Abstract
Background
The influence of circadian disruption from night shift work and poor sleep on allergic rhinitis (AR) risk remains inadequately characterized. The objective of this study was to advance the understanding of how circadian disruption influence AR risk, thereby providing new insights into the prevention and management of AR.
Methods
This prospective cohort study analyzed data from the UK Biobank, comprising 256,945 participants for shift work analysis and 374,672 for sleep pattern assessment. Shift work and sleep behaviors were self-reported at baseline. AR incidents were identified through linked hospital records. Multivariable Cox proportional hazards models were employed to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). Mediation analysis was conducted to assess the potential role of metabolic biomarkers.
Results
Irregular night shift work was associated with an increased risk of AR (HR = 1.21, 95% CI: 1.01–1.45), whereas permanent night shifts showed no clear association. Each categorical decrease in sleep pattern quality corresponded to a 16% increase in AR risk (HR = 1.16, 95% CI: 1.08–1.25). Specific unhealthy sleep behaviors—including evening chronotype, short sleep duration (≤ 6 h), insomnia, and snoring—were also associated with higher AR risk. Small HDL particle concentration partially mediated the association between sleep patterns and AR, accounting for 1.59% of the effect (95% CI: 0.12–5.28).
Conclusion
Both irregular night shift work and poor sleep patterns are associated with an increased risk of AR, with small HDL particles acting as a partial mediator. These findings underscore the importance of maintaining circadian rhythm stability and improved sleep hygiene in the prevention and management of AR.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12889-026-26827-1.
Keywords: Allergic rhinitis, Night shift work, Sleep patterns, Circadian disruption, UK Biobank
Introduction
Allergic rhinitis (AR) is a globally prevalent health condition, affecting approximately 10% to 40% of the population worldwide [1]. It considerably impairs quality of life and places substantial psychological and economic burdens on individuals, families, and society at large [2]. The pathogenesis of AR involves immune dysregulation, and growing evidence suggests that environmental exposures—such as air pollution and climate change—may also contribute to its development [3]. Given the limitations of current AR treatments—which are often costly, exhibit suboptimal efficacy, and are associated with poor adherence [4]—there is a growing imperative to identify modifiable risk factors, especially those linked to lifestyle behaviors, and to implement effective preventive strategies.
Circadian rhythms play a critical role in regulating immune function and have been implicated in various allergic disorders [5, 6]. In contemporary society, night shift work and sleep disturbances are increasingly common and may contribute to chronic circadian disruption, potentially exacerbating allergic inflammation and serving as modifiable risk factors for diseases such as AR [7, 8]. However, current evidence linking night shift work and sleep disorders to allergic outcomes remains limited and predominantly derived from cross-sectional studies. For instance, night shift work has been associated with asthma [9] and chronic spontaneous urticaria [10], but its relationship with AR remains underexplored. While some studies suggest a possible link between evening chronotype and AR [11, 12], sleep is a multi-dimensional behavior. Other key dimensions — such as sleep duration, insomnia, snoring, and daytime sleepiness [13] — have been less extensively investigated in relation to AR. More importantly, existing evidence is fragmented, as few studies have examined these dimensions concurrently or integrated them into a composite sleep pattern to assess their combined association with AR.
To address this knowledge gap, we conducted a prospective analysis using data from the UK Biobank. Our primary aim was to separately examine the associations of night shift work and composite sleep patterns with incident AR, while also independently assessing the roles of specific sleep behaviors, including chronotype, sleep duration, insomnia, snoring, and daytime sleepiness. Additionally, we sought to investigate the combined effect of night shift work and composite sleep patterns on AR risk. Furthermore, to elucidate potential biological mechanisms, we assessed the mediating effects of metabolomic biomarkers in the relationship between composite sleep patterns and AR. Together, these analyses seek to advance the understanding of how circadian disruption influence AR risk, thereby providing new insights into the prevention and management of AR.
Method
Study population
The UK Biobank study is an ongoing prospective cohort study that recruited over 500,000 participants aged 37–73 from across the United Kingdom between 2006 and 2010. At baseline, all participants completed touch-screen questionnaires, underwent comprehensive physical measurements, and provided biological samples. Follow-up assessments have been conducted periodically to update this information. A detailed description of the UK Biobank study design and methodology has been published previously [14]. The study received ethical approval from the North West Multicenter Research Ethical Committee (Ref 11/NW/0382), and all participants provided written informed consent prior to enrollment.
The participant selection process is shown in Fig. 1. First, we restricted the analysis to individuals who were engaged in paid employment or self-employment at baseline (n = 286,028). To ensure data integrity, we then excluded those who self-reported AR (n = 21,119), those diagnosed with AR prior to or within one year after enrollment (n = 226), and those with missing covariate data (n = 7,738, < 5% of the total number), resulting in a final analytical sample of 256,945 participants for the shift work analysis.
Fig. 1.
Flowchart for participant selection. AR, allergy rhinitis; NMR, Nuclear Magnetic Resonance
For the sleep pattern analysis, we again restricted the sample to participants with complete sleep pattern score data (n = 410,575). After excluding individuals who self-reported AR (n = 26,596), those diagnosed with AR before or within one year of enrollment (n = 325), and those with missing covariate information (n = 8,982, < 5% of the total number) to ensure data integrity, the final sample comprised 374,672 participants, among whom 186,299 had complete NMR metabolomics data.
Additionally, to explore potential interactions between sleep patterns and night shift work, we restricted our analysis to participants with complete data on both shift work and sleep patterns (n = 215,869).
Exposure
Shift work is a work schedule that falls outside of the normal daytime working hours of 9am-5pm. This may involve working afternoons, evenings or nights or rotating through these kinds of shifts. Night shifts are a work schedule that involves working through the normal sleeping hours, for instance working through the hours from 12am to 6am.
At baseline, employed participants were asked to report whether their current primary job involved shift work (Data-Field 826). If so, participants were further asked whether their primary job involved night shifts (Data-Field 3426). Based on the responses to these two questions, individuals were categorized into four groups according to their current shift work status: “Day workers”, “Shift work, but never or rarely night shifts”, “Irregular shift work including nights”, and “Permanent night shift work”, similar to the previous study [9]. The specific classification method is shown in Table S1.
During the baseline assessment, data on sleep behaviors—including sleep duration, chronotype, insomnia, snoring, and daytime sleepiness—were collected via a touchscreen-based questionnaire. A healthy sleep pattern was evaluated based on the those five sleep behaviors [15]. Each behavior was classified as either healthy or unhealthy. The specific classification method is shown in Table S2. Participants were assigned a score of 1 for each healthy behavior and 0 for each unhealthy behavior. A composite healthy sleep score was calculated by summing the individual scores across all five behaviors, resulting in a total score ranging from 0 to 5. Higher scores indicate a healthier sleep pattern. Subsequently, overall sleep patterns were categorized based on the healthy sleep score as follows: “Healthy sleep pattern” (a score of ≥ 4), “Intermediate sleep pattern” (a score of 2–3), “Poor sleep pattern” (a score of ≤ 1) [13].
When examining the combined effects of night shift work and sleep patterns, due to the limited number of participants with “Poor sleep patterns”, we merged the “Intermediate sleep pattern” and " Poor sleep pattern” into a single “Unhealthy sleep pattern” group. A joint exposure variable comprising eight subgroups was then created by crossing the four-category shift work variable (“Day workers”, “Shift work, but never or rarely night shifts”, “Irregular shift work including nights”, and “Permanent night shift work”) with the two-category sleep pattern variable (“Healthy” and “Unhealthy”).
Outcomes
The diagnosis of AR was determined using diagnostic codes recorded in primary or secondary positions within hospital inpatient records. This included patients with International Classification of Diseases version 9 (ICD-9) or ICD-10 codes specific to AR (ICD-9 codes: 477.0, 477.8, 477.9; ICD-10 codes: J30.1, J30.2, J30.3, J30.4) [16]. For each participant, the follow-up duration was calculated from the date of enrollment until the earliest occurrence of any of the following events: AR diagnosis, death, loss to follow-up, or the end of the follow-up period. The end of complete follow-up was based on the availability of electronic health record data in the UK Biobank, with the censoring dates being 31 October 2022 for England, 31 August 2022 for Scotland, and 31 May 2022 for Wales [17].
NMR metabolomic biomarkers
EDTA-plasma samples were collected during the baseline assessment between April 2007 and December 2020 [18]. Metabolic biomarkers were quantified using a high-throughput nuclear magnetic resonance (NMR)-based metabolic biomarker profiling developed by Nightingale Health Ltd. The biomarkers cover a wide range of metabolic pathways, including lipoprotein lipids across 14 subclasses, fatty acids and fatty acid composition, as well as various low-molecular-weight metabolites such as amino acids, ketone bodies, and glycolysis metabolites quantified in molar concentration units. Detailed information on these metabolites is provided in Table S3.
Covariates
Based on the previous studies [9, 13] and directed acyclic graphs (Figure S1), our analysis included the following covariates: age (continuous, Data-Field 21022), sex (male/female, Data-Field 31), ethnicity (White/Asian or Asian British/Black or Black British/Mixed/Other, Data-Field 21000), education level (High/Intermediate/Low, Data-Field 6138, re-grouped based on a previous study [19]), Townsend Deprivation Index (TDI, continuous, Data-Field 22189), body mass index (BMI, continuous, Data-Field 21001), smoking status (Current/Previous/Never, Data-Field 20116), alcohol consumption status (Current/Previous/Never, Data-Field 20117), asthma (Yes/No), allergy (Yes/No), and eczema/dermatitis (Yes/No). Among these, asthma, allergy, and eczema/dermatitis were determined based on self-reported data (Data-Field 20002), with the following diagnostic codes: asthma (code 1111), allergy (codes 1374, 1385, 1386, 1563, 1668, 1562), and eczema/dermatitis (code 1452).
Statistical analysis
In the baseline characteristics analysis of participants, continuous variables are presented as mean ± standard deviation (SD), while categorical variables are expressed as numbers (percentages).
We employed Cox proportional hazards models to estimate the associations between night shift work or sleep patterns and the risk of AR, with results reported as hazard ratios (HRs) and 95% confidence intervals (CIs). After fitting the Cox proportional hazards models, the proportional hazards assumption was examined using Schoenfeld residuals, and no variables were found to violate the assumption. Model 1 showed crude associations without adjustment for any covariates. Model 2 was adjusted for age and sex. Model 3 was further adjusted for ethnicity, education level, TDI, BMI, smoking status, and alcohol consumption status. Model 4 included all covariates from Model 3, with additional adjustment for asthma, allergy, and eczema/dermatitis. In all models, “Day workers” or “Healthy sleep pattern” was set as the reference group. To examine the linear dose-response relationship, sleep patterns were modeled as a continuous variable (coded as 1 = Healthy, 2 = Intermediate, 3 = Poor) to estimate the HR and 95% CI for each one-category deterioration. In addition, we conducted a sensitivity analysis in Model 4, in which BMI was winsorized at the 1st and 99th percentiles to reduce the potential impact of outliers on the main findings. In Model 4, we also evaluated the associations between each individual sleep behavior and the risk of AR.
Subsequently, we conducted a joint exposure analysis. Using the " Day workers & Healthy " group as the reference, we evaluated the association between the joint exposure variable and AR risk using fully adjusted Cox proportional hazards models. Additionally, we assessed potential interactions between shift work and sleep patterns by comparing fully adjusted models with and without the interaction term using the likelihood ratio test.
We also conducted subgroup analyses stratified by age (< 60, ≥ 60 years), sex (male, female), ethnicity (White, others), BMI (< 25, 25–30, ≥ 30 kg/m²), smoking status (current, never, previous), alcohol consumption (current, never, previous), TDI (quartiles, Q1–Q4), education level (high, intermediate, low), asthma (yes, no), allergy (yes, no), and eczema/dermatitis (yes, no) to evaluate potential interactions. The fully adjusted Model 4 was used as the primary multivariate model for these analyses. Interaction effects were assessed using the likelihood ratio test, which compared models with and without interaction term. The P-values for interaction were calculated and further adjusted for false discovery rate (FDR). For the NMR metabolomics data, we excluded datasets containing missing values and normalized the raw data using z-score normalization. To identify metabolites that could serve as potential mediators, we applied multiple linear regression models to evaluate the associations between sleep patterns and these metabolites, while Cox regression models were employed to explore the relationships between sleep patterns, biomarkers, and AR [20].Metabolites were selected for mediation analysis if the 95% CIs for both associations (sleep→metabolite β and metabolite→AR HR) excluded the null (0 for β; 1 for HR) and the two associations were directionally concordant (e.g., adverse sleep → higher metabolite → higher AR risk). The proportion mediated (PM) was calculated using the regression-based (RB) method from the “CMAverse” package, and its 95% CI was estimated via the bootstrap method with 500 resamples [21].
All statistical analyses were performed using the R software (version 4.5.0).
Result
Irregular night shift work pattern increased allergic rhinitis risk
Table 1 presents the baseline characteristics of the study participants by night shift work status. Over a median follow-up period of 13.70 years (IQR: 12.98–14.32), a total of 1,800 incident AR cases were identified. The overall mean age of all participants was 52.86 years (SD: 7.09), 47.9% were male, and 94.7% identified as White. Compared with daytime workers, shift workers had a higher proportion of males and a lower proportion of White individuals. They also exhibited higher BMI, higher TDI scores, lower education level, a greater proportion of current smokers, a lower proportion of current alcohol drinkers, and lower baseline prevalence of allergies and eczema/dermatitis.
Table 1.
Baseline characteristics of participants by shift work patterns. (n = 256,945)
| Overall | Day workers | Shift work, but never or rarely night shifts | Irregular shift work including nights | Permanent night shift work | |
|---|---|---|---|---|---|
| n | 256,945 | 212,541 | 21,748 | 16,225 | 6431 |
| Sex (Male %) | 123,021 (47.9) | 98,576 (46.4) | 10,353 (47.6) | 10,133 (62.5) | 3959 (61.6) |
| Age (years, mean (SD)) | 52.86 (7.09) | 53.05 (7.10) | 52.62 (7.05) | 51.26 (6.84) | 51.60 (6.90) |
| Ethnicity (%) | |||||
| White | 243,271 (94.7) | 203,286 (95.6) | 19,965 (91.8) | 14,323 (88.3) | 5697 (88.6) |
| Asian or Asian British | 5018 (2.0) | 3533 (1.7) | 710 (3.3) | 576 (3.6) | 199 (3.1) |
| Black or Black British | 4627 (1.8) | 2924 (1.4) | 560 (2.6) | 802 (4.9) | 341 (5.3) |
| Mixed | 1750 (0.7) | 1359 (0.6) | 189 (0.9) | 149 (0.9) | 53 (0.8) |
| Other | 2279 (0.9) | 1439 (0.7) | 324 (1.5) | 375 (2.3) | 141 (2.2) |
| BMI (kg/m2, mean (SD)) | 27.28 (4.72) | 27.11 (4.66) | 27.82 (4.99) | 28.25 (4.93) | 28.53 (4.87) |
| TDI (mean (SD)) | -1.33 (3.01) | -1.49 (2.92) | -0.63 (3.22) | -0.53 (3.31) | -0.34 (3.32) |
| Education (%) | |||||
| High | 95,788 (37.3) | 86,010 (40.5) | 5290 (24.3) | 3598 (22.2) | 890 (13.8) |
| Intermediate | 86,951 (33.8) | 71,003 (33.4) | 7755 (35.7) | 5871 (36.2) | 2322 (36.1) |
| Low | 74,206 (28.9) | 55,528 (26.1) | 8703 (40.0) | 6756 (41.6) | 3219 (50.1) |
| Smoke (%) | |||||
| Current | 28,179 (11.0) | 21,229 (10.0) | 3076 (14.1) | 2711 (16.7) | 1163 (18.1) |
| Never | 146,082 (56.9) | 122,725 (57.7) | 11,557 (53.1) | 8498 (52.4) | 3302 (51.3) |
| Previous | 82,684 (32.2) | 68,587 (32.3) | 7115 (32.7) | 5016 (30.9) | 1966 (30.6) |
| Alcohol (%) | |||||
| Current | 241,453 (94.0) | 200,765 (94.5) | 19,989 (91.9) | 14,910 (91.9) | 5789 (90.0) |
| Never | 8448 (3.3) | 6301 (3.0) | 975 (4.5) | 780 (4.8) | 392 (6.1) |
| Previous | 7044 (2.7) | 5475 (2.6) | 784 (3.6) | 535 (3.3) | 250 (3.9) |
| Asthma (%) | 25,490 (9.9) | 21,040 (9.9) | 2249 (10.3) | 1563 (9.6) | 638 (9.9) |
| Allergy (%) | 5923 (2.3) | 5023 (2.4) | 464 (2.1) | 311 (1.9) | 125 (1.9) |
| Eczema/dermatitis (%) | 6599 (2.6) | 5598 (2.6) | 494 (2.3) | 371 (2.3) | 136 (2.1) |
Values are presented as number (%) for categorical variables and mean (SD) for continuous variables
n Number of cases, SD Standard deviation, BMI Body mass index, TDI Townsend Deprivation Index
Figure 2 and Table S4 illustrates the relationship between night shift work patterns and the incidence of AR. In the fully adjusted model (Model 4), irregular night shift workers had a higher estimated risk of AR compared to day workers (HR = 1.21, 95% CI: 1.01–1.45), with the confidence interval excluding the null value. For never or rarely night shift (HR = 0.95, 95% CI: 0.80–1.13) and permanent night shift workers (HR = 0.91, 95% CI: 0.66–1.25), the estimates were close to the null value, and the confidence intervals included 1, indicating a lack of clear evidence for an association.
Fig. 2.
Association between shift work patterns and the risk of AR. The forest plot shows HRs and 95% CIs for the relationship between different shift work patterns and AR risk, as estimated by four statistical models with progressive adjustment. Day workers were used as the reference group (HR = 1.00) in all models
Model adjustment strategies: Model 1: Unadjusted; Model 2: Adjusted for age and sex; Model 3: Model 2 + ethnicity, alcohol consumption, smoking status, body mass index (BMI), Townsend Deprivation Index (TDI), and education level; Model 4: Model 3 + asthma, allergy, and eczema/dermatitis.
AR, allergy rhinitis; HR, hazard ratio; CI, confidence interval; BMI, body mass index; TDI, Townsend Deprivation Index.
Unhealthy sleep patterns dose-dependently increase allergic rhinitis risk
Table 2 presents the baseline characteristics of the study participants categorized by sleep patterns. During a median follow-up of 13.64 years (IQR: 12.90–14.28), 2,673 incident cases of AR were recorded. The mean age of the cohort was 56.60 years (SD: 8.07); 44.8% were male and 95.5% were White ethnicity. Compared with those exhibiting a healthy sleep pattern, participants with an unhealthy sleep pattern had higher BMI, higher TDI scores, lower educational attainment, a greater proportion of current smokers, a lower proportion of current alcohol drinkers, and a higher baseline prevalence of asthma, allergies, and eczema/dermatitis.
Table 2.
Baseline characteristics of participants by sleep patterns (n = 374,672)
| Overall | Healthy | Intermediate | Poor | |
|---|---|---|---|---|
| n | 374,672 | 217,983 | 147,635 | 9054 |
| Sex (Male %) | 168,010 (44.8) | 92,768 (42.6) | 70,816 (48.0) | 4426 (48.9) |
| Age (years, mean (SD)) | 56.60 (8.07) | 56.44 (8.19) | 56.83 (7.91) | 56.73 (7.70) |
| Ethnicity (%) | ||||
| White | 357,746 (95.5) | 208,555 (95.7) | 140,714 (95.3) | 8477 (93.6) |
| Asian or Asian British | 6585 (1.8) | 3869 (1.8) | 2512 (1.7) | 204 (2.3) |
| Black or Black British | 5290 (1.4) | 2818 (1.3) | 2283 (1.5) | 189 (2.1) |
| Mixed | 2091 (0.6) | 1129 (0.5) | 890 (0.6) | 72 (0.8) |
| Other | 2960 (0.8) | 1612 (0.7) | 1236 (0.8) | 112 (1.2) |
| BMI (kg/m2, mean (SD)) | 27.42 (4.78) | 26.82 (4.47) | 28.14 (4.98) | 30.05 (5.84) |
| TDI (mean (SD)) | -1.40 (3.03) | -1.55 (2.94) | -1.24 (3.12) | -0.59 (3.37) |
| Education (%) | ||||
| High | 121,021 (32.3) | 75,654 (34.7) | 43,220 (29.3) | 2147 (23.7) |
| Intermediate | 123,202 (32.9) | 71,318 (32.7) | 48,998 (33.2) | 2886 (31.9) |
| Low | 130,449 (34.8) | 71,011 (32.6) | 55,417 (37.5) | 4021 (44.4) |
| Smoke (%) | ||||
| Current | 39,800 (10.6) | 18,787 (8.6) | 19,357 (13.1) | 1656 (18.3) |
| Never | 202,393 (54.0) | 125,671 (57.7) | 72,913 (49.4) | 3809 (42.1) |
| Previous | 132,479 (35.4) | 73,525 (33.7) | 55,365 (37.5) | 3589 (39.6) |
| Alcohol (%) | ||||
| Current | 346,430 (92.5) | 201,867 (92.6) | 136,425 (92.4) | 8138 (89.9) |
| Never | 15,359 (4.1) | 9162 (4.2) | 5792 (3.9) | 405 (4.5) |
| Previous | 12,883 (3.4) | 6954 (3.2) | 5418 (3.7) | 511 (5.6) |
| Asthma (%) | 37,633 (10.0) | 19,925 (9.1) | 16,354 (11.1) | 1354 (15.0) |
| Allergy (%) | 8359 (2.2) | 4683 (2.1) | 3440 (2.3) | 236 (2.6) |
| Eczema/dermatitis (%) | 8950 (2.4) | 5036 (2.3) | 3665 (2.5) | 249 (2.8) |
Values are presented as number (%) for categorical variables and mean (SD) for continuous variables
n number of cases, SD Standard deviation, BMI Body mass index, TDI Townsend Deprivation Index
Figure 3 and Table S5 present the association between sleep patterns and the incidence of AR. A linear dose-response relationship was observed between sleep patterns and AR risk: poorer sleep pattern was associated with higher estimated AR risk. When sleep pattern was analyzed as a continuous variable, each one-category deterioration in sleep patterns was associated with a 16% increase in AR risk in the fully adjusted model (Model 4) (HR = 1.16, 95% CI: 1.08–1.25). Compared with a healthy sleep pattern, a poor sleep pattern was associated with a higher AR risk in the unadjusted model (Model 1) (HR = 1.36, 95% CI: 1.09–1.71). After sequential adjustment for covariates (Models 2–4), this estimate was attenuated but remained positively associated with AR risk in the fully adjusted model (Model 4) (HR = 1.28, 95% CI: 1.02–1.61). Cumulative hazard curves further confirmed a graded increase in AR risk with progressively worse sleep patterns (Fig. 4), demonstrating that individuals with poor sleep patterns experience a higher and faster accumulation of AR risk over time compared to those with intermediate or healthy sleep patterns.
Fig. 3.
Association between sleep patterns and the risk of AR. The forest plot displays HRs and 95% CIs for the association between sleep patterns and AR risk across four statistical models with varying levels of adjustment. The healthy sleep pattern served as the reference group (HR = 1.00) in all models. The “per category decrease” denotes the trend estimate associated with deterioration in sleep pattern quality from healthy to poor. Model adjustment strategies: Model 1: Unadjusted; Model 2: Adjusted for age and sex; Model 3: Model 2 + ethnicity, alcohol consumption, smoking status, BMI, TDI, and education level; Model 4: Model 3 + asthma, allergy, and eczema/dermatitis. AR, allergy rhinitis; HR, hazard ratio; CI, confidence interval; BMI, body mass index; TDI, Townsend Deprivation Index.
Fig. 4.
Cumulative hazard curves for allergic rhinitis (stratified by sleep patterns). Cumulative hazard curves illustrate the incidence of allergic rhinitis over time, stratified by sleep pattern: healthy, intermediate, and poor. The x-axis indicates follow-up time (in days), and the y-axis represents the cumulative hazard
Furthermore, the sensitivity analysis produced results consistent with those of the primary analysis, reflecting the robustness of our findings (Table S6).
The table below the graph shows the number of participants at risk (free of allergic rhinitis) at selected time points for each sleep pattern category. The number of at-risk individuals decreases over time due to incident allergic rhinitis and censoring (including loss to follow-up, death, or end of the study period).
Evening chronotype, short sleep, insomnia, and snoring independently increase allergic rhinitis risk
We further examined the association between 5 sleep behaviors and the risk of AR (Fig. 5). In full adjusted Model 4, compared with participants who were identified as “definitely a ‘morning’ person”, those reporting being “definitely an ‘evening’ person” had a 27% higher risk of AR (HR = 1.27, 95% CI: 1.11–1.46). Short sleep duration (≤ 6 h) was associated with a 12% increased AR risk (HR = 1.12, 95% CI: 1.02–1.22) compared to normal sleep duration (7–8 h), while long sleep (≥ 9 h) showed no clear association (HR = 1.12, 95% CI: 0.97–1.29). A restricted cubic spline model confirmed a non-linear relationship between sleep duration and AR risk (Figure S2). Compared to those who “never/rarely” experienced insomnia, participants reporting “sometimes” (HR = 1.15, 95% CI: 1.04–1.27) or “usually” (HR = 1.23, 95% CI: 1.10–1.37) experiencing insomnia had AR risks that were 15% and 23% higher, respectively. Snoring was also associated with increased AR risk compared with not snoring (HR = 1.10, 95% CI: 1.01–1.19). However, daytime sleepiness was not associated with the risk of AR. Subgroup analyses revealed no clear interaction effects across covariates (Figure S3). Notably, among participants with BMI ≥ 30 kg/m², poor sleep pattern was associated with a markedly elevated AR risk compared to healthy sleep patterns (HR = 1.44, 95% CI: 1.04–2.00).
Fig. 5.
Association between five sleep behaviors and the risk of AR. The forest plot displays HRs and 95% CIs for the association between five sleep behaviors (chronotype, sleep duration, insomnia, snoring and daytime sleepiness) and AR risk across four statistical models with progressive adjustment. All models included the five sleep behaviors simultaneously. The definitely a ‘morning’ person, 7–8 h of sleep duration, never/rarely insomnia, no snoring, and never/rarely daytime sleepiness respectively served as the reference group (HR = 1). Model adjustment strategies: Model 1: Unadjusted; Model 2: Adjusted for age and sex; Model 3: Model 2 + ethnicity, alcohol consumption, smoking status, BMI, TDI, and education level; Model 4: Model 3 + asthma, allergy, and eczema/dermatitis. AR, allergy rhinitis; HR, hazard ratio; CI, confidence interval; BMI, body mass index; TDI, Townsend Deprivation Index
When examining the combined effects of shift work and sleep patterns, we found no overall interaction (P for interaction = 0.68). However, in the joint exposure analysis, we observed that individuals with irregular night shift work and unhealthy sleep pattern had the highest risk of AR (HR = 1.35, 95% CI: 1.02–1.79) compared to day workers with healthy sleep pattern. This risk was notably higher than that observed in day workers with unhealthy sleep pattern (HR = 1.12, 95% CI: 1.00–1.25) and in irregular night shift workers with healthy sleep pattern (HR = 1.10, 95% CI: 0.82–1.48). Detailed results are presented in Table S7.
Identification of Metabolites Mediating the Sleep-AR Association
To explore potential biological pathways linking sleep patterns to allergic rhinitis (AR), we conducted a metabolome-wide mediation analysis. A total of 22 metabolites were identified that were associated with both sleep patterns and AR risk (Figure S4). 20 metabolites with consistent directions of association for sleep patterns and AR were selected and subsequently included in the mediation analysis. Ultimately, 16 metabolites demonstrated mediating effects between sleep patterns and AR risk (Fig. 6). The detailed data are shown in Figure S5. Small HDL particle concentration showed the strongest mediation effect, though its overall contribution was modest (PM = 1.59%, 95% CI: 0.12–5.28). These findings suggest that deteriorated sleep patterns may elevate AR risk partly through increasing small HDL particle levels.
Fig. 6.
Proportion of mediation (PM) by individual metabolites in the association between sleep patterns and AR risk. This bar plot displays the PM and 95%CI for each metabolite in the relationship between sleep patterns and AR risk. Metabolites are arranged in descending order of their PM values. AR, allergy rhinitis; HR, hazard ratio; CI, confidence interval; PM, proportion mediated; HDL, high-density lipoprotein; LDL, low-density lipoprotein; BMI, body mass index; TDI, Townsend Deprivation Index
Discussion
In this large prospective cohort study, we demonstrate that irregular night shift work—though not permanent night shifts—is associated with an increased risk of AR. Furthermore, poor sleep patterns and specific sleep behaviors, including evening chronotype, short sleep duration (≤ 6 h), insomnia, and snoring, were each independently associated with elevated AR risk. Mediation analysis further identified small high-density lipoprotein (HDL) particle concentration as a partial mediator in the sleep–AR relationship.
Circadian rhythm disruption is an established consequence of night shift work and sleep disorders [22], and has been increasingly implicated in the pathogenesis of allergic diseases [23]. The lack of association with permanent night shifts may reflect circadian adaptation over time [24], a phenomenon that has in some contexts been linked to reduced inflammatory responses [25]. Our findings related to specific sleep behaviors further support the relevance of circadian integrity: evening chronotype and short sleep duration may promote circadian misalignment and sleep fragmentation, contributing to immune dysregulation. Experimental studies indicate that sleep disruption can induce systemic low-grade inflammation and alter circadian gene expression in leukocytes, thereby predisposing individuals to allergic sensitization and AR development [26, 27]. In murine models, circadian disruption has been shown to enhance Th2-mediated airway inflammation [28], providing a plausible immunological basis for our epidemiological observations.
A key novel finding was the identification of small HDL particles as a potential metabolic mediator between poor sleep and AR. HDL functionality is highly dependent on particle size and composition, with small HDL often exhibiting pro-inflammatory properties in conditions of metabolic dysregulation [29, 30]. Previous studies have linked elevated small HDL levels to other chronic inflammatory disorders, including osteoarthritis [31]. It is possible that sleep disruption promotes oxidative stress and chronic inflammation, leading to altered HDL metabolism and the accumulation of pro-inflammatory small HDL particles, which in turn may exacerbate allergic inflammation. Nevertheless, the modest mediation proportion (1.59%) suggests that other immune and metabolic pathways are also involved.
Circadian rhythm disruption has been shown to affect adipose tissue dynamics and systemic metabolism, leading to adipocyte hypertrophy accompanied by inflammation and fibrosis, which subsequently contributes to obesity [32, 33]. Although some studies suggest an association between obesity and AR, the directionality of this relationship remains unclear [34]. Our study also revealed a stronger association between poor sleep and AR among participants with obesity (BMI ≥ 30 kg/m²). Obesity is characterized by reduced adiponectin and chronic adipose tissue inflammation, which may synergize with sleep-related circadian disruption to amplify allergic responses [35]. This subgroup effect highlights the potential interplay between metabolic health and sleep-related AR risk.
Although no clear interaction was detected when comprehensively examining the combined effects of shift work and sleep patterns on AR risk, the substantially increased risk of AR among individuals concurrently exposed to irregular night shifts and unhealthy sleep patterns remains clinically meaningful. Therefore, in clinical and occupational health practice, special attention should be paid to this population, and they should be advised to improve their sleep and work schedules promptly to potentially reduce the risk of AR. The relatively small number of participants with complete data on both shift work and sleep patterns, coupled with the consequent reduction in sample size within each subgroup of the interaction analysis, may have limited the statistical power to detect complex interactions. Future studies with larger sample sizes are warranted to confirm the potential interaction between sleep patterns and shift work.
Several limitations must be acknowledged. The observational design precludes causal inference. Self-reported sleep and shift work data are subject to recall bias, and exposure assessment only at baseline may not capture longitudinal variations. We lacked data on lifetime occupational history or second jobs, which could lead to exposure misclassification and potentially underestimate the cumulative health effects of shift work. The UK Biobank’s low response rate (≈ 5%) may limit generalizability [36], and the predominantly White sample necessitates caution in extrapolating results to other ethnicities. Although we adjusted for numerous confounders, residual confounding remains possible.
Conclusion
This study provides robust epidemiological evidence that irregular night shift work and poor sleep patterns are associated with an increased risk of AR, with small HDL particles serving as a partial mediator. These findings underscore the importance of circadian health and sleep quality in the prevention and management of allergic rhinitis. Future research should prioritize mechanistic studies to clarify causal pathways and explore interventions targeting sleep behavior and metabolic health to mitigate allergic disease risk.
Supplementary Information
Acknowledgements
We extend our sincere gratitude to all the contributors who made this study possible. We are especially thankful to the founders and ongoing maintainers of the UK Biobank for their invaluable contribution in establishing and curating this foundational resource.
Abbreviations
- AR
Allergic rhinitisaller
- HRs
Hazard ratios
- CIs
Confidence intervals
- TDI
Townsend Deprivation Index
- SD
Standard deviation
- ANOVA
Analysis of variance
- FDR
False discovery rate
- IQR
Interquartile range
Authors’ contributions
WL was responsible for data visualization, drafted and edited the manuscript. YHW conceived the idea and edited the manuscript. LC analyzed part of data and edited the manuscript. ZXH extracted the data and analyzed part of data. CHC, JQL and HMC reviewed and edited the manuscript. WBL, XR and HL conceived the idea, designed the study, and provided supervision. All authors read and approved the final manuscript.
Funding
This work was supported by grants from the National Natural Science Foundation of China (No. 82101186 to HL), the Natural Science Foundation of Guangdong Province (No. 2022A1515012394 to RX), the Guangzhou Science and Technology Plan Project (No. 2025B03J0086, 2024A04J4630 to HL), and the Kelin New Star Program of the First Affiliated Hospital of Sun Yat-sen University (No. R08036 to HL).
Data availability
This research was conducted using the UK Biobank resource under Application Number [103612]. Due to privacy and ethical restrictions imposed by the UK Biobank, individual-level data from this study cannot be publicly shared. However, researchers can apply for access to the UK Biobank data directly via www.ukbiobank.ac.uk/register-apply. The application process is described online, and approved researchers will be able to replicate the analysis. The analytic code and aggregated data are available from the corresponding author upon reasonable request.
Declarations
Competing interests
The authors declare no competing interests.
Ethics approval and consent to participate
The UK Biobank has obtained ethical approval from the North West Multi-centre Research Ethics Committee (Reference number: 11/NW/0382, https://www.ukbiobank.ac.uk/learn-more-about-uk-biobank/about-us/ethics). This research was conducted in accordance with the Declaration of Helsinki. Informed consent was collected from all study participants via electronic signature.
Consent for publication
Not applicable.
Competing interests
The authors declare that they have no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Wen Lyu, Yihui Wen and Lin Chen are first authors.
Contributor Information
Wenbin Lei, Email: leiwb@mail.sysu.edu.cn.
Rui Xu, Email: xurui@mail.sysu.edu.cn.
Hang Li, Email: lihang27@mail.sysu.edu.cn.
References
- 1.Bousquet J, Khaltaev N, Cruz AA, Denburg J, Fokkens WJ, Togias A, et al. Allergic Rhinitis and its Impact on Asthma (ARIA) 2008. Allergy. 2008;63(s86):8–160. [DOI] [PubMed] [Google Scholar]
- 2.Colás C, Brosa M, Antón E, Montoro J, Navarro A, Dordal MT, et al. Estimate of the total costs of allergic rhinitis in specialized care based on real-world data: the FERIN Study. Allergy. 2017;72(6):959–66. [DOI] [PubMed] [Google Scholar]
- 3.Epstein TEG, Rorie AC, Ramon GD, Keswani A, Bernstein J, Codina R, et al. Impact of climate change on aerobiology, rhinitis, and allergen immunotherapy: Work Group Report from the Aerobiology, Rhinitis, Rhinosinusitis & Ocular Allergy, and Immunotherapy, Allergen Standardization & Allergy Diagnostics Committees of the American Academy of Allergy, Asthma & Immunology. J Allergy Clin Immunol. 2025;155(6):1767–1782.e2. [DOI] [PubMed]
- 4.Cheng X, Zhou Y, Hao Y, Long Z, Hu Q, Huo B, et al. Recent Studies and Prospects of Biologics in Allergic Rhinitis Treatment. Int J Mol Sci. 2025;26(10):4509. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Paganelli R, Petrarca C, Di Gioacchino M. Biological clocks: their relevance to immune-allergic diseases. Clin Mol Allergy. 2018;16(1):1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bechtold DA, Gibbs JE, Loudon ASI. Circadian dysfunction in disease. Trends Pharmacol Sci. 2010;31(5):191–8. [DOI] [PubMed] [Google Scholar]
- 7.Nakao A. Circadian Regulation of the Biology of Allergic Disease: Clock Disruption Can Promote Allergy. Front Immunol. 2020;12:11:1237. [DOI] [PMC free article] [PubMed]
- 8.Nakao A. Clockwork allergy: How the circadian clock underpins allergic reactions. J Allergy Clin Immunol. 2018;142(4):1021–31. [DOI] [PubMed] [Google Scholar]
- 9.Maidstone RJ, Turner J, Vetter C, Dashti HS, Saxena R, Scheer FAJL, et al. Night shift work is associated with an increased risk of asthma. Thorax. 2021;76(1):53–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Huang Y, Jing D, Su J, Huang Z, Liu H, Tao J, et al. Association of Night Shift Work With Chronic Spontaneous Urticaria and Effect Modification by Circadian Dysfunction Among Workers. Front Public Health. 2021;9:751579. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Haldar P, Carsin AE, Debnath S, Maity SG, Annesi-Maesano I, Garcia-Aymerich J, et al. Individual circadian preference (chronotype) is associated with asthma and allergic symptoms among adolescents. ERJ Open Res. 2020;6(2):00226–2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Chen Y, Zhao A, Lyu J, Hu Y, Yin Y, Qu J, et al. Association of delayed chronotype with allergic diseases in primary school children. Chronobiol Int. 2022;3(6):836–47. [DOI] [PubMed]
- 13.Fan M, Sun D, Zhou T, Heianza Y, Lv J, Li L, et al. Sleep patterns, genetic susceptibility, and incident cardiovascular disease: a prospective study of 385 292 UK biobank participants. Eur Heart J. 2020;41(11):1182–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Sudlow C, Gallacher J, Allen N, Beral V, Burton P, Danesh J, et al. UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med. 2015;12(3):e1001779. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Li X, Xue Q, Wang M, Zhou T, Ma H, Heianza Y, et al. Adherence to a Healthy Sleep Pattern and Incident Heart Failure. Circulation. 2021;143(1):97–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Luo P, Ying J, Li J, Yang Z, Sun X, Ye D, et al. Air Pollution and Allergic Rhinitis: Findings from a Prospective Cohort Study. Environ Sci Technol. 2023;57(42):15835–45. [DOI] [PubMed] [Google Scholar]
- 17.Qureshi D, Collister J, Allen NE, Kuźma E, Littlejohns T. Association between metabolic syndrome and risk of incident dementia in UK Biobank. Alzheimer’s Dement. 2024;20(1):447–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Elliott P, Biobank on behalf of U, Peakman TC, Biobank on behalf of U. The UK Biobank sample handling and storage protocol for the collection, processing and archiving of human blood and urine. Available from: 10.1093/ije/dym276. Cited 2025 Sept 7. [DOI] [PubMed]
- 19.Chadeau-Hyam M, Bodinier B, Vermeulen R, Karimi M, Zuber V, Castagné R et al. Education, biological ageing, all-cause and cause-specific mortality and morbidity: UK biobank cohort study. EClinicalMedicine. 2020;29–30:100658. [DOI] [PMC free article] [PubMed]
- 20.Kang Z, Zhang J, Zhu C, Zhu Y, Jiang H, Tong Q, et al. Impaired pulmonary function increases the risk of gout: evidence from a large cohort study in the UK Biobank. BMC Med. 2024;22:606. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Shi B, Choirat C, Coull BA, VanderWeele TJ, Valeri L, CMAverse. A Suite of Functions for Reproducible Causal Mediation Analyses. Epidemiology. 2021;32(5):e20–2. [DOI] [PubMed]
- 22.Nuszkiewicz J, Rzepka W, Markiel J, Porzych M, Woźniak A, Szewczyk-Golec K. Circadian Rhythm Disruptions and Cardiovascular Disease Risk: The Special Role of Melatonin. Curr Issues Mol Biol. 2025;47(8):664. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Deprato A, Maidstone R, Cros AP, Adan A, Haldar P, Harding BN, et al. Influence of light at night on allergic diseases: a systematic review and meta-analysis. BMC Med. 2024;22:67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hansen JH, Geving IH, Reinertsen RE. Adaptation rate of 6-sulfatoxymelatonin and cognitive performance in offshore fleet shift workers: a field study. Int Arch Occup Environ Health. 2010;83(6):607–15. [DOI] [PubMed] [Google Scholar]
- 25.Hedström AK, Åkerstedt T, Klareskog L, Alfredsson L. Relationship between shift work and the onset of rheumatoid arthritis. RMD Open. 2017;3(2). Available from: https://rmdopen.bmj.com/content/3/2/e000475. Cited 2025 Sept 7. [DOI] [PMC free article] [PubMed]
- 26.Besedovsky L, Lange T, Haack M. The Sleep-Immune Crosstalk in Health and Disease. Physiol Rev. 2019;99(3):1325–80. [DOI] [PMC free article] [PubMed]
- 27.Bollinger T, Bollinger A, Oster H, Solbach W. Sleep, Immunity, and Circadian Clocks: A Mechanistic Model. [cited 2025 Sept 7]; Available from: 10.1159/000281827 [DOI] [PubMed]
- 28.Cheng F, An Y, Xue J, Wang Y, Ding X, Zhang Y, et al. Circadian rhythm disruption exacerbates Th2-like immune response in murine allergic airway inflammation. Int Forum Allergy Rhinol. 2022;12(5):757–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Gluba-Sagr A, Olszewski R, Franczyk B, Młynarska E, Rysz-Górzyńska M, Rysz J, et al. High-density lipoproteins. Part 2. Impact of disease states on functionality. Am J Prev Cardiol. 2025 Sept;23:101073. [DOI] [PMC free article] [PubMed]
- 30.Salazar J, Olivar LC, Ramos E, Chávez-Castillo M, Rojas J, Bermúdez V. Dysfunctional High-Density Lipoprotein: An Innovative Target for Proteomics and Lipidomics. Cholesterol. 2015;2015:296417. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Li C, Wu J, Zhang Y, He H, Hu Y, Wei J et al. Plasma lipoprotein subclasses and risk of incident knee osteoarthritis: A population-based cohort study. Osteoarthritis Cartilage. 2025;S1063-4584(25)01094-5. [DOI] [PubMed]
- 32.Xiong X, Lin Y, Lee J, Paul A, Yechoor V, Figueiro M, et al. Chronic circadian shift leads to adipose tissue inflammation and fibrosis. Mol Cell Endocrinol. 2021;521:111110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Ratwani M, Bisht S, Prakash S. Association between sleep disturbance and metabolic dysfunctions in adipose tissue: Insights into melatonin’s role. Biochemical and Biophysical Research Communications. 2025 July 12;770:151978. [DOI] [PubMed]
- 34.Jung SY, Park DC, Kim SH, Yeo SG. Role of Obesity in Otorhinolaryngologic Diseases. Curr Allergy Asthma Rep. 2019;19(7):34. [DOI] [PubMed]
- 35.Medoff BD, Okamoto Y, Leyton P, Weng M, Sandall BP, Raher MJ, et al. Adiponectin Deficiency Increases Allergic Airway Inflammation and Pulmonary Vascular Remodeling. Am J Respir Cell Mol Biol. 2009;41(4):397–406. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Fry A, Littlejohns TJ, Sudlow C, Doherty N, Adamska L, Sprosen T et al. Comparison of Sociodemographic and Health-Related Characteristics of UK Biobank Participants With Those of the General Population. Available from: 10.1093/aje/kwx246. Cited 2025 Sept 7. [DOI] [PMC free article] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
This research was conducted using the UK Biobank resource under Application Number [103612]. Due to privacy and ethical restrictions imposed by the UK Biobank, individual-level data from this study cannot be publicly shared. However, researchers can apply for access to the UK Biobank data directly via www.ukbiobank.ac.uk/register-apply. The application process is described online, and approved researchers will be able to replicate the analysis. The analytic code and aggregated data are available from the corresponding author upon reasonable request.