Background: Infection with human cytomegalovirus (HCMV) is the most common congenital virus infection, affecting about 0.5-2% of newborns. Using DBS on Guthrie cards, it is possible to discriminate congenital from postnatal HCMV-infection. However, a recent European trial revealed serious problems in detection of low HCMV-DNA levels from DBS-filter-cards (Barbi et al., 2008).(7) Objectives: Evaluation of the most sensitive combination of sample size, DNA extraction method and PCR system for the detection of low copy numbers of HCMV-DNA from DBS-filter-cards. Study design: We compared three different manual extraction methods for the detection of HCMV-DNA out of DBS: the QIAmp-blood-Mini-Kit, a heat-extraction-method and traditional phenol-chloroform extraction. Additionally, we tested an automated nucleic acid extraction system (NucliSense Easy-Mag/Biomerieux). Different punch-sizes of DBS spiked with defined HCMV AD169-DNA copy numbers were analyzed. For detection, we used a quantitative in-house-LightCycler-PCR targeting the gB-region using the hybridisation-probe-format. We compared the sensitivity of the real-time-PCR with IE1Ex4-targeted nested-PCR. Results: The highest sensitivity with 200 copies HCMV DNA/ml was achieved using the phenol-chloroform method in combination with the nested-PCR and 6 mm, 3 x 3 mm punches or the whole DBS. The QIAmp-blood-Mini-Kit also showed a very high sensitivity by using the whole DBS and the nested-PCR. Conclusion: These results may have strong implications for retrospective diagnosis of congenital HCMV (cHCMV) infection, since a defined combination of the area of punch, the extraction method, and PCR method determine the probability of detection of viral DNA from DBS according to a logistic model. (C) 2010 Elsevier B. V. All rights reserved.