Abstract
Pathogen cell-free DNA (pcfDNA) in blood and urine is an attractive biomarker; however, the impact of preanalytical factors is not well understood. Blood and urine samples from healthy donors spiked with cfDNA from Mycobacterium tuberculosis, Salmonella enterica, Aspergillus fumigatus, and Epstein-Barr virus (EBV) and samples from tuberculosis patients were used to evaluate the impact of blood collection tube, urine preservative, processing delay, processing method, freezing and thawing, and sample volume on pcfDNA. The PCR cycle threshold (C(T) ) was used to measure amplifiable cfDNA. In spiked samples, the median C(T) values for M. tuberculosis, S. enterica, and EBV cfDNA were significantly lower in blood collected in K(2)EDTA tubes than those in Streck and PAXgene blood collection tubes, and they were was significantly lower in urine preserved with EDTA (EDTA-urine) than in urine preserved with Streck reagent (Streck-urine). Blood and urine samples from TB patients preserved with K(2)EDTA and Tris-EDTA, respectively, showed significantly lower median M. tuberculosisC(T) values than with the Streck blood collection tube and Streck urine preservative. Processing delay increased the median pathogen C(T) values for Streck and PAXgene but not K(2)EDTA blood samples and for urine preserved with Streck reagent but not EDTA. Double-spin compared with single-spin plasma separation increased the median pathogen C(T) regardless of blood collection tube. No differences were observed between whole urine and supernatant and between fresh and thawed plasma and urine after 24 weeks at -80°C. Larger plasma and urine volumes in contrived and patient samples showed a significantly lower median M. tuberculosisC(T) These findings suggest that large-volume single-spin K(2)EDTA-plasma and EDTA-whole urine with up to a 24-h processing delay may optimize pcfDNA detection.