Unraveling Potential Sex‐Specific Effects of Cardiovascular Medications on Longevity Using Mendelian Randomization

Abstract

Background

Establishing the sex‐specific efficacy of cardiovascular medications is pivotal to evidence‐based clinical practice, potentially closing the gender gap in longevity. Trials large enough to establish sex differences are unavailable. This study evaluated sex‐specific effects of commonly prescribed cardiovascular medications on lifespan.

Methods and Results

In a two‐sample Mendelian randomization study, established genetic variants mimicking effects of lipid‐lowering drugs, antihypertensives, and diabetes drugs were applied to genetic associations with lifespan proxied by UK Biobank maternal (n=412 937) and paternal (n=415 311) attained age. Estimates were obtained using inverse variance weighting, with sensitivity analyses where possible. For lipid‐lowering drugs, genetically mimicked PCSK9 (proprotein convertase subtilisin/kexin type 9) inhibitors were associated with longer lifespan, particularly in men (2.39 years per SD low‐density lipoprotein cholesterol reduction [95% CI, 0.42–4.36], P for interaction=0.14). Genetically mimicked treatments targeting APOC3, LPL, or possibly LDLR were associated with longer lifespan in both sexes. For antihypertensives, genetically mimicked β‐blockers and calcium channel blockers were associated with longer lifespan, particularly in men (P for interaction=0.17 for β‐blockers and 0.31 for calcium channel blockers). For diabetes drugs, genetically mimicked metformin was associated with longer lifespan in both sexes. No associations were found for genetically mimicked statins, ezetimibe, or angiotensin‐converting enzyme inhibitors.

Conclusions

PCSK9 inhibitors, β‐blockers, and calcium channel blockers may prolong lifespan in the general population, particularly men. Treatments targeting APOC3, LPL, or LDLR and metformin may be relevant to both sexes. Whether other null findings are attributable to lack of efficacy requires investigation. Further investigation of repurposing should be conducted.

Nonstandard Abbreviations and Acronyms

IHD

ischemic heart disease

MR

Mendelian randomization

PCSK9

proprotein convertase subtilisin/kexin type 9

SBP

systolic blood pressure

Clinical Perspective.

What Is New?

• Our study provides genetic evidence that the lipid‐lowering drugs, PCSK9 (proprotein convertase subtilisin/kexin type 9) inhibitors, and some antihypertensives (β‐blockers and calcium channel blockers) increase lifespan in men but not women.

• Potential lipid‐lowering therapies targeting APOC3, LPL, or possibly LDLR and metformin may also increase lifespan in both sexes.

What Are the Clinical Implications?

• Our findings highlight the importance of development of sex‐specific cardiovascular medications to promote lifespan and address disparities by sex.

Global advocacy for healthy aging calls for closing the gender gap in longevity (lifespan). ¹ , ² Sex differences in lifestyle and health care–seeking behavior likely explain some of the difference. ³ , ⁴ Nevertheless, the importance of accounting for physiological differences between women and men is recognized as an understudied, but crucial, area of clinical research. ⁵ Preventing cardiovascular disease (CVD) deaths is vital to reducing lifespan disparities. ⁶ Identifying drugs with potentially sex‐specific effects could accelerate drug development to promote longevity, such as polypills specific to men and women, thereby contributing to closing the gender gap in longevity globally. Common cardiovascular medications, including lipid‐lowering drugs, antihypertensives, and diabetes drugs, are essential to primary health care. ⁷ Despite being effective in secondary prevention, whether commonly used cardiovascular medications, such as statins, ⁸ , ⁹ antihypertensives, ¹⁰ and metformin, ¹¹ , ¹² , ¹³ reduce all‐cause mortality by sex in the general population is unclear.

In this situation where definitive evidence for primary prevention of all‐cause mortality is lacking, genetic validation of pharmaceutical targets using Mendelian randomization (MR) is increasingly used. ¹⁴ , ¹⁵ MR studies minimize confounding, by taking advantage of the random allocation of genetic variants at conception, which is less affected by socioeconomic position and related attributes, and precedes the onset of disease. ¹⁶ Existing drug target MR studies have rarely considered drug effects on lifespan overall or whether the associations differ by sex. A previous drug target MR study in the general population, rather than people at increased cardiovascular risk, found genetic mimics of PCSK9 (proprotein convertase subtilisin/kexin type 9) inhibitors associated with longer lifespan, whereas genetic mimics of statins and ezetimibe had a similar direction of effect, with CIs including the null, ¹⁷ but the study did not consider men and women separately. To assess whether widely prescribed cardiovascular medications and novel potential cardiovascular treatments enhance lifespan in the general population for men and women, this drug target MR study assessed genetically mimicked effects of lipid‐lowering drugs, antihypertensives, and diabetes drugs on lifespan in women and men. We used ischemic heart disease (IHD) and type 2 diabetes as positive control outcomes.

Methods

Study Design and Setting

The authors declare that the supporting data are available within the article and its online supplementary files or have been made publicly available by the UK Biobank or genome‐wide association study stated in the article. This study is largely a two‐sample MR study using genetic mimics for lipid‐lowering drugs, ¹⁸ , ¹⁹ antihypertensives, ²⁰ and diabetes drugs, ²¹ established in previous drug target MR studies in people of European descent. Lifespan in men and women of European descent (proxied by paternal and maternal attained age) was taken from publicly available sex‐specific genetic summary statistics from the UK Biobank, which recruited half a million people, intended aged 40 to 69 years, from Great Britain during 2006 to 2010. MR has 3 fundamental assumptions: relevance, independence, and exclusion‐restriction. Relevance for drug target studies requires that the genetic variants within the protein‐encoding genes of the drug target are associated with the biomarker targeted by the drug so as to proxy the effects of drug use. Independence requires the genetic variants are unrelated to confounders of the exposure‐outcome association. Exclusion‐restriction means the genetic variants are independent of the outcome given the exposure. ²² Genetic mimics for the cardiovascular medications were based on large and extensively genotyped genome‐wide association studies of relevant biomarkers from samples largely distinct from the UK Biobank for lipid‐lowering drugs, and antihypertensives, whereas mimics for diabetes drugs were from the UK Biobank. ²³ This study was reported according to Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization statement (https://www.strobe‐mr.org/).No ethical approval was required for this study using publicy available summary statistics. Each of the studies included in this investigation had been specifically approved by the Ethics Committees of the original studies, and the participants provided written informed consent; their ethics approval statements can be found in the original publications.

Exposures

Genetically Mimicked Existing and Potential Cardiovascular Medications

Established largely independent genetic variants from the relevant genes for people of European descent were used for each medication. ¹⁸ , ¹⁹ , ²⁰ , ²¹ Associations with the corresponding targets were obtained for the largest available genome‐wide association study apart from the UK Biobank without adjustment for any heritable covariates, such as body mass index. ²⁴ , ²⁵ , ²⁶

Outcomes

Genetic Associations With Lifespan

The main outcome was lifespan in women and men. Lifespan was proxied by maternal and paternal attained age (current age or age at death) for people of White European descent (women=412 937; men=415 311) from the UK Biobank ²⁷ given longevity is partly heritable. ²⁸ Premature parental deaths (mothers, <57 years; and fathers, <46 years) and adopted participants were excluded. ²⁷ Lifespan was previously estimated from parental attained age using Cox proportional hazards regression with Martingale residuals adjusted for age, sex, assessment center, and array type using Bayesian mixed model associations to account for relatedness among participants. The estimates provided were converted into years of life by approximately doubling (women=2.5863 years; and men=2.2869 years) ²⁹ to account for offspring only sharing half their genetic endowment with each parent, and then multiplying by 10 as a verified actuarial rule of thumb. ³⁰

Genetic Associations With IHD and Type 2 Diabetes

The positive control outcomes were IHD and type 2 diabetes. Genetic associations with IHD in women (cases=5801; controls=188 373) and men (cases=15 056; controls=151 964) were obtained from the UK Biobank ³¹ ; and those with type 2 diabetes in women (cases=30 053; controls=434 336) and men (cases=41 846; controls=383 767) were obtained from people of European descent of the Diabetes Meta‐Analysis of Trans‐Ethnic Association Studies (DIAMANTES). ³² Estimates of associations with IHD from the UK Biobank provided as probabilities were converted into odds ratios using an established approximation. ³³

Statistical Analysis

To assess instrument strength (ie, relevance), the univariable F‐statistics were approximated by the squared single‐nucleotide polymorphism (SNP)–exposure association divided by its variance. ³⁴ An F‐statistic <10 indicates a potentially weak instrument. SNPs predicting an exposure that were unavailable for the outcome were replaced by proxies (r ² ≥ 0.8), where available. ³⁵ Palindromic SNPs (coded A/T or C/G) with minor allele frequency >0.42 were aligned on effect allele frequency where possible, and those with noninferrable forward strands were excluded. All available SNPs were aligned on the same effect allele for exposure and outcome. To facilitate comparisons, exposures are standardized (SD).

Inverse variance–weighing estimates were obtained from the Wald estimate (genetic variant on outcome divided by genetic variant on exposure) for 1 SNP or a meta‐analysis of Wald estimates with multiplicative random effects for ≥2 SNPs. To assess possible horizontal pleiotropy (ie, violation of exclusion‐restriction), as sensitivity analyses, 3 complementary methods with different assumptions for validity were used where possible. The weighted median estimate requires at least 50% of the information to be from valid SNPs. ³⁶ The MR‐Egger estimate allows all SNPs to be invalid as long as the Instrument Strength Independent of Direct Effect assumption holds (ie, no genetic confounding exists). The MR‐Egger intercept with P<0.05 indicates the presence of pleiotropy. ³⁴ Estimates (change in outcome per SD lowering in the relevant drug target) with 95% CIs and the MR‐Egger intercept P value with I² _GX are presented. ³⁷ To evaluate if the MR estimates differed by sex, we conducted tests of interaction and considered the direction and magnitude of sex‐specific effect estimates given the P for interaction might not necessarily detect differences even when the estimates seem different. For the interaction test, we derived the z‐statistic based on differences in the sex‐specific estimates and then obtained the 2‐tailed P value, with P<0.05 indicating the presence of a sex difference. Given the previously reported null estimate for genetically mimicked statins on lifespan, ¹⁷ a power calculation (α=0.05; and power=0.8) was performed for statins on lifespan in women and men based on the approximation that the sample size for an MR study is the sample size for exposure on outcome divided by the variance of the exposure explained by the genetic instruments, using an online tool (https://shiny.cnsgenomics.com/mRnd/). The variance for the genetic mimics of statins on low‐density lipoprotein (LDL) cholesterol was obtained using an approximation. ³⁸ Furthermore, as secondary analyses, we included correlated SNPs (ie, both primary and alternative SNPs for the mimics of drug use) as genetic instruments allowing for their correlations.

Statistical analyses were conducted using R, version 4.2.1 (The R Foundation for Statistical Computing, Vienna, Austria) with the R packages: TwoSampleMR ³⁵ for data extraction (if available), harmonization, and obtaining correlations between SNPs; MendelianRandomization for MR estimation; and metafor for sex difference detection.

Results

Genetic Instruments for Mimicking Cardiovascular Medications

Independent genetic variants in the protein‐encoding genes targeted by the cardiovascular medications, as previously published, were used to mimic drug effects, as shown in Tables S1 to S3. Corresponding correlated variants were used in a sensitivity analysis, accounting for their correlations. Genetically mimicked statins explained 0.22% of the variance in LDL cholesterol and had 80% power to detect a difference in life years of 0.68 in women and men per SD reduction in LDL cholesterol.

Associations of Genetically Mimicked Cardiovascular Medications With Lifespan in Women and Men

All the F‐statistics were adequate for the genetic mimics of cardiovascular medications (Table S4). Of the existing lipid‐lowering drugs, genetically mimicked PCSK9 inhibitors were associated with longer lifespan in men (2.39 years per 1‐SD reduction in LDL cholesterol [95% CI, 0.42–4.36]), but not women (Figure [A]), although the sex difference was not significant (P for interaction=0.14). No clear associations of genetically mimicked statins or ezetimibe with lifespan were found (P for interaction=0.94 for statins and 0.10 for ezetimibe) (Figure [A]). For potential lipid‐lowering treatments, genetically mimicked therapies targeting APOC3 or LPL were associated with longer lifespan in both sexes (Figure [A]), with no sex difference for APOC3 but a greater magnitude of effect for targeting LPL in men than women (P for interaction=0.81 for APOC3, 0.87 for LDLR, and 0.02 for LPL). Using correlated SNPs taking their correlations into account showed generally similar patterns of association (Table S4).

Figure . Mendelian randomization (MR) results for cardiovascular medications with lifespan by sex.

The figure displays the genetically mimicked effects of lipid‐lowering drugs (A), antihypertensives (B), and diabetes drugs (C) on lifespan in women and men. ACE indicates angiotensin‐converting enzyme; APOC3, apolipoprotein C3; F‐stat, F‐statistic; HbA1c, glycated hemoglobin; IVW, inverse variance weighting; LDL, low‐density lipoprotein; LDLR, LDL receptor; LPL, lipoprotein lipase; PCSK9, proprotein convertase subtilisin/kexin type 9; P‐interact, P value for sex interaction; SBP, systolic blood pressure; SNP, single‐nucleotide polymorphism; and WM, weighted median.

For antihypertensives, genetically mimicked β‐blockers and calcium channel blockers were associated with longer lifespan in men but not women (Figure [B]), although the sex difference was not statistically significant (P for interaction=0.17 for β‐blockers and 0.31 for calcium channel blockers). Sensitivity analyses for calcium channel blockers using the weighted median and MR‐Egger gave similar findings, with no indication of possible horizontal pleiotropy from the MR‐Egger intercept (Figure [B]). No association of genetically mimicked angiotensin‐converting enzyme (ACE) inhibitors with lifespan was found in women or men (P for interaction=0.80) (Figure [B]). After using correlated SNPs, taking correlations into account also showed similar patterns of association (Table S5).

For diabetes drugs, genetically mimicked metformin was associated with longer lifespan in women and men using inverse variance weighting (Figure [C]), with no sex difference (P for interaction=0.81). Sensitivity analyses using the weighted median and MR‐Egger showed similar findings, with no indication of horizontal pleiotropy from the MR‐Egger intercept (Figure [C]).

Associations of Genetically Mimicked Cardiovascular Medications With IHD and Type 2 Diabetes as Positive Control Outcomes

Genetically mimicked drugs were associated with IHD, type 2 diabetes, or both in both sexes as expected (Tables S6–S8), apart from the fact that ACE inhibitors and β‐blockers were not clearly related to IHD, and β‐blockers were associated with a higher risk of diabetes in women.

Discussion

Principal Findings

This first MR study concerning sex‐specific effects of genetically mimicked cardiovascular medications on longevity, equating to drug use in primary prevention to enhance lifespan, found an existing lipid‐lowering drug (PCSK9 inhibitors) and 2 antihypertensives (β‐blockers and calcium channel blockers) were associated with longer lifespan, particularly in men. Potential lipid‐lowering therapy targeting APOC3, LPL, or possibly LDLR and metformin was associated with longer lifespan in both sexes, with targeting LPL possibly having a greater effect in men than women. Conversely, statins, ezetimibe, and ACE inhibitors were less clearly related to lifespan in women or men. All existing or potential lipid‐lowering therapies affected IHD risk, diabetes risk, or both as expected, except ACE inhibitors and β‐blockers. Apart from highlighting the potential for repurposing to promote longevity, our findings highlight the importance of accounting for sex in drug development and prescription.

Comparison With Other Studies

Consistent with a previous MR study of lipid‐lowering drugs on lifespan in both sexes, ¹⁷ genetically mimicked effects of PCSK9 inhibitors were associated with longer lifespan, but genetically mimicked statins and ezetimibe had no such association. Currently available evidence on PCSK9 inhibitors in primary prevention is minimal. Our findings are consistent with Cochrane reviews mainly in patients at high risk or with established CVD showing protective effects of PCSK9 inhibitors (alirocumab but not evolocumab) ³⁹ on all‐cause mortality but not ezetimibe (with statin or fenofibrate). ⁴⁰ Given the 2 large trials (ODYSSEY OUTCOMES [Evaluation of Cardiovascular Outcomes After an Acute Coronary Syndrome During Treatment with Alirocumab] and FOURIER [Further Cardiovascular Outcomes Research with PCSK9 Inhibition in Subjects with Elevated Risk]) as part of the Cochrane review did not show improved mortality in secondary prevention, ³⁹ we add by showing that beneficial effects of genetically mimicked PCSK9 inhibitors on lifespan are likely specific to men for primary prevention rather than secondary prevention. Our findings for statins, similar to the previous MR study in both sexes combined, ¹⁷ are consistent with the meta‐analysis of statin trials showing statins unrelated to all‐cause mortality at low baseline risk (indicated by 5‐year predicted risk of major coronary events of <5%), ⁸ , ⁹ and a null effect of statins in primary prevention. ⁴¹ As such, benefits of statins may not be evident in the general population for primary prevention, given the large number needed to treat to prevent 1 major cardiovascular event, ⁹ and substantially larger numbers to prevent 1 death. In addition, most statin trials are industry sponsored, and the data curated by the Cholesterol Treatment Trialists' Collaboration remain inaccessible to academia, ⁴² leaving replicability and impartiality of trial results uncertain.

Genetically mimicked effects of β‐blockers and calcium channel blockers, but not ACE inhibitors, were associated with longer lifespan. These findings are consistent with a meta‐analysis on antihypertensives in both sexes that reported different drug classes had broadly similar effects on reducing all‐cause mortality, although only calcium channel blockers might lower the risk, ACE inhibitors had a null effect, and β‐blockers had a negative effect. ⁴³ However, a Cochrane review of head‐to‐head trials showed calcium channel blockers were no different from other drug classes in terms of all‐cause mortality reduction among participants with hypertension. ⁴⁴ We add by showing that beneficial effects of genetically mimicked β‐blockers and calcium channel blockers on lifespan were particularly evident in men.

Genetically mimicked effects of metformin were associated with longer lifespan in both sexes, inconsistent with the follow‐up study of the DPP (Diabetes Prevention Program) trial that observed no change in cardiovascular events or deaths, ¹² which may be subject to confounding. We add by showing that the beneficial effects of genetically mimicked metformin were similar in women and men.

For positive control outcomes, our findings corroborate previous trials ⁸ , ¹¹ , ⁴⁵ , ⁴⁶ and MR studies ⁴⁷ , ⁴⁸ , ⁴⁹ , ⁵⁰ that genetically mimicked cardiovascular medications were generally associated with IHD, diabetes, or both as expected.

Possible Mechanisms

Sex‐specific effects of PCSK9 inhibitors, β‐blockers, and calcium channel blockers may be attributable to the higher risk of cardiovascular mortality in men than women, but it is not clear why similar sex‐specific effects for potential cardiovascular therapies targeting APOC3, LPL, and possibly LDLR were not evident. PCSK9 inhibitors that lower IHD risk in men do not pose diabetes risk, ³⁹ and so may promote lifespan particularly in men. The circulating plasma level of PCSK9, which increases LDL cholesterol, was noted to be higher in women than men, ⁵¹ suggesting better drug response to PCSK9 inhibitors in men, ⁵² coherent with our sex‐specific findings. β‐Blockers pose higher diabetes risk, particularly in women. Both β‐blockers and calcium channel blockers have long been associated with modulating male hormones, ⁵³ , ⁵⁴ which could be relevant to an effect specific to men, with corresponding implications for major causes of mortality. ⁵⁵ Genetically mimicked metformin targeting diabetes had similar effects by sex, consistent with similar metformin efficacy in diabetes reduction in men and women. ¹¹ Metformin has long been of interest as a possible antiaging agent because it has properties similar to a long‐standing target of antiaging research (ie, calorie restriction), and may operate via an inhibitor of a key player in aging, the mechanistic target of rapamycin target complex 1. ⁵⁶ However, the role of metformin has not been established in large trials, but a trial is now getting underway. ⁵⁷ Further drug‐specific investigations are warranted to elucidate the exact mechanisms.

Future Directions

Replication using more functionally relevant genetic instruments for the cardiovascular medications in the future is needed, given relevant genetic mimics for drugs may not have been fully identified. Sex differences in disease vulnerability and length of life undoubtedly occur for a variety of reasons. Here, we focused on cardiovascular and related medications as a potential means of informing this disparity, because CVD is a key contributor to these differences in lifespan. Our findings, if substantiated by triangulation by randomized controlled trials and future drug target MR studies with more functionally relevant SNPs, when available, would foster the drug development specifically for men and women, so as to promote lifespan and reduce disparities.

Strengths and Limitations

Our sex‐specific MR study assessed lifelong effects of major classes of lipid‐lowering drugs, antihypertensives, and diabetes drugs to inform long‐term use of cardiovascular medications to promote real‐world lifespan in the general population (ie, primary prevention at all risk levels), to complement the limited information on this topic from trials. Despite its strengths, this study has limitations. First, the instruments for genetic mimics of a drug target might be weak. We used previously established genetic instruments for cardiovascular medications. These genetic mimics were strongly related to the corresponding biomarkers. Also, all genetic mimics included had F‐statistic >10, which indicates a low possibility of weak instrument bias. ⁵⁸ However, our genetic instrument selection depends on knowledge of protein drug targets, and might have only partially captured all the on‐target effects, especially for drugs (like metformin) with incomplete understanding of the mechanism of action. Moreover, existing drugs for IHD have been suggested to be not necessarily targeting the products of genes that are strongly associated with IHD. ⁵⁸ Nevertheless, the association of genetically mimicked effects of cardiovascular medications with the positive control outcomes of IHD, diabetes, or both were as expected. Second, MR estimates represent lifetime exposures, so for drug targets inevitably include both the nonexposed and exposed periods of drug use. Information on duration of genetically predicted drug use is not available, precluding comprehensive adjustment for drug use; as a result, estimates for commonly used drugs, such as statins, diuretics, and metformin, may be underestimated. In addition, given genetically mimicked drug use is lifelong, whereas use of drugs to prevent and treat chronic disease usually starts in midlife, any off‐target effects of genetically mimicked drugs on lifespan will be lifelong. Notably, statins have several harmful off‐target effects, such as increasing body mass index and basal metabolic rate, ⁵⁹ which reduce lifespan, ⁶⁰ and so would bias estimates for genetically mimicked statins toward the null by including harmful lifelong effects of higher body mass index and basal metabolic rate. Third, lifespan was proxied by parental attained age rather than actual observations. However, longevity is a heritable trait, ²⁸ like other phenotypic traits (eg, susceptibility to chronic disease risk [eg, IHD]). This study assumes the cause composition of mortality is relatively stable, so that findings on the lifespan of people born in the first half of the 20th century are relevant today, which is an inherent limitation of studying longevity. Fourth, genetic mimics of a drug target might affect the outcome via the off‐target biomarkers. We used complementary sensitivity analyses, where possible, which gave a consistent interpretation. Fifth, sample overlap may bias toward the confounded association. ⁶¹ However, our samples for exposure and outcome generally have minimal overlap. Finally, given lifespan was based on the UK Biobank participants, who are healthier than the underlying population, ⁶² our findings are applicable to generally healthy people of European descent only, considering the relevance of a causal factor may vary by setting. ⁶³ Explicating the mechanisms would complement studies assessing relevance in the rest of the global populations.

Clinical Implications

Statins are the mainstay for secondary CVD prevention in patients and those at higher risk. Their role in primary prevention depends on whether enough of the general population would benefit. The UK National Institute for Health and Care Excellence recently recommended wider use of statins for primary prevention of CVD, ⁶⁴ and the US Preventive Services Task Force recommends statin use for people with a ≥10% predicted 10‐year risk of developing CVD. ⁶⁵ Our study accentuates the importance of balancing the long‐term benefits and harms of statin use and the number needed to treat to prevent 1 death in the context of considering medications for use in the general population to prevent aging. Similarly, whether statins should be included in polypills for primary prevention of CVD should involve similar considerations. ⁶⁶

From a clinical perspective, our study suggests some existing commonly used cardiovascular medications prescribed for all patients may be more beneficial to men than women. This also helps clarify the role of cardiovascular medications independent of lifestyle modification in primary prevention given we only used the drug targets aimed at modifying biomarkers, such as LDL cholesterol levels that are specifically attributable to lipid‐lowering drugs rather than LDL cholesterol levels that are collectively attributable to lifestyle modification, lipid‐lowering drugs, or both. Our findings highlight the importance of development of sex‐specific cardiovascular medications to promote lifespan and address disparities by sex. These findings are highly relevant to drug regulatory authorities and pharmaceutical companies to identify promising drug targets for cardiovascular therapy by sex. Sex‐specific drug therapy enables clinicians to make informed decisions with patients in prescription of cardiovascular medications, contributing to promoting longevity for all.

Conclusions

Lipid‐lowering drugs (PCSK9 inhibitors) and antihypertensives (β‐blockers and calcium channel blockers) may prolong lifespan in men more than in women. Potential lipid‐lowering therapies targeting APOC3, LPL, or possibly LDLR and metformin may be beneficial to both sexes. Our sex‐specific MR findings should encourage more sex‐specific drug discovery, investigation, and development, to promote lifespan equity in women and men.

Sources of Funding

This study was supported by the Healthy Longevity Catalyst Awards (Hong Kong) by the US National Academy of Medicine (number 0200010393).

Disclosures

None.

Supporting information

Tables S1–S8.