Abstract Background and Aims Despite the improvement in renal transplant results, the rate of graft loss from the first year has improved poorly, antibody-mediated rejection (AbMR) being the first cause of renal graft loss. Renal transplantation is usually monitored by measuring creatinine, proteinuria, immunosuppressive blood levels and, recently, donor specific antibodies and polyomavirus load. However, currently the only way to detect the alloimmune graft damage is to perform a surveillance o by indication renal biopsy. Non-invasive monitoring of the alloimmune response could contribute to improving renal transplant results. Previous studies have shown that urinary CXCL10 relates to both clinical and subclinical allograft rejection and graft loss. Our aim was to analyze the relationship between urinary CXCL10 and the biopsy findings of 1-year surveillance biopsies. Method From Dec/2013 to Nov/2017 90 surveillance biopsies were performed in our center at 1 year after kidney transplantation. Urine were prospectively collected at 6 and 12 month and urine samples were frozen at -70º. CXCL10 was measured by using human CXCL10/IP-10 DuoSet ELISA kits (R&D Systems) according to the protocols recommended by the company. Biopsies were classified according to Banff criteria. Results Median values of CXCL10 were 51.2 pg/ml (IQR 38.5 – 98.9 pg/ml) at 6 month and 62.8 pg/ml (IQR 42.7 – 99.4 pg/ml) at 12 month after transplantation. Both values were statistically related (r = 0.400, p = 0.012), and there were no significant statistical differences between both (Wilcoxon`s test p = 0.063). CXCL10 did not relate to serum creatinine, glomerular filtration rate, albuminuria, tacrolimus blood levels or prednisone dose at 1 year. Some 16.7% allograft biopsies showed subclinical cellular rejection, 26.7% showed antibody mediated rejection and 20.0% transplant glomerulopathy. Patients with antibody mediated rejection (AbMR) had higher levels of 1-year urinary CXCL10 [55.4 pg/ml (IQR 39.3) vs. 97.4 pg/ml (IQR 103.9), p = 0.021], lower GFR [68.0 ml/min/1.73m2 (IQR 33.5) vs. 43.5 ml/min/1.73m2 (IQR 22.8), p < 0.001] and similar albuminuria (p = 0.079). According to ROC curve, 1-year urinary CXCL10 could discriminate patients with subclinical AbMR (AUC-ROC 0.660, 95%CI 0.511-0.809, p = 0.021). After adjusting by GFR and albuminuria, 1-year urinary CXCL10 in a logarithmic scale was able to predict subclinical AbMR (OR 2.254, 95%CI 1.012-5.021, p = 0.047) by logistic regression analysis. Similarly, patients with subclinical AbMR at 1 year had higher levels of 6-month urinary CXCL10 [45.9 pg/ml (IQR 41.4) vs. 102.3 pg/ml (IQR 1935), p = 0.026]. Moreover, 6-month urinary CXCL10 could discriminate those patients who were going to show subclinical AbMR at 1-year (see figure; AUC-ROC 0.731, 95%CI 0.545-0.916, p = 0.027). After logistic regression, each tertil of 6-month urinary CXCL10 increased the risk of suffering subclinical AbMR more than thrice (OR 3.346, 95%CI 1.184-9.456, p = 0.023). Conclusion Both 6- and 12-month urinary CXCL10 relate to the diagnosis of subclinical AbMR by 1-year surveillance biopsies independently of other variables such as renal function and albuminuria. Of interest, by a non-invasive monitoring of urinary CXCL10 as early as at 6 months we can predict some subclinical histologic findings at 1 year, such as AbMR, that can have a strong impact on the transplant outcome.
Read full abstract