will discuss some findings about the factors that
are correlated to clinical response to clozapine
(Clozaril® , FazaClo® ).
Clozapine is a drug used to treat psychosis. It usually is
used for persons with schizophrenia; sometimes it is used for persons
with schizoaffective disorder, bipolar disorder, or psychosis
associated with Parkinson disease.
Clozapine was developed in the 1960’s. It was the first
so-called atypical antipsychotic medication.
It is not a first-line drug, because of problems with adverse
effects. However, it remains the single antipsychotic drug
that is most likely to produce a favorable response in a patient with
psychosis. That makes it special. Thus, anything we
learn about the way it works is automatically interesting.
The (open-access) article, and a later column in the same journal,
refer to the finding that higher blood lipid levels seem to correspond
to greater clinical response to the medication. That is, the
higher the lipid levels, the greater the positive effect of the drug.
in serum lipids, independent of weight, are associated with changes in
symptoms during long-term clozapine treatment
J Psychiatry Neurosci. 2007 September; 32(5): 331–338.
Investigators have reported that weight gain attributed to clozapine is
associated with its clinical response. However, weight gain is a
nonspecific physiological variable that, in itself, does not explain
the mechanism underlying this relation. Alternatively, other biological
variables that are often associated with weight gain, such as serum
lipids, may assist in explaining this observation. The primary
objective of this study was to determine whether an increase in serum
lipids is associated with improvement in schizophrenia symptoms during
steady state treatment with clozapine.
The data for this study represent a subset of data from a randomized,
double-blinded trial that evaluated subjects with schizophrenia who
demonstrated a poor treatment response to clozapine. While continuing
their clozapine therapy, subjects were randomly assigned to receive
either risperidone 3 mg daily or placebo for 8 weeks. This course of
treatment was followed by an optional (open-label) 18 weeks of
augmentation with risperidone. In the present study, we included all
subjects from the previously reported trial who had fasting lipid
analyses and Positive and Negative Syndrome Scale (PANSS) scores from
days 7 and 63 (n = 55). For the primary analyses, we used multiple
regression to examine the association between serum lipid
concentrations and PANSS scores, after controlling for weight.
The analyses showed that the change in serum lipid concentration
predicted change in symptoms over that of change in weight.
Specifically, an increase in serum triglyceride concentration was
associated with a decrease in the total PANSS score (p = 0.037). In
addition, an increase in either serum total cholesterol concentration
(p = 0.007), serum triglyceride concentration (p = 0.017) or their
combined effect (p = 0.010) was associated with a decrease in PANSS
negative subscale scores.
Elevation of serum lipids is associated with an improvement in
schizophrenia symptoms in subjects treated with clozapine. Although the
mechanism is unclear, serum lipids may play a role in influencing
clozapine’s therapeutic activity.
Earlier, in 2003, Procyshyn had described
an hypothesis: (not open-access)
When drugs enter the blood, they often
form complexes with plasma protein molecules that are present in
abundance. These drug-protein complexes, which primarily involve
albumin, can affect the pharmacokinetic characteristics of the involved
drug. Unfortunately, it is not as widely recognized that drugs can also
form complexes with other plasma proteins including lipoproteins. In
this regard, plasma lipoproteins have also been shown to modify the
pharmacokinetics, tissue distribution, and pharmacological activity of
Recognizing that clozapine is hydrophobic
and associated with hypertriglyceridemia compelled us to examine what
effect dyslipidemia may have on clozapine’s plasma distribution.
Interestingly, in conditions simulating hypertriglyceridemia, clozapine
redistributed itself from the lipoprotein-deficient fraction (where it
is primarily bound to albumin) to the very low-density lipoprotein
fraction. This finding led us to question what effect, if any, this
redistribution of clozapine (secondary to elevated triglycerides) has
on its clinical efficacy or effectiveness. Combining basic science data
with clinical observation, this article addresses the hypothesis that
serum triglycerides may contribute to clozapine’s observed
effectiveness. If our hypothesis is correct, then serum triglycerides
may be the variable that explains why weight gain associated with
clozapine treatment has been documented as a good predictor of clinical
It is well known that drugs bound to plasma proteins are not active;
portion of the drug that is not bound to the
protein is available to do whatever it is that the drug is supposed to
do. Similarly, association of drugs with lipoproteins can
affect the rate and extent to which drugs move throughout the
body and interact with their targets.
Recently, Procyshyn, Horner, and Barr published
a column in the Journal of Psychiatry and Neuroscience
findings. There is no HTML version, but a PDF is openly
serum lipids predict response to clozapine treatment? They
have expanded their hypothesis count to three:
1) Lipids have a direct effect on symptoms
independent of clozapine.
2) The partitioning of clozapine into the LDL and VLDL fraction of
serum creates a “physiological depot” for
clozapine. As such, clozapine is released from the lipoprotein fraction
of serum in a sustained manner comparable to other depot antipsychotic
3) The redistribution of clozapine into the lipoprotein fraction
facilitates its penetration across the blood–brain barrier
either by passive diffusion or by a receptor-mediated process.
Some previous studies have
that the beneficial effect of clozapine is correlated with weight
change. At least one study did not
show this correlation. Let is suffice to say that
there is a vast literature on the metabolic effects of atypical
antipsychotics in general, and clozapine in particular. There
are even articles showing correlations between receptor polymorphisms
and metabolic abnormalities in patients treated with these drugs.
What we have is an extraordinarily complex web of correlations, with
little to show regarding established causation. To complicate
matters even more, we know that the incidence of elevated lipids is
increasing in the general population, regardless of exposure to
antipsychotic medication. So we know that there are
confounding factors in any study we undertake. While this
would be unlikely to effect the outcome of a single study, it could
What I find interesting in all of this, is the notion that the lipid
increases in clozapine-treated patients may not merely be adverse
effects. It could be that there is something therapeutic
about the hypertriglyceridemia, at least in the presence of clozapine.
If this were the case, it could involve either
pharmacokinetic or pharmacodynamic effects.
The study did include measurements of the serum concentrations of
clozapine and its active metabolite, norclozapine:
No correlations were noted between the change in
serum triglyceride concentration and the change in clozapine (r = 0.21;
t50 = 1.54, p = 0.13) or norclozapine (r = 0.13; t50 = 0.91, p = 0.37)
However, this does not rule out the possibility of pharmacokinetic
effects. It would be interesting to see if triglyceride
levels have any discernible effect on receptor occupancy in relation to
What I am really getting to is this: if it can be shown that an
increased triglyceride level does improve the effect of clozapine, but
the improvement is lost when TG is lowered with a statin, would it turn
out that the improvement could be regained by administering fish oil?
I know this kind of thing bothers some people. First, you
give a drug. The first drug has benefit, but also adverse
effects. So you give a second drug that alleviates the
adverse effects, but causes problems of its own. So you give
a third drug to correct the problem caused by the second drug.
But if that is what it takes…
This is all getting to be speculative, obviously. That is
what blogs are for, after all. But it is not entirely
speculative. There is a bit of evidence that fish oil can
help. Specifically, the component known as EPA, or
eicosapentaenoic acid. The evidence is summarized
in a Cochrane review, which unfortunately is now three years old.
The folks at Schizophrenia.com have collected some
It may seem pretty far-fetched to think that lipids (fat) could could
serve an important role in a serious mental illness. But it
is important to get away from the notion that fat is simply a way to
store energy. It serves structural functions, as well as
signaling functions, too.
As the Cochrane people point out, the data are not conclusive.
More research is underway. For example, there is
this study at U Cincinnati: Omega-3
Fatty Acid Deficiency Replacement in Early Schizophrenia.
The odd thing about this study: it uses olive oil as a
placebo. The problem with that, is that we do not know if
there is something particular about EPA that is therapeutic, or if
other types of lipid might have a positive effect. The Procyshyn
study on triglycerides
suggests that other lipids (but not all) might have a positive effect,
at least in the presence of an clozapine. The Cincinnati
study allows concurrent use of antipsychotic medication.
Obviously, it is going to take a while to sort this out.
M. Procyshyn, PhD, Kishor M. Wasan, PhD, Allen E. Thornton, PhD,
Alasdair M. Barr, PhD, Eric Y.H. Chen, MD, Edith Pomarol-Clotet, MD,
Emmanuel Stip, MD, Richard Williams, MB, G. William MacEwan, MD, C.
Laird Birmingham, MD, William G. Honer, MD, and f (2007). Changes in
serum lipids, independent of weight, are associated with changes in
symptoms during long-term clozapine treatment J Psychiatry Neurosci, 32
(5), 331-338 DOI: 17823649
[note, the citation system appears to handle the object identifier
incorrectly. It should not reference the DOI; rather, it
should reference the PMID 17823649. J Psychiatry Neurosci
does not use