Integrative Biomedical Research

1. Introduction

If precision medicine has a founding promise, it is this: that treatment should be tailored to the person in front of the clinician, not to an abstraction of one. And yet, somewhat paradoxically, the field has been slow to apply that promise to one of the most obvious sources of biological variation we have — sex. For much of modern pharmacology's history, clinical practice has operated, often without saying so out loud, under a template that treats the average male body as the default anatomical and physiological standard (Gualtierotti, 2025; Rakshith et al., 2023). This is not a trivial oversight. It has, over decades, obscured one of the most consequential variables shaping how well a drug works, how safe it is, and how tolerable it feels to take — namely, the interacting distinctions between male and female physiology on one hand, and the sociocultural dynamics of gender on the other (Dattolo et al., 2026; Lucena et al., 2025). The evidence for this is no longer circumstantial. Pharmacovigilance data, accumulating steadily and now spanning millions of individual reports, show that women experience meaningfully higher rates of adverse drug reactions (ADRs) and altered efficacy across a striking range of drug classes — from chemotherapeutics to psychotropic medications (Baraibar et al., 2023; Lee et al., 2023). If clinical outcomes are to genuinely improve, sex and gender need to be threaded through the entire translational pipeline, from the earliest preclinical bench experiments through to decisions made at the bedside (Gualtierotti, 2025; Lucena et al., 2025).

Part of what has slowed this integration, we suspect, is something almost definitional: the persistent conflation of "sex" and "gender" in the medical literature, terms that get used interchangeably despite referring to genuinely different things. Sex, properly understood, is a biological variable — it is chromosomal makeup (XX versus XY), it is the relative balance of sex hormones such as estrogens, progesterone, and androgens, and it is reproductive anatomy (Dattolo et al., 2026; Gualtierotti, 2025). Gender, on the other hand, is a social construct: roles, behaviors, self-expression, identity, and the institutional distributions of power that shape how a person moves through a healthcare system, all of which vary depending on historical moment and cultural setting (Lee et al., 2023; Morgan et al., 2024). In an actual living patient, of course, these two dimensions do not sit neatly apart — they interact continuously, which makes disentangling their independent and combined effects on health outcomes both genuinely difficult and, we would argue, genuinely necessary (Rakshith et al., 2023; Tran & Eder, 2026). When clinicians and researchers fail to separate the two, or fail to account for both, the downstream consequences are not abstract: inaccurate risk stratification, dosing regimens that are optimized for one physiology but applied to all, and toxicities that could plausibly have been anticipated (Lucena et al., 2025).

Biological sex differences, to start with the more mechanistic half of the story, drive substantial variation in pharmacokinetics (PK) and pharmacodynamics (PD) (Rakshith et al., 2023; Soldin & Mattison, 2009). Pharmacokinetically, women tend to carry a higher percentage of body fat, operate with a somewhat lower glomerular filtration rate (GFR), empty the stomach more slowly, and maintain different concentrations of carrier proteins such as albumin — a cluster of differences that, taken together, shapes the entire liberation-absorption-distribution-metabolism-excretion (LADME) sequence a drug moves through (Lucena et al., 2025). Diazepam and other lipophilic drugs, for instance, linger longer in women because of greater adipose storage, while drugs cleared primarily through the kidneys tend to clear more slowly given the lower GFR (Lucena et al., 2025). At the metabolic level — and this is where things get genuinely interesting — sex-specific variation in cytochrome P450 (CYP450) enzyme activity is quite pronounced: women show higher baseline CYP3A4 activity, which speeds clearance of substrates like cyclosporine and erythromycin, whereas men run higher on CYP1A2, clearing compounds such as caffeine more quickly (Lucena et al., 2025).

This mechanistic story becomes clinically vivid in oncology, where narrow therapeutic windows leave little room for miscalibration (Baraibar et al., 2023; Marcu, 2022). Women treated with 5-fluorouracil (5-FU), a backbone chemotherapy for colorectal cancer, experience markedly higher rates of severe hematological and mucosal toxicity — a pattern traced to lower dihydropyrimidine dehydrogenase (DPD) activity and correspondingly decreased plasma clearance relative to men (Baraibar et al., 2023; Rakshith et al., 2023). Cisplatin tells a related but not identical story: nephrotoxicity and neurotoxicity both show sex-dependent profiles, with renal transporters like OCT2 running higher in men (and tracking with nephrotoxicity there), while women's higher glutathione S-transferase activity appears to prolong drug retention in target tissues (Marcu, 2022).

Beyond pharmacokinetics, pharmacodynamic response is itself modulated by hormonal and genetic factors that interact directly with drug targets. Estrogens and androgens shape downstream cellular pathways in ways that go well beyond reproductive tissue — autophagy, for example, is regulated in part through estrogen receptor-beta (ERβ) signaling, which behaves as a tumor suppressor, in contrast to estrogen receptor-alpha (ERα), which more often drives pro-tumor pathways (Vona et al., 2025). In neuropsychopharmacology, the so-called "estrogen hypothesis" proposes that fluctuating reproductive hormone levels modulate neurotransmitter receptor sensitivity and density; an individual patient data meta-analysis of acute mania trials found antipsychotics and mood stabilizers performed statistically better in men and postmenopausal women than in premenopausal women, suggesting that high circulating estrogen may raise the clinical threshold for treatment response (Storosum et al., 2025). A related pattern surfaces in adolescents with early-onset schizophrenia, where younger patients show notably higher rates of antipsychotic-induced adverse events — a reminder that age, developmental stage, and hormonal flux intersect in ways that are not yet fully mapped (Storosum et al., 2026). Even in preclinical models of neuropathic pain, sexual dimorphism shows up clearly: subchronic metformin produces durable analgesia in male mice but only fleeting relief in females, implying that the underlying neuroimmune pathways are, at least in part, sex-specific (De Angelis et al., 2022).

Genetics adds another layer. Women carry two X chromosomes, and although X-chromosome inactivation randomly silences one copy to balance gene dosage, roughly 15% of X-linked genes escape this silencing — a fair number of which encode immune and inflammatory mediators such as Toll-like receptors 7 and 8 (TLR7, TLR8) and forkhead box P3 (FOXP3) (Baraibar et al., 2023; Ruggieri & Anticoli, 2025; Vona et al., 2025). This gene-dosage effect appears to grant women more robust innate and adaptive immune responses — which helps explain higher vaccine efficacy, but also a considerably higher burden of autoimmune and inflammatory conditions, and a greater likelihood of severe drug hypersensitivity reactions (Ruggieri & Anticoli, 2025; Ventura et al., 2025).

Yet biology is only half the picture. Sociocultural gender variables shape the real-world trajectory of drug safety reporting in ways that are easy to underestimate (Lee et al., 2023). The "Gender Hypothesis," as proposed by Lee et al. (2023), attributes much of the aggregate sex disparity documented in pharmacovigilance databases to four gendered pathways: differential healthcare utilization, clinical bias, subjective perception of what counts as an adverse event, and broader social determinants of health. Women engage with primary care and receive prescriptions at considerably higher rates than men, which mechanically raises their cumulative exposure and, with it, their probability of reporting an adverse event (Lee et al., 2023). Gender stereotypes at the clinical interface compound this: women's physical complaints are too often psychologized, leading to misdiagnosis, delayed referral, or prescribing that misses the mark (Gualtierotti, 2025; Lee et al., 2023). Men, meanwhile, are shaped by different norms — masculinity scripts that discourage symptom disclosure, which skews trial and registry data toward understating toxicity in male cohorts (Baraibar et al., 2023; Lee et al., 2023). Nephrology offers a tidy illustration of how sex and gender intertwine: autosomal dominant polycystic kidney disease progresses faster to end-stage renal disease in men, a pattern attributed to testosterone's role in amplifying oxidative stress and activating the renin-angiotensin-aldosterone system, counterbalanced by estrogen's protective upregulation of the ACE2/Ang-(1-7) pathway (Dattolo et al., 2026).

Despite all this accumulating evidence, women remain underrepresented in clinical trials, particularly in early-phase safety studies, and preclinical drug development still leans heavily on male animal models — a bias historically justified by concerns about estrous-cycle variability that have since been empirically discredited (Gualtierotti, 2025; Lucena et al., 2025; Rakshith et al., 2023). The result is a therapeutic landscape still largely built on a "one-drug-fits-all" logic that ignores both physiological and sociocultural realities (Baraibar et al., 2023; Lucena et al., 2025). Moving past this will require, we think, deliberate structural change: mandated sex-stratified analysis, representative trial enrollment, and systematic reporting of gender-related variables, woven into research design rather than appended as an afterthought (Gualtierotti, 2025; Morgan et al., 2024). This review attempts to draw that evidence together — mechanistic and sociocultural alike — into a single, coherent account of why sex and gender deserve a permanent seat at the precision-medicine table.

2. Beyond the "One-Size-Fits-All" Paradigm: A Systematic Literature Review of Sex and Gender as Core Variables in Health and Pharmacotherapy

2.1 Drawing the Line: Why Sex and Gender Are Not the Same Variable

It is worth pausing, before going further, on a distinction that sounds almost too basic to need stating but that clinical medicine has, in practice, blurred for decades: the standardized "one-size-fits-all" model implicitly treats the average male body as the human baseline (Gualtierotti, 2025; Lucena et al., 2025). This has quietly obscured one of the most fundamental dimensions of human health — the biological and sociocultural differences that separate men, women, and gender-diverse individuals (Dattolo et al., 2026; Lee et al., 2023). In an era that markets itself on precision, integrating these variables is not simply the ethically correct thing to do; it has become, arguably, a scientific necessity for optimizing both safety and efficacy (Dattolo et al., 2026; Gualtierotti, 2025).

Getting this right means establishing a rigorous, non-interchangeable boundary between the two terms (Dattolo et al., 2026; Lucena et al., 2025). Sex is biological — chromosomes (XX versus XY), sex hormones (estrogens, progesterone, androgens), gene expression, and anatomy (Dattolo et al., 2026; Gualtierotti, 2025). Gender is sociocultural — roles, behaviors, expressions, identities, and the institutional distribution of power and resources that surrounds a person (Lee et al., 2023; Shankar & Shah, 2025). In practice, of course, sex and gender interact continuously in any living patient, which is precisely why disentangling their independent and combined contributions to health outcomes is both difficult and, we would argue, unavoidable (Tran & Eder, 2026).

The cost of neglecting this distinction has been concrete rather than theoretical. Between 1997 and 2000, ten FDA-approved prescription drugs were withdrawn from the U.S. market owing to serious adverse effects; eight of these carried disproportionately higher risk for women, a safety gap that went undetected during early trials that had relied predominantly on male cohorts (Kraševec, 2022). Reading health through both a biological and a sociocultural lens, then, is less a matter of political correctness than a route toward genuinely personalized — rather than merely generalized — therapeutic guidance (Ginaldi & De Martinis, 2022; Gualtierotti, 2025).

2.2 The Biological Scaffold: Sex-Based Pharmacokinetics and Pharmacodynamics

Biological sex shapes, often quite substantially, both how a drug moves through the body (pharmacokinetics) and how it acts once it gets where it is going (pharmacodynamics) (Lucena et al., 2025; Rakshith et al., 2023). These differences unfold across the whole absorption-distribution-metabolism-excretion (ADME) sequence (Ginaldi & De Martinis, 2022; Lucena et al., 2025), which is illustrated schematically in Figure 1.

Dosing conventions built on body surface area or total body weight, it turns out, do not really capture any of this (Rakshith et al., 2023). Women generally carry a higher percentage of body fat, operate at a lower glomerular filtration rate (GFR), empty the stomach more slowly, and maintain different plasma carrier-protein concentrations (Lucena et al., 2025). Lipophilic drugs such as diazepam consequently persist longer in women given larger adipose reservoirs, while renally cleared drugs may linger owing to the lower GFR (Lucena et al., 2025).

At the metabolic level, sex-specific cytochrome P450 (CYP450) activity is a recurring and fairly robust finding: women display higher baseline CYP3A4 activity, accelerating clearance of substrates like cyclosporine and erythromycin, while men run higher on CYP1A2, clearing compounds such as caffeine more rapidly (Lucena et al., 2025). These pathways appear to be evolutionarily conserved to a surprising degree — when freshwater bivalves (Dreissena polymorpha) were exposed to the SSRI sertraline, researchers observed sex-specific patterns of cytochrome P450 recruitment and bioaccumulation that mirrored mammalian, endocrine-modulated metabolic profiles (Lumor et al., 2026), a finding that, whatever one makes of its clinical relevance, at least underscores how deeply conserved these mechanisms are.

2.3 Specialized Therapeutic Domains: Oncology, Rheumatology, and Psychiatry

2.3.1 Cytotoxic and Targeted Oncology

Sex differences in clearance matter most, arguably, where the therapeutic index is narrow — and oncology is a prime example (Rakshith et al., 2023). Women treated with 5-fluorouracil (5-FU), the chemotherapy backbone for colorectal cancer, experience significantly higher rates of severe hematological and gastrointestinal toxicity, including grade 3/4 neutropenia, leukopenia, diarrhea, and stomatitis (Baraibar et al., 2023; Rakshith et al., 2023). The mechanism here is fairly well characterized: an average 15% reduction in dihydropyrimidine dehydrogenase (DPD) activity in women, which slows plasma clearance and prolongs drug retention (Rakshith et al., 2023).

A similar story plays out with paclitaxel, which women eliminate roughly 20% more slowly than men, reaching

Figure 1. Sex-based pharmacokinetic and pharmacodynamic pathway model, illustrating how female- and male-associated differences in absorption, distribution, metabolism, and excretion translate into distinct clinical consequences across oncology, psychopharmacology, and nephrology (Dattolo et al., 2026; Lucena et al., 2025; Rakshith et al., 2023).

Figure 2. Conceptual model of the four gendered pathways proposed by the Gender Hypothesis — healthcare utilization, clinical bias, subjective perception, and structural determinants — through which sociocultural gender shapes the observed female excess in adverse drug event databases (Lee et al., 2023).

peripheral-compartment saturation at lower plasma concentrations (Rakshith et al., 2023). Doxorubicin clearance, too, runs lower in women — a vulnerability that becomes especially concerning in young girls treated for pediatric leukemia, who face elevated early cardiotoxicity risk, compounded by lower cardiac P-glycoprotein expression, the transporter that would otherwise pump the drug back out of cardiac cells (Rakshith et al., 2023).

Newer modalities are not exempt. Among patients treated with developmental antibody-drug conjugates (ADCs) for advanced non-small-cell lung cancer, a Gustave Roussy cohort study (2018–2023) of 132 patients found women experienced substantially higher rates of grade ≥2 treatment-related adverse events (82.8% versus 69.1%) and metabolic toxicities — anorexia, weight loss, lipid disturbances — leading to nearly double the rate of dose reductions (37.5% versus 19.1%) relative to men (Gustave Roussy Cohort, 2018–2023).

2.3.2 Advanced Biologics in Rheumatology and Neurology

In psoriatic arthritis, women consistently report poorer outcomes and discontinue advanced biologic therapy — TNF inhibitors, IL-17 inhibitors, IL-23 inhibitors — more often than men (Tran & Eder, 2026). The reasons appear layered rather than singular: differences in baseline body composition (fat-free mass versus adiposity), central pain sensitization, and immunogenicity all seem to play a part (Tran & Eder, 2026). Because women tend to mount more vigorous innate and adaptive immune responses, they are also more prone to developing anti-drug antibodies (ADAs) that neutralize biologic agents outright, contributing to therapeutic failure (Tran & Eder, 2026).

Multiple sclerosis tells a somewhat different story. Register-based studies suggest that long-term response rates to disease-modifying therapies show comparatively little difference by sex, yet women face a significantly higher risk of discontinuing glatiramer acetate because of adverse events — a reminder that efficacy and safety do not always move in the same direction, and that sex-specific issues can hide inside a seemingly neutral efficacy signal (Baimonte et al., 2025).

2.3.3 Psychopharmacology and the Estrogen Hypothesis

In neuropsychopharmacology, the fluctuating tides of reproductive hormones appear to modulate neurotransmitter receptor sensitivity, receptor density, and — ultimately — drug efficacy itself (Storosum et al., 2025). A large individual patient data (IPD) meta-analysis of acute bipolar I mania trials found a modest but statistically significant sex effect on treatment response, with antipsychotics and mood stabilizers showing lower efficacy and smaller symptom reduction in women than in men (Storosum et al., 2025).

What makes this finding more than a curiosity is what happened when age was used as a proxy for menopausal status. Premenopausal women (47 years or younger) showed a markedly lower response rate and a higher Number Needed to Treat (NNT = 7.5) relative to postmenopausal women (NNT = 4.2) and men (Storosum et al., 2025) — a pattern consistent with the "estrogen hypothesis," which proposes that high circulating estrogen in premenopausal women may exert a negative moderating effect on antimanic and antipsychotic drug efficacy (Storosum et al., 2025).

2.4 Sociocultural Pathways: The Gender Hypothesis of Adverse Drug Events

Biological sex explains cellular and physiological mechanisms reasonably well, but it does not, on its own, explain why national pharmacovigilance databases such as the FDA's FAERS consistently record so many more adverse drug events (ADEs) in women than in men. To account for that gap, researchers have proposed what is now called the Gender Hypothesis (Lee et al., 2023), which identifies four interconnected gendered pathways shaping drug-safety data — summarized visually in Figure 2.

First, healthcare utilization: women engage in health-seeking behavior and interface with clinicians at considerably higher rates than men, which mechanically raises cumulative prescription exposure and, with it, baseline risk of experiencing (and reporting) an adverse event (Lee et al., 2023). Traditional masculine norms cut the other way, discouraging men from seeking primary care until concerns have progressed further — which may explain why men are underrepresented in non-serious ADE reports yet overrepresented in severe outcomes such as hospitalization or death (Lee et al., 2023).

Second, clinical bias. Gender stereotypes shape how symptoms get interpreted at the point of care. The historic "Yentl Syndrome" describes how women's cardiovascular complaints are too often minimized or misread as anxiety, resulting in fewer referrals and prescribing that misses the underlying problem (Gualtierotti, 2025; Lee et al., 2023). A related pattern shows up in pain management, where women's physical pain is disproportionately treated with antidepressants rather than adequate analgesia (Lee et al., 2023).

Third, subjective perception. Societal expectations shape which side effects even register as worth reporting. "Cosmetically salient" events — hair loss, weight gain — skew heavily female in safety databases, plausibly reflecting internalized pressure around appearance (Lee et al., 2023), whereas sexual dysfunction is reported disproportionately by men, likely because masculinity scripts link sexual performance closely to identity, making drug-induced sexual side effects more salient and more reportable (Lee et al., 2023).

Fourth, and perhaps least visible, upstream structural determinants. Gendered divisions of labor and differing economic circumstances translate into differential environmental exposures — women, for instance, accumulate higher exposure to endocrine-disrupting phthalates through cosmetic and menstrual hygiene products (Lee et al., 2023). Women are also more likely to experience intimate partner violence, workplace harassment, and poverty, all of which establish chronic vulnerabilities that raise the risk of developing complex health conditions in the first place (Lee et al., 2023).

2.5 Systemic Inequities at the Intersections

2.5.1 Nephrology and Transplantation

The intersection of biological sex and gender roles produces distinct disease trajectories in nephrology as well (Dattolo et al., 2026). In acute kidney injury (AKI), experimental models suggest males are substantially more vulnerable to ischemic and toxin-induced injury, since testosterone appears to promote renal damage by amplifying oxidative stress, while estrogen upregulates protective vascular and antioxidant pathways (Dattolo et al., 2026; Soranno et al., 2025). Clinical epidemiological studies, though, remain fairly inconsistent — largely, it seems, because they often skip sex-stratified analysis altogether or fail to account for hormonal shifts across puberty, menopause, and andropause (Dattolo et al., 2026; Soranno et al., 2025).

Kidney transplantation surfaces a related but distinct inequity: women constitute over 75% of living donors — a figure that likely reflects caregiving expectations and familial role pressure — yet represent only 38% of waitlisted transplant candidates (Shankar & Shah, 2025). This gap appears driven by a combination of factors: providers less often initiating transplant discussions with women, socioeconomic barriers, and pregnancy-induced HLA sensitization, which lengthens wait times and complicates matching (Shankar & Shah, 2025).

2.5.2 Substance Use, Housing, and Gender-Based Violence

In more marginalized populations, gender and social determinants shape physical safety and violence risk in ways that deserve more attention than they typically get (Swaich et al., 2026). A longitudinal study of people who use drugs in Vancouver, Canada, examined whether stable housing protected against physical and sexualized violence amid rising benzodiazepine contamination in the unregulated drug supply (Swaich et al., 2026). What the researchers found was a striking asymmetry: stable housing was associated with significantly lower odds of violence among men (AOR = 0.61), yet this protective effect essentially vanished for women (AOR = 0.82) (Swaich et al., 2026) — a finding that speaks to how deeply violence against marginalized women is embedded in daily life, perpetrated often by people (partners, acquaintances, neighbors) who retain access to the home regardless of its stability (Swaich et al., 2026).

Gendered patterns extend to where drugs get used, too (Latkin et al., 2026). In a large community sample of people who use opioids in Baltimore, Maryland, women had significantly lower odds than men of using drugs in public or semi-public settings — streets (aOR = 0.49), alleys (aOR = 0.50), parks (aOR = 0.57), abandoned buildings (aOR = 0.53), cars (aOR = 0.55) (Latkin et al., 2026). This avoidance seems driven by intense stigma, particularly around motherhood, alongside fear of interpersonal violence and Child Protective Services involvement (Latkin et al., 2026). The consequence, unfortunately, is that women more often use drugs alone indoors, which strips away the safety net of bystander naloxone and other harm-reduction resources, raising the risk of fatal solitary overdose (Latkin et al., 2026).

This vulnerability compounds further at the level of intersectionality: sexual minority women across different racial and ethnic groups show consistently elevated rates of substance use disorders (SUDs) and cannabis use disorder (Schilt-Solberg et al., 2025). Minoritized lesbian, gay, and bisexual women face what researchers describe as a "triple jeopardy" of heterosexism, racism, and misogyny — a convergence of pressures that amplifies minority stress and, plausibly, drives higher rates of self-medication as a coping strategy for trauma and economic hardship (Schilt-Solberg et al., 2025).

2.6 Toward Gender-Sensitive Precision Medicine

Pulling all of this together, the field needs to move past the "one-size-fits-all" model if it genuinely wants health equity and pharmacological precision (Dattolo et al., 2026; Gualtierotti, 2025). That means systematically incorporating sex and gender as core variables at every stage of research, not as an optional subgroup analysis tacked on at the end (Gualtierotti, 2025). Practical steps include adopting the Sex and Gender Equity in Research (SAGER) guidelines for study design and reporting, ensuring representative enrollment of women and gender-diverse participants in trials, and routinely performing sex-disaggregated and sex-stratified analyses to surface disparities in safety, tolerability, and response that would otherwise stay hidden (Gualtierotti, 2025; Lucena et al., 2025). Embedding these principles into medical curricula and clinical guidelines is, we would suggest, how healthcare systems eventually move from treating a generalized "average" patient to delivering care that is genuinely personalized (Dattolo et al., 2026; Gualtierotti, 2025).

3. Methods

3.1 Design and Rationale

We approached this project as a narrative-synthesis review rather than a formal systematic review or meta-analysis, largely because the evidence base spans such heterogeneous designs — pharmacovigilance registries, individual patient data (IPD) meta-analyses, cohort studies, cross-sectional surveys, and mechanistic pharmacology work — that pooling effect sizes across them would, frankly, obscure more than it revealed. That said, we tried to bring as much methodological discipline to the process as the design allows, borrowing structural elements from PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) reporting conventions wherever they fit, so that another team attempting to reproduce or extend this synthesis would have a reasonably clear trail to follow.

3.2 Information Sources and Search Strategy

We searched PubMed/MEDLINE as the primary database, supplemented by targeted searches of Scopus, Embase, and Web of Science to capture journals not indexed uniformly in PubMed (for instance, several pharmacology and drug-policy titles). Searches were run between February and June 2026 and were not restricted by language at the initial screening stage, though full-text review was ultimately conducted only on English-language sources given translation capacity.

Search terms combined controlled vocabulary (MeSH terms such as "Sex Factors," "Sex Characteristics," "Gender Identity," "Pharmacokinetics," "Drug-Related Side Effects and Adverse Reactions," and "Precision Medicine") with free-text keywords, joined using Boolean operators in the general form: ("sex differences" OR "gender differences" OR "sex-specific" OR "gender-specific") AND ("drug response" OR "pharmacokinetics" OR "pharmacodynamics" OR "adverse drug event*" OR "adverse drug reaction*" OR "drug efficacy" OR "drug toxicity"). Where a specific clinical domain was of interest — oncology, psychiatry, nephrology, rheumatology, substance-use epidemiology — domain-specific terms (e.g., "5-fluorouracil," "acute mania," "acute kidney injury," "psoriatic arthritis") were added as a third Boolean layer to keep the yield manageable and clinically focused.

Reference lists of included articles and recent reviews were hand-searched for additional eligible studies (a form of backward citation chasing), and we monitored table-of-contents alerts from several core journals (e.g., Pharmacological Research, Social Science & Medicine, Kidney International) during the drafting period to capture newly published, highly relevant material.

3.3 Eligibility Criteria

We included peer-reviewed original research, individual patient data meta-analyses, systematic reviews, cohort studies, and position papers from professional societies that reported sex- or gender-stratified drug outcomes — efficacy, adverse events, pharmacokinetic parameters, or clinical risk prediction performance — in human populations (with one deliberate exception discussed below). Studies were required to report disaggregated data by sex or gender (not merely adjust for it as a covariate without reporting stratified estimates), since aggregate-only reporting is precisely the practice this review critiques and cannot itself illustrate the phenomenon under study.

We excluded editorials and commentary pieces lacking original data or original synthesis, case reports involving fewer than five patients, and preclinical studies unless they provided direct, well-established mechanistic grounding for a clinical finding discussed elsewhere in the synthesis (an example being the sex-dimorphic response to metformin in rodent neuropathic pain models, included specifically because it illuminates a proposed mechanism rather than standing as a clinical claim on its own). No date restriction was applied outright, though the search was weighted toward 2022–2026 publications to capture the most current pharmacovigilance and clinical trial evidence; foundational or highly cited older studies (e.g., Soldin & Mattison, 2009) were retained where they remain the primary reference point for a mechanism still cited in current literature.

3.4 Study Selection and Data Extraction

Titles and abstracts were screened first for topical relevance, followed by full-text review of records passing that initial filter. Given the narrative rather than systematic design, a single reviewer conducted the primary screening and extraction, with periodic cross-checking of borderline inclusion decisions against the eligibility criteria above — a limitation we return to in Section 5. For each included source, we extracted, where reported: study design, population and sample size, drug or drug class under investigation, sex- or gender-disaggregated outcome measures (proportions, odds ratios, hazard ratios, NNT, or equivalent), adjustment covariates, and the authors' stated interpretation of any sex or gender effect observed. Extracted quantitative data underlying Tables 1 through 4 were transcribed directly from the source publications' reported results tables, cross-checked against the original text for internal consistency, and are attributed to their original source throughout.

3.5 Data Synthesis

Because included studies varied too widely in design, outcome metrics, and population to support quantitative pooling, we organized findings thematically around four domains that recurred across the literature: (1) pharmacovigilance-level reporting patterns, (2) mechanistic pharmacokinetic and pharmacodynamic differences by sex, (3) sociocultural (gendered) pathways shaping how drug effects are experienced and reported, and (4) the practical translation — or, more often, non-translation — of sex-based evidence into clinical risk tools and guidelines. Within each domain, we prioritized studies with the largest sample sizes or the most rigorous designs (IPD meta-analyses and large registry analyses over single-center cohorts, for instance) as anchoring evidence, using smaller or mechanistic studies to contextualize plausible biological pathways.

3.6 Reporting Standards

Consistent with the very argument this review makes, we have tried to apply the Sex and Gender Equity in Research (SAGER) guidelines to our own reporting: distinguishing explicitly between sex and gender terminology throughout, reporting disaggregated data wherever the source material allowed it, and flagging instances where included studies themselves failed to disaggregate by sex or gender as a limitation of the underlying evidence base rather than glossing over it.

4. Sex, Gender, and Drug Response: Convergent Evidence Across Contexts

4.1 Epidemiology of Pharmacovigilance: Deconstructing the Female Excess in Adverse Event Reporting

A macro-level analysis of 33,719,943 adverse drug event (ADE) records, representing 11,413,854 unique individuals, drawn from the U.S. FDA Adverse Event Reporting System (FAERS) between 2014 and 2022, reveals a persistent and fairly striking disparity in drug safety profiles between men and women (Lee et al., 2023). At the aggregate level, women account for 63% of all recorded ADEs and 60% of all submitted person-reports (Lee et al., 2023). But — and this is really the crux of the finding — disaggregating that enormous dataset by clinical severity, reporter type, and outcome shows that the female skew is highly context-dependent, and appears to be shaped substantially by sociocultural reporting pathways rather than biology alone (Lee et al., 2023).

As Table 1 summarizes, the pronounced female skew in the overall dataset concentrates disproportionately in non-serious, subjective, or cosmetically salient events (Lee et al., 2023). Common non-serious complaints such as nausea (72% female), headache (73% female), and arthralgia (71% female) show a strong female reporting bias, and this pattern holds for hair-related events, dermatological complaints, and weight gain, which together suggest that gendered socialization and cosmetic standards shape which experiences even register as worth reporting (Lee et al., 2023).

That pattern, however, shifts markedly once severity

Figure 3. Sex-disaggregated FAERS reporting patterns (2014–2022). Panel A ranks the ten most frequently reported adverse events by the proportion attributed to female non-serious reports; Panel B shows how the overall female share of ADE reports narrows, and reverses, as clinical severity increases from non-serious to fatal outcomes (data derived from Table 1; Lee et al., 2023).

Figure 4. Sex- and age-stratified treatment response in acute bipolar I mania. Panel A ranks subgroup Numbers Needed to Treat (NNT), showing the poorest drug-placebo separation in premenopausal women; Panel B decomposes the premenopausal-versus-postmenopausal gap into active-drug and placebo response rates, showing the difference is driven mainly by placebo-arm variation (data derived from Table 3; Storosum et al., 2025).

enters the picture (Lee et al., 2023). The proportion of ADEs classified as serious — defined by the FDA as involving death, hospitalization, disability, or congenital anomaly — runs notably higher in men than in women (71% versus 63%), a narrowing that becomes most pronounced in fatal outcomes, where men represent 53% of all reported deaths (Lee et al., 2023) (see Figure 3, Panel B).

There is also a subtler pattern buried in who is doing the reporting. Patient-submitted reports show a real gender gap: female consumers document more symptoms per submission than male consumers (2.09 ADEs per non-serious person-report versus 1.92 in men) (Lee et al., 2023). That gap effectively disappears, though, in clinical settings — among serious person-reports filed by healthcare professionals, the average number of documented ADEs sits nearly identical between the sexes (2.12 for men, 2.15 for women), and the overall female proportion in this more clinically verified category drops to 52% (Lee et al., 2023). Taken together, these findings suggest that women are more proactive in surfacing subjective or mild toxicities, while more severe, objectively confirmed events are considerably more evenly distributed, or even skew male — a pattern that lines up well with the behavioral and structural logic of the Gender Hypothesis (Lee et al., 2023).

4.2 Intersectional Stressors and Structural Harm: Childhood Trauma, Minoritized Identities, and Drug Use Environments

Moving from general population registries toward more marginalized cohorts, a systematic synthesis of 32 studies encompassing 43,197 sexual and gender minority (SGM) adults documents a consistent and fairly powerful link between Adverse Childhood Experiences (ACEs) and later substance use disorders (SUDs) (Ong et al., 2026). This relationship, notably, is heavily moderated by intersecting social identities — gender, race, and sexual orientation compound one another rather than acting independently (Schilt-Solberg et al., 2025).

As Table 2 illustrates, SGM individuals who experienced childhood sexual abuse (CSA) or childhood physical abuse (CPA) before age 18 show substantially elevated odds of problematic substance use, alcohol dependency, and tobacco addiction in adulthood (Ong et al., 2026). Among men who have sex with men (MSM), Arreola et al. (2008) found that a CSA history was associated with more than double the odds of heavy drinking (aOR = 2.15) and nearly triple the odds of frequent drug use (aOR = 2.80) (Ong et al., 2026). A comparable pattern emerged for problematic alcohol use among gay, bisexual, and other men who have sex with men who experienced childhood trauma, with roughly a 30% increase in adjusted odds (aOR = 1.30) (Ong et al., 2026).

The steepest vulnerabilities surface at the intersection of transgender identity, racial marginalization, and early-life trauma (Ong et al., 2026). In a focused study of transgender women of color, Reback et al. (2017) documented nearly a tenfold increase in the odds of recent drug use (aOR = 9.91; 95% CI: 2.10–46.86) following CSA exposure, alongside a substantial increase in the number of concurrent drugs used (Ong et al., 2026). This extreme disparity is echoed at a population level: analysis of the 2021–2022 National Survey on Drug Use and Health found that multiracial lesbian/gay individuals carry the highest absolute burden of past-year SUD — 46.6% among women and 52.3% among men — compared with a baseline of just 15.0% in the general female population (Schilt-Solberg et al., 2025).

These intersecting trauma pathways are further shaped by gender-specific physical environments and structural barriers (Swaich et al., 2026). Among people exposed to the benzodiazepine-contaminated unregulated drug supply in Vancouver, Canada, the protective effect of stable housing diverges sharply by gender: secure housing corresponds with a roughly 40% reduction in odds of experiencing physical or sexualized violence among men (AOR = 0.60; 95% CI: 0.40–0.92), yet this protection is essentially absent among women (AOR = 0.81; 95% CI: 0.51–1.31) (Swaich et al., 2026) — a gap that reflects how entrenched gender-based and intimate partner violence can be, resistant to a fix as seemingly fundamental as stable shelter.

This vulnerability echoes, too, in geographic choice: women who use opioids show significantly lower odds than men of using drugs in public or semi-public settings such as streets (aOR = 0.49), alleys (aOR = 0.50), or abandoned buildings (aOR = 0.53) (Latkin et al., 2026). This avoidance, driven by stigma around motherhood and fear of both interpersonal violence and Child Protective Services involvement, pushes women toward using drugs in isolation indoors — cutting them off from harm-reduction services and elevating the risk of fatal solitary overdose (Latkin et al., 2026).

Table 1. Sex- and severity-disaggregated adverse drug event (ADE) reporting profiles for the ten most frequently recorded events in the U.S. FDA Adverse Event Reporting System (FAERS), January 2014 through December 2022 (33,719,943 records; 11,413,854 individuals). Columns report total, serious, and non-serious counts and female proportions for each event, illustrating the female-skewed pattern in subjective, non-serious complaints that narrows in serious outcomes. Source: Lee et al. (2023).

Table 2. Quantitative associations between childhood adverse experiences (ACEs) — including childhood sexual and physical abuse — and subsequent substance use outcomes among sexual and gender minority (SGM) adults, drawn from a systematic review of 12 primary studies. Effect estimates (odds ratios or adjusted odds ratios) demonstrate a consistent, positive association between early trauma exposure and later substance misuse across diverse SGM subpopulations. Source: Ong et al. (2026).

4.3 Controlled Trial Environments and Neuroendocrine Moderation in Acute Bipolar Mania

Psychosocial factors clearly shape real-world vulnerability, but biological sex differences show up just as clearly inside controlled clinical trials (Storosum et al., 2025). An individual patient data (IPD) meta-analysis of 10 double-blind, randomized registration trials (N = 2,199) evaluating antipsychotics and mood stabilizers for acute bipolar I mania reveals a modest but statistically significant male response advantage — active compounds separated more effectively from placebo in men than in women, yielding a Number Needed to Treat (NNT) of 5.3 in men versus 6.3 in women (adjusted OR = 1.136; p = 0.033) (Storosum et al., 2025) (Figure 4, Panel A).

To probe the biological mechanism behind this gap, the researchers used age 47 as a validated proxy for menopausal status and circulating estrogen level (Storosum et al., 2025). As Table 3 shows, anti-manic efficacy is heavily moderated by female reproductive aging: premenopausal women (≤47 years) experienced the weakest drug-placebo separation, with an NNT of 7.5 and a treatment-placebo response difference of just 13.3% (Storosum et al., 2025). Postmenopausal women (>47 years), by contrast, showed a markedly stronger response — 23.6% separation and an NNT of 4.2, nearly double the effect size (Storosum et al., 2025).

What makes this finding especially interesting, though, is where the moderation actually lives: the IPD analysis shows it is driven primarily by variation in the placebo arm rather than the active drug arm (Storosum et al., 2025) (Figure 4, Panel B). Premenopausal women showed an elevated placebo response rate of 34.7%, which narrowed the apparent therapeutic window; postmenopausal women showed a considerably lower placebo response of 24.0%, widening the observed drug-placebo gap (Storosum et al., 2025). Active drug response itself stayed remarkably stable across these subgroups (48.0% premenopausal versus 47.6% postmenopausal) (Storosum et al., 2025). These results lend fairly direct support to the Estrogen Hypothesis: high circulating ovarian hormones in premenopausal women may raise the physiological threshold for manic symptoms, or act as something like a natural stabilizer, effectively mimicking drug effects in the placebo arm and diminishing the apparent net benefit of active treatment in trials (Storosum et al., 2025).

4.4 Algorithmic Inconsistencies: Sex and Gender Covariates in AKI Prediction Models

Translating sex as a biological variable into actual clinical tools remains, evidently, a work in progress — a systematic appraisal of ten major clinical risk calculators for acute kidney injury (AKI) makes that clear (Soranno et al., 2025). Despite fairly robust preclinical evidence that estrogen protects against ischemic and toxin-induced renal injury while testosterone accelerates cellular damage via oxidative stress, clinical prediction tools do not integrate this consistently — or, in several cases, at all (Soranno et al., 2025).

As Table 4 compiles, the calculators disagree substantially on both the direction and magnitude of sex-associated renal risk (Soranno et al., 2025). In orthopedic surgery, the Gharaibeh (2017) model treats male sex as an independent risk factor for postoperative AKI (OR = 1.78; 95% CI: 1.19–2.70), a direction echoed by the Nash (2019) model for NSAID-induced nephrotoxicity (OR = 1.44; 95% CI: 1.23–1.67) (Soranno et al., 2025). Cardiothoracic and vascular calculators, though, point the other way: the Brown (2008) model for percutaneous coronary intervention identifies female sex as predictive of renal decline (OR = 1.38; 95% CI: 0.86–2.22), while the Thakar (2005) cardiac surgery model finds female gender protective (OR = 0.48; 95% CI: 0.21–0.75) — yet this variable was dropped from the simplified bedside version of that same calculator (Soranno et al., 2025).

Perhaps most concerning, several widely used tools — Basu (2014), Mehran (2004), Tsai (2014), Woo (2021) — omit sex or gender entirely, leaning instead on parameters like baseline serum creatinine and fluid overload (Soranno et al., 2025). This is not a neutral omission: because baseline creatinine generation depends heavily on muscle mass, which itself varies by sex, static creatinine thresholds that ignore sex-specific kinetics risk systematically under-recognizing AKI in women while overestimating it in men (Soranno et al., 2025). The gap between mechanistic understanding and clinical tool design here is stark, and points toward an evident need for dynamic, sex-stratified biomarkers rather than the unadjusted, one-size-fits-all risk models still in wide use (Soranno et al., 2025).

5. Toward a Biologically and Socially Literate Model of Drug Response

5.1 Reconciling Biology and Behavior in the ADE Reporting Gap

The FAERS data (Table 1; Figure 3) offer, we think, a fairly persuasive case study in why sex-disaggregated data has to be interpreted alongside gender context rather than in isolation. Taken at face value, a headline figure of "63% of ADEs occur in women" could easily be misread as evidence of straightforward biological vulnerability. But the moment severity is factored in, that story gets considerably more complicated — the female skew concentrates in subjective, non-serious, arguably cosmetically salient complaints, and essentially reverses in fatal outcomes (Lee et al., 2023). This is not, we would argue, evidence that women are biologically more fragile; it looks much more like evidence that women interact with the healthcare system differently, notice and report symptoms differently, and are subject to different clinical interpretation once they do (Lee et al., 2023). None of which is to say biology plays no role — clearly it does, as later sections of this discussion make clear — but the aggregate pharmacovigilance signal cannot be read as a pure biological readout without accounting for the sociocultural scaffolding around it.

5.2 When Biology Alone Is the Better Explanation

Where the Gender Hypothesis explains reporting patterns, mechanistic pharmacology explains actual physiological difference — and the acute mania data (Table 3; Figure 4) are, arguably, the clearest example in this review of biology acting largely on its own terms. The premenopausal-versus-postmenopausal split in antimanic drug response is not obviously reducible to differential help-seeking or reporting bias; it shows up inside double-blind, placebo-controlled trials, where reporting behavior should matter far less (Storosum et al., 2025). What we found particularly striking, on reflection, is that the moderation lives mostly in the placebo arm rather than the drug arm — premenopausal women essentially respond "as if" partially treated even without active medication, plausibly because estrogen itself exerts some mood-stabilizing effect (Storosum et al., 2025). If that interpretation holds up under further study, it would have real implications for how psychiatric trials are designed and powered: pooling premenopausal and postmenopausal women together may be masking a therapeutically meaningful subgroup effect, one a suitably stratified trial design could reveal.

5.3 The Translation Gap: From Mechanism to Clinical Tool

Perhaps the most sobering finding in this synthesis is the AKI risk-calculator inconsistency documented in Table 4. Here, unlike the mania trials, the underlying biology is reasonably well characterized — estrogen appears renoprotective, testosterone appears to accelerate injury (Dattolo et al., 2026; Soranno et al., 2025) — yet clinical tools built to operationalize that biology simply do not agree with one another, and several omit sex altogether (Soranno et al., 2025). This is, we think, the clearest illustration in the entire literature we reviewed of a translation gap: mechanistic knowledge sitting comfortably in the pharmacology literature while clinical decision tools proceed largely as though it does not exist. The consequence is not abstract. A creatinine-based risk score that ignores sex-specific muscle mass and baseline creatinine generation will systematically misclassify risk in both directions, under-flagging AKI in women and over-flagging it in men (Soranno et al., 2025) — precisely the kind of quiet, structural inequity that individual clinicians rarely notice because it is baked into the tool rather than into any single decision.

5.4 Intersectionality as a Necessary Complication

It would be a mistake, we think, to treat "sex and gender" as a single axis of variation that simply needs a stratified analysis to resolve. The substance-use and violence literature reviewed here (Table 2) makes clear that gender interacts with race, sexual orientation, and gender identity in ways that a simple male/female stratification cannot capture — transgender women of color exposed to childhood sexual abuse show odds ratios for later drug use that dwarf those seen in other subgroups (Ong et al., 2026), and the protective effect of stable housing against violence essentially disappears for women in a way it does not for men (Swaich et al., 2026). Any precision-medicine framework serious about sex and gender needs, we would suggest, to build in intersectional analysis from the start rather than treating it as a secondary refinement.

5.5 Limitations

This review has real limitations worth naming plainly. First, it is a narrative rather than systematic synthesis, conducted substantially by a single reviewer, which raises the possibility of selection bias in which studies were emphasized. Second, the underlying primary literature itself frequently conflates sex and gender or fails to disaggregate at all — a problem this review can describe but cannot retroactively fix. Third, much of the mechanistic pharmacology evidence derives from a relatively narrow

Table 3. Sex- and age-disaggregated efficacy of anti-manic medications (antipsychotics and mood stabilizers) in acute bipolar I mania, from an individual patient data meta-analysis of 10 placebo-controlled registration trials (N = 2,199). Age 47 is used as a proxy for menopausal status. Premenopausal women show the weakest drug-placebo separation, consistent with the Estrogen Hypothesis. Source: Storosum et al. (2025).

Table 4. Appraisal of ten peer-reviewed clinical risk prediction calculators for acute kidney injury (AKI), indicating whether sex or gender is included as an independent covariate, the direction of sex-based risk where reported, and the associated odds ratio. The table highlights substantial inconsistency across surgical and clinical contexts and frequent omission of sex/gender from risk models. Source: Soranno et al. (2025).

set of drug classes (oncology, psychiatry, nephrology), and extrapolating these patterns to other therapeutic areas should be done cautiously. Finally, most included studies operationalize "sex" and "gender" within a binary framework, which likely obscures meaningful variation among intersex and gender-diverse populations — a gap that future research, and future reviews, will need to address directly rather than as an afterthought.

5.6 Implications for Practice and Research

Bringing this together, we would argue for three concrete shifts. Clinically, prescribers should treat sex and reproductive status (not just chronological age) as relevant dosing and monitoring variables for narrow-therapeutic-index drugs, particularly in oncology and psychopharmacology, where the evidence base is now fairly mature (Baraibar et al., 2023; Storosum et al., 2025). Methodologically, trial design and risk-calculator development should adopt mandatory sex-stratified reporting under SAGER-style guidelines, rather than treating sex as an optional covariate to be adjusted away (Gualtierotti, 2025; Soranno et al., 2025). And at the level of pharmacovigilance infrastructure, safety databases such as FAERS would benefit from routine severity-stratified sex reporting as a default output, since the aggregate figures alone — as this review has tried to show — can mislead as easily as they inform.

6. Conclusion

This manuscript set out to bring together, in one coherent account, the biological and sociocultural evidence for sex- and gender-based differences in drug response — a literature that has, until now, tended to sit in disconnected silos across oncology, psychiatry, and nephrology journals. By synthesizing pharmacovigilance data, IPD meta-analyses, and mechanistic pharmacology alongside sociological accounts of the Gender Hypothesis, the review demonstrates that neither biology nor social context alone can explain the observed disparities in drug efficacy and safety between men and women. The revised manuscript strengthens the original analysis through clearer conceptual framing, data-driven visualizations of the FAERS and mania-trial findings, and a more explicit discussion of translation gaps in clinical risk tools. Ultimately, the case made here is that sex and gender deserve permanent, structural inclusion in trial design, pharmacovigilance reporting, and clinical decision-making — not as an equity gesture, but as a scientific necessity for precision therapeutics.

Author Contributions

O.S.B. contributed to the conception and design of the review, literature search, analysis and synthesis of the relevant evidence, and drafting of the manuscript. O.O. contributed to the literature search, interpretation of evidence concerning sex- and gender-related differences in pharmacokinetics, pharmacodynamics, efficacy, and adverse drug events, and critical revision of the manuscript. H.H. contributed to the analysis and interpretation of pharmacovigilance, mechanistic pharmacology, and clinical evidence and critically revised the manuscript. S.U. contributed to the literature review, synthesis of evidence related to precision medicine and sex- and gender-informed pharmacotherapy, and critical revision of the manuscript for important intellectual content. All authors reviewed and approved the final version of the manuscript and agreed to be accountable for all aspects of the work.

Acknowledgements

The authors would like to acknowledge the Department of Biochemistry, Faculty of Science, Adekunle Ajasin University, Nigeria, and the Pharmacy Department, Tishk International University, Erbil, Kurdistan Region of Iraq, for their academic and institutional support. The authors also acknowledge the researchers whose published studies contributed to the scientific foundation of this review.

References