High, Low Magnesium Levels Linked to Dementia Risk

Deborah Brauser

September 20, 2017

Both low and high levels of serum magnesium are associated with dementia risk, suggesting a U-shaped relationship, according to new research.

New analysis of more than 9500 participants from the ongoing Rotterdam Study, none of whom had dementia at baseline, showed that those with low magnesium levels (defined as 0.79 mmol/L or less) and those with high levels (0.90 mmol/L or greater) had significantly increased risk for dementia at an 8-year follow-up assessment.

Dr Brenda C.T. Kieboom

Lead author, Brenda C.T. Kieboom, MD, Department of Epidemiology, Erasmus Medical Center, Rotterdam, the Netherlands, said that the investigators expected to find a linear association, so they were a bit surprised that a high serum magnesium level was linked to dementia risk.

"This has been observed before in one of the case-control studies; however, we do not have a biological explanation yet on how this would work," Dr Kieboom told Medscape Medical News. But she added that low levels are especially concerning.

"As serum magnesium levels are not routinely measured in daily clinical practice, our take-away message is to consider measuring these levels in patients at risk of hypomagnesemia," she added. This includes those who are taking proton-pump inhibitors or diuretics and those with a diet lacking in green vegetables, nuts, and whole grains — "as these foods are rich in magnesium."

However, she cautioned that the findings do not prove a causal relationship between low serum magnesium levels and risk for dementia. "For that, our study needs to be replicated."

The findings were published online September 20 in Neurology.

Limited Data

Dr Kieboom noted that previous experimental data have shown that "magnesium might have a protective effect on learning in rats with dementia."

As for humans, data until now have been limited to just a few small case-control studies "showing conflicting results," plus a small randomized trial that showed improved executive function and working memory in participants with mild cognitive improvement, she reported.

"So we wanted to study if indeed serum magnesium levels associate with dementia risk in people free from dementia at baseline."

To do this, the investigators examined data for 9569 healthy participants in the Rotterdam Study (56.6% women; mean age, 64.9 years). All had given blood samples, which were assessed by the Erasmus Medical Center's Department of Clinical Chemistry using an analyzer from Roche Diagnostics.

The participants were followed for a median of 7.8 years to see whether they developed dementia or not, based on criteria from the Diagnostic and Statistical Manual of Mental Disorders, 3rd edition, revised. At that timepoint, 823 of the individuals had received a diagnosis of all-cause dementia, with 662 diagnosed with Alzheimer's disease.

All participants with the lowest levels of serum magnesium (first quintile) and those with the highest levels (fifth quintile) were compared with the third quintile of participants, who had mid-range magnesium levels (0.84 to 0.85 mmol/L).

After adjustment for multiple factors (including age, sex, educational level, comorbidities, body mass index, alcohol use, and smoking status), both low and high serum magnesium levels were significantly linked to increased risk for all-cause dementia.


Table. Association Between Magnesium Levels and Incident Dementia

Serum Magnesium Level HR (95% CI)
Low 1.32 (1.02 - 1.69)
High 1.30 (1.02 - 1.67)
CI = confidence interval; HR = hazard ratio


The incidence rate of dementia per 1000 person-years was 10.2 for the group with low magnesium levels (95% CI, 8.7 - 11.9) and 11.4 for those with high levels (95% CI, 9.8 - 13.2).

Sensitivity analysis also showed a trend for increased risk for AD among those in the first quintile of magnesium levels (HR, 1.28; 95% CI, 0.97 - 1.69) and those in the fifth quintile (HR, 1.21; 95% CI, 0.92 - 1.58). However, these rates were not statistically significant.

"These results need to be confirmed with additional studies, but the results are intriguing," especially because "current treatment and prevention options for dementia are limited," Dr Kieboom said in a release.

This association between serum magnesium and risk for dementia "is more likely to be U-shaped rather than linear," add the investigators.

They note that one hypothesis for the association may include the "direct effect of neuronal magnesium on regulation of the NMDA [N-methyl-D-aspartate] receptor," while a second pathway may include magnesium influencing dementia risk through oxidative stress.

"Part of the Conversation"

"Detecting magnesium is important mechanistically," Marwan N. Sabbagh, MD, director of the Alzheimer's Disease and Memory Disorders Division at Barrow Neurological Institute, Phoenix, Arizona, told Medscape Medical News when asked for comment.

"People have looked at serum levels and correlation with disease risk. And these data add to a growing body of evidence," said Dr Sabbagh, adding that he was impressed with the large dataset used.

Asked whether screening for magnesium is something that should now be considered, he said "drawing blood for magnesium will be among a broad array of things that will need to be considered."

The current findings "will not hold value by themselves" regarding a diagnostic tool. But, along with detecting tau and amyloid ratios, Dr Sabbagh said he would add checking magnesium in blood samples "to a menu of things to be explored."

"We often look for various biomarker signals to correlate to risk, and this is one among many."

The investigators note that one study limitation is that magnesium levels in blood samples do not always represent total level of magnesium in the body. Dr Sabbagh agreed with this and added that magnesium is "quite ubiquitous" in a normal diet.

"So I don't know that we can extrapolate this to any specific healthcare recommendation quite yet. But it does put us on notice to ask: Do we need to be monitoring our dietary consumption of magnesium? That could be derived in the long term as a possibility on the basis of this," he said.

"I think it's early and speculative at this point, but it can become part of the conversation."

The Rotterdam Study is supported by the Erasmus Medical Center and Erasmus University Rotterdam; the Netherlands Organization for Scientific Research; the Netherlands Organization for Health Research and Development; the Research Institute for Diseases in the Elderly; the Netherlands Genomics Initiative; the Ministry of Education, Culture and Science; the Ministry of Health, Welfare and Sports; the European Commission; and the Municipality of Rotterdam. The study authors and Dr Sabbagh have disclosed no relevant financial relationships.

Neurology. Published online September 20, 2017. Abstract

Follow Deborah Brauser on Twitter: @MedscapeDeb . For more Medscape Neurology news, join us on Facebook and Twitter.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: