Effects of obstructive sleep apnea on circulating immune cells profiles: Evidence from NHANES dataset and Mendelian randomization

Abstract

Objectives  This study aimed to investigate the impact of Obstructive Sleep Apnea (OSA) on immune cell profiles by utilizing both the National Health and Nutrition Examination Survey (NHANES) dataset and a two-sample Mendelian Randomization (MR) method.

Methods  We included 20,732 participants from the NHANES to explore the relationship between symptom defined OSA and peripheral immune cells. Additionally, we conducted a two-sample MR analysis to investigate the effect of OSA on 35 summarized peripheral immune cell properties and 731 immunophenotypes. To ensure the robustness of findings in MR analysis, we conducted a sensitivity analysis.

Results  The NHANES dataset and MR analysis indicated that OSA patients exhibit significantly elevated levels of white blood cells, neutrophils, and granulocytes, along with lower levels of lymphocytes. Furthermore, the MR analysis demonstrated a causal link between genetic predisposition to OSA and ten lymphocyte subtypes. Notably, there is a decrease in CD3 expression on Natural Killer-T (NKT) cells (β_IVW = −0.47, SE = 0.17, p = 0.005), and an increase in lgD expression on lgD + CD38+ B-cells (β_IVW = 0.36, SE = 0.15, p = 0.020) among individuals with a genetic predisposition to OSA. No heterogeneity or pleiotropy was identified.

Conclusion  OSA is associated with elevated neutrophils, granulocytes, and systemic immune-inflammation index. Causally, OSA reduces NKT cells and disrupts lymphocyte homeostasis, impairing B-cell maturation/activation. These immune signatures may serve as clinical biomarkers for guide risk stratification and guide personalized therapies, including targeted immunomodulation.

Level of evidence  This study combined a cross-sectional analysis of NHANES datasets with a Mendelian Randomization (MR) approach utilizing GWAS datasets. The MR analysis provides a level of evidence second only to randomized controlled trials, surpassing that of cohort and case-control studies.

Keywords
Obstructive sleep apnea; Circulating immune cells; National Health and Nutrition Examination Survey (NHANES); Mendelian randomization

Introduction

Obstructive Sleep Apnea (OSA), characterized by breathing interruptions during sleep, impacts an estimated 936 million people globally.¹ OSA results from the collapse of the upper airways, commonly leading to sleep fragmentation and Chronic Intermittent Hypoxia (CIH).² The up-regulation of superoxide ions (oxygen radicals) induced by OSA exposes individuals to a pro-inflammatory condition, leading to immune dysfunction.³

Circulating blood cell phenotypes, vital indicators of immune system activity, are pivotal in the pathogenesis of diseases.⁴ Recent research has revealed that peripheral blood lymphocytes develop and differentiate into distinct subsets, playing crucial roles in regulating immune functions and the development of cancer and cardiovascular events.⁵,⁶ Previous studies have suggested a potential correlation between OSA and increased dysregulation in conventional blood cell counts, e.g., white blood cell, neutrophil, average platelet volume, Neutrophil-Lymphocyte Ratio (NLR), and Systemic Immune-Inflammation Index (SII).⁷-⁹ Moreover, studies have observed an association between OSA and CD3+, CD4+, and CD8+ T-lymphocyte profiles in peripheral blood.¹⁰,¹¹ However, inconsistent findings exist due to limited sample size and clinical heterogeneity.⁷,¹² Furthermore, there might be bidirectional links between sleep and the immune system,¹³ which bias the causal associations between circulating blood cell phenotypes and OSA.

Recent decades, NHANES employs a stratified, multistage probability sampling design to recruit approximately 5,000 U.S. participants annually, representing diverse, non-institutionalized populations across all age groups.¹⁴ This large, population-based cross-sectional sample enhances statistical power for detecting health outcome-risk factor associations. However, the cross-sectional study cannot identify the causality between exposure and outcomes. Impressively, recent Genome-Wide Association Studies (GWAS) have identified hundreds of Single Nucleotide Polymorphisms (SNPs) associated with OSA-related traits or circulating blood cell indices.¹⁵ The Mendelian Randomization (MR) approach leverages the random allocation and immutable nature of the human genome, allowing researchers to examine the causal relationship between exposure and a disease or medical condition.¹⁶ Thus, we conducted MR analysis to validate the findings in the cross-sectional study.

Building on the insights from previous studies, we employed a dual analytical framework to explore the potential relationship: Cross-sectional analysis of the NHANES dataset to identify OSA-immune associations while controlling for confounders; Two-sample MR with GWAS summary statistics to infer causality, leveraging the available genetic databases. This combined approach mitigates weaknesses inherent to observational studies (e.g., reverse causation) while compensating for MR’s reliance on genetic architecture assumptions.

Methods

Ethical approval

The studies involving human participants received approval from the NHANES Institutional Review Board and the Ethics Review Board of the National Center for Health Statistics. All study procedures adhered to the principles outlined in the Declaration of Helsinki. Informed written consent was obtained from all participants and, where applicable, from their legal guardians.

Overall study design

This study was conducted in two stages (Fig. 1A). First, we performed a cross-sectional analysis of the NHANES database using Propensity Score Matching (PSM), during which we carried out a logistic regression analysis to determine the Odds Ratio (OR) and its Confidence Interval (CI) between OSA and peripheral immune cells. Since the cross-sectional study cannot identify the causality between exposure and outcomes, we conducted MR analysis to support our findings in stage 1. The conceptual schematic of the MR research design is illustrated in Fig. 1B: 1) Genetic variants must be strongly linked with exposure; 2) They should be free from potential confounding factors; 3) These genetic instruments should influence the outcome solely through exposure. We reported our findings following the guidelines outlined in the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline (Supplementary Table S1).¹⁷

Fig. 1
Study design. (A) Overall design of NHANES and Mendelian randomization analysis in the study. (B) The conceptual schematic of the MR research design.

Data sources and study population

Every 2-years, the NHANES program collects nationally representative data from approximately 10,000 randomly selected participants. These individuals complete comprehensive health interviews covering demographic, socioeconomic, dietary, and health-related factors, followed by physical examinations and laboratory testing.¹⁴ The 2005-2008 and 2015-2018 NHANES datasets were selected because they are the only survey periods that simultaneously included both questionnaire-based assessments (related to OSA) and laboratory-based peripheral blood cell examinations. A total of 17,133 individuals were excluded due to missing data from the Sleep Disorders Questionnaire or Complete Blood Count, eight were excluded due to missing OSA data, and 1,849 due to being pregnant or being under 18-years of age. Finally, 20,732 participants were included in the stage 1 analysis.

Measures

Outcomes: peripheral immune cells

The clinical hematological parameters examined in this study included Complete Blood Count, including White Blood Cell (WBC), Lymphocyte (Lym), Monocyte (Mono), segmented Neutrophils (Nsg), Eosinophils (Eos), Basophils (Baso), Red Blood Cell (RBC), Hemoglobin (HGB), Hematocrit (HCT), Mean Cell Volume (MCV), Mean Cell Hemoglobin (MCH), Mean Cell Hemoglobin Concentration (MCHC), Red Cell Distribution Width (RDW), Platelet Count (PLT), and Mean Platelet Volume (MPV). In addition, participants' inflammatory markers included NLR and SII, with NLR calculated as the neutrophil/lymphocyte count and SII calculated as platelet count multiplied by the neutrophil/lymphocyte count.¹⁸

Exposure: OSA

The diagnosis was derived from three key criteria in the NHANES Sleep Disorders Questionnaire: (1) Snoring ≥3 nights/week, (2) Nocturnal apneas ≥3 nights/week, and 3) Excessive daytime sleepiness (≥1 episode/month despite ≥7 h nightly sleep). All three criteria were required for classification.¹⁹

Covariates

The covariates were selected based on their established roles in OSA pathogenesis and their potential to confound immune-inflammatory outcomes.²⁰,²¹ Finally, the Covariates adjusted for in this study included age, sex, race, Body Mass Index (BMI), and diagnoses of hypertension, high cholesterol, and diabetes. Diagnoses were defined as participants ever being told by a doctor that they had these conditions.²² These factors influence blood cell parameters through multiple pathways, including age-related immunosenescence, sex hormone-mediated immune regulation, racial differences in baseline profiles, obesity-associated inflammation, and metabolic disorder-induced systemic inflammation.²³,²⁴

Statistical analysis

Independent samples t-tests for continuous variables and Chi-Squared tests for categorical variables were used to compare differences in baseline characteristics between the OSA and non-OSA groups. Multiple imputation was employed to handle missing values, using model estimation and repeated simulations were used to generate a complete set of datasets.²⁵

To analyze the relationship between OSA and peripheral immune cells, the PSM method, based on logistic regression models,²⁶ was applied to adjust for selection bias and potential confounders. To mitigate potential selection bias in patient allocation, we performed Propensity Score Matching (PSM) between the OSA and non-OSA groups using a 1:1 matching protocol with a caliper width of 0.02. This caliper setting was empirically optimized from the theoretical threshold (0.2×σ_logit = 0.116) to achieve optimal covariate balance, as suggested by established diagnostic methods.²⁷ All data in this study were statistically analyzed using R (version 4.3.2).

Mendelian randomization

Data sources

The OSA summary-level data were obtained from a recently published GWAS, which included 16,761 OSA patients and 201,194 controls in the FinnGen study. The characteristics of the genetic data sources used are summarized in Supplemental Table S2.²⁸ OSA was diagnosed using the International Classification of Diseases, 10th edition (ICD-10) and 9th edition (ICD-9) codes (ICD-10: G47.3, ICD-9: 3472A), based on subjective symptoms, clinical examination, and sleep registration using the apnea-hypopnea index of five per hour or respiratory event index of five per hour. Principal covariates, such as age and sex, were adjusted in the association tests for all sources. Astle et al. performed GWAS on the blood cell properties of 173,480 participants of European-ancestry and reported 35 summary blood cell properties, which could be obtained from the IEU OpenGWAS Database (https://gwas.mrcieu.ac.uk). There were 731 immunophenotypes for cellular subpopulation analyses, with 3,757 individuals profiled by flow cytometry using the Sardinian founder population.²⁹ More information on the GWAS datasets for OSA and circulating immune cell phenotypes is shown in Table S2. We examined SNPs associated with confounders or outcomes using LDlink (Https://ldlink.nih.gov/), an alternative method of the Phenoscanner database. SNPs linked to previously identified confounding factors related to circulating immune cells, such as blood pressure and diabetes,²⁴ were not found in LDlink’s analysis. However, SNPs associated with OSA were related to BMI phenotypes. Further details on IVs associated with confounders in reverse MR analyses can be found in Supplementary Table S3.

MR analysis

The study was carried out using the TwoSampleMR (version 0.5.7) software package based on the R platform (version 4.3.2). We extracted genome-wide Single Nuclear Polymorphisms (SNPs) for OSA and hematological traits with a p < 5 × 10⁻⁸ criterion.³⁰ Since the p < 5 × 10⁻⁸ threshold could detect only a few Instrumental Variables (IVs) for several hematological traits, we chose SNPs with p < 5 × 10⁻⁶ as the reverse IVs criterion.³¹ This adjustment enabled robust Mendelian randomization analyses, particularly for the Inverse-Variance Weighted (IVW) method - the gold-standard MR approach that achieves optimal statistical power when adequate valid IVs are available.¹⁷ To ensure that SNPs’ effects on OSA and hematological traits were related to the same allele, we aligned the direction of effects across all SNPs. Additionally, with a window of 10,000 kb, we eliminated SNPs that were in chained disequilibrium (r² < 0.001) and a Minimum Allele Frequency (MAF) of 0.01, and retrieved the remaining SNPs. To ensure robust exposure, we calculated the F-statistic and concluded that an F-statistic of 10 would be sufficient to counteract weak instrumental bias,³² as shown in Supplementary Table S4‒S5. To obtain reliable causal effects, we used Inverse-Variance Weighted (IVW) methods as the principal analytical strategy.³³ More details on MR sensitivity analysis and reverse MR analysis can be found in the previous study.³⁰

Results

Population characteristics of NHANES

A total of 20,732 participants (10,277 men and 10,455 women) were included in the final analysis. Of these, 2,618 participants (12.63%) met the criteria for probable OSA based on self-reported symptoms (snoring, nocturnal apneas, and daytime sleepiness). Compared to non-OSA participants, those with OSA were more likely to be older white men, have a higher BMI, and suffer from hypertension, diabetes, and higher cholesterol levels (p < 0.001).

Association between OSA and hematological parameters

After 1:1 propensity score matching, 2,592 subjects were included in both the OSA and non-OSA groups for the analyses. There were no significant differences between the two groups in terms of age, sex, race, BMI, diagnosis of hypertension, high cholesterol, and diabetes. Before matching, participants with OSA had significantly higher levels of white blood cell, monocyte, segmented neutrophils, eosinophils, basophils, NLR, SII (all p < 0.001), hemoglobin, hematocrit, red blood cell and red cell distribution width (p < 0.05). After adjusting for confounding factors, significant differences in the prevalence of OSA were found in these indicators: white blood cells, monocytes, segmented neutrophils, basophils, mean cell volume, and platelet count (all p < 0.05). In addition, participants with OSA had significantly higher levels of SII (p < 0.05), with a trend towards significance in the NLR (p = 0.091).

Cross-sectional sensitivity analyses

In the cross-sectional section of the study, post-matching Standardized Mean Differences (SMDs) were all < 0.1, confirming successful balancing, while retaining 99% of the OSA cohort for robust statistical power. Covariate balance before and after propensity score matching was visualized in Fig. 2.

Fig. 2
Covariate balance before and after propensity score matching. BMI, Body Mass Index; DM, Diabetes; HCL, High Cholesterol; HTN, Hypertension.

In addition, we employed multiple imputation to handle missing data (e.g., BMI, hypertension, hyperlipidemia) and subsequently performed propensity score matching, resulting in 2,592 well-balanced matched pairs. Complete-case analysis within matched pairs (n = 2,205 non-OSA; n = 2,233 OSA) consistently replicated originally significant hematological findings (WBC, monocytes, neutrophils, MCV, platelets, SII; all p < 0.05), with only MCH showing borderline significance (p = 0.055). Crucially, the minimal discrepancy between approaches (Supplementary Table S6) confirms this section results regarding OSA-immune associations.

MR of OSA and circulating immune cells

The casual effects of genetically predicted OSA on peripheral circulating immune cells were shown in Fig. 3. In general, consistent with the observational study, genetic liability to OSA was associated with an increased count of white blood cells (β_IVW = 0.05, SE = 0.02, p = 0.038) and myeloid white cells (β_IVW = 0.06, SE = 0.02, p = 0.007). Importantly, OSA also increased the count of neutrophils (β_IVW = 0.06, SE = 0.02, p = 0.005) and granulocytes (β_IVW = 0.06, SE = 0.02, p = 0.010). Moreover, there was evidence for a positive effect of OSA on the sum counts of neutrophil and eosinophils (β_IVW = 0.06, SE = 0.02, p = 0.010) and basophil and neutrophils (β_IVW = 0.06, SE = 0.02, p = 0.007), as well as the neutrophil percentage of white cells (β_IVW = 0.06, SE = 0.02, p = 0.008) and the neutrophil percentage of granulocytes (β_IVW = 0.06, SE = 0.02, p = 0.009).

Fig. 3
Forest plot of MR showing estimates of genetically predicted OSA on Peripheral circulating immune cells.

In contrast, although the sum counts of neutrophils and eosinophils were increased, genetically liability to OSA was associated with a decreased eosinophil percentage of white cells (β_IVW = −0.06, SE = 0.02, p = 0.013) and eosinophil percentage of granulocytes (β_IVW = −0.06, SE = 0.02, p = 0.012). Interestingly, we found that OSA decreased the lymphocyte percentage of white cells (β_IVW = −0.06, SE = 0.02, p = 0.012), which is inconsistent with the cross-sectional study. Thus, we extended our analyses by further measuring the causal estimates of OSA on lymphocyte cellular subpopulations.

Table 1
Characteristics of participants with OSA before and after PSM Analysis in the NHANES 2005‒2008 and 2015-2018.

Characteristics	Before Matching				After Matching
	Total	non-OSA	OSA	p	Total	non-OSA	OSA	p
N	20732	18114	2618		5184	2592	2592
Age (y), Mean (SD)	48.7 (18.9)	48.0 (19.1)	53.4 (16.7)	<0.001	53.7 (17.4)	54.1 (18.0)	53.3 (16.7)	0.102
Sex (%)				<0.001				0.889
Female	10455 (50.4%)	9253 (51.1%)	1202 (45.9%)		2370 (45.7%)	1182 (45.6%)	1188 (45.8%)
Male	10277 (49.6%)	8861 (48.9%)	1416 (54.1%)		2814 (54.3%)	1410 (54.4%)	1404 (54.2%)
Race (%)				<0.001				0.081
Non-Hispanic White	8405 (40.5%)	7242 (40.0%)	1163 (44.4%)		2257 (43.5%)	1107 (42.7%)	1150 (44.4%)
Non-Hispanic Black	4486 (21.6%)	3920 (21.6%)	566 (21.6%)		1174 (22.6%)	620 (23.9%)	554 (21.4%)
Mexican American	3641 (17.6%)	3235 (17.9%)	406 (15.5%)		830 (16.0%)	424 (16.4%)	406 (15.7%)
Others	4200 (20.3%)	3717 (20.5%)	483 (18.4%)		923 (17.8%)	441 (17.0%)	482 (18.6%)
BMI (kg/m²)	29.1 (7.01)	28.7 (6.78)	32.0 (7.92)	<0.001	31.6 (7.49)	31.5 (7.47)	31.7 (7.52)	0.351
Diabetes (%)				<0.001				0.347
No	17579 (86.7%)	15557 (87.7%)	2022 (79.6%)		4049 (80.5%)	2037 (81.1%)	2012 (80.0%)
yes	2695 (13.3%)	2177 (12.3%)	518 (20.4%)		978 (19.5%)	475 (18.9%)	503 (20.0%)
Hypertension (%)				<0.001				0.739
No	13573 (65.5%)	12255 (67.7%)	1318 (50.3%)		2647 (51.1%)	1330 (51.3%)	1317 (50.8%)
yes	7159 (34.5%)	5859 (32.3%)	1300 (49.7%)		2537 (48.9%)	1262 (48.7%)	1275 (49.2%)
High cholesterol level (%)				<0.001				0.232
No	13394 (64.6%)	12005 (66.3%)	1389 (53.1%)		2720 (52.5%)	1338 (51.6%)	1382 (53.3%)
Yes	7338 (35.4%)	6109 (33.7%)	1229 (46.9%)		2464 (47.5%)	1254 (48.4%)	1210 (46.7%)
WBC (109/L)	7.31 (3.69)	7.24 (2.49)	7.79 (8.06)	<0.001	7.58 (6.12)	7.38 (3.08)	7.78 (8.09)	0.020
Lym (109/L)	2.23 (2.88)	2.21 (1.49)	2.39 (7.08)	0.194	2.30 (5.29)	2.23 (2.35)	2.38 (7.11)	0.306
Mono (109/L)	0.57 (0.20)	0.57 (0.20)	0.60 (0.21)	<0.001	0.59 (0.21)	0.58 (0.20)	0.60 (0.21)	0.030
Nsg (10^9/L)	4.26 (1.69)	4.22 (1.66)	4.54 (1.86)	<0.001	4.42 (1.76)	4.31 (1.64)	4.53 (1.87)	<0.001
Eos (109/L)	0.21 (0.17)	0.20 (0.17)	0.22 (0.18)	<0.001	0.22 (0.17)	0.22 (0.17)	0.22 (0.18)	0.701
Baso (109/L)	0.05 (0.06)	0.05 (0.06)	0.06 (0.07)	<0.001	0.05 (0.06)	0.05 (0.06)	0.06 (0.07)	0.001
RBC (1012/L)	4.73 (0.51)	4.73 (0.50)	4.75 (0.53)	0.025	4.76 (0.52)	4.76 (0.51)	4.75 (0.53)	0.709
HGB (g/dL)	14.1 (1.56)	14.1 (1.56)	14.2 (1.59)	0.004	14.2 (1.59)	14.2 (1.58)	14.2 (1.59)	0.246
HCT (%)	41.9 (4.30)	41.8 (4.29)	42.1 (4.37)	0.001	42.1 (4.33)	42.0 (4.28)	42.2 (4.38)	0.303
MCV (fL)	88.7 (5.94)	88.7 (5.93)	88.9 (6.02)	0.057	88.8 (6.08)	88.6 (6.13)	89.0 (6.02)	0.029
MCH (pg)	29.9 (2.43)	29.9 (2.43)	30.0 (2.44)	0.300	29.9 (2.50)	29.9 (2.56)	30.0 (2.43)	0.055
MCHC (g/dL)	33.7 (1.05)	33.7 (1.05)	33.7 (1.05)	0.211	33.7 (1.05)	33.7 (1.06)	33.7 (1.05)	0.447
RDW (%)	13.4 (1.37)	13.3 (1.37)	13.5 (1.38)	<0.001	13.5 (1.41)	13.6 (1.46)	13.5 (1.36)	0.211
PLT (109/L)	256 (68.3)	256 (68.0)	256 (70.1)	0.930	254 (70.1)	252 (70.1)	256 (70.1)	0.034
MPV (fL)	8.13 (0.91)	8.13 (0.91)	8.16 (0.91)	0.129	8.16 (0.91)	8.16 (0.90)	8.16 (0.91)	0.940
NLR (%)	2.13 (1.16)	2.12 (1.16)	2.23 (1.16)	<0.001	2.20 (1.14)	2.18 (1.12)	2.23 (1.16)	0.091
SII	548 (346)	545 (346)	573 (346)	<0.001	560 (339)	548 (332)	572 (345)	0.013

Regarding T-cells, the results indicated that OSA was associated with decreased counts of terminally differentiated CD4-CD8- T cells (β_IVW = −0.33, SE = 0.16, p = 0.039), the percentage of CD25++ CD8+ T-cells on CD8+ T-cells (β_IVW = −0.35, SE = 0.16, p = 0.034) and CD3 on Natural Killer T-cells (β_IVW = −0.47, SE = 0.17, p = 0.005). In the analyses of B-cells, there was evidence that genetically determined OSA may decrease the HLA-DR on B cells (β_IVW = −0.50, SE = 0.25, p = 0.041). In contrast, there were evidence that the number of CD20 on IgD + CD38- unswitched memory B-cells (β_IVW = 0.43, SE = 0.21, p = 0.039) and IgD on IgD + CD38+ B-cells (β_IVW = 0.36, SE = 0.15, p = 0.020) was increased. More details on the subsets analysis are provided in Fig. 4 and Supplemental Table S7.

Fig. 4
Forest plot of MR showing estimates of genetically predicted OSA on circulating lymphocyte subpopulations.

Sensitivity analysis and reverse MR analysis

There was no evidence of potential heterogeneity or horizontal pleiotropy in the genetic prediction of OSA and outcome, with comprehensive results detailed in Supplementary Table S3. The visualization of results, including scatter plot, forest plot, funnel plot, and leave-one-out analyses is shown in Figures S1-S23. The reverse MR analysis revealed that all positive results were not biased by bi-directional effects (see Supplementary Table S8).

Discussion

To the best of our knowledge, this study is the first to use cross-sectional NHANES data and two-sample MR with a large genetic database to identify the effects of obstructive sleep apnea on circulating immune cells. Not only the NHANES analyses linked symptom based OSA risk (characterized by self-reported snoring, nocturnal apnea, and daytime sleepiness) to elevated inflammatory markers (e.g., white blood cell, lymphocyte, mean platelet volume), which has been reported in previous meta-analyses, but also have established associations between OSA and novel hematological indices (NLR and SII). Impressively, our study advances the field by providing the first causal evidence, via GWAS-based Mendelian randomization, linking OSA to specific leukocyte alterations (myeloid cells, neutrophils, granulocytes, eosinophils, and lymphocytes). Beyond confirming prior observations, we found that OSA causally disrupts lymphocyte homeostasis - particularly in key immunosurveillance populations such as NKT cells and terminally differentiated CD4-CD8- T-cells. These findings suggest that OSA-associated immune dysfunction may mechanistically contribute to its comorbidities, positioning these immune subsets as potential biomarkers for risk stratification and therapeutic targets (e.g., immunomodulation in OSA-related cancers).

Our findings demonstrated that OSA was causally associated with an increased counts of white blood cells, granulocytes, and neutrophil counts, as estimated by both cross-sectional and MR results, which is consistent with previous observational findings.⁷ Several mechanisms, including chronic hypoxia, sympathetic overactivity, oxidative stress, and endothelial dysfunction, have been implicated in the relationship between OSA and changes in circulating blood cells.³⁴ Regarding the decrease in lymphocytes observed in the MR findings, some researchers have shown that lower lymphocyte counts are associated with hypothalamic-pituitary-adrenal axis activation, increased systemic cortisol production, and altered sleep patterns.³⁵ In addition, our NHANES results after propensity score matching analyses showed that OSA was associated with a higher level of MCV and PLT, which have been less reported in previous studies. Nena et al. suggested that hypoxia may have an activating effect on platelet function, specifically, thrombocytes have Cyclooxygenase (COX) enzymes that can stimulate the synthesis of prostanoids, such as the inflammatory thromboxane A2, making them a potential source of inflammation.³⁶ Last but not least, SII may be a more comprehensive representation of the balanced state of immune and inflammatory conditions,³⁷ and has been recognized as a strong predictor of mortality from cardiovascular diseases.³⁸ The validation of elevated SII levels as indicators of heightened inflammation in OSA was further confirmed in our large population-based study.

Our findings indicate decreased CD3 expression on NKT cells in individuals with OSA, consistent with the research conducted by Gaoatawe et al.³⁹ Their study revealed that patients with severe OSA exhibited an inverse correlation between the number of invariant NKT cells and factors such as apnea-hypopnea index and SpO₂% <90%. A reduced count of NKT cells has been associated with increased vulnerability to atherosclerotic plaque instability and cancer immunity.³⁹,⁴⁰ T-cells and B-cells are crucial in regulating immune responses and inflammatory processes, however, their relationship to OSA remains relatively poorly understood. Our MR analysis revealed that genetically predicted OSA causally increase in Immunoglobulin D (IgD) levels on IgD + CD38+ B-cells and CD20 on IgD + CD38- unswitched memory B-cells. IgD serves as a marker of B-cell maturation, and unswitched memory B-cells are usually at an earlier stage of B-cell development. Our findings suggest that OSA promotes the maturation process of B-cells, indicating a body’s stress-induced pro-inflammatory response. This observation aligns with similar studies that have demonstrated an increase in the expression of CD19+, a cell surface marker typically associated with the development and activation of B-cells.⁴¹ In addition to promoting the maturation of B-cells, the findings also indicate that the reduction in terminally differentiated CD4-CD8- T-cells and the dual pattern of CD45 expression (a decrease in CD45 on CD33- HLA-DR- and an increase in CD45 on granulocytes) suggest a dynamic shift towards activated immune cells in the immune cell composition, potentially signifying alterations in the immune response.¹⁰

From the perspective of pathophysiological mechanisms, emerging evidence suggests that OSA-related pathophysiological mechanisms-including CIH, sleep fragmentation, and sympathetic overactivation-disrupt neuroendocrine-immune homeostasis, leading to impaired innate and adaptive immunity. These alterations may foster a permissive microenvironment for tumorigenesis, accelerate cancer progression, and contribute to therapy resistance in malignancies such as lung cancer.⁴² Go deeper, a key mechanistic link involves TGF-β, a pleiotropic cytokine with dual roles in tumor suppression and promotion. In the tumor microenvironment, TGF-β signaling modulates immune evasion by suppressing cytotoxic T-cell and NK-cell activity (which is consist with our study results) while promoting the expansion of immunosuppressive cell populations (e.g., Tregs, MDSCs).⁴³ In addition, OSA-associated HIF-1α, overexpressed in many cancers, not only drives treatment resistance but also upregulates TGF-β in monocytes and NK cells, inducing an immunosuppressive phenotype.⁴⁴ From Clinical evidences, monocytes from untreated OSA patients exhibit elevated TGF-β secretion, which not only suppresses Natural Killer (NK) cell cytotoxicity but also enhances lung cancer cell migration ‒ a defect reversible upon CPAP-mediated reoxygenation.⁴⁵

From the perspective of clinical standpoint, our findings highlight that OSA-associated immune dysregulation ‒ particularly neutrophil hyperactivation and NKT cell depletion ‒ may mechanistically contribute to comorbidities such as cardiovascular disease and cancer.³⁸,³⁹ Clinically, these immune alterations could serve as biomarkers for risk stratification (e.g., identifying patients with heightened inflammatory or oncogenic potential) and have potential to guide personalized management strategies, including targeted immunomodulation (e.g., neutrophil-inhibiting therapies or NKT cell reconstitution) alongside conventional OSA therapies.

Moving forward, with the advent of artificial intelligence and advanced portable monitoring technologies, farther validation of our findings using objective OSA diagnostic tools (e.g., home sleep apnea testing or polysomnography) in large-scale population studies would be of significant interest. Additionally, well-designed randomized controlled trials should be conducted to systematically evaluate the differential effects of OSA therapies ‒ such as CPAP, surgical interventions, and alternative treatments ‒ on immune modulation, particularly T-cell and B-cell functional dynamics.

The strengths of our study include the use of a large, nationally representative cohort in combination with two-sample MR analyses to investigate the effect of OSA on circulating blood cell phenotypes. MR analyses are less susceptible to bias from reverse causation and confounding than traditional observational studies because they use genetic information as an instrumental variable. Additionally, propensity score matching analyses allowed for the simultaneous inclusion of multiple factors as covariates supported by the large sample size of NHANES. Thus, the consistency between the MR analyses and the observational study added robustness to the findings.

However, this study had several limitations. Firstly, although the cross-sectional findings could be influenced by recall bias due to NHANES’s self-reported symptoms, the combination of key symptoms (snoring, nocturnal apneas, and daytime sleepiness) may help attenuate potential confounding to some extent. Secondly, all GWAS participants in the MR study were of European origin which may limit the generalizability to other populations. Despite the ethnic differences, our results found overlapping and consistent results in NHANES and MR, reinforcing the cross-ethnic impact of OSA on the immunity system. Thirdly, a well-recognized challenge of PSM, particularly with 1:1 matching without replacement, often reduces sample size and statistical power due to unmatched cases. However, our study achieved 99% matching success for OSA cases, maintaining robust statistical power despite using stringent calipers. Although we adjusted for multiple relevant confounders, residual confounding from unmeasured factors may still exist. However, our Mendelian randomization analysis provided evidence supporting causal associations for some key findings. Fourthly, MR approaches for estimating causal effects using instrumental variables assume the exposure-outcome relationship is linear, hence U- or J-shaped effects may be overlooked. Future investigations are needed to identify the non-linear causality between sleep apnea measures and immune cell function. Finally, given the recognized heterogeneity of OSA phenotypes, our MR analysis could not account for potential differences related to daytime sleepiness. Future studies with more detailed phenotypic data are needed to identify the linear and the non-linear causality between sleep apnea measures and immune cell function.

Conclusions

Our findings demonstrated that OSA is characterized by elevated levels of specific peripheral immune cells and SII, a newly established immune-inflammatory index. Importantly, our findings highlight that OSA-associated immune dysregulation ‒ particularly neutrophil hyperactivation and NKT cell depletion ‒ may mechanistically contribute to comorbidities such as cardiovascular disease and cancer. Clinically, these immune alterations could serve as biomarkers for risk stratification (e.g., identifying patients with heightened inflammatory or oncogenic potential) and guide personalized management strategies, including targeted immunomodulation (e.g., neutrophil-inhibiting therapies or NKT cell reconstitution) alongside conventional OSA therapies.

• 1
  These authors contributed equally to this work.
• Funding
  This project was supported by Beijing Natural Science Foundation (L243032), National Natural Science Foundation of China (81970866), and Innovative Health Promotion Strategies for China's New Era (2019-XZ-29).

Data availability statement

Nhanes data are available at https://www.cdc.gov/nchs/nhanes. The datasets analyzed in this Mendelian randomization study are publicly accessible summary statistics, and all summary data can be easily downloaded from the designated website: the IEU Open GWAS project https://gwas.mrcieu.ac.uk/ and https://www.finngen.fi/en.

Acknowledgments

We thank the FinnGen consortium, and the European Bioinformatics Institute (EBI) for providing GWAS data. Meanwhile, we are appreciative of the MRC IEU Open GWAS database

Supplementary data

Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.bjorl.2026.101810.

Appendix A

Appendix A

Abbreviations

• Baso  Basophils
• Eos  Eosinophils
• GWAS  Genome-Wide Association Studies, HGB, Hemoglobin
• HCT  Hematocrit, ICD, International Classification of Diseases
• IVW  Inverse Variance Weighting
• IVs  Instrumental Variables
• Lym  Lymphocyte
• MCH  Mean Cell Hemoglobin
• MCHC  Mean Cell Hemoglobin Concentration
• MCV  Mean Cell Volume
• Mono  Monocyte
• MPV  Mean Platelet Volume
• MR  Mendelian Randomization
• MR-PRESSO  MR Pleiotropy Residual Sum and Outlier
• NKT  Natural Killer T-cell
• Nsg  Neutrophils
• OSA  Obstructive Sleep Apnea-Hypopnea
• PLT  Platelet
• RBC  Red Blood Cell
• RDW  Red Cell Distribution Width
• SNPs  Single Nucleotide Polymorphisms
• WM  Weight Median
• WBC  White Blood Cell

References

• ¹ Benjafield AV, Ayas NT, Eastwood PR, Heinzer R, Ip MSM, Morrell MJ et al. Estimation of the global prevalence and burden of obstructive sleep apnoea: a literature-based analysis. Lancet Respir Med. 2019;7:687-98.
• ² Jordan AS, McSharry DG, Malhotra A. Adult obstructive sleep apnoea. Lancet. 2014;383:736-47.
• ³ Cubillos-Zapata C, Almendros I, Díaz-García E, Toledano V, Casitas R, Galera R, et al. Differential effect of intermittent hypoxia and sleep fragmentation on PD-1/PD-L1 upregulation. Sleep. 2020;43:zsz285.
• ⁴ Parkin J, Cohen B. An overview of the immune system. Lancet. 2001;357:1777-89.
• ⁵ Gutiérrez-Melo N, Baumjohann D. T follicular helper cells in cancer. Trends Cancer. 2023;9:309-2
• ⁶ Hinkley H, Counts DA, VonCanon E, Lacy M. T Cells in Atherosclerosis: Key Players in the Pathogenesis of Vascular Disease. Cells. 2023;12:2152.
• ⁷ Wu M, Zhou L, Zhu D, Lai T, Chen Z, Shen H et al. Hematological indices as simple, inexpensive and practical severity markers of obstructive sleep apnea syndrome: a meta-analysis. J Thorac Dis. 2018;10:6509-21.
• ⁸ Topuz MF, Ture N, Akdag G, Arik O, Gulhan PY. The importance of systemic immune-inflammation index in obstructive sleep apnea syndrome. Eur Arch Otorhinolaryngol. 2022;279:5033-
• ⁹ Yenigun A, Karamanli H. Investigation of the relationship between neutrophil-to-lymphocyte ratio and obstructive sleep apnoea syndrome. J Laryngol Otol. 2015;129:887-92.
• ¹⁰ Li K, Zhuo Y, He Y, et al. T cell receptor repertoire as a novel indicator for identification and immune surveillance of patients with severe obstructive sleep apnea. PeerJ. 2023;11:e15009.
• ¹¹ Dyugovskaya L, Lavie P, Lavie L. Phenotypic and functional characterization of blood gammadelta T cells in sleep apnea. Am J Respir Crit Care Med. 2003;168:242-9.
• ¹² Fan Z, Lu X, Long H, Li T, Zhang Y. The association of hemocyte profile and obstructive sleep apnea. J Clin Lab Anal. 2019;33:e22680.
• ¹³ Zielinski MR, Gibbons AJ. Neuroinflammation, Sleep, and Circadian Rhythms. Front Cell Infect Microbiol. 2022;12:853096-853096.
• ¹⁴ Hayley AC, Williams LJ, Venugopal K, Kennedy GA, Berk M, Pasco JA. The relationships between insomnia, sleep apnoea and depression: findings from the American National Health and Nutrition Examination Survey, 2005-2008. Aust N Z J Psychiatry. 2015;49:156-170.
• ¹⁵ Sidore C, Busonero F, Maschio A, Porcu E, Naitza S, Zoledziewska M et al. Genome sequencing elucidates Sardinian genetic architecture and augments association analyses for lipid and blood inflammatory markers. Nat Genet. 2015;47:1272-81.
• ¹⁶ Smith GD, Lawlor DA, Harbord R, Timpson N, Day I, Ebrahim S. Clustered environments and randomized genes: a fundamental distinction between conventional and genetic epidemiology. PLoS Med. 2007;4:e352.
• ¹⁷ Davies NM, Holmes MV, Davey SG. Reading Mendelian randomisation studies: a guide, glossary, and checklist for clinicians. BMJ. 2018;362:k601.
• ¹⁸ Hu B, Yang X, Xu Y, Sun YF, Sun C, Guo W et al. Systemic immune-inflammation index predicts prognosis of patients after curative resection for hepatocellular carcinoma. Clin Cancer Res. 2014;20:6212-22.
• ¹⁹ Cavallino V, Rankin E, Popescu A, Gopang M, Hale L, Meliker JR. Antimony and sleep health outcomes: NHANES 2009-2016. Sleep Health. 2022;8:373-9.
• ²⁰ Senaratna CV, Perret JL, Lodge CJ, Lowe AJ, Campbell BE, Matheson MC et al. Prevalence of obstructive sleep apnea in the general population: A systematic review. Sleep Med Rev. 2017;34:70-81.
• ²¹ Yeghiazarians Y, Jneid H, Tietjens JR, Redline S, Brown DL, El-Sherif N, et al. Obstructive Sleep Apnea and Cardiovascular Disease: A Scientific Statement from the American Heart Association. Circulation. 2021;144:e56-e67.
• ²² Hong C, Zhu H, Zhou X, Zhai X, Li S, Ma W et al. Association of Blood Urea Nitrogen with Cardiovascular Diseases and All-Cause Mortality in USA Adults: Results from NHANES 1999-2006. Nutrients. 2023;15:461.
• ²³ Christensen JJ, Osnes LT, Halvorsen B, Retterstøl K, Bogsrud MP, Wium C, et al. Altered leukocyte distribution under hypercholesterolemia: A cross-sectional study in children with familial hypercholesterolemia. Atherosclerosis. 2017;256:67-74.
• ²⁴ Siedlinski M, Jozefczuk E, Xu X, Teumer A, Evangelou E, Schnabel RB et al. White Blood Cells and Blood Pressure: A Mendelian Randomization Study. Circulation. 2020;141(16):1307-17.
• ²⁵ Hayati RP, Lee KJ, Simpson JA. The rise of multiple imputation: a review of the reporting and implementation of the method in medical research. BMC Med Res Methodol. 2015;15:30.
• ²⁶ Li L, Greene T. A weighting analogue to pair matching in propensity score analysis. Int J Biostat. 2013;9:215-34.
• ²⁷ Austin PC. An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies. Multivariate Behav Res. 2011;46:399-424.
• ²⁸ Strausz S, Ruotsalainen S, Ollila HM, Karjalainen J, Kiiskinen T, Reeve M et al. Genetic analysis of obstructive sleep apnoea discovers a strong association with cardiometabolic health. Eur Respir J. 2021;57:2003091.
• ²⁹ Orru V, Steri M, Sidore C, Marongiu M, Serra V, Olla S, et al. Complex genetic signatures in immune cells underlie autoimmunity and inform therapy. Nat Genet. 2020;52:1036-45.
• ³⁰ Gao X, Wei T, Xu S, Sun W, Zhang B, Li C, et al. Sleep disorders causally affect the brain cortical structure: A Mendelian randomization study. Sleep Med. 2023;110:243-53.
• ³¹ Liao J, Gao X, Fang T, Li Y, Han D. Obstructive sleep apnea's causal links to depression, well-being, and negative moods: a bidirectional mendelian randomized study. Eur Arch Psychiatry Clin Neurosci. 2026;276(2):543-551.
• ³² Palmer TM, Lawlor DA, Harbord RM, Sheehan NA, Tobias JH, Timpson NJ et al. Using multiple genetic variants as instrumental variables for modifiable risk factors. Stat Methods Med Res. 2012;21:223-42.
• ³³ Burgess S, Scott RA, Timpson NJ, Davey SG, Thompson SG. Using published data in Mendelian randomization: a blueprint for efficient identification of causal risk factors. Eur J Epidemiol. 2015;30:543-52.
• ³⁴ Ryan S, Taylor CT, McNicholas WT. Selective Activation of Inflammatory Pathways by Intermittent Hypoxia in Obstructive Sleep Apnea Syndrome. Circulation. 2005;112:2660-7.
• ³⁵ Acanfora D, Gheorghiade M, Trojano L, Furgi G, Pasini E, Picone C, et al. Relative lymphocyte count: a prognostic indicator of mortality in elderly patients with congestive heart failure. Am Heart J. 2001;142:167-73.
• ³⁶ Nena E, Papanas N, Steiropoulos P, Zikidou P, Zarogoulidis P, Pita E et al. Mean Platelet Volume and Platelet Distribution Width in non-diabetic subjects with obstructive sleep apnoea syndrome: new indices of severity? Platelets. 2012;23:447-54.
• ³⁷ Xu B, Wu Q, La R, Lu L, Abdu FA, Yin G, et al. Is systemic inflammation a missing link between cardiometabolic index with mortality? Evidence from a large population-based study. Cardiovasc Diabetol. 2024;23:212.
• ³⁸ Xia Y, Xia C, Wu L, Li Z, Li H, Zhang J. Systemic Immune Inflammation Index (SII), System Inflammation Response Index (SIRI) and Risk of All-Cause Mortality and Cardiovascular Mortality: A 20-Year Follow-Up Cohort Study of 42,875 US Adults. J Clin Med. 2023;12(3):1128.
• ³⁹ Gaoatswe G, Kent BD, Corrigan MA, Nolan G, Hogan AE, McNicholas WT, et al. Invariant Natural Killer T Cell Deficiency and Functional Impairment in Sleep Apnea: Links to Cancer Comorbidity. Sleep. 2015;38:1629-34.
• ⁴⁰ Ohmura Y, Ishimori N, Saito A, Yokota T, Horii S, Tokuhara S et al. Natural Killer T Cells Are Involved in Atherosclerotic Plaque Instability in Apolipoprotein-E Knockout Mice. Int J Mol Sci. 2021;22:12451.
• ⁴¹ Domagała-Kulawik J, Osińska I, Piechuta A, Bielicki P, Skirecki T. T, B, and NKT Cells in Systemic Inflammation in Obstructive Sleep Apnoea. Mediators Inflamm. 2015;2015:161579-7.
• ⁴² Yuan F, Hu Y, Xu F, Feng X. A review of obstructive sleep apnea and lung cancer: epidemiology, pathogenesis, and therapeutic options. Front Immunol. 2024;15:1374236.
• ⁴³ Romero-Garcia S, Moreno-Altamirano MM, Prado-Garcia H, Sánchez-García FJ. Lactate Contribution to the Tumor Microenvironment: Mechanisms, Effects on Immune Cells and Therapeutic Relevance. Front Immunol. 2016;7:52.
• ⁴⁴ Mallikarjuna P, Zhou Y, Landström M. The Synergistic Cooperation between TGF-β and Hypoxia in Cancer and Fibrosis. Biomolecules. 2022;12(5):635.
• ⁴⁵ Hernández-Jiménez E, Cubillos-Zapata C, Toledano V, Pérez de Diego R, Fernández-Navarro I, Casitas R et al. Monocytes inhibit NK activity via TGF-β in patients with obstructive sleep apnoea. Eur Respir J. 2017;49:1602456.

Edited by

• Edited by
  Dr C Chone.

Publication Dates

• Publication in this collection
  17 Aug 2026
• Date of issue
  2026

History

• Received
  11 Dec 2024
• Accepted
  19 Jan 2026
• Published
  01 June 2026