The aim of the study was to describe the incidence of borderline lesion on kidney graft and the evolution after steroid treatment.
Between 1/2014 and 12/2020, 629 renal transplants were performed in our institution in which, due to the clinical reasons, 803 renal biopsies were performed. 25 of them showed BL lesions according to the Banff classification. (3.1%).
T test and chi square were used to determine statistical differences.
Mean age of these recipients was 46.3 + 16.9 years, and their donor 48.8 + 14.7 years. 19 received a kidney from deceased donor and 6 from living related.  10 were women and 15 men.
All were 1st transplants and no one were hypersensitized. No 0 missmatch was seen but there were 2 patient with 6 MM. All pre-transplant cross-match were negative. 100% of patients received induction, 21 with polyclonal antibodies and 4 with basiliximab. Regarding maintenance immunosuppression, all received corticosteroids and mycophenolate, either sodium or mofetil, while 17/25 received tacrolimus, 6 Belatacept and 2 received sirolimus.
The median follow-up at BL diagnosis was 207 days (range 10 to 1777 days). The temporal distribution was: 5/25 occurred during the 1st month, 12 between the 2nd month and the year and 8 on long term follow up. All the recipients with BL were treated at the same way: 3 intravenous steroids bolus with tapered to baseline levels. Only those that showed no clinical improvement were re-biopsied.
Post-treatment outcome of BL was good in only 13/25 (52%) of the cases with renal function recovery. The remaining 12 evolved as follows: 5 celular rejection (Banff I) and 7 evidence of humoral component with C4d + staining and presence of donor specific antibodies. 2 of them lost their graft due to chronic nephropathy in short time in spite of standard rescue treatment.
In order to find BL evolution predictors, we analyzed some variables as gender, age, MM, type of induction, time to rejection and basal immunosuppression. Only the presence of Belatacept as a maintenance drug was marginally significant (p=0.07) in favor of a better outcome.
Fibrosis score progression were seen on the subset of biopsies belong patients with bad evolution. 
As already described, BL rejection is impressive in its behavior ranging from resolution, even spontaneous, to progression to humoral injury. Although the number of patients involved is too small to draw conclusions, none of the variables analyzed correlated with the outcome, except for the use of Belatacept, which was marginally associated with a better prognosis. The literature supports the use of Belatacept de novo as associated to low incidence of DSA. This could be an explanation to our findings. Prospective study with protocol biopsy are needed to clarify BL behavior.