Impact of aerobic exercise on immune components across healthy and diseased populations: A systematic review and meta-analysis of randomized controlled trials

Abstract

Background

Aerobic exercise may improve immune component quantities in healthy and diseased populations, but its effects across different health conditions and immune systems remain unclear. This review examined its impact on immune components in both populations.

Methods

A search in June 2025 across four databases included randomized controlled studies on aerobic exercise's effects on immune components in adults. Data (M ± SD) were extracted and converted to standardized mean difference (SMD) using random-effect meta-analysis.

Results

Seventeen studies (502 participants) were included. The meta-analysis results showed long-term aerobic exercise increased T-cell counts in diseased individuals (p < 0.05) but not in healthy ones (p > 0.05). It significantly reduced IgA levels in healthy participants compared to non-exercising controls (p < 0.05), potentially counteracting natural temporal increases observed in sedentary populations, without affecting IgG, IgM, leukocytes, neutrophils, or monocytes in either group (p > 0.05). Single-pass exercise did not alter leukocytes, lymphocytes, CD3⁺, or NK cells in healthy participants (p > 0.05).

Conclusions

Long-term aerobic exercise primarily affects adaptive immune components, benefiting individuals with unstable immune conditions. Single-pass exercise has no effect on healthy populations, making long-term interventions more suitable for improving adaptive immunity in unstable immune states.

Systematic review registration

www.crd.york.ac.uk/PROSPERO/, identifier: CRD42024546696.

1. Background

The immune system, a complex network of cells and molecules, safeguards the body against pathogens and malignancies. Improved immune function benefits healthy populations by preventing infections, maintaining systemic homeostasis, and promoting long-term health.1, 2, 3, 4, 5 For patients with compromised immune function, manifestations often include autoimmune dysregulation or diminished capacity to defend against infections. Enhancing immune function can significantly reduce the risk of infections and improve disease management.^4,5 The level of immune component quantity is critical to the immune function, but the relationship between them oftentimes varies. For example, in multiple sclerosis, improvement in immune function is along with the increase in T cells and a decrease in monocytes,^6,7 which the increase in T-cell counts potentially exacerbates immune function in rheumatoid arthritis.^6,8, 9, 10 Therefore, it is critical to regulate the immune component quantify using appropriate strategies for the enhancement of immune function.

Multiple types of strategies to regulate the immune composition have been developed, including nutritional supplementation (e.g., vitamins and probiotics), pharmacological treatments (e.g., interferon),11, 12, 13, 14, 15, 16 and moderate physical exercise. Among them, aerobic exercise, a cost-effective non-pharmacological approach, has increasingly been receiving more and more attention.^12,17,18 Aerobic exercise is a relatively prolonged and low-intensity form of exercise in which oxygen is used as the energy supply.^19,20 It can influence the immune system through the regulation of blood circulation and stress hormone levels, thereby modulating the immune component quantities via cellular and molecular activities (e.g., migration, proliferation, and apoptosis).^8,21, 22, 23, 24 However, the effects of aerobic exercises on immune markers are observed to be inconsistent across studies.^21,22,24, 25, 26, 27, 28 Several studies observed that aerobic exercise could improve the cell counts of healthy participants and participants with conditions (e.g., sickle cell anemia patients),22, 23, 24, 25^,27,29 while other studies reported no such effect.^{8,9,21,26,28,30} Such inconsistency in the observations may be due to the variations in participant characteristics (e.g., healthy cohort versus patients), and the targeting immune component across studies.

Although systematic reviews have investigated the immunomodulatory effects of aerobic exercise,³¹ they have not adequately resolved the controversies concerning the differential effects of acute versus chronic aerobic exercise on immune parameters. Furthermore, there is a lack of meta-analyses quantifying the magnitude of aerobic exercise's impact on specific immune components. Most significantly, no comprehensive comparisons have been made between healthy populations and those with immune dysregulation. In this study, we classified participants into two distinct cohorts based on immune status: healthy individuals (without chronic conditions or persistent immune abnormalities) and clinical populations with chronic immune dysregulation (encompassing conditions characterized by sustained inflammation, immunosuppression, or immune homeostasis disruption). This binary classification was adopted because all included clinical conditions share fundamental features of impaired immune homeostasis, where aerobic exercise has demonstrated comparable immunomodulatory benefits despite disease-specific variations. While recognizing the inherent heterogeneity across conditions, the current landscape of available studies - with limited sample sizes for individual diseases - necessitates this broader categorization to enable meaningful analysis.

Building upon this classification framework, we conducted a systematic review and meta-analysis based on the up-to-date peer-reviewed publications of randomized controlled trials (RCTs) to quantitatively characterize how aerobic exercise differentially modulates specific immune components across these cohorts. This work will ultimately provide critical knowledge for the efficacy of aerobic exercise on immune function across different cohorts. It will define the boundaries for the clinical application of aerobic interventions, specifying which immune components in which populations should be targeted for improvement. Additionally, this will aid in designing future studies in this field.

2. Methods

2.1. Design

This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines.³² Review methods were registered with PROSPERO (CRD42024546696).

2.2. Literature search

The literature search was independently carried out by two researchers (GD and XH). Searches for articles were conducted in four health-related and biomedical databases (PubMed, Web of Science, Medline, and Cochrane Library) from inception to June 6, 2025. Studies were searched in the electronic databases using the following key terms combined by Boolean logic (“AND”, “OR”, “NOT”): (‘immune system’ OR ‘psychoneuroimmunology’ OR ‘neutrophil’ OR ‘iga’ OR ‘immunoglobulin’ OR ‘macrophage’ OR ‘monocyte’ OR ‘leukocyte’ OR ‘lymphocyte’ OR ‘immunity’) AND (‘aerobic∗’ OR ‘aerobic exercise’ OR ‘endurance exercise’ OR ‘aerobic training’ OR ‘endurance training’ OR ‘cardio training’ OR ‘physical endurance’ OR ‘physical exertion’) (Supplementary Material, Table S1). In addition, references of the included studies and previously published reviews were manually searched.³¹

2.3. Selection criteria

All included studies were published articles. The inclusion criteria were carried out according to the PICOS principle: (1) Population: involved human participants without restriction for age, sex, health condition, ethnicity, education level, or socioeconomic status (where reported); (2) Interventions: the intervention group involved at least one session of aerobic exercise, including both acute (single-bout, with outcomes measured ≤24 h post-exercise) and chronic (structured training program lasting ≥2 weeks) interventions; (3) Comparisons: the control group received no intervention; (4) Outcomes: investigated at least one outcome of single-pass or long-term aerobic exercise intervention on innate or adaptive immune component quantities; (5) Study design: randomized controlled trials (RCTs), including crossover RCTs, were included due to their ability to provide robust evidence for acute treatment effects. Studies were excluded if they: (1) involved animal subjects; (2) were published in non-English languages; (3) lacked specific outcome data; (4) were review articles or conference proceedings; (5) represented duplicate publications; or (6) employed quasi-randomized allocation methods (e.g., alternation or birth date assignment), as these approaches may compromise allocation concealment standards. While unpublished studies were identified during the literature search process, only peer-reviewed publications were ultimately included to prioritize methodological validity and data reliability.

2.4. Data extraction

The data extraction process was conducted according to the Cochrane Collaboration Handbook.³³ Two authors (GD and XH) extracted the relevant data from the included studies in a standardized manner including the authors, publication year, sample size, participant characteristics, exercise protocol, and outcome measures. Any disagreement between the two authors was discussed with other two authors (GD and XH) until a consensus was achieved.

The mean and standard deviation of each outcome in post-tests in each study were extracted. For studies that the outcomes were presented as “Mean ± SE/SEM (standard error/standard error of mean)”, calculations were conducted using the following formulas³⁴:

$$SD=SE\times \sqrt{n}$$

If the full-text article data were presented only in a figure format, WebPlotDigitizer (Ankit Rohatgi, 2019, V.4.2; WebPlotDigitizer, Pacifica, CA, USA) was used to extract the data from the figures. If any relevant data were missing, the corresponding author or authors were contacted by the first researcher via email.

2.5. Quality assessment

The quality of the included studies was assessed independently by two authors (GD and XH) based on the principles of the Physiotherapy Evidence Database (PEDro). PEDro was specifically designed for the evaluation of physical therapy interventions, which is appropriate for the type of studies we included here. The reliability and validity of PEDro have been widely validated.³⁵ The PEDro scale includes 11 items, and each study was assessed as either “yes” (score 1) or “no” (score 0) for each of those items. According to the PEDro guidelines, the maximum total score was 10 (item 1 is not used to compute the total score). As blinding (especially participants and investigators) was not easily implemented in exercise intervention trials,³⁶ the methodological quality classification of each article was adjusted with eligibility criteria considered as previously described [sum scores: ≥6 (“high quality, low risk of bias”); scores: 4–5 (“acceptable quality, moderate risk of bias”), and scores: ≤3 (“low quality, high risk of bias”)].36, 37, 38, 39 Any score on which the two authors disagreed was discussed with a third author (GD and XH) until a consensus was achieved.

2.6. Statistical analysis

Revman 5.4 software (Cochrane Collaboration, Oxford, United Kingdom) and Stata version 17.0 (Stata Statistical Software, release 16; Stata Corp., College Station, TX, United States) were used for data synthesis and analysis. For the difference of units between outcome measures, a random-effects model was used to calculate the standardized mean difference (SMD; Hedges' g) of the outcomes, with a 95 % confidence interval (CI), to determine the effect size (ES) of the intervention. ES was classified as trivial (<0.2), small (0.2∼0.49), moderate (0.5∼0.79), or large (>0.8).⁴⁰ For patients, we unified the direction of changes in immune component quantities according to whether the change will improve their immune function, and merged SMD of studies to determine whether aerobic exercise had an improved effect on immune function. Had any trials contained multiple intervention arms, we would have either: (i) combined homogeneous interventions using the Mantel-Haenszel method with sample size adjustment, or (ii) for heterogeneous interventions, selected the most clinically relevant comparison while halving the control group sample size to avoid unit-of-analysis error.³³ Heterogeneity was assessed by measuring the inconsistency (I² statistic) of intervention effects among the trials. The level of heterogeneity was interpreted according to guidelines from the Cochrane Collaboration: trivial (<25 %), low (25∼50 %), moderate (50∼75 %), and high (>75 %).⁴¹ The publication bias was assessed by the funnel plot and Egger's test. If a significant asymmetry was detected, we used the Trim and Fill method for sensitivity analysis of the results.⁴²

3. Results

The flow of the study identification and selection process is summarized in Fig. 1. The systematic search yielded 9613 records: PubMed (n = 1998), Web of Science (n = 4252), Medline (n = 2384), Cochrane Library (n = 975), and Handsearching References (n = 4). A total of 3972 repetitive publications were excluded, leaving 5641 articles. Then, 5549 publications were excluded after reviewing titles and abstracts. After the evaluation of full texts, 75 of the 92 publications were removed, and thus 17 publications were identified and included in the systematic review and meta-analysis. Following full-text screening, all included studies reported complete outcome data for the primary and secondary endpoints of interest. As no missing data requiring additional clarification were identified, no author contact was necessary for data supplementation (see Fig. 2).

Fig. 1.

Study flowchart.

Fig. 2.

Impact of aerobic exercise on the immune component quantities in healthy and diseased participants.

3.1. Participant characteristics

A total of 502 participants with mean ages ranging from 18 to 79.5 years were included (Table 1). Five studies included both men and women participants,^{9,26,28,46,47} nine studies included only men participants,^{8,22,23,25,30,}43, 44, 45^,48 two studies included only women participants,^21,29 and one study did not report the sex of participants.²⁴

Table 1.

Characteristic of included studies (n = 17).

Study	Sample Size	Sex, men/women	Age, mean ± SD, y	Participants Characteristics				Intervention Characteristics								Outcome Measures
				Region	Education level	Socioeconomic status	Physical Condition	Exercise Type	Exercise Mode	Period	Duration	Frequency	Session Duration	Exercise Intensity	Intensity Grade
Baslund et al., 19939	AG (9)	8/1	49 ± 9	Denmark	–	–	Rheumatoid Arthritis	Cycling	Continuous	Long-term	8 weeks	4∼5 times	15 min	161 ± 6 beats/min	Moderate	Leu→, Neu→, Lym→, CD3+→, CD4+→, CD8+→, NK→
	CG (9)	8/1	47 ± 9	Denmark	–	–		None
LaPerriere et al., 199425	AG (7)	14/0	18∼40	America	–	–	Sedentary	Cycling	Continuous	Long-term	10 weeks	3 times	45 min	70∼80 % HRmax	Moderate to High	Leu→, Mon→, Lym→, CD4+↑, CD8+↑
	CG (7)			America	–	–		None
Moyna et al., 199626	AG (32)	32/32	Men: 24.6 ± 2.83 Women: 23.6 ± 2.83	America	–	–	Healthy	Cycling	Continuous	Single-pass	–	–	18 min	55∼85 % VO2peak	Moderate to High	Mon→, Lym→, CD3+→, NK→
	CG (32)			America	–	–		None
Rhind et al., 199623	AG (9)	15/0	25.3 ± 3.2	Canda	College	–	Sedentary	Cycling	Continuous	Long-term	12 weeks	4∼5 times	30 min	65∼70 % VO2max	High	Leu→, Mon→, Lym→, CD3+→, CD4+→, CD8+→, NK↑
	CG (6)		23.4 ± 3.1	Canda	College	–		None
Mitchell et al., 199643	AG (11)	21/0	23.4 ± 7	America	College	–	Healthy	Cycling	Continuous	Long-term	12 weeks	3 times	30 min	75 % VO2peak	High	Lym→, IgA↓, IgG→, IgM→
	CG (10)		20.1 ± 1.9	America	College	–		None
Miles et al., 200244	AG (6)	10/0	32.8 ± 6.9	America	–	–	Moderately trained runners	Running	Continuous	Single-pass	–	–	60 min	78.2 ± 0.8 % VO2peak	High	CD3+→, NK↑ (Immediately after exercise), NK↓ (1.5h after exercise), NK→ (5h and 24h after exercise)
	CG (4)		28.8 ± 7.5	America	–	–		None
Sagiv et al., 200222	AG (15)	25/0	46.7 ± 2	Israel	–	–	Coronary artery disease	Running	Continuous	Long-term	12 weeks	3∼4 times	45 min	124 ± 10 beats/min	Low	Leu→, Neu→, Lym→, CD4+↑, CD8+↑, NK→,
	CG (10)		45.2 ± 2	Israel	–	–		None
Mcfarlin et al., 2003	10	10/0	18–25	America	–	–	Healthy	Cycling	Continuous	Single-pass	–	–	60 min	50 % and 70 % VO2peak	Moderate to High	Leu↑, Neu↑, Lym↑
Lee et al., 200545	AG (9)	18/0	26.8 ± 1.2	Korea	–	–	Healthy	Qigong	Continuous	Single-pass	–	–	60 min	Unclear	–	Leu→, Lym→, NK→
	CG (9)		26.1 ± 1.7	Korea	–	–		None
Martins et al., 200946	AG (22)	8/14	Men: 73.2 ± 4.5 Women: 77.7 ± 7.8	Rural area	–	–	Sedentary	Aerobic gymnastics	Continuous	Long-term	16 weeks	3 times	45 min	0∼4 weeks: 40–50 % HRR 5–8 weeks: 51–60 % HRR 9–12 weeks: 61–70 % HRR 13–16 weeks: 71–85 % HRR	Moderate	IgA→, IgG→, IgM→
	CG (21)	8/13	Men: 78.21 ± 7.68 Women: 79.5 ± 9.1	Rural area	–	–		None
Wang et al., 20118	AG (10)	20/0	21.5 ± 0.7	China	College	–	Sedentary	Cycling	Continuous	Long-term	4 weeks	5 times	30 min	50 % HRR	Moderate	Lym→, CD3+→, CD4+→, CD8+→
	CG (10)		22.9 ± 0.4	China	College	–		None
Weng et al., 201330	AG a (10)	30/0	22.3 ± 0.2	China	College	–	Sedentary	Cycling	Interval	Long-term	5 weeks	5 times	30 min	3-min intervals at 40 % and 80 %VO2max	High	Leu→, Lym→, CD4+→, CD8+→,
	AG b (10)		22.5 ± 1.0	China	College	–			Continuous					60 % VO2max	Moderate
	AG (10)		22.4 ± 0.9	China	College	–		None
Viana et al., 201428	AG (13)	8/5	61 ± 8	United Kingdom	–	–	Chronic kidney disease	Walking	Continuous	Long-term	12 weeks	3 times	30 min	50∼60 % VO2peak	Moderate	Leu→, Neu→, Mon→, Lym→
	CG (11)	7/4	56 ± 16	United Kingdom	–	–		None
Abd El-Kader et al., 201824	AG (30)	Unclear	26.54 ± 6.73	Saudi Arabia	–	–	Sickle cell anemia	Running	Continuous	Long-term	3 months	3 times	45 min	60∼70 % HRmax	Moderate	Leu↑, Neu↑, Mon↑, CD3+↑, CD4+↑, CD8+↑
	CG (30)		24.91 ± 7.32	Saudi Arabia	–	–		None
Yoon et al., 201821	AG (10)	0/20	53.7 ± 3.37	Korea	–	–	Healthy	Running	Continuous	Long-term	12 weeks	3 times	60 min	60–80 % HRR	High	IgA→, IgG→, IgM→, NK→
	CG (10)		52.5 ± 2.68	Korea	–	–		None
Ashem et al., 202029	AG (15)	0/30	45 ± 3.25	Egypt	–	–	Stage 1 breast cancer	Running, cycling, or exercising on elliptical machine	Continuous	Long-term	5 months	3 times	15∼45 min	0∼6 weeks: 60 % VO2max 7∼12 weeks: 70 % VO2max 12 weeks∼: 80 % VO2max	Moderate to High	IgA↑
	CG (15)		45.06 ± 2.98	Egypt	–	–		None
Abd El-Kader et al., 202027	AG (40)	14/26	51.27 ± 5.32	Saudi Arabia	–	–	Chronic primary insomnia	Running	Continuous	Long-term	6 months	3 times	45 min	60∼70 % HRmax	Moderate	CD3+↑, CD4+↑, CD8+↑
	CG (40)	12/28	52.64 ± 4.81	Saudi Arabia	–	–		None

AG, aerobic exercise group; CG, control group; VO₂max, maximal oxygen uptake; VO₂peak, peak oxygen uptake; HR, heart rate; HRR, heart rate reserve; Leu, leukocyte counts; Neu, neutrophil counts; Mon, Monocyte counts; Lym, lymphocyte counts; IgA, immunoglobulin A concentration; IgG, immunoglobulin G concentration; IgM, immunoglobulin M concentration; NK, natural killer cell counts; ↑, Intervention significantly improved the outcome compared with the control group (p < 0.05); →, Intervention induced no significant difference (p > 0.05); ↓, Intervention was significantly less effective than the control.

Eleven studies involved participants without any overt conditions, including sedentary participants (n = 5),^{8,23,25,30,46} moderately trained runners (n = 1),⁴⁴ and participants described as “healthy” (n = 5).^{21,26,43,45,48} Six studies involved participants with conditions, including rheumatoid arthritis,⁹ coronary artery disease,²² chronic kidney disease,²⁸ sickle cell anemia,²⁴ stage 1 breast cancer,²⁹ and chronic primary insomnia.²⁷

3.2. Intervention characteristics

The information on the intervention parameters is included in Table 1. The exercise types of the intervention included cycling (n = 8),^{8,9,23,25,26,30,43,48} running (n = 5),^{21,22,24,27,44} walking (n = 1),²⁸ Qigong (n = 1),⁴⁵ aerobic gymnastics (n = 1),⁴⁶ and one study gave participants the choice of running, cycling or exercising on an elliptical machine.²⁹ One study conducted both interval and continuous exercise modes,³⁰ and the other sixteen studies conducted continuous exercise modes. The exercise intensity of intervention included low intensity (n = 1),²² moderate intensity (n = 7),^{8,9,24,27,28,30,46} moderate to high intensity (n = 4),^25,26,29,48 and high intensity (n = 5),^{21,23,30,43,44} and the intensity was classified according to the study by Garber et al.⁴⁹ The control group in all sixteen studies conducted blank controls (i.e. received no intervention).

Thirteen studies conducted long-term intervention, and another four studies conducted single-pass intervention.^26,44,45,48 The duration of one intervention session varied from 15 to 60 min. For the studies conducted long-term intervention, the weekly frequency of the intervention included three sessions (n = 8),^21,24,25,27, 28, 29^,43,46 three to four sessions (n = 1),²² four to five sessions (n = 2),^9,23 and five sessions (n = 2).^8,30 The overall duration of the interventions ranged from 4 weeks to 6 months.

3.3. Outcome measurements

Immune system is classified into two main branches: the innate and adaptive immune systems.^4,5 The innate immune system offers immediate, non-specific defense through barriers such as skin, along with chemical mediators and cells like macrophages and neutrophils, without retaining memory.^4,5 In contrast, the adaptive immune system targets specific pathogens, which produce antibodies and destroy infected cells, respectively, while developing immunological memory to speed up future responses.^4,5 Outcome measures were categorized according to the innate and adaptive immune components. Immune components inclusive of both types: leukocytes counts (n = 9)^9,22, 23, 24, 25^,28,30,45,48 and lymphocytes counts (n = 11) ^{8,9,22,23,25,26,28,30,43,45,48}; innate immune components: monocytes counts (n = 5),23, 24, 25, 26^,28 neutrophils counts (n = 5),^{9,22,24,28,48} and natural killer (NK) cell counts (n = 7) ^{9,21–23,26,44,45}; adaptive immune components: CD3⁺ cell counts (n = 7),^{8,9,23,24,26,27,44} CD4⁺ cell counts (n = 8),^8,9,22, 23, 24, 25^,27,30 CD8⁺ cell counts (n = 8),^8,9,22, 23, 24, 25^,27,30 immunoglobulin A (IgA) concentration (n = 4),^21,29,43,46 immunoglobulin G (IgG) concentration (n = 3),^21,43,46 and immunoglobulin M (IgM) concentration (n = 3).^21,43,46

3.4. Quality assessment

The results of the quality evaluation of the included sixteen studies are shown in Table 2. All studies were evaluated as high quality.

Table 2.

Quality assessment of included studies (n = 17).

Study	Eligibility criteria	Random allocation	Concealed allocation	Similarity baseline	Participant blinding	Investigator blinding	Assessor blinding	Completeness of follow-up	Intention to treat	Between group comparisons	Point and variability measures	Total score
Baslund et al., 19939	YES	1	0	1	0	0	0	1	1	1	1	6
LaPerriere et al., 199425	YES	1	0	1	0	0	0	1	1	1	1	6
Moyna et al., 199626	YES	1	0	1	0	0	0	1	1	1	1	6
Rhind et al., 199623	YES	1	0	1	0	0	0	1	1	1	1	6
Mitchell et al., 199643	YES	1	0	1	0	0	0	1	1	1	1	6
Miles et al., 200244	YES	1	1	1	0	0	0	1	1	1	1	7
Sagiv et al., 200222	YES	1	0	1	0	0	0	1	1	1	1	6
Mcfarlin et al., 200348	YES	1	0	1	0	0	0	1	1	1	1	6
Lee et al., 200545	YES	1	0	1	0	0	0	1	1	1	1	6
Martins et al., 200946	YES	1	0	1	0	0	0	1	1	1	1	6
Wang et al., 20118	YES	1	0	1	0	0	0	1	1	1	1	6
Weng et al., 201330	YES	1	0	1	0	0	0	1	1	1	1	6
Viana et al., 201428	YES	1	1	1	0	0	0	1	1	1	1	7
Abd El-Kader et al., 201824	YES	1	1	1	0	0	0	1	1	1	1	7
Yoon et al., 201821	YES	1	0	1	0	0	0	1	1	1	1	6
Ashem et al., 202029	YES	1	1	1	0	0	0	1	1	1	1	7
Abd El-Kader et al., 202027	YES	1	1	1	0	0	0	1	1	1	1	7

3.5. Meta-analysis

The implications of immune component quantity changes vary significantly across different diseases. To address this, we adjusted the direction of immune component quantities changes within each disease context to uniformly reflect whether there was an improvement in immune function. Consequently, the meta-analysis focused on whether aerobic exercise enhanced immune system function, rather than merely reporting increases or decreases in immune component quantities. For the healthy population, whose immune characteristics and health markers are relatively consistent, the increases or decreases in immune component quantities were directly interpreted, as their implications are uniform across this group.

3.5.1. Effects of aerobic exercise on immune components containing innate and adaptive components

For healthy participants, five publications showed that the cell counts of leukocytes and lymphocytes were not significantly influenced after long-term aerobic exercises as compared to the control group,^{8,23,25,30,43} and two publications suggested that single-pass aerobic exercises could also not significantly influence cell counts of leukocyte and lymphocytes compared with the control group^26,45 (Table 1).

The results (Fig. 3A) showed that the ES of long-term intervention was non-significant and trivial in monocyte [SMD = −0.12, 95 %CI (−0.86, 0.61), p = 0.75] and lymphocyte counts [SMD = 0.04, 95 %CI (−0.35, 0.43), p = 0.84], non-significant and small in leukocyte counts [SMD = 0.21, 95 %CI (−0.27, 0.69), p = 0.40], without heterogeneity (I² = 0 %, p = 0.45).

Fig. 3.

Results regarding the effects of aerobic exercise on leukocytes and lymphocytes. A: healthy participants [(a) for long-term intervention; (b) for single-pass intervention], B: diseased participants (long-term intervention).

The ES of single-pass intervention was non-significant in leukocyte [SMD = −0.71, 95 %CI (−0.34, 1.75), p = 0.18] and lymphocyte counts [SMD = 0.25, 95 %CI (−0.47, 0.97), p = 0.49], with moderate heterogeneity (I² = 74 %, p < 0.01). The funnel plot (Fig. 4A) and Egger's test (t = 2.13, P = 0.04) indicated evidence of asymmetry, suggesting potential publication bias.

Fig. 4.

Funnel plots of immune components containing innate and adaptive components.

For participants with conditions, one publication showed that long-term aerobic exercises could significantly improve leukocyte counts as compared to the control group,²⁴ while other three publications reported no such effect^9,22,28; three publications suggested that long-term aerobic exercises could not significantly improve lymphocyte counts compared with the control group^9,22,28 (Table 1).

The results (Fig. 3B) showed that the ES of long-term intervention was non-significant and small in leukocyte counts [SMD = −0.28, 95 %CI (−0.81, 0.24), p = 0.29] and lymphocyte counts [SMD = −0.26, 95 %CI (−0.70, 0.17), p = 0.24], with low heterogeneity (I² = 34 %, p = 0.15). The funnel plot (Fig. 4B) and Egger's test (t = 1.54, P = 0.17) indicated no publication bias.

3.5.2. Effects of aerobic exercise on innate immune components

3.5.2.1. Monocytes and neutrophils

For healthy participants, two publications showed that the cell counts of monocytes were not significantly influenced after long-term aerobic exercises as compared to the control group^23,25 (Table 1). The results (Fig. 5A) showed that the ES of long-term intervention was non-significant and trivial in monocyte counts [SMD = −0.12, 95 %CI (−0.86, 0.61), p = 0.75], without heterogeneity (I² = 0 %, p = 0.88). The funnel plot (Fig. 7A) indicated no publication bias. Considering the limited number of studies (n < 10), the Trim and Fill method for sensitivity analysis was conducted [before: SMD = −0.12, 95 %CI (−0.86, 0.61); after: SMD = −0.07, 95 %CI (−0.67, 0.53)], which suggested that the result was robust.

Fig. 5.

Results regarding the effects of long-term aerobic exercise on monocytes and neutrophils. A: healthy participants, B: diseased participants.

Fig. 7.

Funnel plots of innate immune components. A: healthy participants, B: diseased participants.

For participants with conditions, one publication showed significant improvement in monocyte counts as compared to the control group,²⁴ while another publication reported no such effect²⁸; one publication showed significant improvement in neutrophil counts as compared to the control group,²⁴ while other three publications reported no such effect^9,22,28 (Table 1).

The results (Fig. 5B) showed that the ES of long-term intervention was non-significant and trivial in neutrophil counts [SMD = −0.11, 95 %CI (−0.71, 0.49), p = 0.72], and non-significant and small in monocyte counts [SMD = −0.44, 95 %CI (−2.20, 1.32), p = 0.62], with high heterogeneity (I² = 78 %, p < 0.01). The funnel plot (Fig. 7B) and Egger's test (t = 3.25, P = 0.03) indicated evidence of asymmetry, suggesting potential publication bias. The pooled ES (SMD = −0.35) was not robust after the filled meta-analysis, suggesting published studies may underestimate the actual ES.

3.5.2.2. Natural killer cells

For healthy participants, one publication suggested that long-term aerobic exercises could significantly increase NK cell counts as compared to the control group,²³ while another publication reported no such effect.²¹ One publication showed that NK cell counts increased immediately after single-pass aerobic exercises, decreased at 1.5h after the intervention compared with the control group, and there was no significant difference between the two groups at 5h and 24h after the intervention; the other two publications showed non-significant difference as compared to the control groups^26,45 (Table 1). The results (Fig. 6A) showed that the ES of long-term intervention was non-significant and moderate [SMD = 0.49, 95 %CI (−0.19, 1.17), p = 0.16], without heterogeneity (I² = 0 %, p = 0.57). The ES of single-pass intervention was non-significant and small [SMD = 0.22, 95 %CI (−0.12, 0.56), p = 0.24], without heterogeneity (I² = 0 %, p = 0.48). The funnel plot (Fig. 7A) and Egger's test (t = 0.06, P = 0.95) indicated no publication bias, and the Trim and Fill method suggested that the result was robust [before: SMD = 0.29, 95 %CI (−0.01, 0.59); after: SMD = 0.25, 95 %CI (−0.05, 0.54)].

Fig. 6.

Results regarding the effects of aerobic exercise on NK cells. A: healthy participants [(a) for long-term intervention; (b) for single-pass intervention], B: diseased participants (long-term intervention).

For participants with conditions, two publications showed that long-term aerobic exercises could not significantly improve NK cell counts as compared to the control group^9,22 (Table 1). The results (Fig. 6B) showed that the ES of intervention was non-significant and small [SMD = 0.22, 95 %CI (−0.34, 0.77), p = 0.44], with trivial heterogeneity (I² = 13.0 %, p = 0.32). The funnel plot (Fig. 7B) and Egger's test (t = 0.84, P = 0.56) indicated no publication bias, and the Trim and Fill method suggested that the result was robust as the pooled ES remained unchanged [SMD = 0.22, 95 %CI (−0.34, 0.77)].

3.5.3. Effects of aerobic exercise on adaptive immune components

3.5.3.1. T-cells

For healthy participants, one publication showed that long-term aerobic exercises could significantly increase two sub-types counts of T-cell (i.e. CD8⁺ and CD4⁺) as compared to the control group,²⁵ while other three publications reported no such effect.^8,23,30 Four publications suggested no significant effect in CD3⁺ counts after single-pass or long-term aerobic exercise^8,23,26,44 (Table 1). The results (Fig. 8A) showed that the ES of long-term intervention was non-significant and trivial in CD8⁺ [SMD = −0.01, 95 %CI (−0.43, 0.41), p = 0.95] and CD4⁺ counts [SMD = 0.11, 95 %CI (−0.31, 0.53), p = 0.61], non-significant and small in CD3⁺ counts [SMD = −0.20, 95 %CI (−0.87, 0.47), p = 0.56], without heterogeneity (I² = 0 %, p = 1.00). The ES of single-pass intervention was non-significant and small in CD3⁺ counts [SMD = −0.45, 95 %CI (−1.12, 0.23), p = 0.19], with low heterogeneity (I² = 44 %, p = 0.13). The funnel plot (Fig. 10A) and Egger's test (t = −0.01, P = 0.99) indicated no publication bias.

Fig. 8.

Results regarding the effects of aerobic exercise on T cells. A: healthy participants [(a) for long-term intervention; (b) for single-pass intervention], B: diseased participants (long-term intervention).

Fig. 10.

Funnel plots of adaptive immune components. A: healthy participants, B: diseased participants.

For the studies that investigated T-cell counts of participants with conditions, three studies conducted running exercise of low or moderate intensity,^22,24,27 and one study conducted cycling exercise of moderate intensity.⁹ The intervention duration ranged from 4 weeks to 6 months, 3 to 5 sessions per week, and 15–45 min per session. Three publications showed that long-term aerobic exercises could significantly improve two sub-types counts of T-cell (i.e. CD8⁺ and CD4⁺) as compared to the control group,^22,24,27 while another publication reported no such effect.⁹ Two publications suggested significant improvement in CD3⁺ counts,^24,27 while another publication reported no such effect⁹ (Table 1). The results (Fig. 8B) showed that the ES of long-term intervention was significant and moderate in CD4⁺ [SMD = −0.58, 95 %CI (−1.15, −0.00), p = 0.047 (0.05 after rounding by Revman 5.4 software)] and CD3⁺ counts [SMD = −0.74, 95 %CI (−1.12, −0.36), p < 0.01], significant and large in CD8⁺ counts [SMD = −1.05, 95 %CI (−1.40, −0.70), p < 0.01], with moderate heterogeneity (I² = 57 %, p < 0.01). The funnel plot (Fig. 10B) and Egger's test (t = −1.44, P = 0.17) indicated no publication bias.

3.5.3.2. Immunoglobulins

For the studies that investigated immunoglobulin concentrations of participants, one study conducted aerobic gymnastics of moderate intensity,⁴⁶ and two studies conducted running and cycling exercises of high intensity.^21,43 The intervention duration ranged from 8 weeks to 16 weeks, 3 sessions per week, and 30–60 min per session. One publication suggested that long-term aerobic exercises could significantly decrease the immunoglobulin concentration compared with the control group in healthy participants,⁴³ while two studies report no such effect^21,46 (Table 1).

The results (Fig. 9) showed that the ES of long-term intervention was significant and moderate in IgA concentration [SMD = −0.52, 95 %CI (−0.85, −0.19), p < 0.01], and the ES was non-significant and small in IgG concentration [SMD = −0.18, 95 %CI (−0.50, 0.15), p = 0.28] and IgM concentration [SMD = 0.21, 95 %CI (−0.12, 0.53), p = 0.21], with trivial heterogeneity (I² = 5 %, p = 0.39). The funnel plot (Fig. 10A) and Egger's test (t = −0.32, P = 0.76) indicated no publication bias.

Fig. 9.

Results regarding the effects of aerobic exercise on immunoglobulins in healthy participants.

While our analysis pooled data across multiple disease conditions, sensitivity analyses using the leave-one-disease-out method confirmed the robustness of the primary findings (Supplementary Material, Fig. S1). The effect sizes and significance remained stable when each disease subgroup was sequentially excluded, suggesting that no single disease or study disproportionately influenced the overall results.

4. Discussion

This is the first systematic review and meta-analysis to explore the effects of aerobic exercise on the innate and adaptive immune component quantities in healthy individuals and those with diseases. The findings indicate that long-term aerobic exercise improves T-cell counts in diseased patients but has no effect on healthy participants. It also influences IgA concentrations in healthy populations. However, long-term aerobic exercise does not influence the quantities of leukocytes, lymphocytes, neutrophils, monocytes, NK cells, IgG, or IgM in healthy individuals and those with diseases. Cell counts of leukocytes, lymphocytes, CD3⁺ cells, and NK cells were also not influenced by single-pass aerobic exercise intervention in healthy participants.

To investigate the broad immunomodulatory effects of aerobic exercise, we initially categorized diverse chronic conditions under a unified 'diseased' classification. This grouping strategy was justified by two key considerations: (1) all included conditions exhibit chronic immune dysregulation (either inflammatory or immunosuppressive in nature), and (2) aerobic exercise has established immune-modulating effects in these populations. We deliberately excluded acute infections and primary immunodeficiencies to reduce potential confounding factors. Our binary classification allowed us to detect overarching exercise effects applicable to clinical populations with chronic immune dysregulation—a pragmatic first step for future mechanistic studies. However, the pooling of diverse clinical conditions with distinct immune pathologies may obscure disease-specific responses to aerobic exercise, warranting cautious interpretation of the aggregated effects.

The primary findings here are that long-term aerobic exercise can significantly influence the immune markers of adaptive immune system (e.g., T-cells counts and IgA concentration) rather than the innate immune system (e.g., neutrophils, monocytes, and NK cell counts). Additionally, aerobic exercise more readily influences individuals with an unstable immune system and components sensitive to environmental changes. Specifically, individuals with diseases are more likely to be influenced by aerobic exercise than healthy individuals.^{21,25,43,45,46} In healthy individuals, the immune system maintains a state of homeostasis, where immune components and antibody levels remain relatively constant in the absence of pathogenic attacks or significant physiological stressors.^50,51 Aerobic exercise may not sufficiently disrupt this balance to cause noticeable changes in immune parameters.⁵² Additionally, healthy bodies have complex physiological regulatory mechanisms that quickly return immune markers to normal levels after physical activities,^52,53 preventing overreaction by regulating the production and release of immune components. Thus, without additional immune stimuli, the impact of aerobic exercise on immune components in healthy individuals may remain minimal. Furthermore, compared to other immunoglobulin (e.g., IgG and IgM), IgA is more susceptible to the effects of aerobic exercise. IgA, primarily located on mucosal surfaces such as the respiratory and gastrointestinal tracts, plays a crucial role as the first line of defense against pathogens.^46,54,55 Due to its sensitivity to environmental changes, IgA levels are influenced even in healthy individuals with stable immune systems.

T-cells are central to the adaptive immune system, playing a crucial role in mediating immune responses, including recognizing and combating infections and cancer cells, and regulating immune system activities.^56,57 Improved regulation of T-cells (including CD3⁺, CD4⁺, and CD8⁺) in diseased individuals is a direct result of immune activation in response to exercise.58, 59, 60 Unlike healthy individuals whose immune systems maintain homeostasis, diseased individuals often exhibit inactive or overactive immune responses.⁶¹ Aerobic exercise acts as a corrective force that enhances immune function to a more optimal state, thereby providing temporary immunomodulatory benefits to individuals with diseases.^{9,18,22,27,62} This normalization of immune function is particularly advantageous for these individuals. This finding aligns with prior research.58, 59, 60

Our meta-analysis of RCTs demonstrates that long-term aerobic exercise significantly reduces IgA levels compared to non-intervention controls. This result is inconsistent with several previous studies that utilized self-controlled designs and reported no change or an increase in IgA levels following long-term aerobic interventions, potentially confounding the influence of time effects.^54,63,64 By incorporating evidence from randomized controlled trials,^21,43,46 our findings suggest that IgA levels are lower after long-term aerobic intervention than no intervention, which may be a more realistic reflection of the effect of long-term aerobic exercise, suggesting that aerobic exercise might counteract the natural temporal increase in IgA levels observed in healthy participants.⁶⁵ This may be attributed to several physiological mechanisms. Aerobic exercises might directly influence the mucosal and pulmonary environment by increasing respiratory rate and depth during exercises, possibly leading to rapid mobilization and depletion of IgA in these areas, thereby reducing IgA levels.^55,66 Additionally, there might be an increase in the circulation of catecholamines and glucocorticoids during aerobic exercise, which are known to potentially suppress certain immune functions, including the production and secretion of IgA.67, 68, 69 This aligns with the "open window" hypothesis, where prolonged or intense exercise may transiently suppress mucosal immunity, increasing susceptibility to infections in the hours following exercise.^70,71 While our findings suggest a chronic rather than acute reduction in IgA, the underlying mechanisms (e.g., cortisol-mediated suppression) may share similarities with this model.⁷² The decreased IgA levels might represent a chronic adaptation state to optimize for regular stress and physical activity, resulting in a lower but more efficient level of this antibody under non-stressful conditions, which balances the energy and resources between immune surveillance and physical activity demands. However, an alternative interpretation is that sustained aerobic exercise could exert a mild immunosuppressive effect, particularly if recovery periods are insufficient to restore baseline immune function. Future studies are needed to further clarify the effect of aerobic exercise on IgA concentration as only three studies were included in the analysis.^21,43,46

Consistent with previous evidence, we found that long-term aerobic exercise does not influence the cell counts of innate immune system (e.g., neutrophils, monocytes, and NK cell counts). Long-term aerobic exercise does not impact the cell counts of innate immune cells.^73,74 This is largely due to genetic predispositions and intrinsic regulatory mechanisms that maintain immune system homeostasis. Genetic predispositions establish baseline levels for each type of innate immune cell, largely unaltered by external factors like exercise.⁷⁵ The body uses regulatory processes to stabilize these cell counts, including balancing cell proliferation and apoptosis, and redistributing cells between circulating blood and lymphoid tissues.⁷¹ Exercise can transiently increase the mobilization of these cells into the bloodstream, driven by stress hormones and increased blood flow.^76,77 However, this is usually followed by a quick return to baseline due to homeostatic adjustments like increased recruitment to lymphoid organs and higher apoptosis rates.^71,78,79 Although cell counts remain stable, aerobic exercise is known to significantly enhance their functional activity, such as phagocytosis, cytotoxic responses, and antigen presentation, improving immune surveillance and response to pathogens and malignancies.^73,74,76,80

Additionally, our aggregation of results from different measurement points (e.g., immediately after exercise or 2 h post-exercise) indicates that single-pass aerobic interventions do not influence the immune cell counts of leukocytes, lymphocytes, CD3⁺ cells, and NK cells. This fails to obtain a consistent conclusion with previous studies which reported that single-pass aerobic exercise may transiently increase the cell counts of these immune cells, and then progressively reduce them after exercise, until they return to baseline levels.^78,79 Single-pass aerobic exercise has been proven to increase immune cell counts, with the magnitude of the increase being intensity-dependent.^78,81 One of the studies we included involved Qigong as the intervention, which is typically classified as low to moderate intensity and may not be sufficient to induce significant changes.⁴⁵ Additionally, the limited number of studies included in our analysis necessitated the aggregation of cell counts with different measurement points (e.g., immediately after exercise or 2 h post-exercise), introducing variability that complicates the interpretation of our results.^26,44,45 As exercise-induced elevations in cell counts typically return to baseline levels over time, this methodological variability might explain the inconsistency of our findings with previous studies.^78,79

Several limitations should be acknowledged in our study. First, our analysis was limited to quantitative measurements of immune components (cell counts and immunoglobulin concentrations). Importantly, this approach did not assess functional immune parameters such as cytokine production, cytotoxic activity, or proliferative responses, which may better reflect actual immune competence. While our findings provide valuable insights into exercise-induced changes in immune component quantities, they cannot fully characterize functional immune adaptations. Future studies should incorporate both quantitative and functional assessments to obtain a more comprehensive understanding of how aerobic exercise influences immune system function. Additionally, due to the limited number of studies, subgroup analyses to characterize the influences of protocol settings of aerobic exercises (e.g., intervention length, intervention intensity) and participants’ characteristics (e.g., sex or age) on immune system function cannot be completed. Third, due to the limited availability of studies (only one per disease condition included in our analysis), we were unable to perform subgroup analyses to examine disease-specific effects of aerobic exercise. Our findings reflect pooled effects across various chronic conditions rather than disease-specific responses. Future research should build upon these general findings to systematically investigate how aerobic exercise differentially modulates immune function in specific disease contexts. This more granular approach would better inform clinical applications for distinct patient populations with chronic immune dysregulation. Fourth, our meta-analysis incorporated outcome data measured across heterogeneous post-exercise time points (ranging from immediately to 5 h post-exercise). Although this methodological approach strengthened our statistical power and enabled comprehensive evidence integration, it potentially masks critical temporal dynamics in immune responses. Regrettably, insufficient studies reporting standardized time intervals prevented meaningful time-stratified subgroup analyses. Readers should therefore exercise caution when extrapolating these findings to specific post-exercise time windows. Fifth, our study failed to investigate the effects of single-pass aerobic exercise on the immune component quantities and the effects of chronic aerobic exercise on immunoglobulin (especially IgA) in patients due to the insufficient number of existing studies with RCT designs. Studies with more rigorous designs are thus needed to explicitly explore these critical insights into the effects of aerobic exercise on immune system function. Nevertheless, the knowledge obtained from this work will help inform the appropriate design of intervention protocols of aerobic exercise for immune system function in the near future. Last, the inclusion of both sedentary and activity-unspecified control groups may introduce variability, though all groups shared the key criterion of receiving no exercise intervention.

5. Conclusion

Chronic aerobic exercise selectively enhances T-cell counts in immunocompromised individuals while reducing IgA levels in healthy populations, without affecting innate immunity or other immunoglobulins (IgG/IgM). In contrast, acute exercise demonstrates no measurable immunomodulatory effects. These findings highlight the potential of sustained aerobic training (current evidence primarily from running and cycling interventions) to stabilize immune function in clinical populations with dysregulation, though current evidence lacks dose-response data to guide specific exercise prescriptions. Future research should establish standardized protocols with comprehensive immune monitoring to facilitate clinical translation.

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Availability of data and material

The datasets used and analysed during the current study are available from the corresponding author on reasonable request.

Authors’ contributions

DB and QG conceived the idea for the article. GD and XH performed the literature search, data extraction, and data analysis. GD and XH wrote the first draft. The revisions were made by GD and XH with critical input from JH, DB, and JZ. All the figures and tables were made by XH and JH. All authors read and approved the final manuscript.

Funding

No sources of funding were used to assist in the preparation of this article.

Declaration of interest

GD, XH, JH, DB, QG, and JZ declare that they have no competing interests.

Abbreviations

AG

Aerobic exercise group

CG

Control group

VO₂max

Maximal oxygen uptake

VO2peak

Peak oxygen uptake

HR

Heart rate

HRR

Heart rate reserve

Leu

Leukocyte

Neu

Neutrophil

Mon

Monocyte

Lym

Lymphocyte

IgA

Immunoglobulin A

IgG

Immunoglobulin G

IgM

Immunoglobulin M

NK

Natural killer

SD

Standard Deviation

SE

Standard Error

SEM

Standard Error of Mean

SMD

Standard Mean difference

CI

Confidence interval

Appendix A. Supplementary data

The following is the Supplementary data to this article: