Methodological issues relating to surfactant protein A measurement in bronchoalveolar lavage fluid from preterm infants with respiratory distress syndrome

Background: Surfactant protein A (SP-A) has important in vitro functions in surfactant metabolism [1 and 2]. To assess its role in neonatal respiratory distress syndrome (RDS), accurate quantification of its concentration in bronchoalveolar lavage (BAL) fluid is important. Aims: To assess the repeatability and intra-/inter-assay variation of an SP-A ELISA. Methods: Infants, 25–30 weeks gestation, ventilated for RDS and having received exogenous surfactant were candidates for inclusion. Specimens, obtained using non-bronchoscopic BAL, were spun and frozen at −70°C until analysis. A double-antibody (monoclonal anti-human SP-A, rabbit polyclonal anti-human SP-A antibodies) sandwich ELISA technique with goat anti-rabbit peroxidase enzyme was used by a single, experienced investigator running duplicate wells. Results: Fifty-one samples were analysed from 22 infants. Median (interquartile range) gestation: 26 weeks (26–28); birth-weight: 786 grams (704–1070); 11 received natural, 11 artificial surfactant. Median age at BAL collection 4 (2–9) days. Serial dilutions of human SP-A produced standard curves with a mean correlation coefficient of 0.979 (5 ELISA plates) with the coefficient of repeatability for optical densities of 5.4% (n=35). For the same dilution, inter-/intra-assay variation between samples and plates showed no significant difference. SP-A concentration of samples measured in duplicate had a coefficient of repeatability at 1 in 50 dilution of 11% and at 1 in 200 dilution of 10%. However, inter-dilutional limits of agreement for SP-A concentration of a given sample, at 1:50 and 1:200 dilutions respectively, were poor with a mean difference of −68% (95% confidence interval: ±86%). This effect was independent of gestational age, birth-weight and surfactant type, was not effected by block time, antibody concentration, or the addition of human albumin or human IgG to buffer solutions but correlated with postnatal age (r=−0.52, p=0.01) Conclusions: This SP-A ELISA, specific for monomeric SP-A, shows good well-to-well reproducibility and no significant inter-/intra-assay variation in SP-A concentration for a given dilution. However, although no systematic difference is present, SP-A concentrations measured by this ELISA may differ with the dilutional concentration employed. Speculation: Preterm BAL fluid may contain factors interfering with this ELISA. Variability in the molecular form of SP-A in preterm BAL may explain these findings.