Digital PCR (dPCR) is beginning to supersede real-time PCR (qPCR) for quantification of nucleic acids in many different applications. Several analytical properties of the two most commonly used dPCR platforms, namely the QX100 system (Bio-Rad) and the 12.765 array of the Biomark system (Fluidigm), have already been evaluated and compared with those of qPCR. However, to the best of our knowledge, direct comparison between the three of these platforms using the same DNA material has not been done, and the 37 K array on the Biomark system has also not been evaluated in terms of linearity, analytical sensitivity and limit of quantification. Here, a first assessment of qPCR, the QX100 system and both arrays of the Biomark system was performed with plasmid and genomic DNA from human cytomegalovirus. With use of PCR components that alter the efficiency of qPCR, each dPCR platform demonstrated consistent copy-number estimations, which indicates the high resilience of dPCR. Two approaches, one considering the total reaction volume and the other considering the effective reaction size, were used to assess linearity, analytical sensitivity and variability. When the total reaction volume was considered, the best performance was observed with qPCR, followed by the QX100 system and the Biomark system. In contrast, when the effective reaction size was considered, all three platforms showed almost equal limits of detection and variability. Although dPCR might not always be more appropriate than qPCR for quantification of low copy numbers, dPCR is a suitable method for robust and reproducible quantification of viral DNA, and a promising technology for the higher-order reference measurement method.

DNA extraction before amplification is considered an essential step for quantification of viral DNA using real-time PCR (qPCR). However, this can directly affect the final measurements due to variable DNA yields and removal of inhibitors, which leads to increased inter-laboratory variability of qPCR measurements and reduced agreement on viral loads. Digital PCR (dPCR) might be an advantageous methodology for the measurement of virus concentrations, as it does not depend on any calibration material and it has higher tolerance to inhibitors. DNA quantification without an extraction step (i.e. direct quantification) was performed here using dPCR and two different human cytomegalovirus whole-virus materials. Two dPCR platforms were used for this direct quantification of the viral DNA, and these were compared with quantification of the extracted viral DNA in terms of yield and variability. Direct quantification of both whole-virus materials present in simple matrices like cell lysate or Tris-HCl buffer provided repeatable measurements of virus concentrations that were probably in closer agreement with the actual viral load than when estimated through quantification of the extracted DNA. Direct dPCR quantification of other viruses, reference materials and clinically relevant matrices is now needed to show the full versatility of this very promising and cost-efficient development in virus quantification.

Quantitative PCR (qPCR) is an important tool in pathogen detection. However, the use of different qPCR components, calibration materials and DNA extraction methods reduces comparability between laboratories, which can result in false diagnosis and discrepancies in patient care. The wider establishment of a metrological framework for nucleic acid tests could improve the degree of standardisation of pathogen detection and the quantification methods applied in the clinical context. To achieve this, accurate methods need to be developed and implemented as reference measurement procedures, and to facilitate characterisation of suitable certified reference materials. Digital PCR (dPCR) has already been used for pathogen quantification by analysing nucleic acids. Although dPCR has the potential to provide robust and accurate quantification of nucleic acids, further assessment of its actual performance characteristics is needed before it can be implemented in a metrological framework, and to allow adequate estimation of measurement uncertainties. Here, four laboratories demonstrated reproducibility (expanded measurement uncertainties below 15%) of dPCR for quantification of DNA from human cytomegalovirus, with no calibration to a common reference material. Using whole-virus material and extracted DNA, an intermediate precision (coefficients of variation below 25%) between three consecutive experiments was noted. Furthermore, discrepancies in estimated mean DNA copy number concentrations between laboratories were less than twofold, with DNA extraction as the main source of variability. These data demonstrate that dPCR offers a repeatable and reproducible method for quantification of viral DNA, and due to its satisfactory performance should be considered as candidate for reference methods for implementation in a metrological framework.