Abstract

This commentary summarises the data that indicate the early work published on mirtazapine was inaccurate and misleading.  It points out that there are no data to substantiate the claim that mirtazapine is significantly different form mianserin.  Recent data demonstrates there is unlikely to be any meaningful difference between these two drugs, and that the supposed mechanism of action making it different from other antidepressants is mistaken (see table of receptor affinity data).  There are serious doubts about its ability to act as an effective antidepressant because there is no established plausible mechanism by which it could achieve that effect.  If it is an antidepressant, then we require novel ideas to explain how it achieves this effect.

Introduction

I wrote the original version of this commentary many years ago.  For decades Mirtazapine, (aka 6-aza-mianserin) was marketed as a drug whose mechanism of action was claimed to be different (via adrenergic alpha 2 antagonism) to all other antidepressants (including its predecessor and close structural analogue mianserin).

Structures at:

mirtazapine (6-aza-mianserin):

http://pubchem.ncbi.nlm.nih.gov/summary/summary.cgi?cid=4205

Mianserin

http://pubchem.ncbi.nlm.nih.gov/summary/summary.cgi?cid=4184

The ‘6-aza-’ tag simply indicates that one carbon atom is replaced by a nitrogen, in the 6-position.  So, these two compounds are almost identical: slight changes can make big differences, or no difference at all: for physico-chemical data, see [1].

Mirtazapine’s potency, and what effect that has on receptors or neurotransmitters, and what changes that might cause, and whether it is significantly different to mianserin, indeed whether it is an antidepressant at all, are all questions that require quality replicated scientific evidence: but this has never existed.

Since it has always been promoted to doctors as a dual action drug (one proposed to raise both serotonin and noradrenaline) it is instructive to consider this question in detail because it illustrates aspects of continuing debates concerning pharmaceutical companies, scientific integrity, bias in academia, refereeing standards of journals, and related issues.*

Science places the onus on those making novel claims and propositions to adduce evidence to support or ‘prove’ them (prove is a repetitively misused word).  A putative antidepressant effect for mirtazapine was, and still is, a novel claim because there is no established mechanism by which this drug might exert an antidepressant effect.  Science also requires independent replication of evidence (remember ‘cold fusion’).  There are two parts to this process, the first is replicability, that is a description of methodology sufficient for it to be precisely repeatable, the second is the actual replication and the attainment of the same result.  This essential and fundamental requirement of real science is systematically ignored in psychopharmacology.

There are some understandable factors that influence this, but that does not make it excusable.  Drugs are approved following consideration of data given to regulatory authorities by a single interested party, the pharmaceutical company making the drug.  It is rare for independent replication of any part of this evidence to be sought, or attained, prior to approval: and, as this story reveals, not afterwards either.  This takes Psychiatry into a grey area where it is rubbing shoulders with alternative medicine and sustaining itself with partisan evidence and wishful thinking [2-5].

Medical scientific publishing has been subjected to vigorous criticisms because of an appreciation of the lack of objectivity that distorts the scientific method.  Comments from eminent journals create a perspective.  The editor of the BMJ, ‘stepped down’ in 2004, and wrote that ‘Medical Journals Are an Extension of the Marketing Arm of Pharmaceutical Companies’ [6].  Horton, the Lancet editor, stated, ‘Journals have devolved into information laundering operations for the pharmaceutical industry’ [7], and Angell, the NEJM editor, said they were, ‘primarily a marketing machine … co-opting every institution that might stand in its way’ [8].

The evidence is that even meta-analysis and practice guidelines are tainted with bias [9-13].  Favourable evidence about drugs is sometimes manufactured [14-16].  Antidepressant drugs are still being ‘proved’ to work in trials and then found not to work by ordinary clinicians [17] and by subsequent research e.g., most recently, reboxetine [18].  Eyding concluded:

‘reboxetine is, overall, an ineffective and potentially harmful antidepressant. Published evidence is affected by publication bias, underlining the urgent need for mandatory publication of trial data.’ 

Just to reinforce what nonsense meta-analysis produces of research findings, this contrasts with another meta-analytical paper stating there was no difference in efficacy between SSRIs and reboxetine [19]!

See also my essay ‘Why Most New Antidepressants Are Ineffective: And How Pharmaceutical Companies Have Deceived Doctors’, and other commentaries in this section, for further discussion of the issues.

Mirtazapine is the 6-aza- analogue of mianserin, and was introduced as a new antidepressant in the first publication about it, and christened as such in the title ‘… new anti-depressant Org 3770 …’. 

At that juncture there was no published evidence whatsoever of its antidepressant efficacy [20].

This is an extraordinary and egregious example of promoting an idea before presenting any scientific evidence.  Indeed, author instructions in some journals specifically state that the title of an article should not announce the results.  That admonition is especially relevant here, because not only did they announce the ‘result’ but also there was no data whatsoever in the article concerning antidepressant effects.  As Sir Humphrey advised in an episode of ‘Yes, Minister’, ‘Always get rid of the difficult bit in the title – it does less harm there than in the text’ [21].

I have recently published a review of the human pharmacology of mirtazapine and mianserin [22].  In brief, the current research evidence fails to support a meaningful distinction between them.  Their most potent property is histamine receptor H1 antagonism, which is strongly associated with sedation and weight gain [23].  This seems not to be perceived as a positive sales attribute because out of 26 ‘Organon’ papers (i.e., papers that appear to be funded by the pharmaceutical company ‘Organon’, and/or written by employees) on mirtazapine this property receives brief mention in only one of them [24].  The maximum difference in Ki values at the adrenergic alpha2 (α₂) receptor (which is claimed to mediate its putative antidepressant actions), for mianserin & mirtazapine in the Organon papers is two-fold.  That is insufficient to substantiate any claimed difference, because the experimental variation in these measurements is at least 10-fold, as revealed in the PDSP online database (http://pdsp.med.unc.edu/) of receptor affinity data [25] and summarized in my review [22].

It is appropriate to pause and recap on the presentation and manipulation, as well as the meaning and relevance, of the above data.  The most potent pharmacological property of this drug is barely mentioned, even though it is the most potent anti-histamine currently available on the world market and this property cases weight gain.  A minor and almost certainly insignificant difference in α₂ potency is stated dogmatically, and without acceptable evidence, to explain the drugs supposed antidepressant effect.  No-one in the profession appears to have seriously questioned the evidence.  Indeed, calling the data presented ‘evidence’ unjustifiably aggrandizes it.  The full extent and nature of this distortion of the scientific process is revealed more fully in the analysis of the papers presented by Organon discussed below.  It is difficult to interpret the evidence and its presentation as anything other than deceit or incompetence, or a combination of the two.  It is nothing short of a travesty of scientific process.

A huge number of reprints from a supplement to the Journal of Clinical Psychiatry [26] (lucrative for the journal) tabulating data on mirtazapine were purchased at substantial expense, by whom I do not know (but I am definitely not awarding a prize for the best guess), and given to huge numbers of doctors all over the English speaking world. Evidence clearly demonstrates that supplements are of lower standards [27] which in my opinion applies in this case. Figure 1 in this contribution (p 560, ‘Receptor binding profile of mirtazapine’) displays a bar graph of receptor affinities.  This cites a single reference for the data in the graph but has no legend to indicate that it is in fact compiled from different experiments, in different species of animals, and that the work was spread over 6 years and published in several different journals. The table in the reference that is given [28] is itself not the original presentation of the data, it gives no original data either, but presents data ‘adapted’ from 3 further references [29-31]. Those three different references, for anyone who troubles to dig that deep, in turn reveal the use of various preparations including: rat cortex, vas deferens, mouse neuroblastoma, guinea pig ileum etc.  The basis for these ‘adaptions’ is not discussed or elaborated.

The original Figure 1 in the J Clin Psych supplement therefore produces profoundly misleading impressions because it makes invalid comparisons.  It would have been clear at that time that it was not valid or accurate to make such comparisons, between different species and different experiments in different laboratories. In my opinion, it is impossibly difficult to see how those who produced this table could be described as both competent and honest.  The earlier, and only, de Boer (Organon) paper that reported values for both drugs (mianserin and mirtazapine), apparently from the same experiments [32] found that the α₂ affinities were identical.  That paper is no longer referred to at all, and it does not appear in the citations of the J Clinical Psychiatry article [26].  There is no justification for choosing to ignore this contradictory data without comment and explanation.  This exercise is certainly not acceptable science, and at worst, it is disingenuous.  It might be noted that this egregious piece was published in J Clinical Psychiatry, which is regarded as a prestigious and high-profile journal.

My comprehensive analysis of more recent receptor data, concerning mianserin and mirtazapine, was published recently [22] and highlights the lack of significant differences between these two drugs.  It is important to appreciate that even using the latest human cloned receptor techniques the inter-laboratory variation in receptor affinities measured is of the order of five to 10-fold.  This makes it clear that the small differences selected from the early work on mirtazapine are exceedingly unlikely to be meaningful: to assign accuracy to them would be a triumph of hope over experience.

What is known, but not known by the people who need to know what is known, is often the most important thing we need to know.  And indeed, there is a particular piece of information that few, if any, researchers in psychiatry will have become aware of.  It is this: Organon synthesised and tested another drug with A2 antagonist properties (Org 6906) that had the potentially significant advantage of not being a potent histamine H1 receptor antagonist [29]. You might ask what happened to this drug, but we do not know.  Nothing further was published about it, at least nothing referenced in the NLM Pubmed data base.  Obviously, if it was trialed in humans, with negative results, this information would be significant.  But secrecy prevails and the advancement of knowledge counts for less.

In retrospect, it is apparent that not only has there been uncritical acceptance by academics of the evidence for the supposed action of mirtazapine, but also there has been widespread support by influential figures (dubbed ‘key opinion leaders’, KOLs in pharmaceutical company jargon), at least some of whom have been, and probably still are, in receipt of large sums of money [33]. In this context one further peculiarity concerning the publication of this report is notable: the reprint attributes authorship to Hirschfeld, RM. a member of the Department of a noted key opinion leader at Brown University, whereas the pub med database makes it an anonymous contribution http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=pubmed&dopt=Abstract&list_uids=8968311

The idea that mirtazapine is a dual action NaSSA (‘noradrenergic and specific serotonergic antidepressant’) in humans continues to be widely repeated or supported by Key opinion leaders (KOLs) [34-44].  It has no significant serotonergic effects in either humans [22] or animals [45], so this is clearly unlikely to be correct and is ‘spin doctor’ jargon, not science.

The idea that mirtazapine is a dual action NaSSA has Little or no scientific evidence to support it

The history of the investigations concerning mianserin and mirtazapine is one of a lack of scientific rigor in data presentation and a lack of objectivity in analysing results.  This may be accounted for by a predominance of commercial interest combined with lack of independent replication that has been enduringly counter-productive throughout the pharmaceutical industry.  The ultimate irony is that this is driven by secrecy and the belief that each individual company may make a significant and profitable advance before others.  The real outcome is revealed in the history of the last 50 years: a catalogue of complete and abject failures to find more effective antidepressants.  There has been no significant advance (in efficacy) since the discovery of tranylcypromine and clomipramine.

There are many lessons to be learned from this sorry saga.  It will be interesting to see what is revealed as parent company, Akzo-Nobel, sells off the subsidiary Organon, because mirtazapine is a substantial part of its revenue stream, and they have worked hard to maintain that stream.  Substantial damages have been paid by Organon following the USA court settlement about that.

Go to Piribo below to see ‘Maximizing Product Returns Through Reformulation, Old Molecules, New Opportunities’, and for a price, you too can learn some strategies to maintain profits from out of patent drugs!

Recent work establishes the association between serotonergic effects and elevation of serotonin in experiments with rats.  There is also good evidence to extrapolate that to ST in humans[46].  Increasing levels of serotonin are associated with increasing serotonin-mediated side effects, and as they become further elevated, with toxicity, culminating in hyperthermia and death.  The degree to which different drugs exhibit these effects in humans correlates well with their capacity to elevate serotonin levels in rats.

The clinical efficacy of particular drugs in specific conditions, e.g. serotonin reuptake inhibitors (SRIs) in obsessive compulsive disorder (OCD), has been used to correlate alterations in serotonin levels induced by amitriptyline and clomipramine (CMI) with clinical efficacy, and the ability to induce ST[22, 47].

Table 1

Pharmacological profile of mirtazapine and mianserin: uptake inhibition and receptor antagonism

Table notes

The above data is now in PDSP Ki data base. Lower numbers indicate greater potency. The first six rows contain three pairs of data (a [48], b [49], c [29]), each receptor measurements from the same laboratory. Two of these pairs (a, b) of data were discovered by me, and caused to to be entered into the PDSP database, because they were previously ‘unknown’, having been accidentally omitted from one publication (b) and in the other case published only as an abstract (a) and not noticed or quoted by anyone else.  The two notable points are: firstly, that the new human cloned receptor estimates of the relative adrenergic A2 potency of these drugs indicates that mianserin is actually the more potent as an A2 antagonist. This is precisely the opposite to what is required to justify the claim that they made that mirtazapine is better / different. Secondly, it highlights that there is a 5-10 fold variation in measurements between different techniques and laboratories which emphasises that the minor differences found by the pharmaceutical company Organon initially were much too small to justify any speculation that they were different.  Note that the’Organon’ results quoted above [29] are the ones that appear to have been forgotten about, and were not subsequently quoted or referred to, (e.g. other data showing a larger difference was used in the widely disseminated article in the Journal of Clinical Psychiatry).  In view of that, and for the accuracy of the record, I quote the words form their own paper: ‘The binding of [3H]rauwolscine to alpha 2-adrenoceptors was inhibited by (+/-)Org 3770 (i.e. mirtazapine) and mianserin with identical affinity, … . A prominent role for the blockade of alpha 2-adrenoceptors in the therapeutic effects of mianserin and (+/-)Org 3770 in depression is suggested, probably excluding a role of inhibition of the uptake of NA.’

In conclusion

ONE can state simply and confidently that there is neither pharmacological evidence that mirtazapine is likely to be an antidepressant, nor is there substantial clinical trial evidence either. Even more importantly, clinical experience is that it is ineffective as monotherapy for severe depression.

Addendum

One of the many confidence tricks I played during my career was about mirtazapine which I used to teach all the trainees was not an antidepressant (this is back in the 79 to 81 years). It was generally well known in the department that I said that and when someone knew was presenting a case and summarised that the patient had been taking an antidepressant for some weeks and was not improving I would ask what antidepressant. When the reply came back but it was mirtazapine I would respond by saying with an edge of irritation in my voice, no I asked what antidepressant this patient was all not what sedative they were on. Needless to say this caused repeated constellation for the junior staff who were subjected to this joke. I was conducting the professorial ward round shortly before I left, this was Christmas of 1982 and it was the last war round ward round of the year. The above scenario was repeated for the umpteen time and those familiar with the scene breast themselves for another outburst from me. On the spur of the moment I stopped looked at my watch ostentatiously and said and then gave the following account. Look it’s the end of the year and these results are going to be announced in the new year by the Department of health so I guess I can tell you all now. You know I was doing some important research in London before I came here and part of that was my involvement in the biggest and most confidential trial ever done so far in collaboration with the Department of health. It was precipitated by the recognition that it was so difficult to separate placebo effects from active effects in most of these drugs that the only way to solve this problem was to conduct the ultimate placebo trial. Although there was vigourous an extended debate about the ethics of this it was finally agreed the only solution to this problem was to introduce a new drug that was in fact a placebo, not a new drug at all.  That is what mirtazapine is.  The results are going to be announced after the New Year at a special meeting of the Royal College and the department… I would ask you all to say nothing about this until the New Year.

As I left the department after this one the two people followed me exclaiming how extraordinary this was and how exciting it was and…

When they found out they had been comprehensively conned I recognised that it was probably good that I was about to leave!  I remember one or two people who took it very badly. This just illustrates how people get emotionally attached to their intellectual ideas and disproving their thesis is interpreted as a personal affront.