Exploring Sex Differences of Beta-Blockers in the Treatment of Hypertension: A Systematic Review and Meta-Analysis

Aims: In the prevention of cardiovascular morbidity and mortality, early recognition and adequate treatment of hypertension are of leading importance. However, the efficacy of antihypertensives may be depending on sex disparities. Our objective was to evaluate and quantify the sex-diverse effects of beta-blockers (BB) on hypertension and cardiac function. We focussed on comparing hypertensive female versus male individuals. Methods and results: A systematic search was performed for studies on BBs from inception to May 2020. A total of 66 studies were included that contained baseline and follow up measurements on blood pressure (BP), heart rate (HR), and cardiac function. Data also had to be stratified for sex. Mean differences were calculated using a random-effects model. In females as compared to males, BB treatment decreased systolic BP 11.1 mmHg (95% CI, −14.5; −7.8) vs. 11.1 mmHg (95% CI, −14.0; −8.2), diastolic BP 8.0 mmHg (95% CI, −10.6; −5.3) vs. 8.0 mmHg (95% CI, −10.1; −6.0), and HR 10.8 beats per minute (bpm) (95% CI, −17.4; −4.2) vs. 9.8 bpm (95% CI, −11.1; −8.4)), respectively, in both sexes’ absolute and relative changes comparably. Left ventricular ejection fraction increased only in males (3.7% (95% CI, 0.6; 6.9)). Changes in left ventricular mass and cardiac output (CO) were only reported in males and changed −20.6 g (95% CI, −56.3; 15.1) and −0.1 L (95% CI, −0.5; 0.2), respectively. Conclusions: BBs comparably lowered BP and HR in both sexes. The lack of change in CO in males suggests that the reduction in BP is primarily due to a decrease in vascular resistance. Furthermore, females were underrepresented compared to males. We recommend that future research should include more females and sex-stratified data when researching the treatment effects of antihypertensives.


Introduction
Cardiovascular disease (CVD) affects over one billion people globally [1,2]. Hypertension is the most significant contributor to mortality in CVD and is, therefore, the most important and modifiable risk factor to be targeted [3,4]. In CVD, mortality rates differ between sexes [5]. Although hypertension and its CVD risks occur less before menopause compared to age-matched males; after menopause, females rapidly catch up, ultimately exceeding males [6,7].
Weight control and antihypertensive drugs are amongst the most pivotal in controlling hypertension [8]. There are differences suggested in females compared to males in system

Series of Meta-Analysis
The search, inclusion, and exclusion criteria are developed for a series of systematic reviews and meta-analyses. The goal of this series is to assess the effect of the five major groups of antihypertensive drugs on cardiovascular outcomes in females as compared to males. This systematic review and meta-analysis investigates the effect of BBs. Our review was registered in Prospero database with registration number: CRD42021273583.

Literature Search
A literature search from inception (1945) up to May 2020 was conducted in PubMed (NCBI) and EMBASE (Ovid) for studies evaluating the effects of antihypertensive medication on cardiovascular and hemodynamic variables in hypertensive individuals. The search strategy focused on cardiac geometry, heart failure, diastolic dysfunction, myocardial infarction (MI), and cerebral vascular accident (CVA) as detailed in the supplements (Table S1). The search limits used were 'humans' and 'journal article'. The search served to study two objectives: - To determine the representation of females in studies on the effect of antihypertensive drugs on CVD for the past century. -To study differences and similarities between males and females on the effect of antihypertensive medication on cardiac function and structure.
Moreover, studies had to include a mean with a standard deviation (SD), standard error (SE), or a 95% confidence interval (95% CI) of baseline and follow up measurements of the variables of interest. Variables of primary interest were systolic and diastolic blood pressure (SBP and DBP), mean arterial blood pressure (MAP), and heart rate (HR). Variables of secondary interest were cardiac output (CO), left ventricular ejection fraction (LVEF), and left ventricular mass (LVM). Studies also had to report the mean dose or the dose range, as well as treatment duration. Finally, the antihypertensive treatment had to be compared to a reference group (control, placebo, or another antihypertensive medication group). Mean values with SD were requested from the authors by email if articles presented their data differently. All study designs which reported baseline and follow up measurement were included in this systematic review.

Study Selection
After the initial search, studies were screened based on title and abstract. During this selection, other systematic reviews and meta-analyses, literature reviews, case reports, Biomedicines 2023, 11,1494 3 of 28 animal studies, and in vitro studies were excluded. Moreover, articles in other languages than English and Dutch were excluded. The remaining studies were screened for eligibility based on their full text using the eligibility criteria. Studies were excluded if they did not separate outcomes by antihypertensive medication (if participants received more than one antihypertensive medication as an intervention) or did not report the treatment duration and a mean dose or dose range for the antihypertensive medication. Studies with patients undergoing invasive operations, patients who engaged in exercise during measurements, or patients undergoing dialysis or chemotherapy were excluded as well. In case studies that did not stratify outcomes by sex and all other eligibility criteria were met, authors from articles published in 1980 and later were sent an email or were approached via research gate to request sex-specific data. A reminder was sent after two weeks if authors did not reply. Email addresses from either the first author, corresponding author, or head of the department were retrieved from the article, research gate, or using their name and/or institution to search the internet. If no contact details were found or if authors did not respond within three weeks after being sent a reminder, the article was excluded from this systematic review. The reason for exclusion was registered during the fulltext selection. Both selection steps were performed in pairs in a blinded, standardised manner (title-abstract pairs: MA-EV, CD-SL, EL-DM, ZM-JW, and MV-NW; full-text pairs: CD-NW, EL-MV, DM-SL, and EV-JW). Extractions were performed independently by two investigators. Discrepancies were resolved by dialogue or discussion with a third independent investigator.

Data Extraction
Study characteristics (sample size, control group, and study design), anthropometric data (age and ethnicity), intervention characteristics (dose, duration, and method of measurement), and effect measures (mean and SD at baseline and after beta blockage intervention of SBP, DBP, MAP, HR, CO, LVEF, and LVM) were collected in a predesigned format made by the investigators. The study results were separately extracted for males and females. In this systematic review, only blood pressure (BP) data measured using non-invasive methods was extracted. For the other variables, multiple methods were allowed. Moreover, studies had to report a mean at baseline and post-intervention mean including SD for the outcome variables. Studies that only reported a change from baseline without post-intervention mean and SD were not included. Data extraction was performed by two investigators (RA and LK).

Quality Assessment
The included studies were assessed for quality and risk of bias using the Cochrane recommended Risk of Bias 2 (RoB2) tool [17]. Studies were scored with "Low risk of bias", "Some concerns", or "High risk of bias" on five domains including randomisation process, deviations from intended interventions, missing data, outcome measurements, and data reporting. To receive an overall risk of bias judgement of "Low risk of bias", all domains had to receive this judgement. To receive an overall judgment of "High risk of bias", at least one of the domains was scored as such. All other domain score combinations would rate a study with an overall judgement of "Some concerns". The quality assessment was performed by two reviewers (RA and LK) and differences were solved by a third independent reviewer (DM and SL).

Statistical Analysis
If the SE or 95% CI was reported in the article, the SD was calculated according to the Cochrane Handbook for Systematic Review of Interventions [18]. Changes in the cardiovascular and haemodynamic variables from baseline were separately analysed for males and females using a random-effects model as described by DerSimonian and Laird [19]. Because the included studies had some variation in study population and design, the random-effects model was chosen to account for this interstudy variation [20]. Egger's regression test for funnel plot asymmetry was conducted to test for publication bias for each cardiovascular variable [21]. The primary outcome was the mean difference and 95% CI between baseline and follow up of the intervention, visualised in forest plots. The relative change from baseline in percentage including 95% CI was also calculated and reported in parentheses behind the mean difference in the text. The I2 statistic, the ratio between heterogeneity and variability, was calculated as a measure of consistency and expressed as percentage in the forest plots. I2 is able to distinguish heterogeneity in data from solely sampling variance [20]. Interpretation of I2 was based on the guidelines in the Cochrane Handbook for Systematic Review of Interventions [20]. Sources of clinical heterogeneity (compound, treatment duration, and dosage) and methodological heterogeneity (quality of study) were investigated by meta-regression analyses using a mixed-effects model [20]. For the meta-analyses and meta-regression analyses, the meta package in the statistical programme R version 4.0.3. was used [22,23].

Study Selection
The literature search in PubMed and Embase provided a total of 73,867 unique records after removing duplicates ( Figure 1). During the first screening, 58,737 articles were excluded resulting in 15,130 articles that were assessed based on the full text. Of those articles, 14,916 met at least one exclusion criterium and were excluded. For 766 articles (5%), it was not possible to find or access the full text at the university library or online. A total of 1141 articles (8%) had an unsuitable study design. This criterium was met when for example only measurements were taken during exercise, or SBP and DBP were measured intra-arterial. A total of 1058 articles (7%) did not report original research data; these articles were reviews for example. In 1886 articles (13%), no antihypertensives were given to the patients participating. In 2141 articles (14%), antihypertensives were given but treatment results were not stratified by antihypertensive. A total of 1949 articles (13%) were excluded because treatment results were not stratified by sex. A total of 153 articles (1%) did not have reference measurements. A total of 3864 articles (26%) did not contain any measurements of interest. In 536 articles (4%), data were not suitably reported. In 984 articles (6%), there was no information provided regarding either dose, duration, or both. Finally, 438 articles (3%) were excluded because of other complications. At the end of the selection procedure, a total of 214 articles were classified as suitable for inclusion ( Figure 1). Eventually, in 63 of those articles, BBs were administered as treatment.

Systolic Blood Pressure
Data for all parameters are reported as a mean difference and relative change from the baseline in percentages. SBP data are reported in Table 4 (Table 5).
The mean difference for SBP by treatment duration is reported in Table 6. In females, the observed decrease in SBP is greatest in the sub-acute treatment phase, while the decrease attenuates after prolonged treatment. In males, there is no discernible difference between the different phases ( Figures S1, S2 and 3). The scores 'low', 'some concerns' and 'high', have a green, yellow, and red background respectively.    The mean difference for SBP by treatment duration is reported in Table 6. In females, the observed decrease in SBP is greatest in the sub-acute treatment phase, while the decrease attenuates after prolonged treatment. In males, there is no discernible difference between the different phases ( Figures S1, S2 and 3).    . Meta-regression curve of systolic blood pressure (SBP) by beta-blocker treatment duration (days). Every circle represents one article, and the size represents the number of participants included in the study, shown as a small or larger circle.
The mean difference for DBP by treatment duration is reported in Table 6. The observed decrease in DBP is more pronounced after sub-acute treatment in males than in females and similar in chronic treatment ( Figures S3, S4 and 5).
The mean difference for MAP by treatment duration is reported in Table 6. Chronic treatment showed a similar statistically significant change between females and males ( Figures S6, S7 and 6).
The mean difference for MAP by treatment duration is reported in Table 6. Chronic treatment showed a similar statistically significant change between females and males ( Figures S6 and S7 and 6).

Heart Rate
The weighted pre-intervention mean HR in the female population was 76.  (Table 4, Figure S8), changes were comparable in both sexes (p = 0.759). Corrected for publication bias, the mean difference in males was −12.1 bpm (CI, −13.5; −10.7). The heterogeneity was high in females (I 2 = 98%) as well as in males (I 2 = 78%). There was no clinical source of heterogeneity. There was, however, one methodological source for heterogeneity, "Low quality" (p = 0.027) (Table 5).
Biomedicines 2023, 11, x FOR PEER REVIEW 19 of 30 Figure 6. Meta-regression curve of mean arterial pressure (MAP) change in mmHg by BB treatment duration (days). Every circle represents one article, and the size represents the number of participants included in the study, shown as a small or larger circle.

Heart Rate
The weighted pre-intervention mean HR in the female population was 76.  Table  4, Figure S8), changes were comparable in both sexes (p = 0.759). Corrected for publication bias, the mean difference in males was −12.1 bpm (CI, −13.5; −10.7). The heterogeneity was high in females (I 2 = 98%) as well as in males (I 2 = 78%). There was no clinical source of heterogeneity. There was, however, one methodological source for heterogeneity, "Low quality" (p = 0.027) ( Table 5).
The mean difference for HR by treatment duration is reported in Table 6. The observed decrease in HR is similar in sub-acute and chronic treatment for both sexes ( Figures  S9 and S10).

Cardiac Output
The weighted pre-intervention mean CO in the population was 5.1 [4.8; 5.4] L/min. The decrease in CO in males was 0.1 L/min (95% CI, −0.5; 0.2) (% change, −2.9% (95% CI, −9.2; 3.4)) (Table 4 and Figure 7). This decrease was not statistically significant. For CO, no female data were available and as such, no calculations could be made to compare values between sexes. The effect of metoprolol contributed to the heterogeneity (p = 0.044) ( Table   Figure 6. Meta-regression curve of mean arterial pressure (MAP) change in mmHg by BB treatment duration (days). Every circle represents one article, and the size represents the number of participants included in the study, shown as a small or larger circle.
The mean difference for HR by treatment duration is reported in Table 6. The observed decrease in HR is similar in sub-acute and chronic treatment for both sexes (Figures S9 and S10).

Cardiac Output
The weighted pre-intervention mean CO in the population was 5.1 [4.8; 5.4] L/min. The decrease in CO in males was 0.1 L/min (95% CI, −0.5; 0.2) (% change, −2.9% (95% CI, −9.2; 3.4)) (Table 4 and Figure 7). This decrease was not statistically significant. For CO, no female data were available and as such, no calculations could be made to compare values between sexes. The effect of metoprolol contributed to the heterogeneity (p = 0.044) ( Table 5). The mean difference for CO by treatment duration is again reported in Table 6. Data were exclusively available for males, and both acute and chronic treatments showed non-statistically significant changes ( Figure S11).
Biomedicines 2023, 11, x FOR PEER REVIEW 20 of 30 5). The mean difference for CO by treatment duration is again reported in Table 6. Data were exclusively available for males, and both acute and chronic treatments showed nonstatistically significant changes ( Figure S11).
The mean difference for LVEF by treatment duration is reported in Table 6. Sub-acute data between both sexes was not statistically significant. Chronic treatment, however, showed a similar statistically significant change between females and males (Figures S12 and S13).  [27,28,66,76,80]. MD = mean difference.
The mean difference for LVEF by treatment duration is reported in Table 6. Sub-acute data between both sexes was not statistically significant. Chronic treatment, however, showed a similar statistically significant change between females and males (Figures S12 and S13).

Discussion
This systematic review and meta-analysis primarily studied sex differences in effects on hypertension treatment of BBs on BP, HR, and secondary on cardiac geometry and function. BB treatment significantly but comparably lowers BP and HR in both sexes. LVEF increased in both sexes, showing no significant differences between the sexes. However, the rise reached statistical significance only in males. Cardiac geometry did not change appreciably in males and was not reported in females.
BBs are amongst the five classes of antihypertensives recommended to treat hypertension [16]. They significantly reduce the risk of stroke, heart failure, and major cardiovascular events in hypertensive patients [107]. However, sex differences may negatively impact the efficacy of BBs and their protective effects; the anti-ischaemic effect, remaining functional cardiac reserves and survival in patients with hypertension or coronary artery disease using BBs is less in females than in males [26,[108][109][110].
Pharmacologically, the mechanism by which BBs exert their effects is primarily by obstructing B1 receptors on the cardiac myocytes, inhibiting the stimulating effect of ligands, epinephrine, and norepinephrine and with it cardiac automaticity and conduction velocity, translating into negative chronotropic and inotropic effects. In the context of the circulatory system, the blockage of juxtaglomerular B1 receptors reduces renin output, resulting in decreased angiotensin-driven vascular tone and lower aldosteronedependent volume retention. These concurrent modes of action lower BP and reduce cardiac oxygen demand [111]. Oestrogen and progesterone inhibit the cardiac expression of B1-adrenoreceptors, reduce beta-adrenergic mediated stimulation, and are thought to underlie the cardioprotective effects observed in females of reproductive age [112]. In addition, females have up to twice the peak serum concentration and a higher area under the curve regarding metoprolol and propranolol than males [113,114]. This difference is hypothesised to originate from enhanced bowel absorption, smaller distribution volume, and slower clearance via testosterone-affected CYP2D6 metabolisation in females as compared to males [112,114]. Drug exposure in females using oral contraceptives is significantly higher than in females not using oral contraceptives [113,115]. Related to the higher plasma concentration, some studies suggest females exhibit a larger decrease in HR and SBP under BB therapy than males [113,116]. We observed a comparable BP response to BBs in both sexes, while CO does not seem to change significantly. This suggests that BB-induced lower BP is primarily vascular rather than cardiac driven and originates from a reduction in vascular tone, resistance, and circulatory volume instead of a reduction in CO.
Clinically, previous studies comparing the effects of BBs between females and males in several cardiovascular disease states showed conflicting results. On the one hand, a metaanalysis by Olsson et al., including 4353 males and 1121 females on metoprolol therapy after MI, showed a reduction in overall deaths regardless of sex [117]. On the other hand, in heart failure patients, the Metoprolol CR/XL Randomised Intervention Trial in Congestive Heart Failure (MERIT-HF) and the Carvedilol Prospective Randomized Cumulative Survival (COPERNICUS) trial showed, that in contrast to males, the reduction in mortality for females not to be significant [118,119]. In addition, in a study by Fletcher et al., male patients with hypertension treated with BBs had better survival than female patients, which is in line with the findings of the beta-blocker heart attack trial that showed that BB treatment for patients with coronary artery disease was beneficial for males, but not females [108,109]. Nonetheless, in several post hoc analyses of the aforementioned trials, a similar all-cause mortality and all-cause hospitalisation was found between males and females with heart failure with reduced ejection fraction (HFrEF). In the post hoc analysis of the Cardiac Insufficiency Bisoprolol Study II (CIBIS II) study, bisoprolol showed a significant advantage for females over males using bisoprolol for congestive heart failure [120]. Similar results were found in the post hoc analysis of COPERNICUS, MERIT-HF, and CIBIS trials [121][122][123]. In a meta-analysis by Ghali et al., data from the MERIT-HF, COPERNICUS, and CIBIS-II trials were pooled, resulting in similar survival benefits for females and males using BBs with HfrEF [124]. In the Beta-Blocker Evaluation of Survival Trial (BEST), 2708 patients, from which 593 were females and 2115 were males, with a LVEF ≤ 35%, were randomised to bucindolol versus placebo. This study found that within the group of patients that had a nonischaemic aetiology, females had a better survival rate than males. This was the opposite for the group with an ischaemic aetiology. In this group, the males achieved a better survival than females [125]. In a meta-analysis by Shekelle et al., data from BEST, CIBIS-II, COPERNICUS, MERIT-HF, and U.S. Carvedilol were pooled. The meta-analysis shows a significant reduction in mortality for both males and females [126].
Several differences may underlie these conflicting results. First, BBs were underused in females with myocardial infarction. Second, females are underrepresented in most trials. Only 25% of the total patients in these five trials were females, leading to reduced study power. Third, included females were older and had more severe clinical conditions than included males [112]. Fourth, it might be that the response and outcome of several different underlying cardiovascular conditions, MI, and heart failure with or without preserved ejection fraction, may be different, and, consequently, affect observed outcomes when combined in the analysis. The most recent meta-analysis (2016) included 13,833 patients from 11 trials of which 24% are female. Included patients had HFrEF and sinus rhythm. This study shows that irrespective of sex, BBs compared with placebo reduce mortality and hospital admissions in these patients [127]. Interestingly, there was an excess of heart failure exacerbations in the youngest age quarter of females compared to males [127]. This finding is interesting, especially in light of a recently published study in which patients presenting with acute coronary syndrome were investigated [110]. In this study, it was observed that females using BBs for hypertension had a higher chance of presenting with heart failure at hospital admission compared to males.

Strengths and Limitations
There were several strengths to our systematic review and meta-analysis. To our knowledge, this systematic review and meta-analysis is the first to investigate such an extensive number of studies from inception in a systematic way and aimed to include all relevant studies, which resulted in an analysis of the lack of sex-specific outcome measures in antihypertensive studies. The results of this broad search translated into a large pooled patient population of 2161 patients using BBs and having sex-stratified data available. Another strength was that all articles were screened by two independent reviewers. Furthermore, to detect bias, the Cochrane recommended RoB2 tool was used.
Our study also has several limitations. First, the quality assessment of the articles showed that half of the included articles were rated as "High risk of bias". The Risk of Bias 2 (RoB2) tool classifies non-randomised controlled trials as lower than randomised trial. As many of the included studies were prospective non-randomised cohort studies, by using the RoB2 tool, these studies were viewed upon as not within the highest quality and therefore contributing to the risk of bias. However, we cannot rule out an overestimated risk of bias for these studies as the RoB2 tool is developed for RCTs. Second, in many of the included studies, participants received co-medications, next to their study medication. Although we cannot rule out an additional effect on BP control, reported studies observed no significant differences in co-medication between study groups, suggesting comparable effects on both groups. Since most patients commonly use co-medication, this may even contribute to the external validity of our study. Third, only 20.2%of participants were female, lowering the power of findings as compared to males in our study. Many studies did not present their data in a sex-stratified manner. The 20.2%female inclusion rate in our study is not very different from that in large trials, which is around 25% [112]. This may be partly the result of the historical exclusion of pre-menopausal females from clinical trials [128]. Nonetheless, for our primary outcome, we were able to make reliable findings and, therefore, think that the unequal representation of females as compared to males did not affect our results.

Conclusions and Recommendations
In summary, beta-blockers comparably lowered blood pressure and heart rate in both sexes. The lack of changes in cardiac output in males suggests that the reduction in blood pressure is predominantly determined by a reduction in vascular resistance. Furthermore, females were underrepresented compared to males. Future research should have a better female-to-male inclusion ratio and sex stratified data when researching treatment effects of hypertensives.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/biomedicines11051494/s1, Table S1: Literature search: strategy for PubMed (NCBI) and Embase (Ovid) databases. Figure S1: Forest plot of systolic blood pressure (SBP) change in mmHg after sub-acute and chronic beta-blocker use compared to baseline for females. MD = mean difference. Figure S2: Forest plot of systolic blood pressure (SBP) change in mmHg after acute, sub-acute and chronic beta-blocker use compared to baseline for males. MD = mean difference. Figure S3: Forest plot of diastolic blood pressure (DBP) change in mmHg after sub-acute and chronic beta-blocker use compared to baseline for females. MD = mean difference. Figure S4: Forest plot of diastolic blood pressure (DBP) change in mmHg after acute, sub-acute and chronic beta-blocker use compared to baseline for males. MD = mean difference. Figure S5: Forest plot of mean arterial pressure (MAP) change in mmHg after beta-blocker use compared to baseline for females and males. MD = mean difference. Figure S6: Forest plot of mean arterial pressure (MAP) change in mmHg after acute, sub-acute and chronic beta-blocker use compared to baseline for females. MD = mean difference. Figure S7: Forest plot of mean arterial pressure (MAP) change in mmHg after acute, sub-acute and chronic beta-blocker use compared to baseline for males. MD = mean difference. Figure S8: Forest plot of heart rate (HR) change in beats per minute (bpm) after beta-blocker use compared to baseline for females and males. MD = mean difference. Figure S9: Forest plot of heart rate (HR) change in beats per minute (bpm) after sub-acute and chronic beta-blocker use compared to baseline for females. MD = mean difference. Figure S10: Forest plot of heart rate (HR) change in beat per minute (bpm) after acute, sub-acute and chronic beta-blocker use compared to baseline for males. MD = mean difference. Figure S11: Forest plot of cardiac output (CO) change in liter per minute (L/min) after acute and chronic beta-blocker use compared to baseline for males. MD = mean difference. Figure S12: Forest plot of left ventricular ejection fraction (LVEF) change in % after sub-acute and chronic beta-blocker use compared to baseline for females. MD = mean difference. Figure S13: Forest plot of left ventricular ejection fraction (LVEF) change in % after acute, sub-acute and chronic beta-blocker use compared to baseline for males. MD = mean difference. Figure S14: Forest plot of left ventricular mass (LVM) change in grams after beta-blocker use compared to baseline for females and males. MD = mean difference.