Validation of a nomogram for predicting the likelihood of delayed graft function in Australian adult deceased donor renal transplant recipients


To the Editor,

Delayed graft function (DGF), defined as the need for dialysis within 72 h of renal transplantation where the graft subsequently functions, is associated with significant morbidity, including an increased risk of acute allograft rejection, prolonged patient hospitalization, increased health-care costs and poorer graft survival.1–3 Pretransplantation prediction of the likelihood of DGF would therefore be extremely useful for informing patient and organ management decisions. Recently, a nomogram for predicting DGF was developed in the USA, based on United States Renal Data System (USRDS) data extracted from 13 846 adult deceased donor renal transplants performed between 1996 and 1999.4 To date, this nomogram has not been validated in an independent cohort and has not been evaluated outside the USA, where the incidence of DGF (23.7%) is higher than in other countries (e.g. Australia 15%).5

We retrospectively applied the US nomogram to 598 deceased donor renal transplants performed at the Princess Alexandra Hospital between 1994 and 2004 (89% Caucasian, 61% male, recipient age 44.6 ± 14.3 years, donor age 36.4 ± 16.3 years, cold ischaemic time 13.6 ± 4.8 h). Calculation of DGF scores was based on recipient factors (male +7, pretransplant diabetes mellitus +5, HLA mismatches ×1.925, panel reactive antibodies ×0.175, previous transfusions +4, previous transplants +10, pretransplant dialysis +58, no pretransplant dialysis +18, single-organ transplant plus pretransplant dialysis +20) and donor factors (age [years − 13]×5, anoxia +5, cerebrovascular accident +4, hypertension +7, non-beating-heart donor +29, cold ischaemic time hours ×1, terminal serum creatinine concentration µmol/L ×0.065). The overall rate of DGF in the study population was 5.9%, which was much lower than that of the USRDS training sample (23.7%). Higher DGF nomogram scores were associated with a higher risk of DGF (<110, 0%; 110–129, 1.8%; 130–139, 3.4%; 140–149, 10.9%; =150, 17.9%). A cut-off score of 140 was associated with a sensitivity of 74% and a specificity of 71%. The area under the receiver operator characteristic curve was 0.76 (Fig. 1), which compared favourably with the USRDS training sample (0.71).

Figure 1.

Receiver operator characteristic curve for the application of the delayed graft function nomogram to 598 deceased donor renal transplants performed at the Princess Alexandra Hospital between 1994 and 2004.

In conclusion, the USRDS nomogram performed well in the population in the present study despite a very low rate of DGF. Thus, the nomogram may have more universal applicability beyond the initial US study group in predicting DGF and may improve pretransplantation clinical decision making.