One of the most important advances in transplantation medicine has been the recognition that anti-human leucocyte antigen (HLA) antibodies are destructive. Various studies over the past decade have indicated that the alloimmune response, mediated by anti-HLA antibodies, plays a key role in the failure of kidney allografts 1,2; this concept has been extended to heart, lung and composite tissue transplants 3. Although anti-HLA antibodies are considered to be harmful, there is a wide spectrum of graft injury related to these antibodies, ranging from no recognizable damage to florid rejection. The presence or absence and characteristics of donor-specific anti-HLA antibodies (DSA) are therefore crucial to determine patient outcomes following transplantation. Activation of the complement cascade is known to be a key component of antibody-mediated rejection (ABMR), and deposition of C4d is often used as a marker for complement activation 4–6. It is the ability of anti-HLA antibodies to bind complement that determines the cytotoxic potential of these antibodies 5. Therefore, we hypothesized that the presence of DSA and their ability to bind and activate complement would play a key role in solid organ transplant. A total of 1016 consecutive kidney transplant recipients were enrolled into this population-based study 7. All patients were screened for the presence of circulating DSA at the time of transplantation and at 1 year after transplantation, in patients with no immunological event or at the time of rejection using the Luminex assay. The complement-binding capacity of any detected DSA was also analysed. Of the 1016 patients, 1 year post-transplantation, 700 patients (69%) were devoid of circulating DSA and 316 patients (31%) had circulating DSA, 77 of whom (24%) were C1q-positive, i.e. had complement-binding DSA. The patients with complement-binding DSA (representing 7% of the overall population) had the poorest outcome, with a risk of graft loss 11-fold higher than patients with non-complement-binding DSA (C1q-negative) 7. Assessment of allograft biopsies (n = 1016) demonstrated that DSA, both complement (C1q-positive)- and non-complement (C1q-negative)-binding, were capable of inducing injury to the graft. Histologically, the incidence of microvascular inflammation, C4d deposition, early transplant glomerulopathy and interstitial inflammation tubulitis were significantly higher in allografts from patients with DSA (P < 0·05), particularly in those patients with C1q-positive, complement-binding DSA (P < 0·05) 7. These results were confirmed further by molecular microscopy with higher ABMR score, higher endothelial cell injury response and higher numbers of natural killer cells and interferon (IFN)-γ-inducible transcripts observed in patients with C1q-positive, complement-binding DSA. In order to translate our findings into clinical practice, we developed a risk prediction model at 1 year post-transplantation based on all histological and immunological parameters, defined by the current classical diagnosis tools. When C1q-positive, complement-binding DSA were added to the model we observed an increase in the performance and discrimination capacity, and reclassification of patients at lower or higher risk of graft loss 8. Compared with a traditional approach to predicting graft loss, we demonstrate that a new approach, which integrates the capacity of DSA to bind complement, identifies patients at high risk of graft loss and increases our risk stratification performance in kidney transplant recipients. An improved understanding of DSA mechanism of action, better patient identification and construction of risk models will pave the way to specific and personalized treatment for patients undergoing solid organ transplantation. As promising therapeutic agents targeting complement [e.g. a C5 inhibitor (eculizumab) or a C1 inhibitor] are used increasingly in patients undergoing transplantation, this study may serve as a basis for future clinical trials.