The adverse (or side) effects of antidepressants include numerous physical and psychological symptoms and for some patients the burden of these effects may exceed the benefit of medication. One naturalistic study looking at adverse effects in 1,000 patients with a median duration of antidepressant use of one year found that two-thirds reported at least one side effect, with a third having three or more, and with risk of side effects increasing with each year of use (Table 2.2 ). 1
Table 2.2 Adverse Effects of Antidepressants Reported by 1,000 Patients in Primary- or Secondary-Care Settings
| Type of antidepressant | SSRIs(%) | TCA(%) | Venlafaxine(%) | Mirtazapine(%) |
|---|---|---|---|---|
| Type of adverse effect ascribed to medication | ||||
| Sleeplessness | 7 | 5 | 10 | 5 |
| Sleepiness during the day | 21 | 14 | 20 | 30 |
| Restlessness | 9 | 6 | 10 | 12 |
| Muscle spasms, twitching | 9 | 12 | 15 | 7 |
| Dry mouth | 22 | 49 | 23 | 22 |
| Profuse sweating | 20 | 20 | 32 | 14 |
| Sexual dysfunction | 23 | 20 | 31 | 10 |
| Nausea | 10 | 4 | 9 | 5 |
| Constipation | 8 | 20 | 10 | 2 |
| Diarrhoea | 7 | 4 | 5 | 5 |
| Weight gain | 19 | 22 | 17 | 29 |
| Dizziness | 12 | 11 | 19 | 12 |
In surveys of a convenience sample of people who were on antidepressants for a longer period (most for more than three years), and who may not be representative of all antidepressant users, rates of adverse effects were even higher:
Emotional blunting seems to be a common, and dose-dependent, consequence of antidepressant use. 3, 4 Although this effect has sometimes been attributed to the underlying condition a double-blind placebo-controlled study in healthy volunteers found clear evidence of emotional blunting in the patients administered an SSRI for three weeks. 5 One survey found about 46% of patients on antidepressants reported emotional blunting. 6 Some patients report that emotional blunting (which indicates that both positive and negative emotions are blunted) can have detrimental effects on their well-being and relationships. 3 Some observers have suggested that use of antidepressants might undermine a person's autonomy and resilience, increasing their dependence on medical help. 8
Long-term use of antidepressants may cause a greater degree of weight gain than that suggested in short-term trials. In one case-control observational study with almost 2 million patient years of follow-up, with patients taking SSRIs, SNRIs and other commonly used antidepressants such as mirtazapine and tricyclics there was a 30% increased risk of people of normal weight becoming overweight or obese in 10 years of follow-up, compared to people not taking antidepressants. 9 There was also a 30% increased chance of overweight people taking antidepressants becoming obese in 10 years compared with overweight people not taking antidepressants. 9 It is possible that residual confounding might contribute to these associations. The effects were most marked for mirtazapine (50% increased risk of greater than 5% weight gain). 9
Meta-analysis has also found that some antidepressants produce cognitive impairment in healthy controls, on tests of information processing, memory, hand-eye co-ordination, concentration, as well as higher order functions. 10 There was variation between different antidepressants with SSRIs producing between 1% and 16% impairments (where proportions referred to the number of test points where impairment was found), while venlafaxine produced 9% impairment, mirtazapine produced 35% impairment, and older tricyclics producing between 19% and 47% impairment (highest for amitriptyline). 10 These studies are useful in that they exclude confounding by an underlying disorder by studying the effects of antidepressants in healthy controls. Small studies find that MMSE scores (a crude measure of cognition that detects coarse changes in cognitive ability) decreased over consecutive weeks of follow-up in people with OCD given antidepressants. 11 The long-term consequences of these cognitive impairments have not been investigated or quantified.
For older people adverse effects can be more overt. A retrospective cohort study of over 61,000 patients found the following absolute increased risks over 1 year of exposure to SSRIs compared with not being on an antidepressant (adjusted for comorbidities and a range of potential confounding variables):
Absolute risk over 1 year for all-cause mortality was 7.04% for patients not taking antidepressants, 8.12% for those taking TCAs, 10.61% for SSRIs and 11.43% for other antidepressants. 12 This observational research is susceptible to confounding by indication, and residual confounding, so differences in characteristics between patients prescribed different antidepressants could account for some of the associations between them and the adverse outcomes. 12
Most antidepressants have been associated with hyponatraemia, likely due to the syndrome of inappropriate secretion of antidiuretic hormone (SIADH), generally with onset within 30 days of starting treatment. 13 Risk of hospitalisation with hyponatraemia is elevated from 1 in 1,600 in the general population to 1 in 300 for those starting an antidepressant, 14 and is associated with increased mortality (at any severity). 15
There exist no long-term placebo-controlled trials of antidepressants in order to assess their long-term health effects, so observational studies comparing people on and not on antidepressants are the only source of evidence we have for determining long-term health risks, with the issues of confounding such studies entail. One study found that people on SSRIs had an increased hazard ratio of 1.34 at 10 years for risk of cardiovascular diseases, a hazard ratio of 1.73 for mortality due to cardiovascular disease and a hazard ratio of 1.73 for all-cause mortality, and that the effects were dose-dependent. 16 Other antidepressants increased the risk of coronary heart disease by a hazard ratio of 1.99, cardiovascular disease in general of 1.99 and all-cause mortality of 2.20, with some evidence of a dose-response relationship. 16 Although a wide range of confounders were adjusted for, residual confounding cannot be ruled out, especially as baseline depression scores were not available.
There is also evidence that antidepressants may increase risk of dementia. A large nested case-control study of 225,000 people found a dose-response relationship between total exposure to antidepressants and risk of diagnosis with dementia. 17 Those patients with the highest exposure to antidepressants - more than 3 years of daily use of standard doses - had a 34% increased chance of dementia over those patients not exposed at all to antidepressants. Another nested case-control study of 40,000 people found similar results, with antidepressants with the strongest anticholinergic properties (amitriptyline, dosulepin and paroxetine) producing a 10% increased risk of dementia. 18 Other antidepressants (largely SSRIs), with lesser or no anticholinergic effects were also associated with dementia but associations were greater for prescriptions closer to dementia incidence suggesting reverse causation as a possible explanation. 18 Although efforts were taken in these studies to control for confounding variables, there is the possibility that residual confounding may explain some of these associations.
SSRIs and SNRIs inhibit the uptake of serotonin into platelets. Serotonin plays a role in the response to vascular injury, promoting vasoconstriction and morphological changes in platelets leading to aggregation. 19 Depletion of platelet serotonin leads to a reduced ability to form clots and therefore an increase in the risk of bleeding. The relative risk of any bleeding event on an SSRI/SNRI compared with no use is 1.4, with the absolute risk being between 0.5% and 6% (depending on a variety of factors, including length of treatment). 20 SSRIs also increase gastric acid secretion and may therefore increase the risk of peptic ulcer. 21
A population-based study found that SSRIs increase the rate of upper gastrointestinal bleeds (UGIB) with a hazard ratio of 1.97, and low gastrointestinal bleeds (LGIB) with a hazard ratio of 2.96, after adjusting for all relevant risk factors. 22 A meta-analysis of 22 studies concluded that current users of SSRIs are 55% more at risk of UGIB compared with those who do not take SSRIs. 23 The absolute risk of an UGIB while on long-term SSRIs is 2% and of an LGIB it is 1%. 22, 24
Risk of intracranial haemorrhage is increased in people on SSRIs - with an absolute annual risk of 0.42% in older patients. 25 The prevalence of menstrual disorder (unusual or excessive bleeding, irregular menstruation, menorrhagia, etc.) is doubled in women on antidepressants compared with women not taking antidepressants (24.6% vs 12.2%). 26 In addition, the absolute risk of post-partum haemorrhage was 18% for women using SSRIs, compared to 8.7% for women not using antidepressants. 27 The MHRA in the UK issued a warning regarding the use of SSRIs and post-partum blood loss in 2021. 28 Use of SSRIs in the pre-operative period is associated with a 20% increase in inpatient mortality (absolute risk 1 in 1,000). 29 In coronary artery bypass graft (CABG) procedures, there is a 50% increase risk of mortality in users of serotonergic antidepressants compared with non-users. 30
Withdrawal effects from antidepressants occur commonly and can be severe in some people. 31 The likelihood and severity seem to increase with longer term use. 32 This may be a sufficient reason on its own to limit the duration of antidepressant use. Withdrawal symptoms are discussed in detail in subsequent sections.
Withdrawal effects increase in both likelihood and severity with longer term use, probably because the brain and body adapt to a greater degree after longer exposure. 32 For example, in survey data, after 3 months of use, about 25% of patients reported withdrawal effects on stopping, with about 20% of patients reporting moderately severe or severe withdrawal effects. After 3 years of use, more than 60% of patients reported withdrawal effects on stopping, with about 50% of patients reporting moderately severe or severe withdrawal effects. 32 This may factor in to the decision to stop antidepressants sooner rather than later, in order to avoid more severe withdrawal effects.
Sexual adverse effects can include a lack of desire as well as reduced sexual sensation, and failure to orgasm in both genders, 33 and occur in 25% to 80% of patients. 34 It is now recognised that these sexual effects can persist even after cessation of antidepressants in a minority of patients. This is called post-SSRI sexual dysfunction (PSSD), and was recently recognised by the European Medicines Agency. 35, 36 Sexual side effects can negatively affect a person's self-esteem, quality of life and relationships.
Concerns have been raised that long-term use of antidepressants can itself induce dysphoria. 37, 38 This has been thought related to the process of tolerance to these medications, involving serotonin receptor desensitisation, which can 'overshoot' leading to opposite effects to those originally produced by the medications. 37 This has been seen as analogous to opioid-induced hyperalgesia 39 and the increase in anxiety seen in long-term use of benzodiazepines. 40 For example, one observational study found that depressed people who used antidepressants long-term had poorer long-term outcomes compared with non-users or those who used them short-term, even after controlling for baseline depressive severity. 41