Cashews, almonds, and bananas are some of the richest sources of magnesium.
The first author of the study is Dr. Brenda Kieboom, of the Erasmus University Medical Center in Rotterdam, the Netherlands.
Dr. Kieboom and her colleagues measured serum magnesium levels in 9,569 participants aged 64.9 years, on average. The participants did not have dementia at the beginning of the study - that is, between 1997 and 2008. They were clinically followed for 8 years on average, until January 2015.
Low serum magnesium levels were defined as equal to or lower than 0.79 millimoles per liter, and high magnesium levels were defined as equal to or above 0.90 millimoles per liter.
Magnesium levels were divided into quintiles, or fifths; the researchers examined the association between dementia and serum magnesium using the third quintile as a reference.
The researchers adjusted for age, sex, education, risk factors for cardiovascular disease, kidney function, and other comorbidities.
High or low magnesium raises risk by a third
Over the follow-up period, 823 people developed dementia. Of these, 662 were diagnosed with Alzheimer's disease.
As for magnesium levels, both those in the high and the low group were significantly more likely to develop dementia compared with those in the middle group.
More specifically, participants in both the high- and low-magnesium groups had a 30 percent increase in dementia risk compared with their counterparts in the middle group.
The low-magnesium group had 1,771 people, 160 of whom developed dementia. The high-magnesium group comprised 1,748 people, 179 of whom were diagnosed with dementia.
In the middle group, 102 of the 1,387 participants developed dementia.
Strengths and limitations of the study
Firstly, the authors note, the study only used a single measurement of serum magnesium. Although relatively stable over time, magnesium levels do change and such changes may have biased the results.
Secondly, the study did not examine hypomagnesemia or hypermagnesemia, wherein magnesium levels are abnormally low or abnormally high, respectively. Instead, the scientists only focused on normal levels of the mineral.
Finally, the study is purely observational and cannot explain causality. However, the authors note that precautions against this vulnerability were taken.
Namely, Dr. Kieboom and team performed further analyses in which they excluded dementia cases diagnosed in the first 4 years after magnesium measurements were taken. The results were similar, which, the authors write, "[strengthens] the possibility of a causal relationship."
Additional strengths of the research include the long follow-up period and the fact that it was population-based, which reduces the possibility of information bias.
"Furthermore," write the authors, "the detailed assessment of potential confounders and the fact that adjusting for these factors did not alter our effect estimates also strengthens the possibility of a true relationship between serum magnesium levels and dementia, rather than it being the result of other confounders or intermediates."
To the best of their knowledge, this is the first time that such an association has been studied. Therefore, future studies should try to replicate these results in other population samples.
"These results need to be confirmed with additional studies," adds Dr. Kieboom, "but the results are intriguing."
"Since the current treatment and prevention options for dementia are limited, we urgently need to identify new risk factors for dementia that could potentially be adjusted. If people could reduce their risk for dementia through diet or supplements, that could be very beneficial."
Dr. Brenda Kieboom
She also adds that if the results are confirmed, magnesium blood tests could be used to screen those at risk.