In which Sci gets Snarky: Lithium in the Water

Aug 03 2009 Published by under Physiology/Pharmacology

This will otherwise be known as "WTF are you doing, Journal of Medical Hypotheses".
Sci likes Lithium.
lithium flame.jpg
(Lithium burns red. Sci thinks this could have applications for lightsabers if applied correctly)
It's a cool element, interesting in that we've used it over the ages for stuff like gout (which, I hear, is making a comeback), prevention of migraine, blocking the effects of excessive anti-diuretic hormone, and of course for bipolar disorder. But what's really interesting? We don't know how it works. Not a clue. It may raise serotonin levels over time, it may decrease or increase other monoamine neurotransmitter levels (such as dopamine and norepinephrine). But we have no idea HOW this occurs. In this, as with other psychiatric drugs (such as Ritalin, which increases dopamine and norepinephrine and improves attention, or with selective serotonin reuptake inhibitors, which increase extracellular serotonin and alleviate depression), we don't know HOW they work. We just know they WORK (though we're working on that).
And THAT they work is what matters. When you're dealing with someone who is potentially suicidal, or unable to function successfully due to severe psychiatric disorders, you'll take what works and isn't otherwise "bad" for you (that you can tell), and figure out the mechanism later. And some of these drugs, like Lithium, are very old indeed. Lithium has been used to treat mania and depression since the 1870s, and is still used today as one of the most effective treatments for bipolar disorder (a topic which I SWEAR I will cover someday). It goes a long way toward reducing suicide in those with severe bipolar, and can help with mood stabilization as well.
So if it's been used that long, it must be good right? This means it might do everyone some good, right? Like, you could put it in the water, and it'd be fine? Terao, et al. "Even very low but sustained lithium intake can prevent suicide in the general population?" Medical Hypotheses, 2009.

Now, I think we all know that if it's called "the Journal of Medical Hypotheses", these aren't going to be studies which are supported by a lot of data. Really, Sci is not sure why people publish in this journal at all, except for the purposes of giving Sci a good laugh and linking together masturbation and nasal congestion (for the record, I don't really care if this method works, and I'm not interested enough to obtain IRB approval to find out, but I also don't worry that it's dangerous. Well, ok, maybe it is. We all know that some methods of sex are dangerous). I mean, the journal isn't peer reviewed, and if you know anything about rigorous science, you know that the vast majority of us consider peer review to be pretty essential. So really, it's probably rare that you're going to see data sets (though you do see some, but they're not what I'd call complete). So without data sets or peer review, what do you get? You get the somewhat crack-pot musing of scientists when they are bored or drunk (or both), or perhaps when they are just getting desperate for grant ideas. All of this makes for entertaining reading, but please, I beg you, PLEASE, do not take anything you read in the Journal of Medical Hypotheses as the truth. Even if Sci covers it (if I cover it, chances are I think it's funny, but I don't necessarily think it's true).
But then sometimes Sci sees an article (as she did today), and gets angry. Because in Medical Hypotheses, you can suppose anything. You can suppose Lithium in the drinking water. And this, my friends, is some rank levels of crazy up in here.
The authors of the "hypothesis" are relying on correlations between trace levels of lithium in the drinking water and lower rates of suicide in places like Texas and Oita area of Japan. And certainly, it is very possible that there are higher rates of lithium in these areas, and lower rates of suicide. They also note data that even bipolar patients who do not respond well to treatment have lower levels of suicide than those who are bipolar and not treated with lithium. So it appears that lithium treatment correlates with a lower risk of suicide in depressed and non-depressed populations. Got it? Good.
Now let's all take a big, deep breath, and talk about how correlation does NOT equal causation, and about how anecdotes are NOT data. Good? Good.
And now let's come back. There are a LOT of things outside of lithium levels in the drinking water that could correlate with lower rates of suicide in any given area. These could include higher overall rates of mental health, higher rates of treatment for psychiatric disorders, differences in socio-economic status, differences in religion, etc, etc, etc. And as far as bipolar patients, those who are NOT on lithium may have been poor responders. They may also be unmedicated. Both of these things increase the risk of suicide in bipolar patients. Those who are not on lithium may not be receiving as much or as frequent care and follow-up as those on lithium. And one can never discount the placebo effect on mood, it is possible that simply KNOWING you are treated can help you get through some of the more difficult bipolar periods.
But yet. Three studies, all of them correlations, and they want to put lithium in the drinking water. But hey, you might think, lithium's a naturally occurring ion, right? What's the problem?
Well, lithium is indeed naturally occurring, but it's usually in things. Occurring alone as an unbound ion or in the salt form as it is given in the clinic is really pretty rare. It's a very fragile element, and highly reactive, so it's going to bind to things. So even getting IN to the water in a therapeutic form is going to be a challenge. That's problem number one.
Problem number 2 is dosing. Lithium is a VERY hard drug to dose. It's hard to keep drug levels in a constant therapeutic range, and even harder to FIND that therapeutic range in some individuals. There is no proof that low doses below therapeutic range would have any promising effect on the population (except for those two correlative studies). Additionally, there's no proof that lithium can decrease suicide rates among those without psychiatric disorders. And while in low doses it can be a mood stabilizer, and in higher doses can combat forms of mania, it's VERY hard to get those doses right without running into side effects. This is what we call a narrow therapeutic range, and the therapeutic range of lithium is notoriously narrow.
So what are those side effects? Here we get to problem number 3. Luckily, lithium is not deadly in therapeutic doses. The most common side effect is weight gain, which, though not something society needs more of, is more of a long term threat than an immediate one. But another very common side effect is decreases in thyroid hormones.
For those who don't know a lot about thyroid hormones, these are hormones that control metabolism regulation, especially things like weight gain, energy level, and body temperature. Lithium tends to reduce the activity of thyroid hormones, and can cause hypothyroidism, resulting in constant fatigue, constant feelings of cold, increases in appetite (weight gain), constipation, hair loss, and depression.
But it doesn't stop there. Lithium can also have effects on the kidney, causing you to excrete too much sodium and water, and causing dehydration. Dehydration can concentrate lithium levels, which of course can make this worse.
The net result of all these is that patients on lithium must be monitored for their blood levels on a regular basis, to check for kidney and thyroid function, as well as other side effects like weight gain and changes in potassium levels. Not only that, lithium, as I mentioned before, likes to bind to things. This means it also can interact with other drugs. Due to its effects on the kidney, it should NEVER be taken by those who are being treated with diuretics (for things like hypertension, diuretics are a first-line treatment. Lithium has also been known to react very badly to other psychiatric drugs such as haloperidol and other antipsychotics.
Now, constant measurement in people who are bipolar and on lithium is necessary. It's not fun, but it's the drug that works, and for those people, the side effects are often worth it.
But you want to put this in the drinking water?
Now, granted, the authors say they want to use really low doses. But HOW low? If a dose is low enough, it won't have any effect at all. And remember, lithium has a VERY low therapeutic range. What's barely detectable in a grown man might be a big deal in say, a small child or woman. Not to mention the possible combined effects in people who actually ARE on lithium at therapeutic levels. And what about possible interactions with the millions of people on diuretics, or the people on antipsychotics? The side effects of lithium in combination with antipsychotics can be pretty toxic, is there a dose low enough that would even make sense?
I think you can tell where Sci is going with this. This is a bad idea. At best, it's a potentially useless idea, and at worst, it is bad indeed. For hypotheses like these, founded on the bare minimum of evidence and with little regard for things like negative side effects, Sci proposes a new journal. Sci will be editor. Now accepting submissions for "The Journal of Delirium Noctourne."
Terao, T., Goto, S., Inagaki, M., & Okamoto, Y. (2009). Even very low but sustained lithium intake can prevent suicide in the general population? Medical Hypotheses DOI: 10.1016/j.mehy.2009.02.043

23 responses so far

  • I for one am switching to drinking only grain alcohol and rainwater before it's too late!

  • We don't know how it works. Not a clue.

    This is an exaggeration. There certainly are reasonable mechanistic clues concerning lithium action, including, e.g., the inhibition of glycogen synthase kinase 3.

  • Tex says:

    I really hate the kind of correlational 'inductive' reasoning embodied by this Medical Hypothesis article. Any other popularized examples (there are many)?
    To get the ball rolling -- "Marijuana is a gate-way drug." Right, and ice tea is a gate-way drink for alcoholics since almost every alcoholic has had ice tea at some point. Stay away from the ice tea people! Run for the hills!

  • Scicurious says:

    Physioprof, indeed you are correct. I was not aware of that one. There's changes in uptake of myo-inositol as well as changes intracellular pH. Sci stands corrected.

  • ScienceWoman says:

    @Tex: My adamant teetotaler grandfather would not allow soda either. His reasoning? Soda is "training whiskey."

  • "Now, granted, the authors say they want to use really low doses. But HOW low? If a dose is low enough, it won't have any effect at all."
    No, no, Sci, you totally missed the boat on this one. They want to use really really low doses - homeopathic lithium doses. That's totally safe because there won't actually be any Li in the H2O...but the water will remember it and we'll all be less suicidal. Duh.

  • Art says:

    Jacob Wintersmith @1 - "I for one am switching to drinking only grain alcohol and rainwater before it's too late!"
    No sacrifice is too great to protect our 'Precious Bodily Fluids'.

  • daedalus2u says:

    Beg to disagree CPP, the cause of suicide has never been shown to be due to too much activity of glycogen synthase kinase 3. There are some clues that might lead to fruitful investigation pathways, but there really is “no clue” as to the actual mechanism of why lithium reduces suicide or improves bipolar.
    I think the effect of lithium is through regulation of ATP levels and that depression (and also bipolar) are consequences of an ATP level in a certain range (depression low and mania lower still). Lithium acts to increase the ATP level and to make that level more stable.
    Lithium affects a great many pathways; GSK-3 phosphorylates ~50 substrates, there are many other lithium affected pathways. Presumably those lithium affected pathways are all regulated in the absence of lithium and that regulation is changed in the presence of lithium. Exposure to lithium in utero may cause epigenetic programming that then causes a requirement for a certain lithium concentration for proper regulation over that person’s lifetime. Without knowing what those pathways are and how they are regulated by what mechanisms over time, putting lithium in the water is grossly irresponsible.

  • Scicurious says:

    AA: LOL!!!
    Daedalus2u: I don't think CPP is advocating the use of Lithium in the water by any means, he was just mentioning that some of the mechanisms of action of Lithium are known (or at least hinted at).

  • daedalus2u says:

    Sci, I know CPP wasn't advocating putting lithium in the water, but there really isn't an understanding of what properties of lithium improve bipolar by what mechanism(s).
    Just like there isn't a clue as to how SSRIs work for depression. Yes, I know they inhibit serotonin reuptake, but how does that cause resolution of depression? There is not a clue.
    There are clues as to how opiates work, they activate opiate receptors. Opiates are pretty well understood. The understanding of lithium is no where close to that.

  • Neuroskeptic says:

    A while back I wrote about the original paper reporting on the correlation, Ohgami et al, - here.
    I said -
    "The maximum lithium level in the water was found to be 59 micrograms per liter. By contrast, if someone were taking lithium for bipolar disorder, they would on average be taking about 1 gram of lithium carbonate (Li2CO3) per day, which works out at about 100 mg of lithium (and 900 mg of inert carbonate).
    Assuming that people drink about two liters of tap water per day (almost certainly they drink less), that's about 120 micrograms per day. That's 0.12 mg. So about a thousand-fold less than you'd take if you were bipolar.
    Could such a tiny amount of lithium do anything? Well, maybe. It's not impossible that it would have some small effect on suicide rates, but it seems very unlikely that it would account for the difference between a rate of 120 and a rate of 80 which is what the graph seems to imply. The effect is just too large to be credible, if you ask me."

    Which is what I have to say about this paper too.
    Still, don't hate on Medical Hypotheses. The whole point of it is that it runs papers that are highly speculative. It's what it's for. so long as you read them with a big pinch of salt some of them are quite interesting.

  • dd says:

    Hamburgers alleviate starvation, but I hope no one recommends putting them in our water supply. Also, both Texas and Oita have big domed sports stadiums, is this a significant correlation to be studied?

  • arvind says:

    Now let's all take a big, deep breath, and talk about how correlation does NOT equal causation

    O ye, of little faith!! How else do you explain the rise in global temperatures with the fall in pirate population? Huh? Huh??

  • Eric Johnson says:

    Obviously the general tendency is to overstate the degree to which correlation suggests causation. However, in this case I think you may have understated it. If A and B correlate, but without any causal relationship, then the causes of observed variation in B must include variables C, D, E, etc which individually correlate with both A and B. In many cases it is extremely easy to list possibilities for C, D, etc, but in this case I wonder if it isn't a little hard to come up with things that would correlate with both suicide and tapwater lithium (especially if the Li derives from soil/rock and not from the pipes, which I can't say).
    Thus, I would tend to consider this correlation rather suggestive, in itself, if the Li in fact doesn't come from the pipes. But that suggestiveness is not borne out, in my opinion, because I think Neuroskeptic's considerations pretty much dismiss this entire idea.
    Putting Li in the water is no doubt a bad idea regardless.

  • leigh says:

    sample size!!!!! sample size!!!! sample size!1!!!1!!1!

  • David says:

    For those interested, Todd Gould has a really interesting new article proposing that lithium's ability to lessen suicidality has to do with its modulatory actions on impulsivity-related phenotypes (there is an extensive historical literature on the idea that suicide is - essentially - an impulse control problem). This raises all sorts of interesting questions that beg to be answered (drug abuse? ADHD?) ....
    Kovacsics CE, Gottesman II, Gould TD. Lithium's antisuicidal efficacy: elucidation of neurobiological targets using endophenotype strategies. Annu Rev Pharmacol Toxicol. 2009;49:175-98.

  • daedalus2u says:

    It does look like lithium is an essential nutrient with a RDA of ~1 mg per day.
    "In some parts of Texas, for example, tap water Li levels may reach 170 micro g/L, adding about 340 micro g of Li to the daily intake of lithium from foods. In these regions, urinary Li excretions of local residents vary inversely with rainfall, reflecting the dilution of drinking water supplies [18]."

  • Anonymous says:

    Isn't there a study of lithium in El Paso drinking water, which shows that the city of El Paso has a very low suicide, depression rate.

  • Tom Hennessy says:

    Lithium since you seem to be genuinely interested is used to balance the fatty acids in the brain. Simple. THAT is why fatty acids are being found to be of use in anxiety depression and bi-polar.
    Lithium lecithin choline nicotine acetylcholine 'standin' for each other.
    By way of the cholinergic anti-inflammatory pathway.
    I've been working on the connection between this pathway and Amyotrophic Lateral Sclerosis in that two different mouse models have shown a lack of acetylcholine.
    Myasthenia gravis has been shown to be successfully prevented by inclusion of choline containing lecithin and Gravis has been shown to precede the onset of ALS in may cases.

  • Nessie says:

    Neuroskeptic writes,
    "Still, don't hate on Medical Hypotheses. The whole point of it is that it runs papers that are highly speculative."
    Well, that and papers making false and dangerous claims contradicted by the very sources cited. For example, the latest AIDS denialism (Med Hyp, July 2009) from Peter Duesberg, claiming that AIDS hasn't had any impact in Africa. (Because population continues to increase--a fine argument, since it means there are no real medical problems anywhere in the world where there's non-negative population growth.) And that antiretroviral drugs haven't lowered AIDS mortality (quoting a study comparing different ARV regimens, all of which lower mortality drastically).
    Medical Hypotheses is a bizarrely pseudoscientific "journal" where anyone can send in anything and a single editor decides whether to accept or reject the MS. No peer review, not even your standard editorial review, just a yes-or-no decision. If that sounds like a blog, it should, and Medical Hypotheses should be just that: a moderated blog/message board. It would be extremely popular, all the more so for being free, and it wouldn't legitimize cranks with PubMed-listed "publications." As it is, libraries pay several thousand US dollars per year to carry this title and its scientifically unsound weirdness.
    If this concerns you, see the following document, which can be downloaded and sent to your library:
    There's also a new Facebook group (what else?):

  • Meat Robot says:

    As a big fan of lithium, Medical Hypothesis, and Dr. Strangelove, I'm loving this article and thread.
    Regarding lithium, Trevor Young, a bench researcher and psychiatrist at UBC has proposed the possibility (with evidence in vitro and small animals) that lithium may lead to increased neurogenesis, together with some human data suggesting (what else?) increased hippocampal volumes in lithium treated manics, an effect also noted in valproate and carbamazepine treated manics.
    As for Medical Hypothesis, bummer to hear it's not peer-reviewed. I'm loving it less now, though I do think there's got to be some sort of peer-reviewed place for the theorists to play.
    Dr. Strangelove! Just saw a brand new 35 mm print at a local repertory theatre last night. Wow! Keep an eye out for it, as it's a very clean print, as clean as the DVD. My photographer film-buff friend's guess is that this is a brand new transfer from the original 70 mm stock.

  • David says:

    the first (as far as I know) noted correlation between lithium levels in drinking water and mental health was Voors, N C Med J. 1972 v33(7)p597. The French Broad River, near my home, flows past rocks high in lithium, and the downstream areas had markedly lower mental hospital admission rates than the rest of the state.

  • LJ says:

    Lithium, Valporate (Depakote) and other drugs also suppress Toxoplasmosis which some pretty respectable researchers think may play an important role in psychosis, schizophrenia and bipolar - could this be part of lithium's "charm"?

Leave a Reply