In kidney transplant recipients (KTR), C-reactive protein (CRP) has been shown to be associated with increased mortality, but data on this association within the high-sensitivity (hs) range of CRP (<5 mg/l) are lacking. We prospectively studied 710 prevalent and stable KTR over >6 years. We thawed frozen plasma and measured baseline hs-CRP using an ultrasensitive assay. Detailed clinical and demographic baseline characteristics were available for study. We stratified patients by quartile of hs-CRP within the hs range (<5 mg/l), and also included KTRs whose hs-CRP was above the hs range (>5-10 and >10 mg/l). We used multivariate proportional hazards models to test for independent associations. After careful multivariate adjustment, we found a J-shaped association between hs-CRP and mortality. Compared with KTR whose hs-CRP was in the second lowest quartile of hs-CRP (0.06-1.26 mg/l), patients in the lowest quartile (<0.06 mg/l) had more than twice their mortality risk (HR = 2.07; 95% CI: 1.05-4.07), as did patients whose hs-CRP was > or =2.44 mg/l (all HRs >2.27). No association was found between hs-CRP and death-censored allograft loss. In contrast to the general population, the association between hs-CRP and mortality in KTRs is not linear, but J-shaped, suggesting that KTRs with very low hs-CRP may also be at increased risk of death.