Abstract
Aim
Chronic spontaneous urticaria (CSU) affects female more frequently than men, but whether this epidemiological difference is accompanied by a reproducible sex-related laboratory phenotype remains unclear. To evaluate and quantitatively synthesize sex-related differences in laboratory, inflammatory, autoimmune, and autoreactivity biomarkers in CSU.
Materials and Methods
MEDLINE/PubMed, Embase.com, Scopus, and Web of Science Core Collection were searched from inception through 16 June 2026. Eligible studies either directly compared female and male patients with CSU or reported sex-stratified data sufficient for effect estimation. Two investigators independently screened the records and assessed the risk of bias. One investigator extracted the data, and a second investigator independently verified all entries against the source reports. Random-effects pooling was performed using restricted maximum likelihood estimation with Hartung-Knapp inference.
Results
Of 4,083 unique records screened, 25 studies were included. Ten adult or mixed-population studies contributed to the primary quantitative synthesis of autologous serum skin test (ASST) positivity. The pooled odds ratio (OR) for female versus men was 1.57 [95% confidence interval (CI) 0.80–3.09; P = 0.168], with substantial heterogeneity (I2 = 72.3%) and a 95% prediction interval of 0.23–10.53. Leave-one-out analyses did not materially change the estimate. A pediatric-inclusive sensitivity analysis also remained non-significant (OR 1.73, 95% CI 0.96–3.11). Individual studies suggested enrichment in females for selected findings related to thyroid autoimmunity, coagulation, platelets, basophil activation tests, and osteopontin, whereas inflammatory, eosinophil, and IgE results varied across populations and treatment settings.
Conclusion
Sex-related laboratory differences in CSU appear to be biomarker- and population-specific rather than components of a uniform female or male laboratory phenotype. The pooled ASST analysis did not demonstrate a statistically robust sex difference. Standardized sex-stratified reporting is required before sex-specific laboratory or treatment algorithms can be justified.
INTRODUCTION
Chronic spontaneous urticaria (CSU) is a mast cell-driven disease characterized by the recurrent occurrence of wheals, angioedema, or both for longer than 6 weeks in the absence of a specific external trigger.1, 2 Beyond visible skin lesions, CSU imposes a substantial burden through itch, sleep disruption, impaired quality of life, reduced work productivity, and repeated healthcare use.2, 3 Female are affected more frequently than men, and recent large-scale data suggest that the female predominance becomes particularly evident during adulthood and is accompanied by greater disease burden and poorer control in midlife.4, 5 These observations suggest that biological sex is not merely an epidemiological descriptor but may modify the inflammatory and autoimmune architecture of CSU.
Current concepts place type I autoallergy and type IIb autoimmunity at the center of CSU pathogenesis.6 Type I autoallergy involves IgE antibodies against self-antigens, whereas type IIb autoimmune CSU is characterized by functional IgG antibodies that activate mast cells and basophils and are directed against IgE or its high-affinity receptor. The latter endotype is commonly associated with a more severe and treatment-refractory clinical course, basopenia and eosinopenia, low total IgE, elevated inflammatory markers, positive autologous serum skin testing, and positive basophil functional assays.6-9 Because these tests reflect distinct biological domains, including systemic inflammation, coagulation activation, circulating granulocyte patterns, thyroid autoimmunity, immunoglobulin profiles, and functional serum autoreactivity, sex-specific evaluation may provide further valuable insights that cannot be inferred from clinical outcomes alone.
Sex-related immune regulation may be influenced by X-chromosome-linked immune genes, sex hormones, reproductive transitions, body composition, and differences in autoimmune susceptibility.10, 11 Kasperska-Zajac et al.11 summarized experimental and clinical observations suggesting that estrogens may enhance mast-cell and basophil histamine release, whereas progesterone and testosterone may exert inhibitory effects on mast-cell secretion; they also highlighted possible modulation of urticaria during the menstrual cycle, pregnancy, menopause, and exposure to exogenous hormones. However, much of this evidence is experimental, case-based, or derived from small cohorts, and lower dehydroepiandrosterone sulfate concentrations have been reported in chronic urticaria irrespective of sex and autologous serum skin test (ASST) status.11 Thus, biological plausibility has outpaced direct clinical evidence in supporting a reproducible female-versus-male laboratory phenotype. A broader systematic review identified 26 reports addressing sex and gender differences across epidemiology, clinical characteristics, comorbidities, treatment, and quality of life, but laboratory and functional autoreactivity findings were scattered and were not quantitatively synthesized.4 The present study therefore aimed to identify and critically appraise evidence on female-versus-male differences in laboratory, inflammatory, autoimmune, and autoreactivity biomarkers among patients with CSU and, where data permitted, to quantitatively synthesize these differences.
MATERIALS AND METHODS
Protocol and Reporting Framework
This evidence synthesis was conducted and reported in accordance with the PRISMA 2020 statement.12 An operational protocol specifying eligibility criteria, outcome domains, deduplication rules, screening procedures, and sensitivity strata was finalized before independent title/abstract screening. The protocol was not prospectively registered.
Eligibility Criteria
We included cohort (observational), cross-sectional, case-control, registry, and interventional studies that provided baseline or pretreatment data for patients with CSU and reported a direct comparison between female and male patients for at least one laboratory, serological, hematological, inflammatory, coagulation, autoimmune, or functional autoreactivity outcome. Studies using the historical term chronic idiopathic urticaria were eligible when the population was clinically compatible with CSU and did not predominantly consist of inducible urticaria or urticarial vasculitis. Pediatric studies and treatment-selected cohorts were retained in prespecified sensitivity strata.
We excluded acute urticaria, isolated chronic inducible urticaria, urticarial vasculitis, hereditary or acquired bradykinin-mediated angioedema, case reports, narrative reviews, and studies that reported sex only as a covariate, without providing a direct female-versus-male laboratory estimate or sufficient data to calculate one. Studies comparing each sex with healthy controls, without a within-CSU female-versus-male comparison, did not contribute to the focused synthesis. When multiple reports described overlapping populations, the most complete report was retained and secondary reports were linked but not counted separately.
Information Sources and Search Strategy
MEDLINE/PubMed, Embase.com, Scopus, and Web of Science Core Collection were searched from database inception through 16 June 2026. The PubMed, Scopus, and Web of Science searches were run on 15 June 2026, and the Embase search was run on 16 June 2026. Search concepts combined chronic urticaria terminology and sex/gender terms with laboratory biomarkers, including inflammatory, autoimmune, coagulation, hematological, and functional autoreactivity biomarkers. No limits regarding date, language, human subjects, or publication type were applied. Full electronic search strategies for all databases, together with the search dates and result counts, are provided in Supplementary Table 1. Search screenshots and raw NBIB/RIS exports were retained for audit purposes. The 26 reports included in the previous systematic review were reconciled against the master database using DOI, PMID, author/year, and title. Fifteen were already represented in the database set, and 11 were unique citation-only records. The 11 unique records were screened using the same eligibility criteria; none advanced to full-text assessment.4
Study Selection and Data Extraction
The four database exports were merged and deduplicated using a prespecified Python-based workflow. Exact PMID matches were merged; exact normalized title plus first-author matches were merged when publication years were compatible; DOI matches were merged only when titles were compatible; and exact cross-database title matches were merged after checks of notice type and year. Retractions, corrections, and expressions of concern were kept separate from the original articles. Records from conference supplements that shared an issue-level DOI were not merged when their titles differed. The 6,760 raw database records were reduced to 4,072 unique publications, and the deduplicated library was exported to a structured screening workbook. Two investigators independently screened all 4,072 database records and the 11 unique citation-only records. Records considered potentially eligible by either investigator underwent full-text assessment. Differences at the title/abstract or full-text levels were resolved through discussion and consensus; no third adjudicator was required. One primary exclusion reason was recorded for each full-text exclusion, and companion publications were linked to prevent double-counting.
A standardized extraction form captured study design, country, setting, recruitment period, diagnostic terminology, age range, treatment status, inclusion and exclusion criteria, total sample size, female and male sample sizes, assay method, threshold definition, timing of sampling, and sex-stratified outcome data. For binary outcomes, event and total counts were extracted. For continuous outcomes, means and standard deviations were preferred; medians and measures of dispersion were retained without forced conversion when distributional assumptions were uncertain. Reconstructed denominators were explicitly labeled and restricted to sensitivity analyses. One investigator extracted all eligible data using the standardized form, and the second investigator independently verified each extracted entry against the source report before analysis. Discrepancies were resolved through discussion and consensus.
Outcomes
The prespecified outcome domains were: (i) systemic inflammatory and coagulation markers, including CRP, ESR, D-dimer, fibrinogen, and leukocyte indices; (ii) circulating eosinophil and basophil counts or categorical eosinopenia/basopenia; (iii) total IgE and immunoglobulin profiles; (iv) thyroid and other autoantibodies; and (v) functional or surrogate markers of autoreactivity, including the ASST, autologous plasma skin test (APST), basophil activation test (BAT), and basophil histamine release assay. Emerging biomarkers, such as osteopontin, were summarized in a separate exploratory domain. Laboratory abnormalities that are expected to differ physiologically by sex, such as hemoglobin levels or red blood cell counts, were not interpreted as CSU-specific signals unless supported by disease-specific evidence.
Risk-of-bias Assessment
Two investigators independently assessed risk of bias using the revised JBI analytical cross-sectional tool, with design-specific interpretation for cohort, registry, case-control, treatment-selected, and legacy-CU reports because the study synthesized baseline or pretreatment sex-stratified laboratory comparisons.13 Overall study-level judgments were categorized as low-risk, some concerns, or high-risk using predefined decision rules, without calculating a numerical quality score. Disagreements were resolved through discussion and consensus.
Ethics Committee Approval and Informed Consent
This study was based exclusively on a secondary analysis of aggregate data extracted from previously published reports. No new data were collected from human participants, no intervention was performed, and no individual-level identifiable patient information was accessed. Therefore, ethics committee approval and informed consent were not required.
Statistical Analysis
Female constituted the index group, and men constituted the reference group. Binary outcomes were expressed as odds ratios (ORs) with 95% confidence intervals (CIs). Continuous outcomes measured on the same scale were summarized as mean differences. Random-effects models were estimated using restricted maximum likelihood, and Hartung-Knapp CIs were applied when at least three clinically and methodologically compatible studies were available. Heterogeneity was reported using Cochran’s Q, I2, τ2, and a 95% prediction interval. The stability of the primary synthesis was examined using leave-one-out analysis.
Planned sensitivity analyses examined the inclusion of pediatric cohorts in the ASST quantitative synthesis. Treatment-selected, phenotype-restricted, legacy-CU, and reconstructed-data studies were retained as contextual sensitivity strata and were interpreted separately when clinical or methodological compatibility was insufficient for pooling. For individual studies containing a zero cell, a continuity correction of 0.5 was added to all four cells of the corresponding 2 × 2 table. Small-study effects for the 10-study primary ASST analysis were assessed in an exploratory analysis using a funnel plot and Egger regression; these findings were interpreted cautiously because only 10 studies were available. Outcomes represented by only two clinically heterogeneous studies were not considered confirmatory pooled analyses. Data checks were performed using IBM SPSS Statistics for Windows, Version 25.0 (IBM Corp., Armonk, NY, USA). Deduplication and quantitative synthesis computations, including the REML random-effects models, Hartung-Knapp inference, prediction intervals, leave-one-out analyses, and Egger regression, were performed in Python (version 3.13) using SciPy (version 1.17.0).
RESULTS
Study Selection and Characteristics
The four database searches identified 6,760 records: 1,588 from MEDLINE/PubMed, 2,778 from Scopus, 666 from Web of Science, and 1,728 from Embase. Conservative within-database deduplication removed 2,688 records, leaving 4,072 unique database records. Reconciliation of the 26-report prior-review citation map identified 15 records already represented in the master database and 11 unique citation-only records; all 11 were excluded at the title/abstract stage because they did not provide an eligible female-versus-male CSU laboratory comparison. Across all sources, 4,083 unique records were screened, 4,053 were excluded, and 30 reports were sought. Two reports were not retrieved; 28 were assessed; 3 were excluded; and 25 studies were included (Figure 1). Five reports from the previous broad review overlapped with the final evidence set, and no additional eligible study was identified through citation reconciliation.4 Study characteristics, key sex-related laboratory and autoreactivity findings, and database-search and all-source study-selection yields are summarized in Tables 1-3, respectively.
Both reviewers independently judged all 25 included studies to be at low risk of bias, with no unresolved disagreements remaining after consensus. Study-level overall judgments and key methodological considerations are presented in Supplementary Table 2. No study was excluded on the basis of the risk-of-bias assessment.
Systemic Inflammatory, Coagulation, and Hematological Biomarkers
Signals of greater systemic inflammatory, coagulation, or hematological activation in female were not uniform across the CSU population. In severe antihistamine-refractory CSU, elevated D-dimer was more frequent in female (50/88 vs. 16/44; OR 2.30, 95% CI 1.09–4.85).14 In the CURE registry, elevated CRP did not differ significantly between female and men overall, although a higher prevalence was reported among female aged 51–65 years. An approximate effect estimate reconstructed from the reported age-stratified data was retained only as a sensitivity result.5 In an angioedema-restricted cohort, elevated CRP and/or ESR occurred in 10/20 female and 1/14 men (OR 13.00, 95% CI 1.42–119.08), but sparse events and phenotype selection limited generalizability.15 A large historical database showed elevated mean platelet volume in 2,652 of 8,472 female and in 1,010 of 4,306 men (OR 1.49, 95% CI 1.37–1.62).16 Elevated CRP and eosinophil distributions were similar between sexes in the Preis cohort.17 Other eosinophil, leukocyte, and neutrophil-to-lymphocyte findings were inconsistent or not fully effect-estimable.18-20
Autoreactivity and Autoimmune-Associated Biomarkers
CD63 BAT positivity was observed in 41/56 female and 15/31 men (OR 2.92, 95% CI 1.16-7.31).21 Ten adult/mixed studies contributed directly calculable ASST data (10,22–30). The REML random-effects model yielded a pooled OR of 1.57 for female versus men (Hartung-Knapp 95% CI 0.80–3.09; P = 0.168), with substantial heterogeneity (I2 = 72.3%; τ2 = 0.592; Q = 32.54, df = 9, P < 0.001) (Figure 2). The 95% prediction interval was 0.23–10.53. Leave-one-out pooled ORs ranged from 1.30 to 1.78; each CI crossed unity (Supplementary Figure 1). Inclusion of the pediatric Sahiner and Godse cohorts produced a 12-study sensitivity estimate that also crossed unity (OR 1.73, Hartung-Knapp 95% CI 0.96–3.11; P = 0.064; I2 = 67.6%);31, 32 Özçeker et al.33 likewise reported no significant pediatric sex association but did not provide extractable sex-by-ASST cells. Exploratory Egger regression did not detect funnel-plot asymmetry (intercept -1.43; P = 0.312), although statistical power was limited (Supplementary Figure 2).
Emerging and Subgroup-Specific Biomarkers
In Liu et al.,19 female had higher serum osteopontin than men (5.41 ± 1.58 vs. 4.39 ± 0.77 ng/mL; MD 1.02 ng/mL, 95% CI 0.21–1.83; P = 0.025), and osteopontin correlated with UAS7 in female but not men. Collet reported autoimmune thyroid disease in 8/16 female and none of 29 men (corrected OR 59.0, 95% CI 3.08–1130.09),34 while Najib reported composite thyroid autoimmunity in 50/146 female and 7/42 men (OR 2.60, 95% CI 1.08–6.28).35 Confino-Cohen et al.16 and Chanprapaph et al.26 showed female enrichment for several autoimmune serological markers, whereas Asero10 found no significant sex difference in thyroid autoantibody prevalence within an ASST-positive subgroup.10 Wan and Wu36 reported anti-thyroid antibodies without extractable sex-specific denominators. In treatment-selected cohorts, IgE findings were inconsistent: high total IgE was more frequent in men in the Preis cohort, whereas mean total IgE was higher in female in the smaller Sirufo cohort.17, 25
DISCUSSION
Principal Findings
Across 25 studies, sex-related laboratory differences in CSU were not uniform. The primary quantitative synthesis of ASST positivity across 10 adult/mixed cohorts produced a female-directed, non-significant pooled estimate accompanied by substantial heterogeneity and a wide prediction interval. These findings indicate that sex-associated laboratory signals in CSU are likely to depend on the biomarker pathway, age, phenotype, and treatment context, rather than constituting a single reproducible female or male laboratory profile.
Evidence Base and Methodological Quality
The present study addressed a more focused laboratory question than the previous broad review of sex and gender differences in CSU. Consequently, only five of its 26 reports overlapped with the final 25-study evidence set, and citation reconciliation identified no additional eligible studies.4 This distinction is important because studies focusing on epidemiology, clinical burden, comorbidities, or treatment response cannot substitute for direct sex-stratified laboratory data. Although all included studies were judged to be at low risk of bias, these judgments do not eliminate limitations of the wider evidence base, including historical disease terminology, selective laboratory testing, underrepresentation of men, and incomplete sex-specific reporting.
Systemic Inflammatory, Coagulation, and Hematological Biomarkers
Inflammatory and hematological differences were conditional rather than universal. The female-associated D-dimer signal arose in severe treatment-refractory CSU, whereas CRP was similar overall in the much larger CURE registry and differed only in midlife female.5, 14 The large CRP/ESR effect in angioedema-restricted disease is biologically plausible, but remains hypothesis-generating because of sparse events and phenotype selection.15 The robust MPV association in the Confino-Cohen database supports greater platelet activation in female, yet MPV is nonspecific and may reflect background sex differences, comorbidity, or testing patterns rather than a CSU-specific mechanism.16 Discordant eosinophil, leukocyte, and NLR findings further suggest modification by disease activity, endotype enrichment, treatment exposure, assay timing, and age.17-20 Taken together, these results argue against a fixed sex effect on systemic inflammation and instead support an interaction between sex and clinical context.
Autoreactivity and Autoimmune-Associated Biomarkers
Functional autoreactivity and autoimmune-associated biomarkers varied markedly across studies. The pooled ASST estimate was directionally female-associated, but the CI crossed unity, substantial heterogeneity was observed, and the prediction interval encompassed effects favoring either sex. The opposing estimates from large or informative studies, including Asero10 and Caproni et al.30 explain why a simple universal sex effect is not credible.10, 30 Leave-one-out analysis confirmed that no single study determined the conclusion, and pediatric inclusion did not materially change it.31-33 By contrast, CD63 BAT positivity was higher in female in one directly stratified cohort.21 This apparent difference is not contradictory: ASST, APST, BAT, HRA, and BHRA capture overlapping but non-equivalent biological processes and should not be treated as interchangeable measures. The current evidence therefore supports possible sex enrichment in selected autoimmune or autoreactive endotypes rather than a general female excess in autoreactivity.
Emerging and Subgroup-Specific Biomarkers
As a novel biomarker, osteopontin was higher in female and correlated with disease activity only in female, providing a biologically coherent link among immune-cell recruitment, tissue remodeling, and mast-cell-associated inflammation, but this finding currently rests on a single small study.19 Female enrichment for clinical thyroid autoimmunity and autoimmune serology was observed across legacy and contemporary cohorts,16, 26, 34, 35 although Asero10 and Wan and Wu36 illustrate that not all thyroid-related comparisons are concordant or quantitatively complete.10, 36 These findings align with broader evidence that type IIb autoimmune CSU and autoimmune comorbidity cluster in selected patients and may influence treatment behavior.37, 38 Hormonal mechanisms remain plausible context rather than proof: experimental and clinical observations suggest that estrogens may enhance mast-cell or basophil activation, whereas progesterone and testosterone may be inhibitory, but available CSU data rarely account for menstrual phase, pregnancy, menopause, or exogenous hormone exposure.11 The opposing IgE findings in the Preis and Sirufo treatment-selected cohorts further demonstrate that sex-related IgE patterns are not uniform and may depend on cohort selection, laboratory definitions, and treatment context.17, 25 Taken together, these findings indicate pathway- and population-specific rather than uniform sex effects.
Sex, Treatment Response, and Prognosis
Evidence that sex independently modifies treatment response or prognosis remains limited and internally inconsistent. In a 250-patient real-world omalizumab cohort, men achieved higher UCT scores at early follow-up, but 12-month outcomes were comparable between sexes, suggesting a difference in response kinetics rather than sustained efficacy.17 A smaller 42-patient study found clinical remission in both sexes during treatment but reported a sex-associated difference in recurrence after omalizumab withdrawal; the small sample and limited adjustment preclude firm conclusions.25 At the broader disease level, the CURE registry documented greater burden and poorer control in female, particularly in midlife.5 In addition, thyroid autoimmunity was more frequent in female, while anti-TPO positivity predicted a longer time to remission and APST positivity was linked to greater treatment difficulty; these biomarker-outcome associations should not be interpreted as proof of a direct sex effect.26 Taken together, these data do not justify sex-specific treatment selection or dosing. Prospective studies should test sex as an independent predictor and an interaction term while jointly accounting for age, hormonal status, baseline disease activity, autoimmune endotype, and biomarker profile.
Study Limitations
Strengths of this study include a comprehensive four-database search; formal reconciliation of the prior citation map; independent screening and quality appraisal; source-level verification of sex-stratified data; and a planned distinction between the primary adult/mixed synthesis and the pediatric, treatment-selected, phenotype-restricted, and legacy-CU sensitivity strata. The primary quantitative synthesis used REML random effects, Hartung-Knapp inference, a prediction interval, leave-one-out analysis, and an exploratory assessment of small-study effects. Limitations are equally important: male subgroups were often small; assay thresholds and disease definitions varied; two conference reports were unavailable; several outcomes were represented by only one or two studies; favorable within-study risk-of-bias judgments could not overcome the indirectness, selective testing, or clinical heterogeneity of the wider evidence base. Future cohorts should prespecify sex-stratified analyses, report complete denominators and dispersion measures, standardize biomarker thresholds, and capture age, menopausal status, hormonal exposure, autoimmune endotype, disease activity, and treatment status.
CONCLUSION
Current evidence does not support different routine laboratory panels or treatment algorithms solely on the basis of sex. Clinicians should continue guideline-based, clinically directed testing, while recognizing that female with severe, midlife, or autoimmune-featured CSU may be enriched for selected inflammatory, coagulation, platelet, thyroid, or autoreactivity signals, whereas IgE findings remain inconsistent across treatment-selected settings. For research, sex should be treated as a prespecified biological and analytical variable rather than merely adjusted away. In summary, the available literature supports sex-aware phenotyping and standardized reporting, but not sex-specific diagnostic or therapeutic algorithms; larger prospectively designed studies are required to determine whether any of these signals can improve endotyping, treatment selection, or prediction of remission.


