BACKGROUND: Pre-transplant donor-specific anti-human leucocyte antigen (HLA) antibodies (DSAs) are associated with impaired kidney graft survival while the clinical relevance of non-donor-specific anti-HLA antibodies (nDSAs) is more controversial. The aim of the present paired kidney graft study was to compare the clinical relevance of DSAs and nDSAs. METHODS: To eliminate donor and era-dependent factors, a post hoc paired kidney graft analysis was performed as part of a Dutch multicentre study evaluating all transplantations between 1995 and 2005 with available pre-transplant serum samples. Anti-HLA antibodies were detected with a Luminex single-antigen bead assay. RESULTS: Among 3237 deceased donor transplantations, we identified 115 recipient pairs receiving a kidney from the same donor with one recipient being DSA positive and the other without anti-HLA antibodies. Patients with pre-transplant DSAs had a significantly lower 10-year death-censored graft survival (55% versus 82%, P=0.0001). We identified 192 pairs with one recipient as nDSA positive (against Class I and/or II) and the other without anti-HLA antibodies. For the patients with nDSAs against either Class I or II, graft survival did not significantly differ compared with patients without anti-HLA antibodies (74% versus 77%, P = 0.79). Only in patients with both nDSAs Class I and II was there a trend towards a lower graft survival (58%, P = 0.06). Lastly, in a small group of 42 recipient pairs, 10-year graft survival in recipients with DSAs was 49% compared with 68% in recipients with nDSAs (P=0.11). CONCLUSION: This paired kidney analysis confirms that the presence of pre-transplant DSAs in deceased donor transplantations is a risk marker for graft loss, whereas nDSAs in general are not associated with a lower graft survival. Subgroup analysis indicated that only in broadly sensitized patients with nDSAs against Class I and II, nDSAs may be a risk marker for graft loss in the long term.