Depression and cardiovascular disease (CVD) may become the two biggest disease burdens by 2020 and their association has attracted significant research interest. Healthy cardiac activity involves a high degree of beat-to-beat variability, mediated by the parasympathetic (vagus nerves) and sympathetic nervous system that slows and accelerates heart rate, respectively. Reduced heart rate variability (HRV), characterised by reduced vagal and increased sympathetic influences to the heart, is a robust predictor of fatal or near-fatal cardiac arrhythmia, even in relatively healthy individuals without a prior history of CVD (Reference Dekker1). A random sample of 900 subjects from a population study of 14 672 men and women without coronary heart disease (CHD) (Reference Dekker1) reported that low HRV measured from a 2-min electrocardiogram (ECG) was associated with an adverse cardiovascular risk profile and elevated risk of incident CHD and death from several causes 6 years later. For instance, relative risks of CVD death in participants with measures of low HRV ranged from 1.98 to 3.44, suggesting that reduced HRV may be a sensitive indicator of future mortality that may precede the manifestation of CVD.

Increased heart rate variability may help to foster calm behavioural states and facilitate social engagement (Reference Porges2). HRV is an index of beat-to-beat changes in heart rate and may reflect an individual's capacity to engage and interact with their environment. Although research has traditionally focused on paired antagonism between parasympathetic versus sympathetic innervation to the heart, the polyvagal theory (Reference Porges2) distinguishes between the myelinated vagus which facilitates facial expression, vocalisation and listening (i.e. social communication) from the unmyelinated vagus which is linked to immobilisation behaviours including feigning death, bradycardia, vasovagal syncope and behavioural shutdown. This theory highlights a bidirectional relationship between spontaneous social engagement and bodily states that is mediated by the myelinated vagus. Patients with depression display significant social impairments including lack of facial expressiveness and deficits in emotional prosody. Patients may also display reduced HRV, although many studies have focused on patients with CVD (Reference Kemp, Quintana, Gray, Felmingham, Brown and Gatt3). The polyvagal theory (Reference Porges2) suggests that these social impairments and autonomic dysregulation such as reduced heart rate variability may be linked. (See (Reference Brown, Barton and Lambert4) for a more general review of the association between depression and abnormalities in the cardiovascular system.) It is therefore possible that successful treatment of depression may also alleviate cardiovascular impairment.

A recent meta-analysis examined the impact of depression and a variety of antidepressant treatments on HRV (Reference Kemp, Quintana, Gray, Felmingham, Brown and Gatt3). Depressed patients have reduced HRV – this was particularly apparent when using nonlinear measures of HRV – and further HRV decreases with increasing depression severity. Unlike the majority of prior studies that had been conducted in cardiac patients, this study focused on patients without CVD, highlighting the need for clinicians to be mindful of the cardiovascular system in patients without such history. It was also found that while tricyclic antidepressants (TCAs) decreased HRV, serotonin-reuptake inhibitors (SSRIs) and other antidepressants including mirtazapine and nefazodone had no significant impact despite patient response to treatment. Although these findings re-inforce the need for caution over prescribing TCAs, particularly in patients with CVD, they indicate that SSRIs have relatively benign HRV-mediated cardiovascular effects (Reference Reid and Barbui5).

These findings are further supported by a recent prospective study (Reference Hamer, Batty, Seldenrijk and Kivimaki6) on 14 784 community-dwelling adults without a history of CVD, which reported an elevated risk of fatal and non-fatal CVD events (CVD events included CVD death, non-fatal myocardial infarction, coronary surgical procedures, stroke and heart failure) in users of tricyclic (TCA, n = 324) but not SSRI medications (n = 299) over an 8-year follow-up. The authors reported a 35% increased risk in CVD events in TCA users after accounting for a variety of covariates, including depression and anxiety. Other covariates included age, sex, socioeconomic group, marital status, physical activity, smoking, alcohol, body mass index, CVD medication and hypertension. The adverse effects of TCAs probably relate to increased anticholinergic effects at the sinoatrial node of the heart, involving blockade of the muscarinic acetylcholine receptors. Although tricyclic antidepressants are associated with greater affinity to muscarinic receptors, SSRIs are more benign, with the exception of paroxetine, which has considerable antimuscarinic potency (Reference Richelson7). Blockade of these receptors leads to decreased parasympathetic activity, disinhibition of sympathetic nervous system activity, tachycardia (rapid heart rate) and reduced HRV.

However, others have recently reported that while HRV is reduced in depression, these effects are primarily driven by antidepressants – including TCAs, serotonin and noradrenaline reuptake inhibitors (SNRIs) and SSRIs – and that autonomic function may recover after their discontinuation (Reference Licht, Geus, Dyck and Penninx8). These authors reported that new users of TCAs (n = 12) displayed a decrease in respiratory sinus arrhythmia (RSA; an indicator of HRV that controls for respiration rate) of 23 ms over the 2-year assessment period; new users of SNRIs (n = 23) displayed a decrease of 12 ms and new users of SSRIs (n = 74), a decrease of 7 ms, compared to persistent non-users of medication. This study gives reason for concern over the impact a variety of antidepressants may have on the cardiovascular system, with the authors suggesting that SSRIs may decrease net vagal effects through various relay nuclei of the parasympathetic nervous system, while the antivagal effects of TCAs and SNRIs may be located in the heart itself. It is important to note, however, that this study (Reference Licht, Geus, Dyck and Penninx8) had a number of limitations highlighting the need for further research. Although this study was characterised by a large sample size (n = 2114), specific antidepressant groupings were characterised by much smaller sample sizes as noted above. Other limitations included recruitment of patients with a prior history of CVD, failure to examine the impact of a variety of confounds on specific antidepressant classes, collapsing across multiple SSRI medications some of which (e.g. paroxetine) may have stronger anticholinergic effects and inappropriate use of analysis of covariance (Reference Kemp, Quintana and Malhi9).

In conclusion, although clinicians are advised to avoid prescribing tricyclic antidepressants to depressed patients with a history of CVD given unfavourable cardiovascular effects, contradictory reports highlight the need for a series of studies to clarify the impact of depressed mood and treatment with second-generation antidepressants on heart rate variability, especially considering that the noradrenergic and serotonergic systems are implicated in depressed mood (i.e. the monoamine-deficiency hypothesis (Reference Belmaker and Agam10)) as well as antidepressant treatment (i.e. augmentation of serotonin by the SSRIs and noradrenaline and serotonin by the SNRIs).