Should we use oral magnesium supplementation to improve sleep in the elderly? Article reviewed: K. Held, I.A. Antonijevic, H. Künzel, M. Uhr, T.C. Wetter, I.C. Golly, A. Steiger, H. Murck, Oral MG²⁺ supplementation reverses age-related neuroendocrine and sleep EEG changes in humans, Pharmacopsychiatry, 35 (2002), 135-143

Objectives: Magnesium (Mg) has been recognized as potentially affecting sleep and sleep-related neuroendocrine function and shown in prior work by these authors to alter EEG sleep characteristics. This study aims to determine if oral Mg supplement also improves sleep and related neuroendocrines, particularly in the elderly (age 60-80). Study design: Double-blinded, randomized, placebo-controlled crossover study. Study population: Healthy elderly subjects, age 60-80 (average 68.1, standard deviation 5.7), six women and six men. Methods: Two 20-day treatment periods, separated by 2 weeks, involved either a placebo or Mg administered as effervescent tablets given in a fixed escalation of the 10 mmol dose, first only in the morning for 3 days, then in the morning and noon for 3 days, and finally three times a day for the remaining 14 days of the treatment period. At the end of each treatment period a polysomnogram (PSG) was obtained for two nights, the first for accommodation without any actual recording and the second for the study data. Sleep was recorded from 11 p.m. to 7 a.m. the next day. An indwelling intravenous cannula was used to obtain repeated blood samples every 30 min from 8 to 10 p.m. and every 20 min from 10 p.m. to 7 a.m. for analyses of adrenocorticotropic hormone (ACTH), cortisol, renin and aldosterone. Arginine-vasopressin (AVP) and angiotensin II (ATII) were measured from the blood samples taken every hour. The blood analyses of ACTH, cortisol, renin and aldosterone were evaluated using standard area under the curve calculated separately for samples from the first (11:00 p.m. - 2:20 a.m.) and the second (2:20 - 7:00 a.m.) parts of the night. Results: The analyses of the PSG for Mg compared to the placebo treatment condition showed significantly more slow-wave sleep (averages ± standard deviations: 16.5 ± 20.4 vs. 10.1 ± 15.4 min.), but no significant change in total sleep time or amount of wake time during the night. The spectral analyses of the sleep EEG during the Mg treatment showed greater power in the delta and sigma frequency ranges (differences in both of about 0.5 standard deviations) but no significant differences in the theta, alpha and beta ranges. The analyses of the nocturnal hormones for the Mg compared to the placebo treatment found significantly lower Cortisol and higher renin for the first part of the night and significantly higher renin and aldosterone for the second part of the night. Aldosterone was also higher for the Mg condition in the first part of the night, but the difference was not statistically significant (f = 2.8, n = 9 - blood analyses data were lost on three subjects). Serum analyses of Mg levels failed to show any significant difference between the treatment conditions. Adverse effects associated with the Mg treatment included soft stools for all subjects and slight edema in the extremities for one subject. Conclusions: The authors note that aging produces decreases in slow-wave sleep and in delta and sigma (spindle frequency) spectral power and that these three measures are improved with Mg treatment in their sample of older adults. Similarly, aging is associated with increased nocturnal cortisol release and with decreased activity of the renin-angiotensin-aldosterone system. The latter is shown by decreased plasma renin not associated with high blood pressure as well as reduced aldosterone concentration. The Mg treatment reversed the aging effect on both of these measures. The authors note that the increase in aldosterone might account for the slow-wave sleep increase and for the changes in cortisol, both of which have been shown to occur with experimentally induced increases in aldosterone. It is noted, however, that the increase in aldosterone was only marginally significant for the first part of the night. The authors also suggest that the differences observed could have resulted from the Mg effects on both the glutamatergic and GABAergic systems. They further argue that the changes with aging partly reversed by Mg treatment match those for depression, and the Mg treatment might thus serve as a mood stabilizer.