Kidney transplant model may identify patients at lower rejection risk

After a kidney transplant, the patient faces a difficult trade-off: protect the new organ with immunosuppressive drugs while accepting their severe side effects. Scientists at Northwestern Medicine have developed a risk-stratification model that may identify patients whose donor-kidney mismatch carries a lower risk of rejection.

The immune system can make antibodies against mismatched human leukocyte antigens, or HLAs—proteins carried by the donor kidney. The team analyzed data from more than 3,000 recipients at five transplant centers in the U.S., Belgium, the Netherlands and Spain, looking beyond a simple count of mismatches to the biology of the HLA proteins involved.

Using measures of evolutionary divergence and amino-acid physicochemical qualities, the researchers found that HLA-DQ alleles—different forms of the HLA-DQ family—fell into two main groups that diverged during evolution. Mismatches across those groups were linked to a higher risk of de novo donor-specific antibodies, meaning antibodies newly generated against the donor. The researchers then divided the alleles into six functional groups and developed a qualitative model associated with antibody-mediated rejection and graft loss.

Concretely, the model could help doctors decide which patients might be candidates for less intensive immunosuppressive medication after transplantation, while keeping higher levels for those at greater risk. It could also guide the frequency of immune monitoring for donor-specific antibodies and invasive biopsies, as well as the safety of a stepwise reduction in medication. These are proposed clinical uses, not outcomes already demonstrated in routine care.

The next validation step is geographic: the researchers plan to validate the model with patient data from other regions. Anat Roitberg-Tambur, the study’s senior author, said the immediate post-transplant period is where risk stratification could have the greatest practical effect, but the model must first show that its biological risk hierarchy holds beyond the centers that generated the analysis.