AMR was better than DSA strength (diagnostic odds ratio 16.3 vs. 6.4, respectively). Furthermore, a multivariable logistic regression showed that C1q+ DSA was a risk factor for AMR (OR = 16.80, P = 0.001), while high MFI DSAs were not. Graft survival was lower in high MFI C1q+ DSA in comparison with patients with C1q− high or low MFI DSA (at 6 years, 38%, 83% and 80%, respectively; P = 0.001). Both DSA strength and C1q-binding ability assessment seem valuable for improving pretransplant risk assessment. Since DSA C1q-binding ability was a better predictor of AMR and correlated with graft survival, C1q-SAB may be a particularly useful tool.

In kidney transplantation, the impact of delayed graft function (DGF) on long-term graft and patient survival is controversial. We examined the impact of DGF on graft and recipient survival by accounting for the possibility that death with graft function may act as a competing risk for allograft failure. We used data from 1281 adult primary deceased-donor kidney recipients whose allografts functioned at least 1 year. The probability of graft loss occurrence is overestimated using the complement of Kaplan-Meier estimates (1-KM). Both the cause-specific Cox proportional hazard regression model (standard Cox) and the subdistribution hazard regression model proposed by Fine and Gray showed that DGF was associated with shorter time to graft failure (csHR = 2.0, P = 0.002; sHR = 1.57, P = 0.009), independent of acute rejection (AR) and after adjusting for traditional factors associated with graft failure. Regarding patient survival, DGF was a predictor of patient death using the cause-specific Cox model (csHR = 1.57, P = 0.029) but not using the subdistribution model. The probability of graft loss from competing end points should not be reported with the 1-KM. Application of a regression model for subdistribution hazard showed that, independent of AR, DGF has a detrimental effect on long-term graft survival, but not on patient survival.