β-Blockers and the Risk of New-Onset Depression: Meta-analysis Reassures, but the Jury Is Still Out

J Clin Psychiatry 2021;82(3):21f14095

To cite: Andrade C. β-blockers and the risk of new-onset depression: meta-analysis reassures, but the jury is still out. J Clin Psychiatry. 2021;82(3):21f14095.
To share: https://doi.org/10.4088/JCP.21f14095

© Copyright 2021 Physicians Postgraduate Press, Inc.

β-Blockers (BBs) comprise a group of about 2 dozen drugs that are competitive antagonists at β-adrenergic receptors. BBs are classified in different ways. For example, propranolol, nadolol, timolol, and pindolol are classical, first generation, nonselective BBs; they block both β₁ and β₂ adrenoceptors. Atenolol, metoprolol, esmolol, and bisoprolol are newer, second generation, selective blockers of β₁ adrenoceptors. Labetalol, nebivolol, and carvedilol are examples of third generation BBs; they have additional properties besides β₁-blocking selectivity. For example, nebivolol has vasodilation activity, and labetalol and carvedilol block α₁ adrenoceptors in addition to β₁ adrenoceptors. Finally, drugs such as propranolol are lipophilic and readily cross the blood-brain barrier, whereas drugs such as atenolol are hydrophilic and have low central nervous system activity.¹

Indications

BBs are important in several different contexts in patients with cardiovascular disorders such as ischemic heart disease (IHD), cardiac failure, cardiac arrhythmias, and hypertension.^2,3 BBs are also prescribed for many neuropsychiatric indications, with rather discouraging results for most anxiety disorders⁴ and akathisia^5,6 and with good results for performance anxiety,⁷ essential tremor,⁸ and migraine prophylaxis.⁹ BBs have also been used to successfully attenuate lithium tremor.¹⁰ Less popularly, BBs have been used in attempts to reduce agitation in dementia¹¹ and aggression in schizophrenia and other psychiatric disorders.^12,13

β-Blockers and Depression

BB (pindolol) augmentation of antidepressant drugs has been suggested to accelerate treatment response in depression.¹⁴ However, for over half a century, BB treatment has also been associated with new-onset depression.¹⁵ A decade ago, Luijendijk and Koolman¹⁶ identified 22 observational studies of the association between BBs and incident depression; they described a pattern where the publication of one study identifying a significant association was followed by the publication of several other studies that refuted the association. Surprisingly, a recent population-based observational study that used data extracted from Danish registers found that 4 of 15 BBs were actually associated with a reduced risk of depression.¹⁷

One observational, registry-based study has suggested that BBs are associated with an increased risk of suicide¹⁸; however, the analysis in this study did not adjust for potential confounders, and a subsequent study found no increase in suicide risk associated with BB treatment but, unexpectedly, an increased suicide risk associated with current use of angiotensin-receptor antagonists.¹⁹

In a systematic review and meta-analysis of randomized controlled trials (RCTs), Ko et al²⁰ identified 15 studies conducted in patients (pooled N > 35,000) with cardiovascular disorders. All trials recruited at least 100 patients, each, and all followed patients for at least 6 months, reporting depressive symptom outcomes. The authors found that BBs were associated with a significant increase in sexual dysfunction (number needed to harm [NNH] per year of treatment, 199) and fatigue (NNH per year of treatment, 57), but not with a significant increase in depressive symptoms.

A Recent Meta-Analysis

The relationship between BB treatment and the risk of depression has at last been investigated in a comprehensive, PRISMA-compliant, PROSPERO-registered meta-analysis. In this meta-analysis,²¹ the authors searched electronic databases (including clinical trial registries), reference lists, and other sources for double-blind, parallel group RCTs of non-topical BB treatments, used in monotherapy or as add-on therapy for any nonpsychiatric indication in adolescents and adults. RCTs were included only if they were at least 2 weeks in duration. The risk of depression and other carefully and well-operationalized psychiatric adverse events was compared between BB and control groups using random effects models. The risk of fatigue/tiredness was examined as a positive control because this adverse effect is known to be associated with BBs.

The review identified 285 RCTs of 24 BBs. These included 200 RCTs with an active comparator arm, 73 RCTs with a placebo control arm, and 12 RCTs with both active and placebo control arms. In these RCTs, 288 BB samples were compared with active drugs and 87 were compared with placebo in a total of 361 unique BB samples (pooled N = 53,533). The most common active comparators were α₁-blockers, calcium channel blockers, renin-angiotensin system blockers, and thiazide and thiazide-like diuretics.

Most of the RCTs had been conducted for the indication arterial hypertension (197 trials, N = 32,042); other frequent indications included myocardial infarction, heart failure, and angina pectoris. Migraine was the commonest noncardiovascular indication (19 RCTs; N = 1,313). Only 53 RCTs had used an active method to assess adverse events. As many as 157 RCTs had industry support. Only 151 of the 285 RCTs could be included in the various quantitative analyses.

Important findings from the meta-analysis,²¹ published in the main paper and in a 110-page supplement, are presented in Table 1, Table 2, and Table 3. In summary, BBs were not associated with an increased risk of depression or withdrawal due to depression relative to either placebo or active controls. However, BBs were associated with an increased risk of fatigue/tiredness relative to placebo as well as relative to some groups of active controls. BBs were also associated with an increased risk of unusual dreams, relative to placebo.

Interpretation of the Results

On the surface, this very large meta-analysis²¹ of placebo arm and active arm parallel group RCTs provides strong evidence for the absence of risk of new-onset depression following initiation of BB treatment. So, what is the reason for the longstanding clinical concern that BB treatment is associated with an increased risk of developing depression?

There are 2 explanations. One is that it is well known that IHD and depression are each associated with an increased risk of the other,^22,23 and IHD is an important indication for the initiation of BB treatment. So, confounding by indication may explain clinical observations of new-onset depression after initiation of BBs in patients with IHD. The second explanation is that tiredness/fatigue and unusual dreams, which Riemer et al²¹ found to be significantly more common with BBs, are common symptoms of depression and may be misinterpreted and misreported as depression.

Riemer et al²¹ wrote in their abstract that insomnia and sleep disorder were other psychiatric adverse events that were possibly related to BB therapy. Whereas neither adverse event emerged as a significant finding in the main quantitative analyses for either placebo- or active-controlled comparisons, there were some significant or near-significant associations with BB treatment in sensitivity and subgroup analyses. So, insomnia and sleep disorder may additionally contribute to the possibility of nonspecific adverse events being misinterpreted as depression.

Critical Comments

Although this meta-analysis²¹ is an important milestone for research in the field, it is not without limitations. The first limitation is that the meta-analysis included a large number of short-term trials, some of which (eg, the 2- to 4-week trials) may have been too short to allow for the emergence of depression after treatment with BBs. These trials could therefore have biased the research findings toward the null hypothesis. A second limitation is that in most of the RCTs, investigators did not elicit adverse events systematically but merely recorded adverse events that were reported; because such an approach could underestimate the incidence of adverse events, the findings of the meta-analysis could, therefore, again have been biased toward the null hypothesis.

A third and important limitation is that there is no report of whether depression as a clinical diagnosis was formally recorded as an adverse outcome in any of the RCTs; in fact, knowing how RCTs are usually conducted, it is unlikely that this was done. So, whereas we know that depression as a symptom or possibly a syndrome (as operationalized in the Supplementary Materials, Table S4, by the authors of the meta-analysis) was not more common with BBs, we have no way of knowing whether depression as a clinical diagnosis occurred more frequently with BBs or not. The real question is whether patients receiving BBs suffer clinically relevant depression, such as a major depressive episode, and not whether they complain of depression as a symptom.

Perhaps the most important limitation is that we do not know what the findings with commonly used BBs are. These BBs include, for example, propranolol, atenolol, and metoprolol. Why is this important? Consider, for example, the interesting findings in Tables S15 and S19 in the supplementary materials²¹ that propranolol was associated with depression in 858/3800 (22.6%) patients and with fatigue/tiredness in 1683/4651 (36.2%) patients. Was the strikingly high incidence of depression with propranolol greater than the incidence of depression in the comparator arms? We do not know the answer.

It is important to know whether the risks are different for different BBs because different BBs have different pharmacokinetic and pharmacodynamic properties and may be preferred in different clinical contexts. Returning to propranolol, for example, in a systematic review and meta-analysis²⁴ of 9 RCTs (pooled N = 1,175) of propranolol compared with labetalol, oxprenolol, enalapril, atenolol, diltiazem, hydrochlorothiazide, and ketanserin for the treatment of hypertension, depression was found to occur significantly more frequently in propranolol-treated patients than in control patients (Mantel-Haenszel test, P = .039; mistakenly stated as P = .39 by the authors). This is definitely grounds for concern, and an issue that was not addressed by Riemer et al.²¹

The last word surely remains to be said. In this context, it has perhaps appropriately been stated that information from case reports should be carefully considered when RCTs have not been adequately designed to detect adverse outcomes.²⁵

Parting Notes: β-Blockers and Bronchoconstriction

BBs are used with caution or are avoided in patients with pulmonary disease, especially bronchial asthma, because of the risk of respiratory adverse events, especially bronchoconstriction.²⁶ However, an analysis of linked data from the UK Clinical Practice Research Datalink found that cardioselective BBs, unlike nonselective BBs, were not associated with the risk of moderate or severe exacerbation of asthma in patients with cardiovascular disease and comorbid asthma.²⁷ Furthermore, in a systematic review and multiple treatment comparison meta-analysis of 23 observational studies and 14 RCTs, Gulea et al²⁸ found that only propranolol was associated with reduced forced expiratory volume during the first second (FEV1); otherwise, relative to no treatment, BB treatment was actually associated with a reduced risk of acute exacerbation of chronic obstructive pulmonary disease events (hazard ratio, 0.77; 95% CI, 0.70–0.85). A qualitative analysis of the data in this study suggested no detrimental effect of BBs on quality of life, all-cause hospitalization, and mortality outcomes. It therefore appears that cardioselective BBs may be safe and may even improve outcomes in patients with comorbid cardiovascular and pulmonary disease; however, the BBs will need to be introduced in low doses and uptitrated gradually, and patients will need to be carefully monitored, especially during the initial weeks.²⁶

Published online: June 1, 2021.

Each month in his online column, Dr Andrade considers theoretical and practical ideas in clinical psychopharmacology with a view to update the knowledge and skills of medical practitioners who treat patients with psychiatric conditions.

Department of Clinical Psychopharmacology and Neurotoxicology, National Institute of Mental Health and Neurosciences, Bangalore, India (candrade@psychiatrist.com).
Financial disclosure and more about Dr Andrade.