Monday, September 5, 2011

Intriguing Links Between Depression and Cholesterol

Today's paper comes from the Journal of Clinical Psychiatry: Serum Lipoproteins Improve After Successful Pharmacologic Antidepressant Treatment: A Randomized Open-Label Prospective Trial.  Turns out that depression researchers are thinking about the whole body and not just the brain.  My basic premise is that it is all connected, folks.  Inflammation, heart disease, diabetes, obesity, metabolic syndrome, depression, IBS… Let's lump, shall we?

(Music - Ravel, Jeux d'eau - right click to open in new tab)

So the truth of the matter is that people suffering from depression have lower overall cholesterol than average.  I know, crazy, right?  Here's where I really blow your mind - people with major depressive disorder, despite the lower cholesterol, have higher rates of death from heart disease, whether or not you had heart disease before or after the diagnosis of depression.

But anyone in the primal/paleo community will know that total cholesterol doesn't mean much - we want to know about HDL and the subfractions of LDL - the big fluffy new fresh LDL are not associated with increased risk, whereas the old, rancid, small dense LDL are certainly associated with increased risk of heart disease.  Here is a quote from the article (though I'm not entirely sure I agree with the pathology description):

"small dense" LDL particles, resulting from packing of LDL particles with higher amounts of triglycerides, have a higher propensity to be oxidized, to be trapped in the subendothelial space, and, subsequently, to form the seed of an atherosclerotic plaque."
In favor of further lumpage, it is known that depression is associated with insulin resistance, which is associated with more small, dense LDL, and thus higher triglycerides and higher apolipoprotein B.   Successful antidepressant treatment will just so happen to to improve insulin sensitivity and all those bad metabolic markers.

In the paper, lipoprotein composition of various folks with major depressive disorder were recorded and compared to healthy controls.  Then depressed subjects were randomized to treatment with either mirtazapine (an effective antidepressant known to cause weight gain) or venlafaxine (another antidepressant less likely to cause the same).

Results:  not surprisingly, total cholesterol was lower in depressed patients compared to healthy controls (surely the total cholesterol amount and how important robust cholesterol levels are to the brain has nothing to do with that.  I'm sure it is just a confounder.) Both HDL and LDL were significantly lower.  There was also a higher ratio of nasty, small, dense LDL particles to HDL particles  in depressed patients compared to controls.  Depressed people are much more likely to have rancid LDL lingering in their bloodstream.  To me, not surprising.

Let's add the psychotropics - predictably, folks taking mirtazapine gained some weight, and folks taking venlafaxine lost some weight (serotonergic antidepressants such as venlafaxine will tend to cause short-term weight loss, likely though increasing satiety signals to the hypothalamus, of course some folks consider the hypothalamus to be of secondary importance with respect to weight gain or loss).

So what happened with the interesting subfractions of lipoproteins among the antidepressant treated groups?  With fat-inducing mirtazapine (which causes weight gain through a central histamine or anticholinergic mechanism, most likely), total cholesterol and triglycerides increased.  Under venlafaxine treatment, total cholesterol remained stable and triglycerides decreased.  With both groups (and these are both powerful antidepressants who tend to actually work and help people feel a bit better - the study showed about 60-65% response, which is typical for these agents), the HDL levels improved, and HDL to LDL ratio and apolipoprotein B (a measure of old, dense, rancid cholesterol) decreased.  The mirtazapine group gained weight (as expected), the venlafaxine group did not.  In both groups, responders had a slight increase in total cholesterol.

What do we learn?  Depressed folks who responded to antidepressant treatment tend to increase their HDL levels and had favorable changes decreasing old small dense LDL.  Oddly enough, those on the fattening mirtazapine had similar good lipid changes (if the medicine worked) compared to the short-term slimming venlafaxine.  (Editorial note:  inflammation is inflammation, my friends, and an anticholinergic or histamine mechanism will make you fat regardless of how inflamed you are - and while most of this work has been done with antidepressant medication, I have linked studies in the past showing that therapy has also been shown to be anti-inflammatory, though to be honest it has not been looked at robustly).

Wildly enough, this study says "to the best of our knowledge, the composition of LDL particles has not been studied previously in depressed patients."  WHAT?  Small dense LDL vs large fluffy LDL was only discovered, like, 20 years ago, right?  Mental health research is always lagging! And antidepressants might actually be anti-inflammatory as suspected… dare I say like statins?  Both classes of drugs have evidence for some modest benefit in certain situations, and major drawbacks.


  1. Could it be that antidepressants can be antimicrobial?

    I know SSRIs are anti-fungal agents ( What if the positive responders are the folks with Candida infections?

  2. "the total cholesterol amount and how important robust cholesterol levels are to the brain...", maybe it's the other way around? how important the brain is to a robust cholesterol level?

    A more functioning brain will make more cholesterol (and release more into the blood stream) than a compromised brain?

  3. Interesting study and post.
    Cholesterol in autism, and in particular some suggestion that it might be 'low' in some people is fast becoming a talking point and the suggestion of supplementation currently under investigation:
    @Paul Jaminet. There is a wider question about lots of different pharmacotherapies and their various modes of actions. Although perhaps slightly outside of Dr Dean's current post, this paper provides some interesting speculation on schizophrenia, dopamine and parasitosis:

  4. I was with you until here: "(Editorial note: inflammation is inflammation, my friends, and an anticholinergic or histamine mechanism will make you fat regardless of how inflamed you are)."

    Can you explain this one just a bit more?

    Another question re: inflammation... Even though SSRIs cause diarrhea in a lot of folks, they are still thought to be helpful for IBS and anti-inflammatory in general? The diarrhea secondary to SSRIs is not a kind of inflammation?

  5. Also, here's the Maes piece which seems apt here:,%20cytokine%20hypothesis.pdf

    I think you refer to it elsewhere.

    If the main anti-depressant function of anti-depressants is not about serotonin retention but about reducing inflammation, then do statins help with depression?

    The use of NAC (n acetyl cysteine) for anxiety may also apply here?

    Also the stuff about niacin and inflammation:

  6. Paul - Since folks with depression and anxiety have underactivity or overactivity of certain neurotransmitters that seems to respond to successful treatment with the antidepressants that act on those neurotransmitter system, it is the mist likely mechanism, though no one really knows. I get cheeky thoughts along those lines (not infections so much, as I tend to think of infections as fire accelerants rather than the core fuel, though we disagree on that) with respect to the atypical antipsychotics. The best-working atypicals for schizophrenia seem to have the worst metabolic side effect profile, causing weight gain, diabetes, etc. Clozaril being the main example here. I always wonder if the metabolic disaster is part of how the medicine is somehow altering the psychotic symptoms. I can't explain it precisely - they tend to raise triglycerides and make one more insulin resistant rather than raising cholesterol per se…

    Scott - I'm not sure how much a functioning brain turns over cholesterol compared to a less-functioning one. Interesting thought. The synapses are hotbeds of turnover, the myelin sheaths have very slow turnover.

    Paul - very interesting, as usual! And I'm not surprised, given the data with respect to dementia and psychosis that the somewhat similar neuronal killing fields pathology of autism would also be correlated with low cholesterol. More evidence the best theoretical treatment for autism could be bone broth and plenty of saturated fat?

    Cate - I was probably being too flippant and too lumpy. So far as the statin question, the literature is a little thin, but statins are very effective cholesterol-lowering agents but don't seem to cause (as much) depression, suicide, etc. as the pre-statin cholesterol-lowering drugs probably did - I wonder if the anti-inflammatory effect is the reason. Even cholesterol-lowering *diets* have been correlated with increasing depression - so it is funny that statins are somewhat spared.

  7. First of all, I find your blog really interesting. Being I'm a sociology/psychology major, my spin om some things are a little different.

    By the way, I've been on both Effexor (Venlafaxin) and Zoloft etc in the past and part of the slimming effect of Effexor is the fact that it also increases noradrenalin in the brain, which, being a stimulatory neutotransmitter, lessens the need and probability of turning to food for stimulation. Ie candy etc becomes redundant; at least for me I noticed that, compared to for example Paxil or Zoloft which made me just dull and wanting to eat candy all the time (probably to get my catecholamines and motivation up).

    Your opinion?

  8. Just imagine:

    "Ovator (Cholesteryl Palmitate) can help to control symptoms of depression. Ask your doctor if Ovator is right for you."

  9. Jens - effexor is primarily an SSRI with a 120:1 activation of the serotonin receptor compared to the norepi receptor - though at doses higher than 225mg things can change - clinically I've found the weight changes to the various agents to be highly variable, though I outlined the overall studies in this post -