Institute for Preventive Medicine, Nutrition, and Cancer Folkhälsan Research Center, Fin-00014 University of Helsinki, Finland Theodore Fotsis, M.D., Ph.D. Laboratory of Biological Chemistry, University of Ioannina Medical School GR-45110 Ioannina, Greece
An article published by Lim et al. (1) in JCEM concerning postmenopausal women with and without osteopenia describes an estrogen profiling method without giving enough details to allow the reader to evaluate the methodology. The values described are 260 times higher than those we obtain in postmenopausal women, and the relations between the estrogens show that the method is not reliable.
We have the following comments on the method and the results:
1. The method is very short. According to our experiences it is not possible to measure the estrogens in urine by gas chromatography-mass spectrometry (GC-MS) after such a brief purification. In fact, there is no elimination of any compounds known to interfere with estrogen assays by GC-MS. The extract will contain neutral steroids and various dietary phenolic compounds having much higher concentrations than the estrogens and interfering with the assay.
2. Only one internal standard is included, and this has only two deuterium atoms, invalidating sensitive and accurate assays by GC-MS with selected ion monitoring. The absolute minimum number of deuterium atoms required is 3, and preferably deuterated internal standards should be used for all or most compounds (2, 3). The polarity and retention time of the internal standard used in the study by Lim et al. differs very much from those of some of the measured estrogens, making quantitation inaccurate. The ideal situation is to have deuterium-labeled conjugated internal standards. Such standards are, however, not available. At present we use deuterated unconjugated standards for most of the compounds (4).
3. There are no details on retention times of the estrogens measured or of other estrogens known to be present and that may interfere. No chromatograms are shown. Neither are there details with regard to the ions monitored in GC-MS-SIM (selected ion monitoring) measurements nor reference to a methodological communication of the authors.
4. The authors have measured many estrogens that have never been
identified or quantified in urine before in nonpregnant
individuals. This would have needed some data on the identity.
Based on the amounts measured, mass spectra could have been obtained
for many of the compounds. To our knowledge, some of the
metabolites have never been identified in urine of pregnant women
(6-hydroxyestriol, 4-methoxyestradiol, 6-dehydroestrone,
6
-hydroxyestradiol), nor has their presence in pregnancy urine ever
been postulated.
To be able to compare our results with theirs (Table 1), we have converted their values to
excretion of estrogens in nmol/24 h by using the lowest and highest
reference value (8 and 15 mmol/24 h) for creatinine in urine in women
in Helsinki University Central Hospital. The lowest creatinine value (8
mmol/24 h) is probably closest to the true ones, because the women are
postmenopausal and their muscle mass is relatively low. Most values are
grossly overestimated, including those which we have not measured, such
as 16,17-epiestriol, and 17
-estradiol, which in our experience occur
in such small amounts that their quantification is impossible by
GC-MS-SIM in postmenopausal urine. The 16
-hydroxyestrone values are
very high, but this is not surprising because numerous compounds
interfere if the purification method is insufficient, making the
calculation of the ratio 16
-hydroxyestrone/2-hydroxyestrone
unreliable. Also the 16-ketoestradiol values are very high. Only the
values for estrone and estriol are closer to ours, being 100% higher
than our values. The estriol values are lower than the
16
-hydroxyestrone values, which we have never seen during almost 30
years of method development. 16
-Hydroxyestrone is a precursor of
estriol, and the conversion is rapid and not rate-limiting in estrogen
metabolism (our observation). The ratio of estrone to estradiol in the
Lim study is about 1:1, a ratio which is never obtained using specific
and accurate methods (2, 5, 6). Total estrogen values are at least 10
times too high.
|
Footnotes
Address correspondence to: Herman Adlercreutz, Department of Clinical Chemistry, PB 60, Institute for Preventive Medicine, Nutrition & Cancer, Folkhälsan Research Center, University of Helsinki, Finland FIN-00014.
Received May 14, 1998.
References