To the Editor: I find the article by Andreas Ströhle, M.D., et al. (1) very puzzling. The authors studied seven women and three men with panic disorder and examined them in relation to 10 age- and sex-matched comparison subjects. The steroid levels were given as the mean and SD in nanomoles per liter, with the sexes combined.

We were not told in what phase of the menstrual cycle the women were studied; one must assume, since the SDs for the steroids were relatively small—implying that the values for men and women were similar—that the women must have been studied in the follicular phase of their cycles. Or perhaps some of them were postmenopausal, since the age range was quite wide (37.2 years, SD=10.2).

However, the levels of progesterone for the comparison subjects (24.1 nmol/liter, SD=2.6), and, indeed, for all the subjects, are clearly in the luteal range!—far above the well-established value of about 1 nmol/liter in men, in postmenopausal women, and in women in the follicular phase of the cycle (2–4). Luteal levels range from 6 to 64 nmol/liter. The research method used was gas chromatography/mass spectrometry; one was referred to a previous article by the same group (5). In that article, the mean level of progesterone in eight healthy comparison men was 7.2 nmol/ml, while the mean level of allopregnanolone was about 4 nmol/liter. The authors’ levels of these two steroids are thus inconsistent with their own previous data.

Their progesterone values are in serious disagreement with well-established data, obtained mostly through radioimmunoassay. Their values for almost all of the steroids measured are much higher than those found by a colleague and me by using a combination of high-performance liquid chromatography and radioimmunoassay (6). While gas chromatography/mass spectrometry holds great promise for providing more sensitive methods for detection of neuroactive steroids, which are present in very low concentrations in the blood, one cannot ignore the large differences in the results obtained to date that have not been explained.