B-100 Impact of underfilling blood collection tubes on plasma cell-free DNA and protein assays

Abstract

Liquid biopsies enhance early cancer detection by enabling “multiomic” interrogation of several analyte classes simultaneously. For optimal performance, blood collection tubes (BCTs) must preserve these analytes closely to their in vivo state. Many BCTs use formaldehyde-based fixation to preserve cellular structure via cross-linking biomolecules, but this may introduce bias in some analytes. Further, manufacturers caution against underfilling the BCTs, which increases the preservative-to-sample ratio, but low-fill conditions are common in clinical practice, and their impact on multiomic assays remains largely uncharacterized. In this study, we established two distinct experimental systems to separately evaluate low-fill impact on genomics and proteomics assays using the Streck cell-free DNA (cfDNA) tubes. Genomic Assay: To simulate different fill levels (1–10 mL), varying preservative doses were mixed with fixed masses of sheared cancer-cell-line DNA, unmethylated contrived cfDNA, and a methylation internal positive control (mIPC) in a 10 mL reaction volume. Following 1 or 3 days of incubation at room temperature, DNA was extracted from uniform input volumes (4 mL) and enzymatically converted for CpG methylation assessment by targeted sequencing. Methylation integrity was assessed using mIPCs. Proteomic Assay: Whole blood from 8 donors was drawn into BCTs at different fill levels (2.5-10 mL) and stored for 1, 3, or 7 days before plasma separation. We then evaluated the relative abundance of 1,034 plasma proteins using an immuno-proteomic profiling platform and further validated the observed trends using a quantitative immunoassay platform targeting three proteins of interest. In both experimental systems, preservative-to-sample ratios were also verified by measuring BCT components: potassium and formaldehyde. Potassium and formaldehyde assay data confirmed preservative-to-sample ratios were elevated under simulated or true low-fill conditions. Genomic Impact: Despite an elevated ratio, DNA extraction yield, size distribution, and methylation rate of mIPCs remained the same across all fill levels. Proteomic Impact: As expected, low-fill tubes showed a broad decrease in protein concentrations, consistent with the volumetric dilution effect by the liquid preservative. On day 1, 96% of significantly different proteins (151 out of 157) were reduced in 5 mL tubes as compared to 10 mL tubes. By Day 7, the reduction in protein concentrations in 5 mL versus 10 mL tubes was significantly more pronounced for platelet and RBC markers than for non-platelet/RBC proteins (platelet protein odds ratio: 2.99), suggesting that higher fixative concentrations reduced cellular lysis and protein release. Quantitative immunoassays confirmed that negative bias varied significantly across analytes. For example, Complement C5 (-5%), MBL (-15%), and Complement Factor 1 (-50%) exhibited distinct reductions compared to control fill levels by Day 3. Underfilling Streck BCTs showed no evidence of affecting cfDNA methylation, suggesting their robustness for genomic applications even at low-fill conditions. Conversely, plasma protein showed negative bias driven by at least two factors: volumetric dilution from the preservative and enhanced stabilization of blood cells, which reduces the background protein release compared to standard fill tubes. These results underscore the need to independently evaluate tube-fill effects for each multiomic analyte in liquid biopsy tests.