In the special article and classic article last year's April 2015 PL issue, evidence was reviewed regarding antidepressants and suicide, and PL concluded there was little benefit if any with those agents, but there was probably a small amount of harm. Lithium, on the other hand, was shown to have clear and large benefit for suicide prevention. The current article of the April 2015 issue reviewed a meta-analysis of lithium studies which demonstrated that benefit, and readers are encouraged to revisit that article if they want to see that evidence for themselves.
In this special article, we plan to extend that prior discussion by focusing on research on very low doses of lithium in drinking water, and its apparent anti-suicide benefits.
We think about certain minerals as being essential for life: there's calcium and magnesium, which seem needed for normal heart rhythm. There's sodium and potassium, needed for cell membrane function. These are all basic ions, part of the periodic table of chemical elements. They are essential for life; but we don't think of them as toxic or dangerous. Yet, too much magnesium or potassium causes cardiac arrhythmia leading to immediate death. Too much calcium can cause delirium, and too much sodium causes swelling of the brain.
Lithium also is an essential mineral, needed for viability of life. If you have none, zero, then the organism’s ability to live is harmed. We all know too much is toxic, and so it is often seen as dangerous. Yet we don't seem to fear potassium or calcium in the same way.
All drugs are toxic; only the dosing and indication makes them therapeutic. That’s the old maxim, dating back to Paracelsus and later William Osler. But the maxim applies not just to “drugs”; it applies to the chemical elements themselves, the basic building blocks of biological life. We should say instead:
All elements of chemistry are toxic; it’s only their amount and combinations that make them therapeutic.
First, let’s appreciate that lithium is essential for life. Chemists and geologists study such matters by exposing animals to a diet deficient or wholly lacking in a certain element. They then compare those animals to others who are fed a diet rich in the element. If one group dies more quickly than the other, then the element is deemed to be essential for viability of the animal.
With lithium, this research has been conducted with worms, mice, rats, and goats. It turns out that animals deprived of any lithium have much shorter lifespans than animals who eat the usual amount of lithium in the diet. These animal studies provide context for studies of effects of lithium in drinking water.
Lithium seeps from rocks into the ground, where it is taken up into water, or into vegetables. It is consumed by animals and humans through eating vegetables or meats of other animals who have eaten vegetables. Eggs, for instance, have a notable amount of lithium. All this depends of course on the source of such food. Since food is shipped to different parts of the world, a better measure of local lithium supply is the drinking water, which tends to come from local sources.
In the last 50 years, about ten different studies have been conducted of the biological, medical, and social effects of lithium in drinking water. To give a sense of these studies, let’s take the first from 1970, which looked at lithium levels in the 100 largest American cities. It found lower overall mortality in areas with higher lithium levels; much of this benefit was related to less death from cardiovascular disease. Another 1970 study compared different counties in Texas. More recently, a 2009 study in Japan looked at over 1 million people, and 4 reports in 2011-2012 came from Japan, England, and Austria, and a new Texas study was published in 2013. These studies mainly focus on suicide, and, rather consistently, they find lower suicide rates in those regions where lithium levels are higher in water compared to those regions with lower lithium levels.
The most obvious question in such population studies is the issue of confounding bias, as discussed in the April 2015 PL issue. What other factors in these populations could affect suicide rates? Could it be that the lithium association is spurious because of those other factors? As described last year in PL, there are many known risk factors for suicide, such as depression and social isolation, which cannot be assessed in such large national studies. Some proxy measures can be assessed, though, such as living in cities versus rural areas, or measures of poverty and unemployment rates in areas being studied. Some of the recent studies controlled for those social factors. Still, since randomized analysis is impossible in studies of entire countries, we will be left with some uncertainty in such large-scale analyses. With that caveat, controlling for some social confounding factors, there still is an association between more lithium and less suicide.
The fact that lithium has been proven to prevent suicide in randomized clinical trials, as reviewed in the meta-analysis in last year’s April PL issue, is the main scientific rationale for accepting the likely validity of that same association in the multiple studies of lithium in drinking water.
In other words, the water studies don't “prove” that lithium prevents suicide. We already know that is the case with the meta-analysis of RCTs of bipolar illness. What the water studies strongly suggest is that very low doses of lithium could be sufficient for such suicide prevention purposes. We now turn to the meaning of the concept of “very low dose” lithium.
If this association is real, then we can turn to characterizing what it might mean in clinical settings. Clinicians who prescribe lithium are used to thinking of standard doses as around 900 mg/d, which produces blood levels of about 0.6-1.0, which is proven therapeutic for acute mania or prophylaxis of bipolar illness. Lower doses are seen as too “low”; thus, for clinicians a dose of 300 mg/d of lithium carbonate would be seen as “low” or “subtherapeutic.” Most would consider that dose clinically ineffective.
However, when you start to think about lithium doses from the perspective of the normal diet, rather than prescribing pills for bipolar disease, then the concepts of “low” and “high” begin to change. In the normal diet, we consume about 1 mg/d of elemental lithium. In the water studies, “high” amounts of lithium meant more than that amount. Thus, about 5 mg/d of elemental lithium would be considered high by dietary standards. This kind of dose produced apparent suicide prevention benefit.
It is important to translate elemental lithium into lithium compounds. For lithium carbonate, 100 mg/d of that compound is equivalent to 18.8 mg/d of elemental lithium. Thus, the “high” dietary amount of 5 mg/d of elemental lithium translates into about 25 mg/d of lithium carbonate. This dose would be considered vanishingly low by clinicians who would think of 300 mg/d as “low”. In fact, 300 mg/d of lithium, which is about 56 mg/d of elemental lithium, is quite high. It is 56 times more than the normal amount consumed in the diet. The “standard” lithium dose of 900 mg/d of lithium carbonate translates to 169 mg/d of elemental lithium, which is extremely high, using the standards of naturally consumed lithium.
In other words, if the suicide prevention studies of lithium are correct, clinicians may be justified in rethinking their conception of lithium dosing, and coming to the conclusion that there is almost no minimum to lithium dosing. Any amount of lithium is better than none, at least for suicide prevention.
This perspective would be justification for giving very low doses of lithium, as discussed in the Clinical Tip below, for suicide prevention.
Another aspect to the water studies is that they showed suicide prevention in the general population, not just people with bipolar illness. Thus, they can provide justification for considering lithium in all suicidal persons, irrespective of diagnosis. Given the risks of toxicity and side effects, if lithium is effective in very low doses for suicide, such dosing could allow for more feasible use in a wide range of populations at risk for suicide, such as military veterans and athletes with chronic traumatic encephalopathy. In other words, don't think of lithium for suicide prevention as only being limited to bipolar illness.
A word of caution regarding “parasuicide”, such as self-mutilation and self-cutting. The meta-analysis of RCTs showed that lithium did not reduce parasuicidal behavior. Thus, these considerations apply to serious suicide risk, such as overdoses or hanging or such behavior, and do not apply to persons who only engage in self-mutilation, such as in many cases of borderline personality.