Capillary blood self-collection for high-throughput proteomics
El-Sabawi et al.
The finding, in our words
Capillary blood self-collection yielded protein concentrations that matched venipuncture for only a minority of analytes, with 34 of 230 proteins showing acceptable correlation and low variability, while the remainder were more variable; this suggests capillary sampling can support decentralised proteomics for a subset of proteins but may not be universally interchangeable with venipuncture.
A paraphrase to the Library’s standard, never the abstract. The source is one link away and is always the authority.
Self-collected capillary blood samples posted by standard mail showed high concordance with couriered venous samples for HbA1c, creatinine and lipid parameters, supporting their use for remote monitoring in diabetes management and screening.
Plasma lipidomes from self-collected upper-arm capillary blood were statistically indistinguishable from paired venous plasma, most classes at r=0.95–0.99, validating at-home capillary self-collection for lipidomics.
TAP II capillary serum, both professionally collected and self-collected, correlated with venous serum at R > 0.9 for eight analytes including ALT, AST, cholesterol, HDL and triglycerides. Creatinine correlated acceptably but carried a consistent negative bias, and carbon dioxide, potassium and albumin did not reach acceptable agreement: the useful reading is that the panel has to be validated, not the device alone.
Patients self-collected upper-arm capillary blood with press-activated devices including TAP II; autoantibody and CRP results agreed with venous at r≥0.98, were rated less painful, and 98.6% got enough blood within two attempts.
The Tasso SST capillary blood micro-sampling device generated reproducible serum protein measurements, but its concordance with venous samples was analyte-specific: CRP, ferritin, IL-6 and PCT showed strong agreement, while D-dimer, IL-1β and IL-1Ra did not. Self-collection at home with delayed processing led to significant concentration differences for several proteins, underscoring that user proficiency and timely sample handling are critical for decentralised diagnostic applications.