Use of Micro-HPLC-MS/MS Method to Assess Diurnal Effects on Steroid Hormones

To the Editor:

Steroid metabolism is known to show diurnal variation for both cortisol and testosterone starting as early as 2–3 months of age (1). Cortisol shows significant diurnal variation, with higher values in the morning than in the evening, as has been documented by several groups. The measurement of midnight cortisol concentration is performed routinely as a screening test for Cushing syndrome, and dysregulation of diurnal variation is associated with disease pathology. Brambilla et al. (2) showed diurnal variation in testosterone for men 30–40 years of age; concentrations were 20%–25% lower at 1600 than at 0800 and the difference declined with age, showing a 10% difference at 70 years of age. Less information is available for other steroids, particularly where measurement has been performed with the newer, specific LC-MS/MS methodologies.

Our objective was to demonstrate and define the extent of steroid diurnal concentration fluctuations that potentially necessitate strict adherence to time of blood sample draw and requirement of separate time-dependent reference intervals with our recently reported multisteroid profile using minimal sample volume (3). triple-quadrupole MS coupled with an atmospheric pressure photoionization source and Agilent 1200 Infinity series HPLC, including isotope dilution with deuterium-labeled internal standard for each analyte. A protein crash method was used for sample preparation. We added 75 μL acetonitrile containing internal standards to 50 μL plasma. The sample was then spun down to collect protein into a pellet. Supernatant (75 μL) was diluted with 250 μL water, and an aliquot was injected onto a Poroshell 120 ECC18 column. After washing, the steroids were eluted with a methanol gradient. We used multiple reaction monitoring (MRM)1 to measure the steroid hormones cortisol, cortisone, progesterone, testosterone, androstenedione, 11-deoxycortisol (11DOC), 17- α-hydroxyprogesterone (17OHP), and corticosterone.

The use of volunteer samples reported in this study was approved by the National Institute of Diabetes and Digestive and Kidney Diseases/National Institute of Arthritis and Musculoskeletal and Skin Diseases institutional review board (approved protocol NCT00428987). Morning samples (minimum n = 20) were collected between 7:30 AM and 8 AM, and paired midnight samples were taken between 11:30 PM and midnight. Healthy participants were not taking any medications or supplements, had stable body weight (within 3%) for at least 30 days before testing, and were generally healthy with no indication of current or past disease during a history and physical examination by a medical provider. Additionally, all women were premenopausal and in the early follicular phase during in-patient admission, on the basis of their menstrual cycle histories. Paired-sample Student t-tests were performed between AM and PM values for all the measured steroid hormones. Statistical analyses were performed with GraphPad Prism version 6 and MedCalc version 14.8.1. Significant values were determined at P < 0.05.

Our findings revealed significant diurnal fluctuations for 11DOC, corticosterone, cortisone, androstenedione, and 17OHP, all of which showed significantly higher values in the morning than in the evening (Fig. 1). Progesterone showed no significant difference for values taken at the 2 different times of day for men (P = 0.1039) and women (P = 0.2089). Additionally, women displayed no significant difference in AM vs PM testosterone values (P = 0.6771), whereas males did display a significant difference (P < 0.001).

Fig. 1.

Box-and-whisker plots of am versus pm values for men and women. (A), 11DOC. (B), Corticosterone. (C), 17OHP. (D), Cortisol. (E), Cortisone. (F), Androstenedione. (G), Testosterone (men only). (Mean is indicated by small square markers and median by bar in box.)

The literature has clearly documented variation in testosterone and cortisol concentrations with time of day; this study shows similar diurnal patterns for other clinically relevant and commonly measured steroids, with the exception of progesterone. Our findings demonstrate the important need for standardization of blood collection times for many steroid hormones and reference intervals that are specific for the time of day. Incorrect clinical interpretation of steroid hormone values by use of reference intervals that are not specific for time of day may lead to patient mismanagement or delay in treatment. We recommend that where relevant, time-specific reference intervals be available in all clinical laboratories measuring and reporting steroid hormone concentrations. We recommend that optimal times for consideration include 8 AM (range 8 AM–10 AM), 8 PM (8 PM–9 PM), and midnight (10 PM–midnight) and that age- and sex-related reference intervals be developed at these time points. Reference interval studies reported recently (4) and critiqued (5) in Clinical Chemistry have failed to adequately control for the diurnal fluctuations found in this study. Reference interval values generated for these steroid hormones without due consideration of the effects of diurnal variation are of dubious value.

1. Nonstandard abbreviations:

MRM

multiple reaction monitoring

11DOC

11-deoxycortisol

17OHP

17α-hydroxyprogesterone