A comparison has been made of the efficiency of high performance liqui
d chromatographic (HPLC) protocols and a human platelet-based bioassay
to quantify parthenolide, a pharmacologically active sesquiterpenoid
in Tanacetum parthenium (Feverfew). The optimum protocol, in terms bot
h of concentration of parthenolide measured and of reproducibility of
results, involved re-suspension of an acetone extract of the plant mat
erial in 100% ethanol prior to isocratic HPLC. The mean parthenolide c
oncentration in a leaf extract dissolved in ethanol from a glasshouse-
grown clone (1.52+/-0.13%; n=10) was significantly greater (P=0.01) th
an in an extract dissolved in acetonitrile (1.23+/-0.25%; n=10). progr
essive reduction of the organic phase in the re-extraction solvent by
replacement with phosphate-buffered saline resulted in concomitantly l
ess parthenolide being measured in the samples. The mean parthenolide
concentration measured by HPLC in extracts employing 3% ethanol in pho
sphate-buffered saline as re-extracting solvent, as used previously in
bioassays, was significantly (P<0.0001) lower than obtained with 100%
ethanol. Parthenolide concentrations in extracts with 3% ethanol were
also highly variable as reflected by coefficients of variance of up t
o 15%. The protocol using 100% ethanol as re-extracting solvent was op
timal for rapid and reproducible assessments of parthenolide content i
n small (>20 mg dry weight) samples from glasshouse-grown and in vitro
cultured materials of Feverfew.