[HTML][HTML] Comparative analysis of hemagglutination inhibition titers generated using temporally matched serum and plasma samples

GN Defang, NJ Martin, TH Burgess, EV Millar… - PloS one, 2012 - journals.plos.org
GN Defang, NJ Martin, TH Burgess, EV Millar, LNA Pecenka, JR Danko, JC Arnold…
PloS one, 2012journals.plos.org
Influenza-specific hemaggluitination inhibition (HAI) antibody titer, an indicator of immunity
to influenza, is often used to measure exposure to influenza in surveillance and
immunogenicity studies. Traditionally, serum has been the specimen of choice for HAI
assays, but a desire to reduce the amount of blood collected during studies and the
availability of plasma in archived sample collections warrant the evaluation of plasma for
HAI titer. Therefore, the relationship between serum and plasma HAI titer values is of great …
Influenza-specific hemaggluitination inhibition (HAI) antibody titer, an indicator of immunity to influenza, is often used to measure exposure to influenza in surveillance and immunogenicity studies. Traditionally, serum has been the specimen of choice for HAI assays, but a desire to reduce the amount of blood collected during studies and the availability of plasma in archived sample collections warrant the evaluation of plasma for HAI titer. Therefore, the relationship between serum and plasma HAI titer values is of great interest. Here, we compare HAI titers determined on temporally matched serum and plasma (citrated and heparinized) using influenza A and B viruses. Bland-Altman plots, McNemar's test, and geometric coefficient of variation were used respectively for evaluating agreement, correlation and variability in the serum-plasma titer results. We observed a high degree of agreement (80.5%–98.8%) and correlation (r = 0.796–0.964) in the serum and matched plasma titer values although plasma titers were generally lower than corresponding serum titers. Calculated seropositive (HAI ≥40) rates were higher using serum titers than with plasma titers, but seroconversion rates were unaffected by sample type. Stronger agreement and decreased variability in titers were seen between serum and citrated plasma than between serum and heparinized plasma. Overall, these data suggest that serum or plasma can be used in serodiagnostic HAI assays, but seropositive rates may be underestimated using plasma HAI titers. The type of anticoagulant present in plasma may affect HAI titer values and warrants further investigation.
PLOS