Androgenetic alopecia risk in allergic rhinitis: A cohort study of immune and antihistamine effects

Abstract

Background

Androgenetic alopecia (AGA), though traditionally attributed to androgens and genetics, is increasingly linked to immune and inflammatory pathways. Allergic rhinitis (AR) and AGA may share overlapping mechanisms, including prostaglandin signaling abnormalities. Second-generation H1-antihistamines (sgSH), used to treat AR, have anti-inflammatory and immunomodulatory properties and may influence hair follicle biology. This study aimed to determine the association between AR and the risk of AGA and to explore whether the use of sgSH impacts this risk.

Methods

This retrospective cohort study used the Taiwan National Health Insurance Research Database (NHIRD) between 2000 and 2021. Totally, 539,397 patients diagnosed with AR were included and categorized into sgSH users and nonusers. AGA incidence was estimated using Cox proportional hazards models.

Results

AR patients had a higher risk of AGA (adjusted hazard ratio [aHR]: 1.81, 95% confidence interval [CI]: 1.69–1.93, p < 0.001). Among AR patients, sgSHs users had a lower risk of AGA than nonusers (aHR: 0.23, 95% CI: 0.20–0.26), with a dose-dependent effect (p < 0.001). The strongest effect seen in patients <30 years old (aHR: 0.18, 95% CI: 0.15–0.21, p < 0.001).

Conclusions

This study observed that AR is associated with an increased risk of AGA and that sgSH use was statistically associated with a lower incidence of AGA, especially in younger individuals. However, the possibility that the observed associations reflect unmeasured confounding, rather than direct biological effects, cannot be ruled out. Future research should further clarify the complex relationship between allergic diseases and hair loss from multiple perspectives and assess the possibility of sgSH as a supplementary measure in the prevention or treatment of AGA. From a clinical perspective, our findings underscore the importance of heightened awareness of early hair loss in patients with chronic AR.

Introduction

Allergic rhinitis (AR) and androgenetic alopecia (AGA) are common chronic conditions that substantially affect the quality of life of patients. AR is well known to disrupt sleep and daily functioning, often resulting in daytime fatigue, reduced concentration, and decreased work or academic performance. In parallel, AGA has been consistently linked to psychological distress, including lowered self-confidence, social discomfort, and depressive symptoms. When these 2 conditions coexist, their combined impact may extend beyond the effects of either disorder alone, potentially amplifying both physiological strain and psychosocial stress.

AR is a chronic inflammatory disease associated with an overactive immune response, and its prevalence varies with age and geographic location.^1,2 Based on the Visual Analogue Scale (VAS), the disease was classified into mild and moderate-to-severe forms. Its pathological mechanism is primarily driven by T helper 2 (Th2) cells, involving the secretion of interleukin (IL)-4, IL-5, and IL-13 and promoting immunoglobulin E (IgE) production and eosinophil activation.3, 4, 5 Furthermore, the condition may also be influenced by factors such as genetic predisposition, environmental pollution, and climatic conditions.6, 7, 8 AGA is a hair loss disorder that is predominant globally, with prevalence increasing with age and impacting almost 50% of certain ethnic populations. AGA is often regarded as a complex condition influenced by androgens and characterized by hereditary predisposition. However, the exact pathogenesis of AGA remains to be fully elucidated. Growing evidence shows that its etiology involves factors beyond androgen signaling, such as local inflammatory responses, perifollicular fibrosis, and impaired energy metabolism of hair follicles.⁹

In addition to local nasal symptoms, AR patients often exhibit systemic immune dysfunction, including elevated peripheral blood eosinophils, inflammatory cytokine activation, and overlapping immune features with comorbidities such as asthma.¹⁰ These findings suggest that AR is not merely a local disease but may reflect systemic immune dysregulation. Similarly, the development of AGA has been linked to perifollicular low-grade micro-inflammation.⁹ Spatial transcriptomics and multiplex immunohistochemistry have revealed significant CD4⁺ T-cell infiltration within the immune microenvironment of the upper hair follicle in AGA scalps compared to controls. This infiltration manifests as a skewed Th2-type response, suggesting a localized Th2-mediated micro-inflammatory imbalance.¹¹ These collective insights suggest that Th2-related cells and inflammatory mediators—particularly mast cells and histamine—may constitute a shared immunopathological background between AR and AGA. Moreover, genome-wide association study have shown shared susceptibility loci between AGA and other autoimmune disorders, indicating that immune regulatory mechanisms may contribute to the pathogenesis of AGA.¹² Immune cell infiltration around the hair follicles of individuals with AGA, particularly T cells and macrophages, potentially engaging with hair follicle epithelial cells, resulting in hair follicle shrinkage and degeneration.¹¹ Among immune-related pathways, prostaglandins have garnered significant interest due to their shared role in the 2 previously mentioned conditions. Prostaglandin D2 (PGD₂) is a key mediator implicated in AR;^13,14 it is produced by mast cells and recruits eosinophils via the CRTH2 receptor, worsening nasal congestion and mucus production.¹⁵ PGD₂ also suppresses hair follicle development by maintaining follicles in the catagen phase, while prostaglandin E2 (PGE₂) and prostaglandin F2 alpha (PGF_2α) promote hair growth.16, 17, 18, 19 In summary, these suggest a possible mechanistic overlap between AR and AGA.

Despite the high prevalence of AR and AGA and their shared potential for immune-mediated pathogenesis, investigations into their comorbidity are sparse. Literature to date has largely prioritized the association between alopecia areata (AA) and allergic conditions.20, 21, 22 Consequently, the interplay between AR and AGA remains poorly understood. In addition, there is a lack of research exploring integrated anti-inflammatory approaches to address these 2 conditions, which may share common pathways of chronic immune activation. Second-generation selective H1-antihistamines (sgSH) are extensively used for the treatment of AR owing to their advantageous safety profile and effective H1 receptor antagonism. These medicines have immunomodulatory actions through antigen presentation inhibition, Th2 cytokine production suppression, and reduced eosinophil infiltration, thereby achieving anti-inflammatory outcomes.23, 24, 25 Furthermore, preclinical investigations indicate that cetirizine may diminish PGD₂ expression while elevating PGE₂ levels, and these alterations correlate with prolonged hair length and enhanced hair density.^26,27 These results suggest that sgSH may influence hair follicle biology and affect the risk of AGA in individuals with AR. Therefore, this study aims to conduct a longitudinal cohort study to investigate the association between AR and AGA and to evaluate the association between sgSH use and the risk of AGA development among AR patients, thereby elucidating the potential epidemiological link between these conditions and pharmacological interventions.

Materials and methods

Data sources and study design

The National Health Insurance Research Database (NHIRD), one of the Taiwanese government's National Health Insurance (NHI) programs, provided the data analyzed in this study.²⁸ This research examined medical claims data from the NHIRD Longitudinal Generation Tracking Database 2005 (LGTD2005), obtained from a random sample of 2 million beneficiaries spanning the years 2000–2021.

This study was approved by the Committee on Research Ethics [CMUH112-REC1-117(CR-2)]. To protect patient privacy, the NHI anonymizes data from patients and caregivers before distributing it for research purposes. Consequently, the Research Ethics Committee decided to waive the need for informed consent. The study was conducted in accordance with the Declaration of Helsinki of the World Medical Association.

Study population and exposure selection

Patients whose index dates fell between 2000 and 2020 were selected in this study. This ensured that each case had a minimum 1-year observation period. Moreover, to ensure the accuracy of the disease outcome, certain patients were excluded from the study. First, patients with preexisting alopecia or chronic skin diseases diagnosed before the start of the follow-up period, were excluded.²⁷ Second, patients who had received chemotherapy within 1 year before the follow-up period were excluded.^29,30 Third, patients who had used medications known to potentially cause alopecia within 3 months before the end of the follow-up period were excluded.^29,31 Finally, patients who had used systemic corticosteroids for more than 14 days within 2 months before the start of the follow-up period were excluded.³⁰ Ultimately, a total of 1,685,579 participants were included in this study (Fig. 1).

Fig. 1.

Participant selection. sgSH, second-generation H1-antihistamine

In this study, the study population comprised patients diagnosed with AR using the International Classification of Diseases, Ninth Revision Clinical Modification (ICD-9-CM) code 477, and International Classification of Diseases, Ten Revision Clinical Modification (ICD-10-CM) code J30. However, to ensure accurate subject recruitment, the comparison cohort was composed of individuals who had never been previously diagnosed with AR. The index date for the AR group was established as the initial diagnostic date of AR, while for the control group, a randomly selected day during the research period was determined as the index date. After recruiting the 2 cohorts independently, propensity score matching created a 1:1 match cohort with a 0.25 caliper tolerance. Matching was based on sex, age, index date (year), season distribution of index date, comorbidities, and medication use.

Outcomes selection

The primary outcome measure for this study was the number of newly identified cases of AGA during the follow-up period. Diagnoses were made by a dermatologist using the ICD-9-CM code 704.00, and ICD-10-CM code L64. The trial endpoint was defined as the earliest occurrence of the following events: AGA diagnosis, withdrawal from the NHI program, or study completion (December 31, 2021).

Medication exposure as a predictor

Moreover, patients with AR who used sgSH agents throughout the follow-up period were identified using anatomical therapeutic chemical (ATC) codes (Supplemental Table 1). The medications were subgrouped according to their cumulative dosage using tertiles: <1,285, 1285–4,379, and ≥4380 mg. The individuals' daily doses of the sgSH medications were calculated using the World Health Organization's defined daily dose (DDD) for each drug. Subsequently, individuals were categorized into 3 groups based on cumulative usage (<45 DDD, 45–139 DDD, and ≥140 DDD), allowing for the assessment of the effects of AGA under various usage conditions.

Potential confounding factors

The potential confounding factors in this study included variables such as season distribution of index date, age, sex, urbanization level, insurance amount, comorbidities, and medication usage. All specific codes are listed in Supplemental Table 1.

Statistical analysis

All data were statistically analyzed using SAS version 9.4 software (SAS Institute, Inc., Cary, NC, USA) and R version 4.3.1 (R Core Team, Vienna, Austria). Standardized mean differences (SMDs) were used to evaluate the distribution of demographic features between the control and AR cohorts were assessed. To assess the risk of AGA, the Cox proportional hazards regression models were used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs). For medication exposure, cumulative doses and DDDs were utilized to evaluate dose-response relationships, accompanied by corresponding tests for trend. All models were adjusted for age, sex, and other potential confounding variables to mitigate the influence of confounding variables on AGA. Furthermore, stratified analyses were conducted to explore effect variations across different subgroups, and interaction tests were performed to evaluate whether the associations were modified by age, sex, and comorbidities. To ensure the robustness of the findings and minimize potential diagnostic misclassification, multi-dimensional sensitivity analyses were conducted for validation. Kaplan–Meier analysis was used to determine the probability of AGA occurrence throughout the follow-up period. The log-rank test was used to assess disparities between the groups. The statistical tests conducted were two-sided, and p-values of 0.05 or lower were deemed statistically significant. Log–Log plots (Supplemental Figs. 1 and 2) were used to test the proportional hazards hypothesis.

Results

Study population characteristics

Table 1 summarizes the demographic and clinical characteristics of the 2 cohorts after propensity score matching. Each AR and control cohort comprised 539,397 individuals. All baseline characteristics were well-balanced between the 2 cohorts (SMD <0.1) after matching. The mean follow-up period for AGA was 11.48 years in the AR group and 12.45 years in the control group.

Table 1.

Characteristics of allergic rhinitis and comparison cohorts.

Characteristics	n (%)		SMD
	Comparison cohort (n = 539,397)	Allergic rhinitis cohort (n = 539,397)
Season
Spring	139,476 (25.86)	142,443 (26.41)	0.013
Summer	123,379 (22.87)	120,643 (22.37)	0.012
Autumn	132,773 (24.62)	129,906 (24.08)	0.012
Winter	143,769 (26.65)	146,405 (27.14)	0.011
Age, years			0.081
<30	183,133 (33.95)	204,185 (37.85)
≥30	356,264 (66.05)	335,212 (62.15)
Mean ± SD	38.49 ± 18.46	37.31 ± 19.62	0.062
Sex			0.062
Male	275,206 (51.02)	258,412 (47.91)
Urbanization levelsa
1 (highest)	309,567 (57.39)	309,307 (57.34)	0.001
2	181,659 (33.68)	179,680 (33.31)	0.008
3	39,413 (7.31)	41,373 (7.67)	0.014
4 (lowest)	8758 (1.62)	9037 (1.68)	0.004
Unknown	309,567 (57.39)	309,307 (57.34)	0.001
Insurance amountsb, NT$
<20,000/unknown	345,914 (64.13)	340,463 (63.12)	0.021
1–19,999	133,496 (24.75)	132,856 (24.63)	0.003
20,000–39,999	59,987 (11.12)	66,078 (12.25)	0.035
≥40,000	345,914 (64.13)	340,463 (63.12)	0.021
Comorbiditiesc
Hypertension	173,139 (32.1)	171,077 (31.72)	0.008
Hyperlipidemia	166,400 (30.85)	175,595 (32.55)	0.037
Diabetes mellitus	100,720 (18.67)	100,081 (18.55)	0.003
Thyroid disorders	50,733 (9.41)	61,756 (11.45)	0.067
Systemic lupus erythematous	1650 (0.31)	2204 (0.41)	0.017
Rheumatoid arthritis	24,254 (4.5)	28,938 (5.36)	0.040
Vitiligo	1285 (0.24)	1601 (0.3)	0.011
Hematologic malignancies	3638 (0.67)	3949 (0.73)	0.007
Mental disorder	204,628 (37.94)	228,259 (42.32)	0.089
Eczema	38,987 (7.23)	48,320 (8.96)	0.063
Psoriasis	6166 (1.14)	8376 (1.55)	0.036
Crohn's disease	15,022 (2.78)	19,200 (3.56)	0.044
Inflammatory bowel disease	1294 (0.24)	1652 (0.31)	0.013
Malnutrition	6885 (1.28)	8002 (1.48)	0.018
Anemia	49,958 (9.26)	56,140 (10.41)	0.038
Alcohol use	13,037 (2.42)	12,214 (2.26)	0.010
Smoking	21,680 (4.02)	22,789 (4.22)	0.010
Polycystic ovary syndrome	6868 (1.27)	9683 (1.8)	0.042
Obesity	9098 (1.69)	11,357 (2.11)	0.031
Tinea capitis	1588 (0.29)	1907 (0.35)	0.010
Medicationsc
Methotrexate	1283 (0.24)	1597 (0.3)	0.011
Hydroxychloroquine	3829 (0.71)	4945 (0.92)	0.023
Sex hormones and modulators of the genital system	37,363 (6.93)	46,903 (8.7)	0.066
Follow-up time, year
Mean ± SD	11.48 ± 5.65	12.45 ± 5.41	0.177

SD, standard deviation; SMD: standardized mean difference. Values are expressed as means ± SD or number (percentage).

^aDefined at the beginning of the follow-up period.

^bThe average value during the follow-up period.

^cDefined before the survival date

Associations between AR and AGA

Table 2 demonstrates the correlation between the likelihood of AGA occurrence in the AR and control groups. After controlling for confounding factors, the AR cohort was associated with a higher risk of AGA than the control group (adjusted HR [aHR]: 1.81, 95% CI: 1.69–1.93). The AR group exhibited a significantly higher cumulative incidence rate than the control group during the follow-up period, based on the log-rank test results (p < 0.001; Supplemental Fig. 3). Moreover, the age-stratified analysis revealed a statistically substantial interaction between AR and an increased risk of developing AGA (p for interaction <0.001).

Table 2.

Stratified analysis of androgenetic alopecia incidence and hazard ratios in the allergic rhinitis and comparison cohorts.

Population		Study group	AGA	PY	Incidence ratea	Adjusted HRb (95% CI)	p-value
Total		Comparison (n = 539,397)	1296	6,190,012	0.21	1.00 (reference)	–
		AR (n = 539,397)	2478	6,717,392	0.37	1.81 (1.69–1.93)	<0.001
Age	<30 years	Comparison (n = 183,133)	774	2,347,191	0.33	1.00 (reference)	–
		AR (n = 204,185)	1428	2,835,384	0.50	1.59 (1.46–1.74)	<0.001
	≥30 years	Comparison (n = 356,264)	522	3,842,822	0.14	1.00 (reference)	–
		AR (n = 335,212)	1050	3,882,009	0.27	2.12 (1.91–2.36)	<0.001
	p for interaction						<0.001
Sex	Female	Comparison (n = 264,191)	546	3,006,464	0.18	1.00 (reference)	–
		AR (n = 280,985)	1141	3,484,232	0.33	1.82 (1.65–2.02)	<0.001
	Male	Comparison (n = 275,206)	750	3,183,548	0.24	1.00 (reference)	–
		AR (n = 258,412)	1337	3,233,160	0.41	1.80 (1.64–1.97)	<0.001
	p for interaction						0.699
Comorbidity	No	Comparison (n = 176,172)	681	2,100,955	0.32	1.00 (reference)	–
		AR (n = 149,475)	1079	1,869,370	0.58	1.83 (1.66–2.01)	<0.001
	Yes	Comparison (n = 363,225)	615	4,089,057	0.15	1.00 (reference)	–
		AR (n = 389,922)	1399	4,848,022	0.29	1.81 (1.64–1.99)	<0.001
	p for interaction						0.245
Medication	No	Comparison (n = 498,317)	1239	5,733,563	0.22	1.00 (reference)	–
		AR (n = 487,840)	2339	6,072,357	0.39	1.81 (1.69–1.94)	<0.001
	Yes	Comparison (n = 41,080)	57	456,449	0.12	1.00 (reference)	–
		AR (n = 51,557)	139	645,035	0.22	1.76 (1.29–2.40)	<0.001
	p for interaction						0.855

AGA, androgenetic alopecia; AR, allergic rhinitis; CI, confidence interval; HR, hazard ratio; PY, person-year.

^aper 1000 person-years.

^bCox regression models were adjusted for season, age, sex, urbanization level, insurance amount, comorbidities (hypertension, hyperlipidemia, diabetes mellitus, thyroid disease, systemic lupus erythematous, rheumatoid arthritis, vitiligo, hematologic malignancies, mental disorders, eczema, psoriasis, Crohn's disease, inflammatory bowel disease, malnutrition, alcohol use, smoking, polycystic ovary syndrome, obesity, tinea capitis), and medications (methotrexate, hydroxychloroquine, sex hormones and modulators of the genital system).

Associations between patients with AR receiving second-generation H1-antihistamines and AGA

Table 3 presents the correlation between the risk of AGA and the use of sgSH medications in patients with AR, including the effect of this risk in various stratified groups. After adjusting for confounding factors, the risk of AGA was found to be lower in the sgSH-treated group than in the untreated group (aHR: 0.23, 95% CI: 0.20–0.26). Furthermore, stratified analyses revealed that AR patients aged less than 30 years and those aged 30 years or older who received sgSH treatment had a significantly reduced risk of AGA than those who did not (aHRs: 0.18 and 0.29, respectively). This evidence that patients with AR who start using sgSH agents before the age of 30 are less likely to develop AGA in the future than those who start after the age of 30 (p for interaction = 0.017).

Table 3.

Stratified analyses of androgenetic alopecia risk in patients with allergic rhinitis receiving or not receiving second-generation H1-antihistamines.

Population		sgSH	AGA	Adjusted HRa (95% CI)	p-value
Total		Without (n = 22,449)	293	1.00 (reference)	–
		With (n = 516,948)	2185	0.23 (0.20–0.26)	<0.001
Age	<30 years	Without (n = 5337)	151	1.00 (reference)	–
		With (n = 198,848)	1277	0.18 (0.15–0.21)	<0.001
	≥30 years	Without (n = 17,112)	142	1.00 (reference)	–
		With (n = 318,100)	908	0.29 (0.24–0.34)	<0.001
	p for interaction				0.017
Sex	Female	Without (n = 10,479)	120	1.00 (reference)	–
		With (n = 270,506)	1021	0.22 (0.18–0.27)	<0.001
	Male	Without (n = 11,970)	173	1.00 (reference)	–
		With (n = 246,442)	1164	0.24 (0.20–0.28)	<0.001
	p for interaction				0.582
Comorbidity	No	Without (n = 8013)	166	1.00 (reference)	–
		With (n = 141,462)	913	0.24 (0.20–0.28)	<0.001
	Yes	Without (n = 14,436)	127	1.00 (reference)	–
		With (n = 375,486)	1272	0.20 (0.17–0.24)	<0.001
	p for interaction				0.636
Medication	No	Without (n = 21,172)	282	1.00 (reference)	–
		With (n = 466,668)	2057	0.23 (0.20–0.26)	<0.001
	Yes	Without (n = 1277)	11	1.00 (reference)	–
		With (n = 50,280)	128	0.15 (0.08–0.29)	<0.001
	p for interaction				0.324

AGA, androgenetic alopecia; CI, confidence interval; HR, hazard ratio; sgSH, second-generation H1-antihistamine.

^aCox regression models were adjusted for season, age, sex, urbanization level, insurance amount, comorbidities (hypertension, hyperlipidemia, diabetes mellitus, thyroid disease, systemic lupus erythematous, rheumatoid arthritis, vitiligo, hematologic malignancies, mental disorders, eczema, psoriasis, Crohn's disease, inflammatory bowel disease, malnutrition, alcohol use, smoking, polycystic ovary syndrome, obesity, tinea capitis), and medications (methotrexate, hydroxychloroquine, sex hormones and modulators of the genital system)

Supplemental Fig. 4 displays the cumulative impact of varying dosages of sgSH agents on the risk of AGA in patients with AR. Patients with AR who used low-dose (<1285 mg), medium-dose (1285–4379 mg), and high-dose (≥4380 mg) treatments experienced a significant reduction in the likelihood of developing AGA compared to those who did not receive any treatment. For patients using low-dose sgSH medications, the aHR was 0.36 (95% CI: 0.32–0.41), for medium-dose it was 0.19 (95% CI: 0.16–0.22), and for high-dose it was 0.12 (95% CI: 0.10–0.13). Trend analysis indicated a statistically significant association between cumulative dosage and the risk of AGA (p trend <0.001). Moreover, further stratified analyses yielded consistent findings (p trend <0.001). sgSH may exhibit superior prophylactic efficacy against AGA in patients with AR under the age of 30, followed by those with AR but without any concurrent medical conditions (p for interaction <0.001). Nevertheless, the stratified analyses for different sexes revealed comparable risk ratios for AGA (p for interaction <0.001).

Furthermore, when considering the cumulative use of sgSH agents measured in DDDs over time, consistent results were observed. Compared to patients with AR who did not use sgSH agents, the aHR were 0.32 (95% CI: 0.28–0.37, p < 0.001), 0.20 (95% CI: 0.17–0.23, p < 0.001), and 0.16 (95% CI: 0.13–0.18, p < 0.001) for cumulative use of less than 45, 45–139, and ≥140 DDDs, respectively (Fig. 2). Similar findings were observed in other subgroups categorized based on specific characteristics, showing a notable decrease in the incidence of AGA with an increase in the cumulative DDD usage (p trend <0.001).

Fig. 2.

Forest plot of cumulative defined daily dose of second-generation H1-antihistamines and the risk of androgenetic alopecia in patients with allergic rhinitis. CI, confidence interval; DDD, defined daily dose; HR, hazard ratio; sgSH, second-generation H1-antihistamine. ^aCox regression models were adjusted for season, age, sex, urbanization level, insurance amount, comorbidities (hypertension, hyperlipidemia, diabetes mellitus, thyroid disease, systemic lupus erythematous, rheumatoid arthritis, vitiligo, hematologic malignancies, mental disorders, eczema, psoriasis, Crohn's disease, inflammatory bowel disease, malnutrition, alcohol use, smoking, polycystic ovary syndrome, obesity, and tinea capitis), and medications (methotrexate, hydroxychloroquine, and sex hormones and modulators of the genital system). All p-values were <0.001

Supplemental Fig. 5 displays the cumulative incidence curves of AGA in the AR cohort estimated using the Kaplan–Meier technique. Throughout the observation period, the occurrence rate of AGA in the group of patients with AR who were treated with sgSH agents was lower than in the untreated group (log-rank test, p < 0.001; Supplemental Fig. 5A). Furthermore, there was a direct correlation between the cumulative use of sgSH agents and a decrease in the overall occurrence of AGA. This correlation remained consistent over time, as indicated by the log-rank test (p < 0.001; Supplemental Fig. 5B and C).

Sensitivity analysis

Supplemental Table 2 presents multi-dimensional sensitivity analyses conducted to validate the robustness of the study findings and minimize potential diagnostic misclassification bias. To enhance diagnostic precision, the definition of AR exposure was refined by excluding cases of non-allergic rhinitis (such as vasomotor rhinitis or structural nasal disorders). Additionally, the outcome of AGA was restricted to cases confirmed by dermatologists, while excluding non-AGA hair loss types such as alopecia areata or telogen effluvium. The results demonstrated that after applying these stringent inclusion criteria, the association between AR and the subsequent risk of AGA increased to an aHR of 3.13 (95% CI: 2.56–3.82, p < 0.001). Notably, within this rigorous cohort, the protective effect of sgSH remained stable and significant (aHR: 0.37; 95% CI: 0.26–0.53, p < 0.001).

Furthermore, to ensure temporality and examine the robustness of the results across various exposure definitions, individuals who developed AGA within 1 year after the index date were excluded. In this sensitivity analysis, sgSH exposure was redefined as stable use within the first year following AR diagnosis, encompassing patients with at least 2 prescription records or with cumulative use exceeding 30, 90, and 180 days. Supplemental Table 3 indicated that after excluding lag-period cases, the aHR for those with at least 2 prescriptions in the first year was 0.40 (95% CI: 0.33–0.47, p < 0.001). When defined by cumulative use exceeding 30, 90, and 180 days, the respective aHRs were 0.39 (95% CI: 0.33–0.48, p < 0.001), 0.36 (95% CI: 0.27–0.49, p < 0.001), and 0.24 (95% CI: 0.12–0.47, p < 0.001). These findings are consistent with a potential association between AR-related inflammatory processes and the development of AGA, and they support the robustness of the study results across multiple analytical assumptions.

Discussion

This large population-based cohort study found a significant correlation between AR and an increased incidence of AGA. These results suggest that AR, as a systemic Th2-mediated allergic disease, may be associated with perifollicular micro-inflammation by increasing the systemic inflammatory load and the release of prostaglandins (such as PGD₂), thereby accelerating the transition of hair follicles from the anagen to the telogen phase. Notably, our findings demonstrate that sgSH treatment is associated with a significantly reduced risk of AGA among AR patients. This association not only lends support to the inflammatory hypothesis of AGA pathogenesis but also suggests the potential utility of systemic anti-inflammatory interventions in mitigating the risk of hair loss.

While AGA was traditionally attributed to androgenic factors, increasing evidence indicates that follicular micro-inflammation plays a vital role in its pathogenesis. Histopathological studies have demonstrated perifollicular inflammatory cell infiltration in approximately 50%–71% of scalp samples from AGA patients, characterized by the activation of lymphocytes, histiocytes, and CD4⁺ T cells specifically at the follicular bulge 32, 33, 34, 35. Further investigations into the immune microenvironment have revealed increased infiltration of mast cells and γδ T cells in alopecic regions, accompanied by the aberrant activation of innate immune mechanisms. This includes the upregulation of Toll-like receptor pathways and the inflammasome component CASP1, the latter of which facilitates the release of pro-inflammatory cytokines such as IL-1β and IL-18.^36,37 Moreover, the elevation of hub immune-related genes, such as MMP9, alongside the downregulation of BMP2 and THBS1, suggests that inflammation not only triggers immune signaling but also leads to extracellular matrix remodeling and dermal papilla dysfunction.^36,38,39 Although epidemiological literature directly exploring the comorbidity of AR and AGA remains limited, the findings of the present study are consistent with previous observations regarding the impact of atopic inflammation on follicular health.20, 21, 22^,40,41 Li et al noted that house dust mite allergy is associated with the severity of AA, suggesting a universal detrimental effect of systemic Th2 inflammation on hair follicles.⁴⁰ Similarly, Kridin et al observed that patients with AA have a significantly higher risk of asthma, atopic dermatitis, AR, and allergic conjunctivitis compared to controls, with the association strengthening as the number of comorbidities increases.²⁰ This association was further confirmed to be bidirectional in a systematic review.²¹ Collectively, these findings support the role of inflammatory and immune dysregulation—both innate and adaptive—in driving follicle miniaturization and disrupting the hair cycle. As another classic Th2-mediated condition, AR may create an unfavorable background for hair growth within the scalp microenvironment through similar immune dysregulation pathways, thereby exacerbating the systemic inflammatory load. This association may stem from atopic cytokines (eg, IL-4 and IL-13) and elevated concentrations of PGD₂, mediators known to interfere with follicular signaling, promote the transition of follicles into the telogen phase, and induce the miniaturization process.16, 17, 18, 19

Within this framework, the interplay between inflammatory and hormonal pathways³⁷ may provide a novel perspective on the progression of AGA. While the core mechanism of AGA involves the binding of androgens to their receptors to inhibit hair follicle growth, chronic inflammation elicited by AR may play a pivotal role in triggering and enhancing receptor susceptibility. Specifically, inflammatory mediators produced during AR, such as PGD₂ and Th2 cytokines, may alter the follicular microenvironment, inducing heightened sensitivity of hair follicle cells to androgens. Among these, PGD₂ is recognized as a crucial molecular link between allergic responses and hair follicle biology. As a signature lipid mediator in mast cell and Th2 reactions, PGD₂ has been shown to be significantly upregulated in the balding scalps of AGA patients, serving as a key factor in the local inhibition of hair growth.¹⁶ Furthermore, research has indicated that PGD₂ can activate the AKT signaling pathway via its receptor, DP2 (GPR44), which promotes both the expression and nuclear activity of the androgen receptor within dermal papilla cells. This process simultaneously induces the expression of TGF-β1 and other hair-inhibiting or catagen-promoting factors, thereby shaping a follicular microenvironment that is hypersensitive to androgens.⁴² Consequently, even in the absence of elevated systemic androgen levels, hair follicles may be more prone to initiating the catagen phase and the miniaturization process due to the amplification of androgen receptor signaling. Collectively, these lines of evidence suggest that the Th2-mediated inflammation and PGD₂-related signaling associated with AR are not merely comorbid phenomena; rather, they may contribute to the development and progression of AGA by remodeling the follicular microenvironment and amplifying androgen receptor signaling.

Beyond inflammatory pathways, both genetic and environmental determinants likely orchestrate the interplay between AR and AGA. While both conditions exhibit significant heritability, genome-wide association studies have yet to confirm shared susceptibility loci. Nevertheless, the Wnt/β-catenin signaling pathway emerges as a compelling molecular bridge linking these 2 disorders. Evidence underscores that microRNA miR-29a-3p can compromise nasal epithelial barrier integrity by directly targeting CTNNB1 (the gene encoding β-catenin), while the crosstalk between Wnt/β-catenin and TGF-β signaling is concurrently implicated in the pathogenesis of AGA.^6,43 Such systemic epithelial barrier dysfunction may manifest across both the nasal mucosa and scalp skin, not only facilitating allergen intrusion but also precipitating follicular growth arrest. Furthermore, environmental stressors (such as pollutants and psychosocial strain) may interface with genetic susceptibility via neuro-immune axes, potentially triggering dysbiosis within both nasal and scalp microenvironments.6, 7, 8^,44 This process may be further modulated by systemic regulation through the gut-immune axis, fueling the persistent release of inflammatory mediators and ultimately driving tissue remodeling, characterized by mucosal hyperplasia or follicular atrophy.⁶ Collectively, these multifaceted factors likely exert synergistic effects alongside primary inflammatory pathways. Although direct epidemiological or mechanistic evidence linking these specific circuits remains to be fully elucidated, it is plausible that in predisposed individuals, systemic stress-induced neuro-immune activation exacerbates both mucosal allergic inflammation and perifollicular micro-inflammation, thereby compounding the severity of both AR and AGA.

Given that sgSH are capable of suppressing mast cell activation and Th2-mediated inflammatory responses, accumulating evidence suggests that targeting inflammatory pathways may represent a feasible therapeutic strategy. The concept of systemic anti-inflammatory intervention has gained increasing attention in prior studies. The conventional anti-androgen agent finasteride, beyond its established role in inhibiting dihydrotestosterone, has also been shown to downregulate inflammasome-associated pro-inflammatory cytokines, such as IL-1β, within the scalp.^36,37 These observations imply that reducing systemic or local inflammatory burden may confer a broader benefit in maintaining hair follicle homeostasis. Prior research suggests that certain antihistamines, particularly cetirizine and levocetirizine, may affect hair growth by modulating prostaglandins. Prostaglandins are critical to hair follicle physiology.⁴⁵ PGE₂ and PGF_2α promote hair growth, while PGD₂ inhibits it.16, 17, 18, 19 Studies show that cetirizine may reduce PGD₂ release and increase PGE₂ production, thereby modifying the hair follicle microenvironment and supporting hair development.^26,27 These changes have been associated with increased hair density, larger hair diameters, and prolonged anagen phase duration in small-scale clinical studies.29, 30, 31 In vitro evidence further supports that levocetirizine inhibits the PGD₂-GPR44 pathway, promoting dermal papilla cell proliferation.^46,47 In summary, modulating prostaglandin levels may provide a new path for treating AGA, and cetirizine and levocetirizine have features that are appropriate for patients with AGA. Conversely, while other antihistamines—such as fexofenadine, loratadine, ebastine, mizolastine, azelastine, and desloratadine—exhibit anti-inflammatory properties, there is currently no direct evidence that they modulate PGD₂, PGE₂, or other prostaglandins, and their relationship with hair growth remains unclear. As a result, existing research has largely focused on cetirizine and its derivatives, leaving the effects of other sgSH medications largely unexplored. This study investigated the impact of several sgSH agents on AGA risk to address this knowledge gap. As shown in Supplemental Table 4, all medications—except ebastine and mizolastine—were significantly associated with a lower incidence of AGA. These epidemiologic findings suggest that sgSH may share protective properties that extend beyond prostaglandin modulation alone. This observation further supports the potential relevance of systemic anti-inflammatory approaches in the context of AGA prevention. In conclusion, our results call for a reevaluation of the potential roles of various antihistamines in hair physiology. Future studies should investigate whether these agents exert protective effects by modulating immune and inflammatory pathways, potentially improving therapeutic strategies for AGA.

As indicated in Table 1, psychiatric disorders, hypertension, hyperlipidemia, and diabetes mellitus manifested a high prevalence within the study population. To more thoroughly elucidate the risk-benefit profile of sgSH, we conducted supplementary stratified analyses focusing on these specific comorbidities. Supplemental Table 5 demonstrate that the protective efficacy of sgSH against AGA remained uniform across all strata, with interaction tests yielding no statistically significant differences (p for interaction >0.05). These findings suggest that the clinical utility of sgSH persists in patients with concurrent chronic conditions. Notably, age-stratified analyses showed that sgSH agents had weaker protective effects in individuals over 30 years old compared to those aged 30 or younger (p for interaction = 0.017). Several factors may explain this age-related difference.⁴⁸ Hair follicle structures in younger individuals remain highly adaptable, and their immune environment may be more responsive to pharmacological modulation.⁴⁹ Early-onset AGA may reflect a phenotype with heightened immunologic involvement,⁵⁰ making this group more receptive to agents with anti-inflammatory or immunomodulatory properties. Additionally, chronic inflammation tends to worsen with age, reducing the efficacy of pharmacologic interventions and contributing to microscopic follicular inflammation—a proposed mechanism of AGA.^51,52 Although further research is needed to clarify the underlying mechanisms, these findings may have meaningful implications for clinical practice. Given the observed biological plausibility linking AR and AGA, clinicians may consider maintaining heightened awareness of early signs of hair loss when managing patients with chronic AR. Early clinical assessment—such as simple visual inspection or inquiry into family history—could facilitate timely identification before irreversible follicular miniaturization occurs. Moreover, in individuals with coexisting AR and AGA, the utilization of sgSHs, which have a favorable safety profile, for long-term inflammatory control may offer potential ancillary benefits for hair health while effectively managing respiratory symptoms.

This study has several limitations that warrant consideration. First, as an observational analysis, this retrospective cohort study is limited by the nature of claims-based records, which may involve inaccuracies in diagnostic coding and potential disease misclassification. Second, the lack of specific clinical and biological data, including inflammatory biomarkers (such as IgE levels), immunological profiles, prostaglandin levels, scalp histological assessments, hair density measurements, or standardized evaluations of allergy severity. This limits our ability to directly correlate the laboratory severity of atopy with the degree of follicular miniaturization. Third, although we estimated exposure based on cumulative dose and observed a significant dose–response trend (p-trend <0.001). It must be acknowledged that prescription records alone cannot definitively ascertain actual medication adherence. Fourth, despite adjusting for many confounders, the presence of residual confounding remains a possibility. Key genetic and lifestyle data, such as family history, stress levels, dietary intake, or body mass index, were uncaptured in the administrative records but may simultaneously affect to the pathogenesis both AR and AGA. Finally, given potential differences in the prevalence of alopecia and treatment practices between Taiwan and other countries, the generalizability of our findings to other populations may be limited. Despite these limitations, this study employed rigorous inclusion criteria and statistical approaches—including stratified analyses, sensitivity analyses, and adjustment for multiple confounders—to maximize the robustness of the results and mitigate potential biases. Moreover, this is the first study to investigate this association using a nationwide Taiwanese population database, with the longest follow-up period reported to date, providing preliminary clinical evidence for the epidemiological link between allergy inflammation and AGA.

In conclusion, we employing a large-scale retrospective cohort study to found that AR is associated with an increased risk of AGA and that sgSH use may be linked to a reduced incidence of AGA in individuals with AR. These findings highlight potential connections among immunological pathways, prostaglandin signaling, and hair follicle development. Importantly, the possibility that the observed associations reflect residual confounding, rather than direct biological effects, should be considered. Although these medications are FDA-approved for allergic conditions but not for the prevention or treatment of AGA. Therefore, the use of H1-antihistamines in this context is considered off-label and exploratory. Future research should explore to further elucidate the complex relationship between allergic diseases and AGA from multiple perspectives. Genetic linkage studies may help identify potential shared susceptibility loci underlying both conditions. In addition, in-depth immunological analyses—particularly examining cytokine patterns—should be conducted in AGA patients with and without concomitant atopic dermatitis to assess the specific effects of systemic anti-inflammatory interventions on the progression of AGA. Finally, larger-scale translational studies, prospective longitudinal cohort investigations, and randomized controlled trials are warranted to clarify the underlying pathophysiological mechanisms through biomarker analyses and to evaluate the potential of sgSH as a preventive or adjunctive therapeutic strategy for AGA.

Ethics statement

This study was reviewed and approved by the Committee on Research Ethics, China Medical University Hospital Clinical Trial Center; approval: CMUH112-REC1-117(CR-2). To protect patient privacy, the NHI anonymizes data from patients and caregivers before distributing it for research purposes. Consequently, the Research Ethics Committee decided to waive the need for informed consent. The study was conducted in accordance with the Declaration of Helsinki of the World Medical Association.

Data availability statement

The data underlying this article were provided by Taiwan's Ministry of Health and Welfare Data Science Center (HWDC) under license/by permission. Data are available at https://dep.mohw.gov.tw/DOS/np-2500-113.html with permission from HWDC.

Author contributions

Han-Wei Zhang: Conceptualization, Investigation, Methodology, Project administration, Software, Supervision, Validation, Writing–review & editing, Ya-Ting Chang: Conceptualization, Investigation, Methodology, Project administration, Supervision, Validation, Writing–original draft, Writing–review & editing, Yi-Jie Kuo: Conceptualization, Investigation, Methodology, Project administration, Supervision, Validation, Writing–original draft, Writing–review & editing, Yu-Pin Chen: Conceptualization, Investigation, Methodology, Project administration, Supervision, Validation, Writing–review & editing, Yu-Shan Lin: Formal analysis, Investigation, Methodology, Software, Visualization, Writing–original draft, Fuu-Jen Tsai: Data curation, Resources.

Disclosure of the use of generative AI and AI-assisted technologies

The authors used ChatGPT (GPT-5, OpenAI) solely for English grammar and fluency improvement. The authors reviewed and approved all AI-assisted text and take full responsibility for the scientific content.

Funding sources

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of competing interests

None declared.

Appendix A. Supplementary data

The following is the Supplementary data to this article.