Lithium in Tap Water Does Not Dilute Bipolar Disorder, Dementia Risk

Batya Swift Yasgur, MA, LSW

May 30, 2018

The presence of lithium in groundwater does not reduce the risk for bipolar disorder or dementia, new research shows.

Investigators used data on lithium concentrations from more than 3000 drinking water wells across the country and diagnostic data from Medicare, Medicaid, and private insurers to analyze the relationship between groundwater lithium concentrations and diagnoses of bipolar disorder and dementia.

After accounting for confounding variables such as demographics and healthcare resources, they found no association between exposure to groundwater lithium and the risk for bipolar disorder or dementia.

"The associations found in previous studies between lithium concentrations, dementia, and bipolar disorder were produced by imbalances in covariate distribution — differences in access to quality healthcare or diagnostic practices, for example," lead author Robert Gibbons, PhD, Blum-Riese Professor of Biostatistics, University of Chicago, Illinois, told Medscape Medical News.

"The variation in demographics and healthcare resources led not to a reduction in the rate of underlying disorders but to the likelihood that the disorders would be diagnosed," he said.

The study was published online May 23 in JAMA Psychiatry.

Intriguing Investigation

Lithium is a "naturally occurring trace element in groundwater" and a "cornerstone therapy" for bipolar disorder. In addition, it may have a role in the treatment of dementia, the authors write.

A Danish study found an inverse association between the presence of lithium in drinking water and dementia. Another study conducted in the United States also found that lithium exposure was associated with lower rates of mental health disorders.

However, it is possible that the association may be due to other factors, given the fact that rates of mental health diagnoses "vary substantially with local healthcare resources and demographics," the authors suggest.

"This study grew out of a graduate class I taught in statistical applications and biostatistics," said Gibbons.

"I saw the Danish paper about lithium and dementia, and although it was really cool, I was skeptical about the findings. So I suggested to the class that we investigate to see if we could reproduce the findings here in the US, and my students took the bait and did a great job," he said.

The investigators used US Geological Survey data from 1992 to 2003 relating to more than 3000 water wells.

They identified psychiatric diagnoses from the inpatient hospital, long-term care, and other therapy claim files from an array of databases. Data included information for Medicare and Medicaid recipients as well as privately insured patients.

To prevent "spurious causal inference from inadequate adjustment for confounders," the researchers repeated their analysis for three "negative control" outcomes that have no known link to groundwater lithium: major depressive disorder (MDD), myocardial infarction (MI), and prostate cancer.

Data from several additional databases pertaining to geographic variation in factors affecting health status and care in the United States were used to evaluate county-level healthcare resources and demographics. The data pertained to census population, median age, education, race, and ethnicity.

The researchers used a mixed-effects Poisson regression model with inverse probability of treatment weighting, with weights based on county-level healthcare resources "designed to give the low-lithium counties the same distribution of healthcare resources as the high-lithium counties."

They also controlled for sex, payer, and county-level demographics.

A sensitivity analysis examined the association of lithium as a continuous variable (restricted cubic spline with 5 knots). The analysis controlled for county-level demographic and healthcare resources.

"We had been studying statistical approaches to confounding in observational data. And, as we had heard for years about the potential association between lithium in drinking water and rates of bipolar disorder, we thought it would be an intriguing investigation," Gibbons said.

Biologically Dubious

The researchers analyzed claims data for 4,227,556 adults living in 174 counties (n = 3,046,331 with private insurance, 261,461 with Medicare Supplemental, and 919,764 with Medicaid).

Of these, 40,4662 patients (9.6%) lived in 1 of 32 counties with high lithium concentrations (defined as >40 μg/L).

The mean and median lithium concentrations were 27.4 μg/L and 11.1 μg/L, respectively (IQR, 3.7 - 23.6 μg/L).

Mean lithium exposures in low- and high-lithium counties were 6.0 μg/L and 141.3 μg/L, respectively (P < .001).

Unadjusted prevalence rates for bipolar disorder and dementia were found to be significantly lower in high-lithium counties compared to low-lithium counties. However, high-lithium counties had fewer physicians and healthcare resources as well as smaller, younger, less educated, and more Hispanic populations.

After adjustment for county-level demographics and healthcare resources, high lithium concentrations were not found to confer any significant preventive benefit for bipolar disorder, dementia, or the negative controls (MDD, MI, and prostate cancer).

The authors note that therapeutic lithium doses are "orders of magnitude larger than groundwater lithium concentrations, making a true causal relationship between groundwater lithium and mental health biologically dubious."

"A patient would need to drink more than 1000 L of water a day to get even the lowest reported effective therapeutic dose of lithium, which is 150 mg," Gibbons commented.

Final Word?

Commenting on the study for Medscape Medical News, J. John Mann, MD, Paul Janssen Professor of Translational Neuroscience and director, Molecular Imaging and Neuropathology Division, Department of Psychiatry, Columbia University/New York State Psychiatric Institute, New York City, who was not involved with the study, said that the researchers "used much more sophisticated statistical techniques than those used in previous studies to rule out an association between lithium in groundwater and lower risk of serious illnesses such as bipolar disorder."

These statistical methods enabled them to "rule out potential confounding factors that may lead to spurious results," he said.

"On and off, for many years, people have been putting out the idea that lithium in the water supply plays a helpful role in terms of bipolar disorder prevention, meaning that it's like fluoride in water, which can transform dental health," he noted.

Because this study "shows the benefit of better design, it is more definitive, and to some degree, we can lay this matter to rest."

Gibbons added, "The most important take-home message is that the beauty of randomization and randomized controlled trials means that we can control for all sorts of confounding factors and selection bias because of the balance brought about through randomization."

This is particularly important with observational studies, which can "often lead to the appearance of an effect that may not actually be there."

The study authors received funding from the National Institutes of Health, the Agency for Healthcare Research and Quality, and the National Research Service. Dr Mann has disclosed no relevant financial relationships.

JAMA Psychiatry. Published online May 23, 2018. Abstract

For more Medscape Psychiatry 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.