We compared the data from four growth hormone (GH) immunoassays for an
alyzing 24-h GH profiles in four apparently normal subjects and four o
bese subjects (508 serum samples), The detection limit was 0.02 mu g/L
for one immunochemiluminometric assay (ICMA), 0.1 mu g/L for two IRMA
s, and 0.4 mu g/L for one RIA. All GH pulses with a peak ICMA value >1
mu g/L were detected by each of the other methods. Overall, the corre
lation coefficient between the values obtained with all four assays ex
ceeded 0.90. However, for GH concentrations less than or equal to 0.25
mu g/L, acceptable concordance (r(2) greater than or equal to 0.80) w
as reached only between the ICMA and one IRMA; between the ICMA. and t
he RIA, concordance was acceptable only for GH concentrations greater
than or equal to 10 mu g/L. In the normal subjects, the percentage of
undetectable values was 0% with the ICMA but 29% with one of the IRMAs
; in obese subjects, the corresponding values were 12% and 38%.