Certain antidepressants become associated with particular adverse effects, sometimes without justification. An example is the TCA lofepramine, which became linked to potential hepatotoxicity during the first few years of its launch; the first reported case was in October 1987 (Macphee et al. 1987) and a further 57 reports had been received by the UK Committee on Safety of Medicines 2 months later. A case report featuring withdrawal of lofepramine and exposure to the metabolite desipramine implicated the ‘parent drug’ (i.e. lofepramine) as being responsible for hepatotoxicity (Lack et al. 1990). However, a subsequent case series involving elderly patients revealed that the effects of lofepramine on hepatic function were only transient (Kelly et al. 1993), and a further investigation demonstrated that there was no difference between lofepramine and placebo in the incidence of abnormal liver function tests (Tan et al. 1994). No further published reports about hepatic dysfunction with the drug have appeared.

There is much variability in the quality of case reporting of adverse drug reactions; a systematic review of all case reports published in a single psychopharmacology journal over 25 years found that only 7 out of 40 reports of presumed adverse reactions were deemed sufficiently robust for ascribing possible or probable causality (with no correlation between the quality of the report and its impact as assessed by citations) (Talat et al. 2013). Similarly, presumed hepatic reactions to SNRI antidepressants (primarily duloxetine and venlafaxine) have been reported 17 times between 1999 and 2012; but only two reports involved re-exposure, only two investigated possible ‘dose-response’ relationships, and only three confirmed presumed hepatic reactions with a biopsy (Pradeep et al. 2004). High likely causality scores were only seen in reports of acute hepatitis in patients treated with venlafaxine (Pradeep et al. 2004), of asymptomatic transaminitis in a patient with known Gilbert’s syndrome treated with venlafaxine (Phillips et al. 2006), and of acute hepatitis attributed to venlafaxine in a patient who was also taking five other medications, and who had recently stopped a phytomedicine (Feinberg 2010).

Findings from clinical trials and pharmacoepidemiological studies of duloxetine suggest that the incidence of elevated liver function test results in patients without hepatic disease is generally low (being somewhat higher in patients with known liver pathology) and not significantly greater than when treated with placebo (Wernicke et al. 2008a). In a large pharmacoepidemiological investigation based on 1.55 million years of patient exposure, there were 406 reported cases of duloxetine-associated hepatic problems, equating to a cumulative reporting rate of 0.008 % for all hepatic events and a risk of severe hepatic injury in 0.7/100,000 exposed person-years (Wernicke et al. 2008b). However, independent analysis of the FDA adverse event reporting system, and of the i3 Aperio (health insurance) database, suggests that the reporting of hepatic events with duloxetine may be disproportionately greater, when compared to other antidepressants (Strombom et al. 2008).

The novel antidepressant agomelatine has been shown to be hepato-protective against paracetamol-induced liver damage in rats (Karakus et al. 2013). However in human studies, elevated AST and/or ALT levels were seen in 1.4 % of patients taking 25 mg per day, and 2.5 % of patients prescribed 50 mg per day. Agomelatine is therefore contraindicated in patients with pre-existing hepatic disease, and regular monitoring (at baseline and at weeks 3, 12, and 24 of treatment) of hepatic function is required (McAllister-Williams et al. 2010). One case of fulminant hepatic failure in a patient with pre-existing fatty liver disease has been described (Gruz et al. 2014), and the German regulatory agency BfArm has received 58 cases of ‘hepatotoxic adverse drug reactions’ (Gahr et al. 2013), emphasising the need for further research into the effects of agomelatine on hepatic function.

Prenatal exposure to antidepressant drugs has been suggested to be associated with a number of adverse effects in the newborn infant, and several variables need to be considered when deciding whether or not to treat a depressed pregnant woman. One potential adverse effect that should be considered is persistent pulmonary hypertension of the newborn (Kieler et al. 2012), (see Table 8.3 and Table 8.4) a relatively rare condition (with an estimated prevalence of 1.9 per 1,000 live births) (Walsh-Sukys et al. 2000), in which pulmonary vasculature in the infant fails to ‘relax’ resulting in poor oxygenation (Grigoriadis et al. 2014). Symptoms range in severity from mild respiratory distress to hypoxia which requires urgent intensive medical care (Jong et al. 2012). Study findings suggesting an association between maternal use of SSRIs during pregnancy and persistent pulmonary hypertension in the newborn led both Health Canada and the FDA to issue advice to clinicians alerting them to the potential adverse effect (FDA US 2012). However more recent findings have been inconsistent (Occhiogrosso et al. 2012), with some studies reporting no association (Andrade et al. 2009; Wilson et al. 2011), others some association (Kieler et al. 2012; Chambers et al. 2006), and others indicating differential effects depending on the stage of the pregnancy when exposure occurred (Kieler et al. 2012; Chambers et al. 2006) (See Table 8.2). The FDA has stated, ‘given the conflicting results from different studies, it is premature to reach any conclusion about a possible link between SSRI use in pregnancy’ and persistent pulmonary hypertension in the newborn (FDA 2012).

Risk factors for persistent pulmonary hypertension (Grigoriadis et al. 2014) include certain congenital malformations, premature birth, meconium aspiration, maternal obesity, and caesarean section delivery (Occhiogrosso et al. 2012; Koren and Nordeng 2012a, b), but many of the studies of SSRIs did not exclude or control for other known risk factors and more than one risk factor may be a prerequisite for developing the condition (Occhiogrosso et al. 2012; Galbally et al. 2012). In a systematic review and meta-analysis of the associations between SSRI treatment and persistent pulmonary hypertension, Grigoriadis and colleagues found an increased risk if exposure to SSRIs occurred in late (but not early) pregnancy: study design, congenital malformation, and meconium aspiration were not significant effect modifiers, but the possible moderating effects of caesarean section, preterm birth, and maternal obesity could not be examined (Grigoriadis et al. 2014). There was a statistically significantly pooled odds ratio of 2.5 for exposure to SSRIs in late pregnancy, though the absolute risk is low; between 286 and 351 women in late gestation would have to be treated with an SSRI for there to be one additional case of persistent pulmonary hypertension in a new born baby. Therefore the meta-analysis indicates that less than 1 infant in 100 will develop persistent pulmonary hypertension following prenatal exposure to SSRIs. Although this is a serious condition with death rates between 5 % and 10 % when associated with other conditions (such as congenital malformations, meconium aspiration, sepsis, and idiopathic disease), it can be managed successfully (Koran and Nodding 2012). The mortality of infants with persistent pulmonary hypertension who have been exposed to SSRI is not established, though one study suggests 9.1 % (3 out of 33 infants) of the infants who were exposed to an SSRI died, whereas 9.5 % (183 of 1,935 infants) who were not exposed to an SSRI died (Kieler et al. 2012); however the disparity in group size makes this finding hard to interpret.

Depression is common in older people with around 10–15 % of those living in the community being affected by depressive symptoms (McDougall et al. 2007; Beekman et al. 1999). Adverse drug events are more common in elderly patients, due to higher rates of comorbid illness, age-related physiological changes, and polypharmacy (Cadieux 1999). Despite this they are underrepresented in randomised controlled trials, as these typically exclude older people and those with comorbid conditions (Giron et al. 2005). Several observational studies have investigated adverse outcome associated with antidepressants (Reid and Barbui 2010), though few have been dedicated specifically to an older population.

This paucity of data led Coupland and colleagues to undertake a large cohort study of antidepressant use in older people (aged 65 years and older), involving a total of 60,746 UK patients. The most commonly prescribed antidepressant drugs were SSRIs (54.7 % of prescriptions), followed by TCAs (31.6 % of prescriptions), other antidepressants (13.5 %), and MAOIs (0.2 %) (Coupland et al. 2011); the latter were excluded from analyses due to the small number of prescriptions. All classes of antidepressant were associated with significantly increased risks of all-cause mortality, nonfatal self-harm, falls, fractures, and upper gastrointestinal bleeding, compared with patient groups not taking antidepressants: SSRIs and the group of ‘other’ antidepressants were associated with greater risks of stroke and/or transient ischaemic attack and of epilepsy/seizures, and SSRIs were also associated with increased risk of myocardial infarction and hyponatraemia (Coupland et al. 2011) (see Tables 8.1 and 8.2).

Associations with seven adverse outcomes differed significantly between drug classes. SSRIs were associated with significantly higher rates of all-cause mortality, stroke/transient ischaemic attack, falls, fracture, epilepsy/seizures, and hyponatraemia. No significant difference was found for nonfatal self-harm. The group of ‘other’ antidepressants had significantly higher rates for all-cause mortality, nonfatal self-harm, stroke/transient ischaemic attack, fracture, and epilepsy/seizures when compared with TCAs. No significant difference was found for falls or hyponatraemia. TCAs prescriptions were not associated with significantly higher rates of adverse outcomes than either SSRIs or the group of ‘other’ antidepressants.