β‐Blockers and risk of neuropsychiatric disorders: A systematic review and meta‐analysis

Abstract

Aims

This systematic review and meta‐analysis aimed to evaluate the association between β‐blocker use and neuropsychiatric adverse events, specifically focusing on short‐term outcomes.

Methods

A comprehensive literature search identified studies reporting neuropsychiatric outcomes in patients using β‐blockers, including randomized controlled trials and observational studies. Relative risks (RR) and 95% confidence intervals (CIs) were calculated for outcomes such as dizziness, insomnia, nightmares, drowsiness and delirium.

Results

Our analysis revealed that β‐blocker use was significantly associated with an increased risk of dizziness (RR 1.72, 95% CI [1.39–2.14]; I ² = 1%, 14 studies) compared to placebo. Lipophilic β‐blockers, especially propranolol, showed an even greater risk of dizziness (RR 3.13, 95% CI [1.44–6.84]; I ² = 0%, three studies). Propranolol was also associated with increased insomnia risk compared to placebo (RR 1.13, 95% CI [1.00–1.28]; I ² = 0%, five studies). Our data did not show statistically significant increases in the reports of nightmares and somnolence. Other adverse effects, including drowsiness, sleep disturbances, hallucinations and delirium, were noted.

Conclusions

Our findings suggest a significant association between β‐blocker use and an increased risk of neuropsychiatric adverse events, particularly insomnia and dizziness with higher risks associated with lipophilic β‐blocker use. Given the ambiguity surrounding dizziness and its classification as a neuropsychiatric effect, our findings are exploratory, and we cannot exclude a potential cardiovascular origin for dizziness. Most studies (75%) were published before the CONSORT statement in 1996, indicating potential reporting limitations and a lack of recent research. Additionally, 60% of studies had a high risk of bias, underscoring the need for more rigorous and contemporary investigations into the neuropsychiatric implications of β‐blocker use.

1. INTRODUCTION

β‐blockers are a group of medications that target β‐adrenergic receptors to manage a wide range of cardiovascular and non‐cardiovascular conditions, including tachycardia, hypertension, myocardial infarction, congestive heart failure, cardiac arrhythmias, coronary artery disease, hyperthyroidism, essential tremor, aortic dissection, portal hypertension, glaucoma and migraine prophylaxis. Additionally, β‐blockers are used off‐label for the treatment of anxiety‐related disorders. ¹ Approximately 30 million adults use β‐blockers in the United States alone. ² Although β‐blockers have proven effective for treating various conditions, including heart failure with reduced ejection fraction, they have also been associated with adverse drug events such as hypotension, bradycardia and hyperkalaemia, and neuropsychiatric adverse events such as depression, insomnia, hallucinations and dizziness due to β‐adrenoreceptor blockade. ³ , ⁴ , ⁵ , ⁶ While several meta‐analyses have explored the association between β‐blockers and depression, few have specifically investigated their association with other neuropsychiatric adverse events. An early meta‐analysis of four heart failure trials, including 10 082 patients, showed that β‐blockers vs. placebo was associated with a significant relative increase in reported dizziness (RR 1.37; 95% CI 1.09–1.71). However, this analysis was restricted to heart failure patients, limiting its generalizability to other populations. ⁷

Furthermore, a recent systematic review explored the association between neuropsychiatric consequences and lipophilic β‐blockers. ⁸ The study focused primarily on the mechanistic, pharmacokinetic and pharmacodynamic properties of β‐blockers to provide insight into potential neuropsychiatric effects but did not comprehensively evaluate their clinical outcomes. The review suggests that lipophilic β‐blockers penetrate the blood–brain barrier more easily than hydrophilic β‐blockers, which may result in higher neuropsychiatric adverse events. Therefore, our study aims to further investigate and synthesize existing evidence to better understand the relationship and quantify the magnitude of risk between β‐blockers and neuropsychiatric adverse events, other than depression, across different cardiovascular and non‐cardiovascular conditions. We have included dizziness in this study as a neuropsychiatric adverse event based on clinical guidelines and the established pharmacological profiles of β‐blockers. We categorize dizziness as a central nervous system (CNS) adverse effect of these medications, following recognized classifications and guidelines. The Medical Dictionary for Regulatory Activities (MedDRA) and prescribing sources, such as UpToDate, classify dizziness under “nervous system disorders” for these drugs. ⁹ Product monographs for lipophilic β‐blockers, including propranolol and metoprolol, explicitly recognize dizziness as a potential CNS adverse effect. ¹⁰ , ¹¹ , ¹² While MedDRA also classifies dizziness under cardiac and vascular disorders, noting that it may arise from cardiovascular factors such as hypotension and bradycardia, the CNS effects of β‐blockers—particularly those that can cross the blood–brain barrier—play a significant role in this symptom. Mechanistic studies suggest that lipophilic β‐blockers may be more likely to induce neuropsychiatric symptoms, including dizziness, due to their pharmacological properties and effects on brain function. ¹³ This mechanism may contribute to a range of neuropsychiatric symptoms, suggesting that the dizziness experienced could be related to changes in mood, sleep patterns and overall brain function. ¹⁴ Additionally, we will examine whether lipophilic β‐blockers pose a higher risk of inducing neuropsychiatric adverse events compared to hydrophilic β‐blockers.

2. METHODS

2.1. Protocols and registration

We followed the Preferred Reporting Items for Systematic Review and Meta‐Analysis (PRISMA) protocol. ¹⁵ This review has been registered with the International Prospective Systematic Reviews Registry (PROSPERO) under the number CRD42023485078.

2.2. Eligibility criteria

Eligibility criteria for study selection included: (1) outpatients aged 18 years or older; (2) intervention group comprising patients taking β‐blockers; (3) control groups, which included patients not taking β‐blockers or those taking a placebo, another β‐blocker or another antihypertensive medication; (4) the presence of neuropsychiatric adverse events other than depression (studies, where depression was the only neuropsychiatric adverse event, were not included in this systematic review since several meta‐analyses have extensively explored the association between β‐blockers and depression; ¹⁶ , ¹⁷ (5) the following types of studies were included: randomized controlled trials (RCTs), non‐randomized trials, cohort studies, cross‐sectional studies, case–control studies, crossover studies, case reports, case series and pharmacovigilance studies; and (6) studies published in the English language.

The following exclusion criteria were used: (1) studies including pregnant or lactating women; and (2) reviews, editorials and grey literature (e.g., conference abstracts, institutional reports and unpublished theses).

2.3. Data source and search strategy

We systematically searched Medline and Embase for all relevant articles reporting an association between β‐blocker use and neuropsychiatric side effects other than depression up to 5 August 2024. The following keywords were used for beta‐blockers: “Beta‐blocker, atenolol, acebutolol, betaxolol, bisoprolol, carteolol, esmolol, metoprolol, penbutolol, nadolol, nebivolol, pindolol, propranolol, timolol, sotalol, carvedilol, oxprenolol, labetalol.” For searching neuropsychiatric adverse events, we included the following keywords: “Neurotoxicity, encephalopathy, altered mental status, confusion, cognitive impairment, behavioural changes, mood swing, agitation, speech or language problem, hallucination, drowsiness, delirium, nervousness, aggression, insomnia, nightmares, dizziness, giddiness, disorientation, somnolence, transient alteration of awareness.”

The choice of study outcomes was derived from the System Organ Classes (SOCs) in MedDRA, specifically from the categories of “Nervous System Disorders” or “Psychiatric Disorders”. Dizziness, however, also appears under the “Cardiovascular Disorders” SOC in MedDRA, given its potential association with both neuropsychiatric and cardiovascular aetiologies. The search strategy for each database is provided in Appendix Table S1 in the Supporting Information.

2.4. Study selection

Following the initial search strategy, two independent reviewers (L.E. and R.P.) utilized the inclusion and exclusion criteria to screen titles and abstracts and evaluated the full texts of relevant studies for eligibility. Any disagreements were resolved by consensus or a third reviewer (F.T.M.).

Studies that did not meet the inclusion criteria were excluded, and the reasons for exclusion were outlined in a flow chart (Figure 1).

FIGURE 1.

PRISMA flow diagram of study selection. The diagram illustrates the identification and screening process to determine eligible studies from the databases Medline and Embase. Studies were included if the study design was a randomized control trial, cohort study, case–control study, case reports, case series or cross‐sectional study; the study population was adults 18 years or older; the intervention was β‐blocker use; the comparator was other antihypertensive or no β‐blockers. (indexed from 1946 to October 12, 2023].

2.5. Data extraction

Data were extracted from included studies and managed using a Word document. We then extracted the following information from the articles in a standardized form, including the first author's name, publication year, population characteristics, intervention/exposure, comparator/control group and outcome (one of the neuropsychiatric adverse events listed above in the search strategy section), as well as effect estimates and confidence intervals. When the measure of association was not reported, we calculated the composite risk ratios using the Mantel–Haenszel chi‐squared test. Five reviewers grouped into pairs (L.E. and R.P.; F.A. and A.J.; L.E. and A.M.O.), independently extracted data from relevant articles. Any disagreements were resolved by consensus or a third reviewer (F.T.M.).

2.6. Risk of bias in individual studies

Five reviewers, grouped into two pairs (L.E. and R.P.; F.A. and A.J., L.E. and A.M.O.), independently assessed the risk of bias for each included study. Any disagreements were resolved by consensus or a third reviewer (F.T.M.).

Version 2 of the Cochrane Risk of Bias for Randomized Trials (RoB2) tool was used to appraise RCTs. If an RCT exhibited concerns in the randomization process, deviations from the intended intervention, missing outcome data, inadequacy in the measurement of the outcome or selective reporting, it was classified as having an overall high risk of bias. If the RCT did not provide sufficient information to make a definite judgement in these bias domains, the study was classified as having an unclear risk of bias. If a study was scored as high or unclear risk of bias in one domain, the overall article was classified as high or unclear, respectively. If a study was classified as low risk across all domains, the overall article was deemed to have a low risk of bias.

For cohort studies, the Risk of Bias in Non‐randomized Studies—of Exposures (ROBINS‐E) tool was used. The modified Downs and Black checklist was used to appraise case–control and cross‐sectional studies. These studies were assigned scores ranging from 0 to 28 and categorized into four quality levels: excellent (scores 26–28), good (scores 20–25), fair (scores 15–19), and poor (scores <14) All reviewers who completed the risk of bias assessment convened to ensure uniformity and clarify uncertainties throughout the evaluation process.

2.7. Statistical analysis

We calculated the summary risk ratios and 95% confidence intervals using a random‐effects model (the method of DerSimonian and Laird) due to differences between study populations and interventions. ¹⁸ Meta‐analysis was conducted if three or more studies were available for the same outcome of interest and the same comparator group. ¹⁹ Pooled estimates were also obtained for two studies where the results were sufficiently similar. ²⁰ A 0.5 continuity correction was applied to include data from a study in which one of the arms had zero events. This continuity correction allows for the inclusion of zero‐event trials while maintaining analytic consistency. ²¹ The heterogeneity was quantified using the I ² statistic, which indicates the percentage of variability in effect estimates attributed to heterogeneity rather than sampling error. If the I ² statistic exceeded 75%, we considered the heterogeneity to be high. Funnel plots were created and visually inspected with the eyeball test to assess publication bias in case at least 10 studies were included in a meta‐analysis. ²²

All analyses were performed using RevMan 5.4 Review Manager software. ²³

3. RESULTS

3.1. Literature search results

A total of 4706 studies were retrieved from the databases Embase and Medline. After removing 1317 duplicates, 3389 studies remained for the title and abstract screening. During this screening, 3092 articles were excluded, leaving 297 articles eligible for full‐text screening.

During the full‐text review, 226 articles were excluded, and the reasons for exclusion are provided in Figure 1. Finally, 71 studies were eligible for data extraction. Following a comprehensive reference search, an additional 18 articles were incorporated into the study. The literature was published from 1972 to 2020. The process of literature inclusion is shown in Figure 1, and details of the search strategy are provided in Table S1 in the Supporting Information.

3.2. Characteristics of included studies

Data from 53 047 adults across the 89 included studies (range: 10–10 000) were analysed in this review. Propranolol, carvedilol, metoprolol and atenolol were the most frequent β‐blockers reported in the included studies.

Of the 89 studies included in the review, 44 were randomized controlled trials (RCTs), 10 were non‐randomized clinical trials, seven were cohort studies, five were crossover trials, three were cross‐sectional studies, two were pharmacovigilance studies, 14 were case reports, and four were case series studies. Most RCTs (75%) included were published before the introduction of the Consolidated Standards of Reporting Trials (CONSORT) statement in 1996. A summary of the characteristics and risk of bias of included studies evaluating the use of β‐blockers and the Risk of Neuropsychiatric Disorders is shown in Table S2 and S3 in the Supporting Information.

3.3. β‐Blockers and dizziness

Dizziness was reported in 33 RCTs, comprising a total of 9133 participants. These trials focused on patients who either received a β‐blocker as both an intervention and comparator or who were given a β‐blocker as an intervention while others received another antihypertensive drug or a placebo as a comparator group.

Compared to other antihypertensive medications, β‐blockers were not associated with a statistically significant higher risk of dizziness (RR 0.87, 95% CI [0.39–1.93]; I ² = 57%, 10 RCTs) ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² (Figure 2). There was no suggestion of a publication bias in this meta‐analysis (Figure S1 in the Supporting Information).

FIGURE 2.

Forest plot, a pooled measure of the risk of dizziness, and risk of bias assessment in 10 RCTs comparing β‐blockers to other antihypertensives with a similar indication. The meta‐analysis included adults treated with β‐blockers (n = 1010) or a different antihypertensive drug (n = 1021). The pooled risk ratio for dizziness was 0.87 (95% CI [0.39–1.93], I ² = 57%). The inverse variance was used to determine risk ratio (RR) and a random‐effects model was used to assess pooled treatment effects across studies included in each analysis. Red (−) icons indicate a high risk of bias, yellow (?) icons indicate an unclear risk of bias, and green (+) icons indicate a low risk of bias.

β‐Blockers compared to placebo were associated with a statistically significant higher risk of dizziness (RR 1.72, 95% CI [1.39–2.14]; I ² = 1%, 14 RCTs) ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ (Figure 3). There was no suggestion of a publication bias, as shown in Figure S2 in the Supporting Information).

FIGURE 3.

Forest plot, a pooled measure of the risk of dizziness, and risk of bias assessment in 14 RCTs comparing β‐blockers to a placebo with a similar indication. The meta‐analysis included adults treated with β‐blockers (n = 3702) or a placebo (n = 1238). The pooled risk ratio for dizziness was 1.72 (95% CI [1.39–2.14], I ² = 1%). The inverse variance was used to determine risk ratio (RR) and a random‐effects model was used to assess pooled treatment effects across studies included in each analysis. Red (−) icons indicate a high risk of bias, yellow (?) icons indicate an unclear risk of bias, and green (+) icons indicate a low risk of bias.

Compared to other β‐blockers, propranolol, a high lipophilic β‐blocker, was also associated with a statistically significant higher risk of dizziness (RR 2.36, 95% CI [1.27–4.40]; I ² = 0%, three RCTs) ⁴⁷ , ⁴⁸ , ⁴⁹ (Figure 4).

FIGURE 4.

Forest plot, a pooled measure of the risk of dizziness, and risk of bias assessment in three RCTs comparing propranolol to other β‐blockers with a similar indication. The meta‐analysis included adults treated with propranolol (n = 151) or other β‐blockers (n = 272). The pooled risk ratio for dizziness was 2.36 (95% CI [1.27–4.40], I ² = 0%). The inverse variance was used to determine risk ratio (RR) and a random‐effects model was used to assess pooled treatment effects across studies included in each analysis. Red (−) icons indicate a high risk of bias, yellow (?) icons indicate an unclear risk of bias, and green (+) icons indicate a low risk of bias.

In another meta‐analysis comparing propranolol to placebo, propranolol was also associated with a statistically significant higher risk of dizziness (RR 3.13, 95% [1.44–6.84]; I ² = 0%, three RCTs ⁴⁰ , ⁴¹ , ⁴³ (Figure 5). Characteristics of RCTs included in this study are shown in Table S2 in the Supporting Information.

FIGURE 5.

Forest plot, a pooled measure of the risk of dizziness, and risk of bias assessment in three RCTs comparing propranolol to placebo with a similar indication. The meta‐analysis included adults treated with propranolol (n = 185) or a placebo (n = 141). The pooled risk ratio for dizziness was 3.13 (95% CI [1.44–6.84], I ² = 0%). The inverse variance was used to determine risk ratio (RR) and a random‐effects model was used to assess pooled treatment effects across studies included in each analysis. Red (−) icons indicate a high risk of bias, yellow (?) icons indicate an unclear risk of bias, and green (+) icons indicate a low risk of bias.

3.4. β‐Blockers and insomnia

Insomnia was reported in 22 RCTs comparing 10 121 individuals. These trials focused on patients who either received a β‐blocker as both an intervention and comparator or who were given a β‐blocker as an intervention while others received another antihypertensive medication or a placebo as a comparator group. Compared to other antihypertensive medications, β‐blockers were not associated with a statistically significant higher risk of insomnia (RR 0.97, 95% CI [0.25–3.78], I ² = 0%, three RCTs) ²⁴ , ²⁹ , ⁵⁰ (Figure 6).

FIGURE 6.

Forest plot, a pooled measure of the risk of insomnia, and risk of bias assessment in three RCTs comparing β‐blockers to other antihypertensive drugs with a similar indication. The meta‐analysis included adults treated with β‐blockers (n = 225) or other antihypertensive drugs (n = 190). The pooled risk ratio for insomnia was 0.97 (95% CI [0.25–3.78], I ² = 0%). The inverse variance was used to determine risk ratio (RR) and a random‐effects model was used to assess pooled treatment effects across studies included in each analysis. Red (−) icons indicate a high risk of bias, yellow (?) icons indicate an unclear risk of bias, and green (+) icons indicate a low risk of bias.

In another study, β‐blockers compared to placebo were associated with a slight increase in the risk of insomnia with (RR 1.12, 95% CI [0.99–1.26], I ² = 0%, 14 RCTs) ³⁶ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁵ , ⁴⁶ , ⁵¹ , ⁵² , ⁵³ , ⁵⁴ , ⁵⁵ (Figure 7). There was no suggestion of publication bias as shown in the funnel plot (Figure S3 in the Supporting Information).

FIGURE 7.

Forest plot, a pooled measure of the risk of insomnia, and risk of bias assessment in 14 RCTs comparing β‐blockers to a placebo with a similar indication. The meta‐analysis included adults treated with β‐blockers (n = 5115) or a placebo (n = 2843). The pooled risk ratio for insomnia was 1.12 (95% CI [0.99–1.26], I ² = 0%). The inverse variance was used to determine risk ratio (RR) and a random‐effects model was used to assess pooled treatment effects across studies included in each analysis. Red (−) icons indicate a high risk of bias, yellow (?) icons indicate an unclear risk of bias, and green (+) icons indicate a low risk of bias.

Compared to placebo, propranolol a high lipophilic β‐blocker was also associated with a slight increase in the risk of insomnia with (RR 1.13, 95% CI [1.00–1.28], I ² = 0%, five RCTs) ⁴⁰ , ⁴¹ , ⁴³ , ⁵¹ , ⁵⁵ (Figure 8). Characteristics of RCTs included in this study are presented in Table S2 in the Supporting Information.

FIGURE 8.

Forest plot, a pooled measure of the risk of insomnia, and risk of bias assessment in five RCTs comparing propranolol to placebo with a similar indication. The meta‐analysis included adults treated with propranolol (n = 2243) or a placebo (n = 2205). The pooled risk ratio for insomnia was 1.13 (95% CI [1.13–1.28], I ² = 0%). The inverse variance was used to determine risk ratio (RR) and a random‐effects model was used to assess pooled treatment effects across studies included in each analysis. Red (−) icons indicate a high risk of bias, yellow (?) icons indicate an unclear risk of bias, and green (+) icons indicate a low risk of bias.

In a retrospective cohort study of 4603 patients newly diagnosed with hypertension who started a β‐blocker, the risk of insomnia within 30 days was higher in patients starting propranolol, a highly lipophilic β‐blocker, compared to those starting non‐propranolol β‐blockers. The adjusted odds ratio (aOR) was 2.12 (95% CI [1.61–2.94]). The risk was even greater when comparing propranolol to bisoprolol (aOR 3.22, 95% CI [2.00–5.26]) or atenolol (aOR 2.17, 95% CI [1.52–3.03]). ⁵⁶

A summary of the remaining studies (i.e., two crossover trials, two non‐randomized trials, one cross‐sectional study) reporting insomnia is also provided in Table S2 in the Supporting Information.

3.5. β‐Blockers and nightmares

Four RCTs reported nightmares. Compared to other antihypertensive medications, the risk of nightmares was higher in β‐blocker users but did not reach statistical significance (RR 2.80, 95% CI [0.47–17.41], I ² = 0%, two studies) (Figure 9). Characteristics of the RCTs included in this study are shown in Table S2 in the Supporting Information.

FIGURE 9.

Forest plot, a pooled measure of the risk of nightmares, and risk of bias assessment in two RCTs comparing β‐blockers to other antihypertensive drugs with a similar indication. The meta‐analysis included adults treated with β‐blockers (n = 108) or other antihypertensive drugs (n = 109). The pooled risk ratio for nightmares was 2.80 (95% CI [0.45–17.41], I ² = 0%). The inverse variance was used to determine risk ratio (RR) and a random‐effects model was used to assess pooled treatment effects across studies included in each analysis. Red (−) icons indicate a high risk of bias, yellow (?) icons indicate an unclear risk of bias, and green (+) icons indicate a low risk of bias.

A large pharmacovigilance study comparing the risk of reporting nightmares among different β‐adrenoceptor antagonists showed that pindolol (adjusted reporting odds ratio [aROR] 2.82, 95% CI [2.19–3.61]) and metoprolol (aROR 1.89, 95% CI [1.66–2.16]) had the highest risk of nightmare reporting. Compared to low lipid solubility β‐adrenoceptor antagonists, the use of moderate or high lipid solubility β‐adrenoceptor antagonists was significantly more associated with nightmare reports (aROR moderate vs. low 1.72, 95% CI [1.47–2.00] and aROR high vs. low 1.84, 95% CI [1.53–2.22]).

Table S2 in the Supporting Information also provides a summary of the remaining studies (i.e., one crossover study, one non‐randomized trial and five case reports) reporting nightmares.

3.6. β‐Blockers and somnolence

Somnolence was noted in 11 RCTs comparing 5499 individuals. Compared to placebo, β‐blockers were associated with a slight increase in the risk of somnolence (RR 1.13, 95% CI [0.73–1.74], I ² = 40%, six studies) ⁴⁴ , ⁴⁵ , ⁵⁵ , ⁵⁷ , ⁵⁸ , ⁵⁹ (Figure 10).

FIGURE 10.

Forest plot, a pooled measure of the risk of somnolence, and risk of bias assessment in six RCTs comparing β‐blockers to placebo with a similar indication. The meta‐analysis included adults treated with β‐blockers (n = 2451) or placebo (n = 929). The pooled risk ratio for somnolence was 1.13 (95% CI [0.73–1.74], I ² = 40%). The inverse variance was used to determine risk ratio (RR) and a random‐effects model was used to assess pooled treatment effects across studies included in each analysis. Red (−) icons indicate a high risk of bias, yellow (?) icons indicate an unclear risk of bias, and green (+) icons indicate a low risk of bias.

3.7. β‐Blockers and other neuropsychiatric side effects

Neuropsychiatric side effects associated with β‐blockers were documented across various study designs. Drowsiness was observed in one crossover trial, one RCT and one non‐randomized trial. ²⁶ , ⁶⁰ , ⁶¹ Sleep disturbances, including various disruptions to sleep patterns, were reported in one crossover trial, one RCT, two non‐randomized trials, and one cohort study. ³² , ⁶² , ⁶³ , ⁶⁴ , ⁶⁵ and metoprolol was linked to sleep disturbances in 21% of examined case reports. ⁶ , ⁶⁶ , ⁶⁷ Hallucinations, including visual and auditory manifestations, were documented in one crossover trial, ⁶⁸ one RCT, ⁴⁹ two case series ⁶⁹ , ⁷⁰ and in 50% of case reports analysed. ⁷¹ , ⁷² , ⁷³ , ⁷⁴ , ⁷⁵ , ⁷⁶ , ⁷⁷ Delirium was documented in 36% of the case reports reviewed ⁵ , ⁷¹ , ⁷² , ⁷⁶ , ⁷⁸ and in one case series. ⁷⁹ Of note, in the case of hallucinations and delirium, propranolol was predominantly implicated in most of the case reports reviewed. The characteristics of the studies reporting other neuropsychiatric side effects can be found in Table S2 in the Supporting Information.

4. DISCUSSION

Our systematic review and meta‐analysis aimed to comprehensively analyse the existing evidence regarding the association between β‐blocker use and neuropsychiatric adverse events. We focused on short‐term outcomes and excluded depression from our investigations to narrow our scope to immediate concerns. Additionally, we made this decision because previous meta‐analyses have extensively studied β‐blocker‐associated depression. Our study documented many adverse neuropsychiatric effects associated with β‐blockers, such as dizziness, insomnia, nightmares, drowsiness, sleep disturbances, hallucinations, somnolence and delirium.

Dizziness was the most frequent neuropsychiatric effect reported in our study. Our meta‐analysis revealed a 72% higher risk of dizziness in patients using β‐blockers compared to placebo. However, when comparing β‐blockers to other antihypertensive medications, we did not observe a significant difference, although there was moderate heterogeneity (I ² = 57%), indicating variability in the results. This suggests that the risk of dizziness may vary across different types of β‐blockers and other antihypertensive drugs.

While we recognize that dizziness may arise as a cardiovascular disorder—resulting from conditions such as hypotension, particularly in patients using antihypertensive medications— our findings suggest that dizziness could also be a potential neuropsychiatric adverse event. By blocking beta‐1 adrenergic receptors, β‐blockers reduce the force of cardiac contraction and slow heart rate, leading to lower blood pressure. This reduction in cardiac output may decrease cerebral perfusion, particularly in patients with pre‐existing cardiovascular conditions, potentially resulting in dizziness. However, we argue that the dizziness observed in our study is more likely a neuropsychiatric disorder rather than a purely cardiovascular one. This hypothesis is supported by several factors. First, propranolol, a highly lipophilic β‐blocker, was associated with more than twice the risk of dizziness compared to less lipophilic beta‐blockers. Lipophilic β‐blockers can cross the blood–brain barrier, leading to CNS effects that result in dizziness rather than solely cardiovascular effects. In contrast, hydrophilic β‐blockers, which do not easily penetrate the CNS, are associated with a lower risk of dizziness, highlighting the role of drug lipophilicity.

Second, dizziness often co‐occurs with other neuropsychiatric symptoms such as confusion, insomnia and nightmares. This clustering of symptoms points towards a central nervous system origin, as these effects are commonly associated with CNS disturbances. If dizziness were primarily due to cardiovascular mechanisms, such as hypotension, we would expect it to occur independently of these neuropsychiatric symptoms.

Lastly, the underlying mechanisms for these neuropsychiatric effects are not fully understood, but several proposed mechanisms suggest a dose‐related effect. One potential explanation is the presence of beta receptors in the brain, which can affect mood and sleep. Additionally, beta‐blocking agents may exhibit affinity for serotonin (5‐HT) receptors, further influencing these neuropsychiatric outcomes. ⁸⁰ Beta‐1 blockade might also impact sleep by inhibiting sympathetic signalling to the pineal gland, resulting in decreased night‐time levels of melatonin. ⁸¹ These changes in sleep patterns could exacerbate feelings of dizziness, as disturbed sleep is known to impair cognitive and vestibular function. In summary, while we recognize that dizziness can arise from cardiovascular factors, the evidence in our study suggests that the dizziness associated with β‐blockers, particularly the more lipophilic agents, is more likely to stem from CNS effects. ¹³ The drugs' pharmacokinetics, the co‐occurrence of symptoms and the proposed underlying mechanisms involving complex interactions within the CNS support this neuropsychiatric origin of dizziness. However, we acknowledge that interpreting dizziness in this context is complex and should be approached with caution. Our study does not provide sufficient evidence to definitively attribute dizziness to a neuropsychiatric origin. Future mechanistic and pharmacoepidemiological research should explore this distinction more thoroughly.

Insomnia was another commonly reported outcome. Though there was no statistically significant increased risk of insomnia compared to other antihypertensive medications, a slight increase in risk was observed when β‐blockers were compared to placebo. In particular, propranolol was associated with a notable increase in insomnia risk compared to both placebo and non‐propranolol β‐blockers. This finding is corroborated by a retrospective cohort study indicating that lipophilic β‐blockers are more likely to disrupt sleep due to their greater ability to cross the blood–brain barrier.

While less frequently reported, nightmares showed a trend towards increased risk in patients using β‐blockers. Though the meta‐analysis did not find a statistically significant association, a large pharmacovigilance study identified that moderate and highly lipophilic β‐adrenoceptor antagonists, such as pindolol and metoprolol, were significantly more likely to be linked to nightmares compared to those with lower lipid solubility.

Somnolence was also a reported side effect in the included studies. Compared to placebo, β‐blockers were associated with a slight increase in the risk of somnolence, though this was not statistically significant. However, the findings suggest a possible trend towards increased somnolence, particularly with specific β‐blockers, warranting further investigation into this side effect.

Our review highlighted other neuropsychiatric side effects, including dizziness, insomnia, nightmares and somnolence. Although the risk estimates for these events were generally not statistically significant, the findings suggest that certain β‐blockers, particularly propranolol, might be more likely to contribute to such adverse effects due to their CNS penetration.

Taken all together, our findings are consistent with an earlier meta‐analysis that showed that β‐blockers vs. placebo was associated with a 37% increase in the risk of dizziness. Furthermore, our results support the findings of a recent review indicating that lipophilic β‐blockers are more likely to penetrate the blood–brain barrier than hydrophilic β‐blockers, potentially leading to a higher incidence of CNS‐related adverse events. However, given the ambiguity surrounding dizziness and its classification as a neuropsychiatric effect, our findings are exploratory, and we cannot exclude a potential cardiovascular origin for dizziness.

Our study has several limitations. Firstly, most of the articles included were published before the introduction of the CONSORT statement in 1996. This might have resulted in inadequate reporting of trials and outdated findings, indicating a notable gap in recent research on this topic. As a result, our conclusions may not accurately represent the current clinical landscape. Secondly, most included studies exhibited a high risk of bias, potentially compromising our findings' validity. This highlights the need for more rigorous and contemporary research in this area to ensure the generation of robust evidence. Finally, our focus on short‐term adverse events excluded investigations into potential long‐term neuropsychiatric adverse events of β‐blockers. Understanding the chronic effects of these medications on mental health could provide valuable insights, highlighting the necessity for future studies to explore this aspect comprehensively.

5. CONCLUSION

Our review provides comprehensive insights into the association between β‐blockers and neuropsychiatric adverse events. β‐blockers were linked to a broad spectrum of adverse outcomes, including dizziness, insomnia, nightmares, somnolence, drowsiness, hallucinations and delirium. The risk of these events varied based on the specific characteristics of β‐blockers, particularly liposolubility, with high lipophilic agents such as propranolol more frequently associated with adverse CNS effects.

Our findings underscore the complex impact of β‐blockers on neurological and psychiatric functions. Although these adverse events are not uniformly observed across all studies, they appear to be significant in certain contexts. Therefore, clinicians should carefully consider individual patient characteristics, such as susceptibility to neuropsychiatric conditions, when prescribing β‐blockers. In particular, populations vulnerable to CNS adverse effects, such as those with a history of mental health disorders, may require closer monitoring or alternative treatment options. In this context, dizziness, one of the most frequently reported neuropsychiatric effects, deserves particular attention. While dizziness can arise from cardiovascular factors, our findings suggest that it may also represent a potential neuropsychiatric adverse event associated with β‐blockers, particularly lipophilic agents, and it is more likely to stem from CNS effects. The proposed mechanisms underlying these neuropsychiatric effects include the influence of beta receptors in the brain, interactions with serotonin (5‐HT) receptors and potential disruptions in melatonin production due to beta‐1 blockade. These factors indicate that clinicians should be vigilant regarding dizziness and other CNS‐related symptoms in patients prescribed β‐blockers. Given the ambiguity surrounding dizziness and its classification as a neuropsychiatric effect, our findings are exploratory, and we cannot rule out a potential cardiovascular origin for dizziness. This distinction highlights the need for further investigation into the mechanisms underlying β‐blocker‐induced dizziness.

However, our review also highlighted significant limitations in the available literature, such as outdated studies and a high risk of bias in many of the trials included. Some findings may suffer from these shortcomings, which call for more contemporary and methodologically rigorous studies. Future investigations should aim to clarify the potential neuropsychiatric origins of dizziness in β‐blocker users, and further distinguish between short‐term and long‐term effects to better inform clinical decision‐making.

AUTHOR CONTRIBUTIONS

Lujain Ez Eddin: Conceptualization; data curation; formal analysis; investigation; methodology; writing—original draft; writing—review and editing. Rebecca Preyra: Data curation; investigation; writing—review and editing. Fatemeh Ahmadi: Data curation; investigation; writing—review and editing. Atefeh Jafari: Data curation; investigation; writing—review and editing. Mohammad Ali Omrani: Data curation; investigation. Flory T. Muanda: Conceptualization; methodology; supervision; validation; writing—review and editing.

CONFLICT OF INTEREST STATEMENT

The authors declare that they have no competing interests to disclose. This research was conducted in the absence of any financial, personal or professional relationships that could be construed as potentially influencing the work presented in this manuscript.

Supporting information

Table S1. Literature search.

Table S2. Characteristics of studies included in the systematic review.

Table S3. Risk of bias of studies included in the systematic review.

Figure S1. Funnel plot for assessing publications bias in 10 studies comparing beta‐blockers vs. other antihypertensive drugs and the risk of dizziness.

Figure S2. Funnel plot for assessing publication bias in 14 studies comparing beta‐blockers vs. placebo and the risk of dizziness.

Figure S3. Funnel plot for assessing publication bias in 14 studies comparing beta‐blockers vs. placebo and the risk of insomnia.

Eddin LE, Preyra R, Ahmadi F, Jafari A, Omrani MA, Muanda FT. β‐Blockers and risk of neuropsychiatric disorders: A systematic review and meta‐analysis. Br J Clin Pharmacol. 2025;91(2):325‐337. doi: 10.1111/bcp.16361

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available in the Supporting Information.