Reproducibility of the circadian rhythms of serum cortisol and melatonin in healthy subjects [...], 2003, Selmaoui et al

[ScoutB]

Reproducibility of the circadian rhythms of serum cortisol and melatonin in healthy subjects: a study of three different 24-h cycles over six weeks
Selmaoui, Brahim; Touitou, Yvan

Abstract
Plasma melatonin and cortisol are characterized by a marked circadian rhythm, but little information is available about the reproducibility and stability of these rhythms over several weeks in the same subjects. This study examined the characteristics of these rhythms in 31 healthy human subjects 20 to 30 years of age. They were synchronized with a diurnal activity from 0800 to 2300 and nocturnal rest. They participated in three 24-hour sessions (S1, S2, and S3): S2 took place two weeks after S1 and S3 4 weeks after S2. Blood samples were taken during each session at 3-hour intervals from 1100 to 2000 and hourly from 2200 to 0800.

Comparison of the circadian rhythms between groups used repeated measures 2-way ANOVA, the cosinor method, and Bingham's test. Intraindividual variations were compared by the cosinor method and Bingham's test. The groups did not differ, but a slight difference in the amplitude or acrophase of individual circadian rhythms was observed in 5 of 31 subjects for melatonin and 1 of 31 for cortisol. The circadian means did not differ over the three sessions. These results show that the circadian profile of cortisol and melatonin are highly reproducible over a six-week period, in both individuals and groups. Our study clearly shows that these hormones can be considered to be stable markers of the circadian time structure and therefore useful tools to validate rhythms' synchronisation of human subjects.


Web | DOI | Life Sciences | Paywalled | Sci-hub

Note: this is a 2003 study, on healthy controls. I just wanted to have their graphs here to link to.

 

[ScoutB]

I was trying to get a sense of what cortisol does throughout the day normally, as a reference point for all the claims being made about it in ME/CFS. This study had a few sample graphs of individuals which I found informative.

All young male students, and selected to have a regular sleep pattern:

Thirty-one healthy young men, aged 20 to 30 years, were selected after routine clinical and laboratory examination. Selection criteria included: regular sleep habits, no chronic disease or disability, no recent acute illness, no night work, no travel to another time zone within the preceding 2 months, and no smoking. They were instructed to eat balanced meals and to abstain from consuming alcohol and coffee for 24 hours before and during each experimental session. They were synchronized with a diurnal activity from 0800 to 2300 and nocturnal rest. The selection of subjects was made based on questionnaire. The selected subjects had regular sleep habits from 2300 to 0800 ( F 1 hour).

Protocol:

Each subject participated in three 24-hour sessions (S1, S2 and S3), S2 two weeks after S1,and S3 4 weeks after S2. Since all subjects were students and recruited at the faculty of medicine Pitié-Salpêtrière in Paris, France, we managed the protocol to avoid periods of examinations and holidays. Catheters were placed in the antecubital vein for the 24-hour period, and the first blood sample was taken at 1100. Subjects slept in darkness between 2300 and 0800 and spent their waking time in the laboratory (diurnal activity from 0800 to 2300 and nocturnal rest). During the waking time the subjects were exposed to the natural light since they were free to move in the hospital during the daytime period.

Fig. 1. Circadian rhythms of melatonin (top) and cortisol (bottom) in three different 24-h sessions spaced two weeks apart between the 1st and 2nd session then 4 weeks apart between the 2nd and 3rd session: S1 (open circle), S2 (black circle) and S3 (black triangle). Each time point is the mean SEM of 31 subjects.



The average in Fig 1 makes cortisol look pretty consistent. But the authors also shared the individual measurements of 6 of the 31 subjects and I found the variability there useful for getting a sense of what cortisol might be doing each day.

Fig. 3. Circadian rhythm of cortisol in individual subjects sampled in three 24-h cycles: S1 (open circle), S2 (black circle) and S3 (black triangle).




In short: group averages seem to smooth over a fair bit of complexity.

 

[Hutan]

If you sample cortisol just at 8 am, as many studies have done, even in just these six participants, the values would be all over the place, from 100 to nearly 500. Even an hour shift in regular wake up times would have a material impact on reported cortisol levels.

(note that some of the charts have a y axis max of 400 and some of 500)

(it's odd that the values for 9 and 10 am are missing - they are some of the most important values, potentially covering peak values, I would have thought. )

 

[ScoutB]

Yeah, their sampling method is weird: they took samples every 3 hours during the day, then hourly at night (I assume to be less cumbersome to the subjects). But they could have made that more obvious on the graphs -- it means the daytime data is relatively scrunched up compared to the night data.

[blood samples were taken] at 3-hour intervals from 1100 to 2000 and hourly from 2200 to 0800.

Was surprised they didn't try to include the morning jump too.

Even an hour shift in regular wake up times would have a material impact on reported cortisol levels.

Yes, the steep drop at 8pm also had me thinking about the 8pm sampling of Demitrak et al.