Abstract
BACKGROUND:
In the case of depression (and other psychiatric disorders), a huge number of scientists have been trying for decades to establish and postulate the disease-based drug concept for antidepressants and these efforts have not born fruits.
OBJECTIVE:
To show that the discussions about statistical significance of the efficiency of the antidepressants over placebo are non-productive and irrelevant.
METHOD:
Researching the history of the onset of a chemical imbalance theory in the brain and the launch of antidepressants according to the drug-based disease model was the basis for doubting the appropriateness of using antidepressants in the treatment of depression.
RESULTS:
Antidepressants (AD), as their name suggests, are supposed to address a disease-specific model to reverse the neuropathological basis of depression. However, science has not constituted the theory, using various brain imagining techniques or measuring the concentration of serotonin in various body liquids, that depression is characterized by the lack of serotonin in synapses and then created a substance that would block the reuptake of this neurotransmitter or in some other ways elevate its content.
CONCLUSION:
Even the reductionistic phase of the AD story, the mechanism of action, failed to be consistent so it is unclear how the largest meta-analysis of about 21 antidepressants by Andrea Cipriani et al. appeared in The Lancet.
Introduction
Clinical neuro-pharmacologist and retired clinical psychiatrist Ken Gillman wrote: “…caveats and limitations rightly expressed by the authors about the validity and generalisability of their data and results (published in the largest meta-analysis of about 21 antidepressants by Andrea Cipriani et al. [1]) will soon, as usual, be forgotten in the oversimplified implementation resulting from insufficiently informed or critical use of ‘guidelines”’ [2].
In the world in which Max Weber’s idea of instrumental rationality has been turned upside down: “rather than in purposes seeking the most effective means, it now consists in the means that seeks, and more often than not finds, suitable applications” [3]. We can easily read this thought as: rather than in disease seeking the most effective remedy, it now consists in the remedy that seeks (and often finds) a suitable disease. It is possible to imagine this idea as a groundwork for creating a syntagma of a drug-based disease [4]. First SSRI AD fluoxetine was launched in 1987, more than two decades after the monoamine hypothesis of depression had been put into words. Namely, in the late 1960s, a post-mortem study showed a diminished level of serotonin in depressive suicides [5] but the idea that serotonin levels might be lowered in depression was omitted [6] and tricyclic antidepressants continue to be a drug of choice for treating severe depression. When in 1980s benzodiazepines as addictive drugs were attempted to be supplanted by the serotonergic drug buspirone, this idea flopped. Instead, SSRIs were marketed for treating depression despite being weaker than older tricyclic ADs. That strategy was a great success, central to which was the belief that SSRIs restore the serotonin level in synapses to normal. This notion later metamorphosed into the definition that SSRIs cure chemical imbalance in the brain.
Birth of mythology
Fluoxetine was created and established to block the serotonin transporter and presynaptic reuptake and therefore elevate serotonin content in synapses [7]. Accordingly, the drug-centred model was created: depression is a lack of serotonin in synapses. The disease-centred model advocates that antidepressants work because they reverse the processes in the disease that give rise to the appearance of depression-characteristic symptoms. Until now, nobody has proven this model [8]. On the other hand, “the ‘drug-centred’ model suggests that far from correcting an abnormal state, as the disease model suggests, psychiatric drugs induce an abnormal or altered state. Psychiatric drugs are psychoactive substances, like alcohol and heroin. Psychoactive substances modify the way the brain functions and by doing so produce alterations in thinking, feeling and behaviour. The drug-centred model suggests that the psychoactive effects produced by some drugs can be useful therapeutically in some situations. They do not do this in the way the disease-centred model suggests by normalising brain function. They do it by creating an abnormal or altered brain state that suppresses or replaces the manifestations of mental and behavioural problems” [9].
The ‘drug-centred’ story of depression began in the 1950s with a doubtful experiment with reserpine. It postulated the paradigm that the decreased level of serotonin was the cause of depression disorder. The theory still persists even though a single proof to support the theory has not yet been found. Its life will also be additionally preserved by the article published in The Lancet in February 2018. Immediately after the publication of the largest meta-analysis, The Times in the article entitled “More people should get pills to beat depression” wrote for the Sunday Times: “Millions of sufferers would benefit, doctors told. More than a million extra people should be offered antidepressants after the largest study of its kind concluded that they worked, experts have said. Patients and GPs must not be squeamish about treating mental health problems with drugs, according to scientists who found that every one of 21 common antidepressants was better than a placebo”.
The National Health Service (www.nhs.uk) under the title “Big new study confirms antidepressants work better than placebo” wrote: “Antidepressants are highly effective and should be prescribed to millions more people with mental health problems, researchers declared last night”, reports the Mail Online and continues the article in a more sober way: “However, that does not mean they are “highly effective” – it means people are more likely to see their symptoms improve if they take an antidepressant than if they take a placebo”.
Despite the fact that Cipriani et al. wrote that summary effect sizes were “mostly modest” [1], some popular media concluded the article had proven that antidepressants work, any debates over these drugs are now over and declared that “happy pills” are here again. The soundbite “antidepressants work” will remain in public mind and, unfortunately, even in the minds of psychiatrists and GPs. Nevertheless, the defined daily doses are rising and in the USA, UK, Iceland, Australia and Portugal are over 100 per 1000 inhabitants. Most people are put on these drugs in primary care, after a very brief visit and without clear symptoms of clinical depression. Usually there is some improvement, and it is often based on the passage of time or placebo effect [10,11]. The value of Cipriani’s study must be questioned again for many reasons. Firstly, the drawback of randomized controlled trials is the selection of population by a contract research organisation which provides high responders, who are additionally selected by the run-in phase of trials. There is only a few ten percent match between participants in trials and patients with a prescription for AD in their hands. Secondly, the duration of trials also does not correspond to the duration of administrating the drug. Whereas in 522 trials in the article the exposure lasts 4–8 weeks, less than 6% of the population takes AD for 8 weeks or less and 26% for 2 months-2 years, 24% 2–5 years, 19% 5–10 years and 25% takes AD for more than 10 years respectively (in the period 2011–2014) [12]. Side effects and risk factors of long term use are an important drawback and are rarely mentioned. Some lesser known effects are depicted below.
Dementia
Depression and dementia often exist together. Both are neurocognitive disorders portrayed by cognitive impairment [13]. The early-life depression is a risk factor for depression. There is increasing evidence that early-life depression is a risk factor for developing dementia in adulthood, and depression could be a consequence or symptom of dementia [14]. Despite the close similarity between dementia and depression, there is an unclear connection between treatment with ADs and dementia. But a study by Chee-Kin Then et al. [15] suggested that dementia might be the iatrogenic effect of antidepressants use. Since the hazard ratio (HR) of depression per se for dementia is 1.59, the HR for depression patients on antidepressants is 2.42 and 4.05 for non-depressive patients on antidepressants. Since antidepressants are widely prescribed besides depression also for off-label uses such as back pain [16], insomnia [17] or migraine [18], there is a considerable risk for developing dementia for these users.
Weight gain and type 2 diabetes mellitus
There is a potential weight gain in the population during widespread utilisation of AD [19]. An average self-reported weight gain, based on repeated surveys, was 1.4 kg (2.5%) among nonusers and 2.5 kg (4.3%) among users of ≥200 defined daily doses (P trend < 0.0001). In these data, continuous use of antidepressant medication was associated with an increased relative risk of type 2 diabetes, although the elevation in absolute risk was modest [20]. The use of SSRI was also connected to decreased secretion of insulin in nondiabetic participants and an elevated risk of insulin dependence in type 2 diabetics in older population [21]. Due to the high prevalence of type 2 diabetes and AD use it is important to be cautious when prescribing these agents to patients at risk for diabetes mellitus [22].
Increased suicide rates
A recent paper by Khan et al. [23] reported a decreased rate of suicide attempts. But that study used flawed statistical analysis and was inappropriate to study the problem. It is known that the suicidal rate is highest during the first weeks of acute treatment [24–27], whereas Khan’s study was based on patient exposure years. As Hengartner and Plöderl suggest [28], suicide risk is elevated during the first weeks after treatment initiation and is close to zero thereafter, not cumulatively or over a long period of time. In almost all of the 522 papers ranking from 1979 to 2016 individuals with suicidal ideation or those who collected more than 5 points from item 10 of MADRS were excluded. Participants in those trials were a highly selected group and by no means corresponded to the population who were administered ADs. It is meaningful to ask how GPs, who prescribe most ADs, can evaluate a suicidal risk of every patient after a very short examination time on average. Do they also exclude those with suicidal potential?
Akathisia
It is known that akathisia, a very distressing condition, is associated with an increased risk of impulsive behaviour and suicidal ideation [29,30]. Akathisia may be caused by various medications, including antipsychotics, which are the main causative agents and are present in 45% of the patients using first generation antipsychotics [31], antiemetics and antidepressants (the authors rated 55% of the cases as being “probably” related to the antidepressant in question [32]). Suicide and homicide can be generated by people experiencing akathisia, when SSRI type of antidepressants were given to healthy adults with no prior history of a mental condition [33]. Most (99.8%) of akathisia subjects using antidepressants have at least one variant allele for the three cytochrome P450 genes: CYPs 2C9, 2C19 and 2D6, whereas there are 81% with at least one variant allele between randomly selected primary care patients. The odds ratio in this collation is 19.3 [34]. Since 9–10% of the general population are 2D6 poor metabolizers [35], one-third of the individuals express heterozygous genotype CYP2C9*1/*2 or CYP2C9*1/*3 and are regarded as poor metabolizers [36] and 2C19 poor metabolizers are more spread among people with a Middle East ethnical backgrounds (13–16%) than in Caucasians (1–2%), who have a high risk for developing akathisia, which is a risk factor for suicide and violent or aggressive behaviour and akathisia-related homicides [17]. The need for identifying and managing akathisia thus cannot be overemphasized.
Pregnancy
As study from Nordic countries suggests that approximately 2–8% of pregnant women receive antidepressants [37] and its use has been constantly rising in the last few decades [38]. Comparisons of adjusted hazard ratios for specific psychiatric disorders between mothers who used antidepressants before and during pregnancy (continuation group) and mothers who discontinued with antidepressants just before pregnancy (discontinuation group, hazard ratio referenced as = 1) reveal the highest hazard ratio for mood disorder (hazard ratio = 2.76) and for neurotic, stress-related and somatoform disorder (1.62). Hazard ratios for any psychiatric disorder were highest in the antidepressant continuation group. In conclusion, the in utero exposure to antidepressants was associated with an increased risk of psychiatric disorders [39].
Car accidents
At this point car accidents should also be mentioned. Epidemiological studies have shown a positive correlation between AD exposure and elevated traffic accident risk. A study in the Netherlands shows that in different psychotropic medicine group users, SSRIs exhibit the largest odds ratio (OR), namely 2.03. This unexpectedly high OR can be partially a consequence of intentional car accidents and are therefore associated with the risk of suicide [40]. Similar results were obtained in a Taiwanese study [41].
Conclusion
Given these circumstances, it is obvious that debates around effectiveness and odds ratios in antidepressants therapy, considering the methodological flaws, reanalysis of a network meta-analysis, which are the main themes in articles that put under question Cipriani’s article, are irrelevant. Even the name itself, namely ‘antidepressants’, is wrong because it suggests that these medications correct the biological basis of depressive disorder, which is far from the truth. We can speculate that not only akathisia, which is connected to CYP P450 poor metabolizers, but also other side effects are a consequence of high concentrations of antidepressants in blood due to slow metabolism with cytochrome P450 enzymes. Additionally, can we even speculate that fast metabolizers have the lowest antidepressants concentrations in blood and therefore the weakest side effects and consequently the highest decrease on the Hamilton Depression Rating Scale?
Finally, Eric Nestler, the Dean for the Academic and Scientific Affairs and Director of the Friedman Brain Institute at the Icahn School of Medicine at Mount Sinai in New York, wrote: “…after more than a decade of PET studies (positioned aptly to quantitatively measure receptor and transporter numbers and occupancy), monoamine depletion studies (which transiently and experimentally reduce brain monoamine levels), and genetic association analyses examining polymorphisms in monoaminergic genes, there is little evidence to implicate true deficits in serotonergic, noradrenergic, or dopaminergic neurotransmission in the pathophysiology of depression. This is not surprising, as there is no a priori reason that the mechanism of action of a treatment is the opposite of disease pathophysiology” [42]. Furthermore, regarding devastating side effects and despite many biases, which raise the effectiveness of these drugs, the odds ratios are still only statistically significant, the risk/benefit ratio is far on the risk side. Therefore we can say that prescribing antidepressants to patients is not in line with the Hippocratic Oath and Cipriani’s article should be retracted.
Footnotes
Conflict of interest
None to report.