We compared the results of chimerism analyses with real-time SNP-PCR to those obtained by the classical STR-PCR method in 135 hematopoietic stem cell transplantation recipients. Using 10 different SNP gene loci, the SNP-PCR method was able to discriminate patient from donor cells in 125 of 135 cases (93%), whereas the use of 11 different STR gene loci with the STR-PCR analysis using agarose or polyacrylamide gel resolution resulted in accurate donor-host discrimination in all patients. Of the 470 analyzed samples we found in 74% concordant results for both chimerism methods. In all 26% discordant cases the SNP-chimerism method showed mixed chimerism (MC), whereas the STR-method found complete chimerism (CC). As a consequence, the SNP-PCR chimerism analysis method detected a MC prior to the occurrence of relapse significantly earlier than the STR-PCR chimerism method (120 vs. 30 days, P < 0.007). The probability of relapses was significantly higher in patients with increasing MC (70%) compared to 30% in patients with CC (P < 0.00001) associated with a significantly shorter overall survival in patients with increasing MC. The multivariate Cox model showed that chimerism analsis with SNP-PCR was the only significant risk factor predicting relapse (RR 6.08, P < 0.0001).Furthermore, we analyzed the chimerism status in male recipients with a female donor in 580 samples of 134 patients using quantitative real-time PCR of Y-chromosome-specific sequences and compared the results with interphase XY-fluorescent in situ hybridization (FISH). MC without signs of relapse was detected in 35% of samples using quantitative real-time PCR of Y-chromosome-specific sequences. The detected Y-DNA amounts were low compared to the amounts detected in 104 samples of 42 patients with leukemic relapse at the time of analysis (P < 0.0001). Quantitative real-time PCR of Y-chromosome-specific sequences detected therefore an increasing MC with high residual host DNA amounts approximately 143 days (mean) prior to the occurrence of relapse. By comparing the results of Y-chromosome PCR with the XY-FISH analysis we found concordant results in 73% in patients with myeloablative regimens. The XY-FISH could detect 12 relapses, whereas the Y-chromosome PCR detect 36 relapses by MC (P < 0.005). Residual host cells gradually decreased during the posttransplant period from a mean of 5.4 ng (first months) to 0.5 ng (above 5 years) without evidence of relapses. The probability of relapses was significantly higher in patients with increasing MC (100%) compared to 8% in patients with CC (P < 0.00001) associated with a significantly shorter overall survival in patients with increasing MC. The multivariate Cox model showed that chimerism analysis of Y-chromosome-specific sequences is an important risk factor for relapse (RR 17.0, P < 0.0001). We conclude that the use of real-time SNP or Y-PCR may be superior to the STR-PCR or interphase XY-FISH methods in detecting patients who are at high risk for relapse after transplant.