Doppler spectrum analysis in the differential diagnosis of renal transplant dysfunction

Cyclosporin A (CyA) nephrotoxicity and rejection of a renal allograft each demands a specific therapy. This study was designed to establish the capability of Doppler spectrum analysis to diagnose either one of these causes during renal dysfunction. Between October 1989 and October 1991 we performed echo-Doppler examinations in 209 renal transplant recipients on a routine basis during the first three months after transplantation. Echo-Doppler examinations during periods of renal dysfunction were analyzed. A total of 93 periods of renal dysfunction, retrospectively due to rejection (n=40) or CyA toxicity (n=53), occurred in 70 patients during the study period. A control group consisted of 82 patients with normal functioning grafts. When compared to the control group, the Doppler features (in segmental arteries) of the rejection group showed significant lower frequency shifts [Fmax (Hz) 1637±423 vs. 1436±465; p<0.05; Fdia (Hz) 582±180 vs. 458±225; p<0.05], a shorter deceleration time of the Doppler spectrum [Tdown (ms) 340±100 vs. 276±102; p<0.05], and a higher Resistance Index (RI 0.64±0.08 vs. 0.68±0.13; p<0.05). Doppler spectra during CyA toxicity showed only a significantly longer acceleration time [Tmax (ms) 123±36 vs. 139±40; p<0.05]. The capability of differentiation between the two causes was assessed with ROC analysis of single Doppler features, stepwise regression and canonic discriminant analysis on a set of Doppler features and with manual selection of several features with extreme values. ROC analysis yielded maximum sensitivity and specificity for the diagnosis of rejection using Tdown (sensitivity 65%; specificity 68%). Stepwise regression and canonic discriminant analysis of a set of features rendered a sensitivity and specificity of 73% and 64%, respectively. Explorative selection of extreme Doppler feature values showed that 18 of the 40 grafts with rejection had values that were only seen in 2 cases with CyA toxicity (positive predictive value 90%; sensitivity 45%; specificity 96%). In half of these cases Doppler features preceded the clinical diagnosis of rejection by a median of 4 d. In conclusion, Doppler spectra are influenced by rejection and CyA toxicity in specific ways. The Doppler features, however, do not enable definite differentiation between rejection and CyA toxicity in all cases. Some changes in Doppler spectra are only seen in cases of rejection and thus enable positive identification of grafts with rejection, often earlier than clinical signs indicate rejection. A normal Doppler spectrum does not exclude rejection as the cause of renal dysfunction.