Comparative effectiveness of oral nutritional supplements in preventing respiratory tract infections among adults: a systematic review and network meta-analysis

Summary

Background

Different nutritional supplements may prevent respiratory tract infections (RTIs) in adults, but their comparative effectiveness remains unclear. We aimed to evaluate the effectiveness of oral nutritional supplements in preventing RTIs and reducing their symptom duration and severity.

Methods

In this systematic review and network meta-analysis, we searched for randomized controlled trials (RCTs) assessing oral micronutrients, flavonoids, probiotics, or synbiotics in adults for the prevention of RTIs in PubMed, Embase, Cochrane Central Register, and ClinicalTrials.gov from inception to February 25, 2025. Studies with participants under 18 years, immunodeficiencies, non-oral administration, or a therapeutic rather than preventive focus were excluded. We performed network meta-analyses (NMA) using frequentist random-effects models to compare interventions through direct and indirect evidence. We used the CINeMA framework to rate the overall certainty of evidence. The primary outcome was the incidence of RTIs. The secondary outcomes were upper respiratory tract infections (URTIs) or common cold, COVID-19 or influenza, RTI symptom duration, RTI symptom severity, and adverse events. The study is registered with PROSPERO, CRD420250653276.

Findings

We identified 120 eligible trials, among which 107 trials involving 101,751 adults were included in the network meta-analysis. Compared with placebo, catechin (RR = 0.79, 95% CI: 0.66, 0.95; n = 5; high certainty), Bifidobacterium animalis (RR = 0.79, 95% CI: 0.63, 0.99; n = 3; moderate certainty), and multi-strain probiotics (RR = 0.90, 95% CI: 0.82, 0.98; n = 16; moderate certainty) were the most effective interventions for reducing RTI incidence. For COVID-19 or influenza prevention, high-dose vitamin D was highly effective (RR = 0.66, 95% CI: 0.51, 0.86; n = 9; moderate certainty). Catechin (MD = −2.64 days/RTI, 95% CI: −4.92, −0.35; n = 2; moderate certainty) and multi-strain probiotics (MD = −0.97 days/RTI, 95% CI: −1.78, −0.16; n = 4; high certainty) most effectively shortened RTI symptom duration. Multi-strain probiotics (SMD = −0.33, 95% CI: −0.51, −0.14; n = 6; moderate certainty) also showed superior performance in alleviating RTI symptom severity. None of the interventions increased the risk of adverse events versus placebo.

Interpretation

Based on the current evidence, catechin, B. animalis, and multi-strain probiotics show relatively better effects in preventing adult RTI compared to other nutritional supplements. For COVID-19 or influenza prevention, high-dose vitamin D supplementation may be more effective. This study is limited by the lack of head-to-head comparisons between interventions (particularly across major categories) and heterogeneity from varied administration methods (e.g., capsules, beverages, mouthwashes), impacting the consistency and overall estimation of the NMA. Future high-quality studies are needed to provide direct evidence on the optimal intervention type, frequency, and form/dose of administration to guide clinical practice effectively.

Funding

This work was supported by the Ningbo Top Medical and Health Research Program and the Zhejiang Provincial Disease Prevention and Control Science and Technology Plan.

Research in context.

Evidence before this study

We conducted a preliminary search of PubMed, Embase, Cochrane Library, and Web of Science databases, reviewing evidence from their inception to February 17, 2025 on the role of nutritional supplements in preventing respiratory tract infections (RTIs) among adults, with no language restrictions. Search terms included combinations of (flavonoid∗ OR flavonol OR probiotic∗ OR synbiotic∗ OR “multiple micronutrient” OR “micronutrient supplementation” OR “vitamin D” OR cholecalciferol OR “25-hydroxyvitamin D” OR “1,25-dihydroxyvitamin D” OR “vitamin C” OR “ascorbic acid” OR “vitamin A” OR retinol OR retinoids OR “beta-carotene” OR carotenoids OR “vitamin E” OR zinc) AND (“respiratory tract infection” OR “respiratory infection” OR “common cold” OR influenza OR pneumonia OR COVID-19 OR SARS-CoV-2) AND (adult∗) AND (“randomized controlled trial” OR RCT OR “clinical trial” OR “controlled trial” OR randomized). Eligible participants were adults aged 18 or older, taking oral nutritional supplements, and free of immune-related diseases. Approximately 20 systematic reviews and meta-analyses were identified, all using traditional pairwise meta-analysis with a narrow focus on specific supplement types. No network meta-analysis (NMA) was found that systematically compared the full range of nutritional interventions for RTI prevention.

Added value of this study

Our systematic review and NMA comprehensively assessed the efficacy and safety of oral nutritional supplements, including low-, moderate-, and high-dose vitamin C, standard- and high-dose vitamin D, vitamin E, zinc, flavonoids, probiotics, and synbiotics, for preventing RTIs in adults. Incorporating 107 randomized controlled trials (RCTs) with 101,751 participants, this study provides a substantially larger evidence base than previous reviews. We utilized the CINeMA framework to evaluate the confidence in NMA effect estimates and presented clear, intuitive rankings of intervention performance across outcomes. Additional analyses included subgroup comparisons by age (<65 vs ≥65 years) and meta-regression and sensitivity analyses considering factors such as participant age, sex, follow-up duration, outcome measurement methods, high-intensity training status, and risk of bias. Our results indicate that catechin, Bifidobacterium animalis, and multi-strain probiotics are among the most effective for reducing RTI incidence, while high-dose vitamin D shows potential for preventing COVID-19 and influenza.

Implications of all the available evidence

This review indicates that certain nutritional supplements are effective in preventing RTIs among adults. This study is limited by the absence of head-to-head comparisons between interventions, particularly across major categories, and by heterogeneity arising from diverse administration methods, which affects the consistency and overall estimation of the NMA. Future high-quality studies are needed to provide direct evidence on the optimal intervention type, frequency, and administration form or dose to effectively guide clinical practice.

Introduction

Respiratory tract infections (RTIs) are defined as diseases of the respiratory system causing infections of the sinuses, throat, airways, and lungs, categorized into upper respiratory tract infections (URTIs) and lower respiratory tract infections (LRTIs). The incidence of RTIs has been increasing annually, imposing a significant disease burden (URTI: 225,505.7 cases per 100,000 population per year [201,156.4–253,739.5]; LRTI: 4770 cases per 100,000 population per year [4510–5040]).^1,2

Interest in the potential for nutrient supplementation to reduce the risk of RTIs has increased since the emergence of the COVID-19 pandemic³ and the resurgence of seasonal influenza globally.⁴ Nutrient supplementation has long been considered a potentially effective strategy for preventing RTIs.⁵ Supplementing with micronutrients (including vitamins A, C, D, E, zinc, selenium, iron, etc.) and plant-based antioxidants (such as flavonoids and polyphenols) can reduce the risk of RTI by activating the immune system, reducing oxidative stress, providing anti-inflammatory effects, and maintaining respiratory tract barrier function.^3,6 Probiotics and synbiotics can mediate the balance and diversity of the gut microbiota, which plays a significant role in the host's response to many viral infections, either by promoting the attachment of viruses to host cells or by supporting antiviral immunity (such as interferons and innate immune activation).^6,7

In recent years, many systematic reviews and meta-analyses of randomized controlled trials (RCTs) have explored the effectiveness of oral nutrient supplements in preventing RTIs,^4,8, 9, 10, 11, 12, 13, 14, 15, 16 suggesting beneficial effects of flavonoids and probiotics, while the effects of vitamins A, C, D, E, and zinc remain controversial. No review has simultaneously examined the comparative effects of all these interventions. Although primarily affecting children under 5 years of age, RTIs continue to have high incidence and mortality rates in adults, particularly among middle-aged and elderly individuals over 50 years old.^1,2 However, fewer systematic reviews have focused on the adult population.^4,12 Additionally, there are still limited meta-analyses on the role of nutrient supplementation in preventing COVID-19 and influenza.17, 18, 19, 20 Therefore, we conducted a systematic review and network meta-analysis (NMA) of RCTs to investigate the effectiveness and safety of these interventions in reducing the incidence and severity of RTIs in adults.

Methods

Search strategy and selection criteria

We conducted a systematic review and NMA following the Preferred Reporting Items for Systematic Reviews-Network Meta-Analyses (PRISMA-NMA) guideline.²¹ The study protocol was registered on PROSPERO (CRD420250653276). The changes made in this study compared to the PROSPERO registration along with the reasons are detailed in Table S1.

Electronic databases, including PubMed, Embase, Cochrane Central Register, and ClinicalTrials.gov, were searched from their inception to February 25, 2025 without language restriction. Reference lists of included original studies, reviews, and meta-analyses were also examined to identify additional relevant studies. The detailed search strategy is provided in eMethods 1 in the Supplement. Following the initial search, five researchers (Z.Z., X.Z., Y.C., B.Z., and Y.C.) independently screened titles and abstracts in duplicate. Subsequently, three researchers (Z.Z., X.Z., and Y.C.) performed full-text screening, assessing eligibility based on the Population, Intervention, Comparison, Outcome, and Study Type (PICOT) criteria.

We included randomized clinical trials that investigated the preventive effects of oral micronutrients, flavonoids, probiotics, or synbiotics on the incidence, duration, or severity of respiratory tract infections in adults aged ≥18 years. Micronutrient supplements encompassed vitamin A (including alpha- and beta-carotenes), vitamin C (classified into low-dose VC: average daily dose ≤500 mg, moderate-dose VC: average daily dose 501–1000 mg and high-dose VC: average daily dose >1000 mg), vitamin D (classified into standard-dose VD: average daily dose <2000 IU, and high-dose VD: average daily dose ≥2000 IU), vitamin E (including alpha- and beta-tocopherols), and zinc. The grouping of VC and VD was primarily based on previous researches^4,10 and data availability. Flavonoid supplements were defined as oral interventions with flavonoids including catechins and quercetins as the primary biologically active components. Probiotics were classified at the species level into single probiotics such as Lactobacillus casei (L. casei), Lactobacillus rhamnosus (L. rhamnosus), Lactobacillus helveticus (L. helveticus), and Bifidobacterium animalis (B. animalis), and multi-strain probiotics (combinations involving multiple strains of Lactobacillus ssp., Bifidobacterium ssp., Streptococcus ssp., Enterococcus ssp., and/or Saccharomyces boulardii).^12,22 Synbiotics were defined as combinations of probiotics (single or multiple strains) and prebiotics (nondigestible carbohydrates fermented by gut microbiota).

Exclusion criteria included studies with incomplete data (lacking information on the age of the included population, specific types of interventions, number of cases, duration of symptoms, or symptom severity scores), reviews, commentaries, editorials, animal or in vitro studies, and observational studies; Studies with a mean or median participant age <18 years (due to differences in immune function, lifestyle, and the types and dosages of nutritional supplements applicable to children and adolescents compared to adults); Non-oral administrations such as nasal sprays, injections, and intestinal colonization; Studies investigating therapeutic effects rather than preventive effects; or cohorts of hospitalized patients, individuals with immunodeficiencies, autoimmune or other immune-related disorders, or those receiving immunomodulatory therapy.

Data analysis

Data extraction and risk-of-bias assessments were independently conducted by two reviewers (Z.Z. and X.Z.). Disagreements were resolved through group discussion. A standardized Excel template was used to extract the following: author, year, country, population characteristics, sample size, follow-up duration, sex, mean age, interventions, outcomes, and outcome ascertainment. In cases of overlapping data, the study with the most comprehensive dataset was selected.

Risk of bias was evaluated using the Cochrane Risk of Bias 2 tool for randomized trials.²³ Small-study effects were assessed using comparison-adjusted funnel plots in the network, along with Egger's and Begg's tests for statistical significance.²⁴ If publication bias is identified, we further correct pairwise comparisons with over 10 studies using the trim-and-fill method. The certainty of evidence was evaluated using the CINeMA (Confidence in Network Meta-Analysis) framework, which adapts the GRADE approach for network meta-analyses, addressing within-study bias, reporting bias, indirectness, imprecision, heterogeneity, and inconsistency (see eMethods 2 in the Supplement for details).^25,26

The primary outcome was the overall incidence of RTIs, defined as the ratio of individuals reporting one or more RTIs to the total number of participants, including URTI, LRTI, common cold, or influenza. Data were collected regardless of whether the diagnosis was based on self-report, clinical assessment, or laboratory confirmation.

Secondary outcomes were categorized as follows: (1) URTI/common cold included cases specifically described in the original studies as “URTI” or “common cold.” Studies reported with a non-specific term such as “RTI” without further details were not included in this subgroup. (2) COVID-19/influenza included cases specifically identified as “COVID-19” (e.g., confirmed by PCR or antigen test), “influenza” (e.g., laboratory-confirmed or clinically diagnosed), or “influenza-like illness (ILI)”. ILI was defined according to the criteria provided in the original study, typically involving fever plus cough or sore throat. (3) Symptom duration was measured as the mean duration of symptoms per RTI episode. (4) Symptom severity was assessed by the mean severity score per RTI episode. (5) Adverse events were defined as any mild-to-severe intervention-related adverse reactions.

To obtain robust estimates, we included in the NMA only interventions supported by at least three studies or data points. We calculated risk ratios (RRs) with 95% confidence intervals (CIs) for the incidence of RTI, and odds ratios (ORs) were used for adverse events. For symptom duration and severity, mean differences (MDs) and standardized mean differences (SMDs) with Hedges' g correction for small studies and corresponding 95% CIs were used, respectively. All pooled estimates were calculated based on intention to treat analysis. Primary and secondary outcomes were analyzed using frequentist NMAs based on a graph-theoretical approach, implemented with the netmeta package in R version 4.3.3 (R Project for Statistical Computing).

Ranking probabilities were estimated using the surface under the cumulative ranking curve (SUCRA) and rankograms.²⁷ Heterogeneity was quantified with within-design and between-design Q statistics and I². The transitivity assumption was evaluated by comparing key study characteristics and assessing local consistency (net splitting of direct evidence from head-to-head studies vs indirect evidence inferred through common comparators) and global consistency (design-by-treatment interaction test).^28,29

Subgroup and sensitivity analyses were performed for the primary outcome to explore differences in RTI risk between adults aged <65 years and those ≥65 years, as well as changes in effect sizes after excluding high risk-of-bias studies, those with a “no treatment” control group, or high-intensity-trained population (including athletes and military recruits). Meta-regression analyses were conducted within a Bayesian framework, examining the influence of confounders such as participant age (<65 vs ≥65 years), proportion of females, follow-up duration (weeks), self-reported RTI (yes vs no), high-intensity-trained population (yes vs no) and risk of bias of studies. These models utilized 50,000 burn-in iterations and 100,000 additional simulations with four chains of different initial values via the Markov Chain Monte Carlo method, obtaining medians and 95% credible intervals (CrIs). The Deviance Information Criterion (DIC) was used to complement the assessment of the model differences. Additionally, to investigate the nonlinear dose-response relationship between vitamin D and C supplementation and RTI risk, a one-stage, nonlinear random-effects dose-response meta-analysis was performed using the dosresmet package, with knots placed at the 10th, 50th, and 90th percentiles of the exposure distribution.³⁰

Role of the funding source

The funders of this study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. All authors had full access to the data in the study and the corresponding author (Yi Chen) had final responsibility for the decision to submit for publication.

Results

An initial search yielded 8696 citations, which required abstract screening. Subsequently, 543 reports underwent full-text screening, and 120 clinical trials31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60^,61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90^,91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120^,121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150 were included in the systematic review. After excluding 13 trials for reasons (Table S3), 107 involving 101,751 participants were included in the final NMA (Fig. 1). These trials included 36 on vitamin D, 22 on vitamin C, 4 on zinc, 10 on flavonoids (6 catechin and 4 quercetin), 40 on probiotics (24 single probiotic and 16 multi-strain probiotics), and 4 on synbiotics. Six studies had three arms, two studies had four arms, and the remaining studies were all two-arm RCTs. The median (IQR) of the duration of follow-up was 12 (8–25.7) weeks (Table S2).

Fig. 1.

Flowchart according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guideline.

In the 120 RCTs, 50 had a low risk of bias, 58 had some concerns, and 12 had a high risk of bias. Detailed information on the risk of bias assessment is provided in Table S4 of the Supplement. The certainty of the evidence for the relevant outcomes was classified as high, moderate, low, or very low, and the interventions were categorized as among the best (statistically significant difference vs placebo and >1 other treatment), among the intermediate (statistically significant difference vs placebo), and among the worst (no statistically significant difference vs placebo).

RTI incidence

Of the 107 RCTs, 99 with 99,583 adults involving 15 preventive nutritional interventions addressed the primary outcome of RTI incidence (Fig. 2A). Of the 22 direct comparisons, 16 involved 2 studies or more. Comparisons of L. rhamnosus, multi-strain probiotics, moderate-dose VC, high-dose VD, and standard-dose VD, each versus Placebo, revealed significant heterogeneity (I² > 50%). The three comparisons between high-dose VD, standard-dose VD, and placebo also exhibited local inconsistency (local P-inconsistency < 0.05). Global inconsistency was insignificant (P = 0.304) (Table S10 in Supplement).

Fig. 2.

Network of eligible comparisons. (A) RTI Incidence. (B) URTI/Common Cold. (C) COVID-19/Influenza. (D) Duration of RTI symptoms. (E) Severity of RTI symptoms. (F) Adverse events. The size of the nodes corresponds to the number of adults randomly assigned to that intervention. Interventions directly compared are connected by a line; the thickness of the line represents the number of trials evaluating that comparison. B. animalis: Bifidobacterium animalis; L. casei: Lactobacillus casei; L. rhamnosus: Lactobacillus rhamnosus; L. helveticus: Lactobacillus helveticus; RTI: respiratory tract infection; URTI: upper respiratory tract infection; PRO: probiotics; SYN: synbiotics; MultiProb: multiple-strain probiotics; VC-low: low-dose VC (average daily dose ≤ 500 mg); VC-moderate: moderate-dose VC (average daily dose of 501–1000 mg); VC-high: high-dose VC (average daily dose >1000 mg); VD-standard: standard-dose VD (average daily dose <2000 IU); VD-high: high-dose VD (average daily dose ≥2000 IU).

Fig. 3 summarizes the main findings for each outcome, presenting the network estimates for each intervention compared to placebo and their 95% confidence intervals. Compared with placebo, Catechin (RR = 0.79, 95% CI: 0.66, 0.95; SUCRA = 71.4; high certainty), B. animalis (RR = 0.79, 95% CI: 0.63, 0.99; SUCRA = 70.0; moderate certainty), and multi-strain probiotics (RR = 0.90, 95% CI: 0.82, 0.98; SUCRA = 51.3; moderate certainty) were among the most effective interventions in reducing RTI incidence. SYN (RR = 0.64, 95% CI: 0.55, 0.74; SUCRA = 92.6; low certainty) and zinc (RR = 0.69, 95% CI: 0.50, 0.94; SUCRA = 83.8; low certainty) may be among the most effective but lack certainty. High-dose VD (RR = 0.94, 95% CI: 0.89, 1.00; SUCRA = 35.5; moderate certainty) and Moderate-dose VC (RR = 0.93, 95% CI: 0.86, 1.00; SUCRA = 41.5; moderate certainty) were associated with intermediate effectiveness (Fig. 3, Fig. 4A, and Fig. 5). Fig. 4A shows the comparative effectiveness and certainty for all pairwise comparisons. Figure S1 shows the number of participants and effect size of each study. Table S5 provides specific details of the CINeMA certainty assessment for the RTI incidence outcome.

Fig. 3.

Network meta-analysis results sorted based on the CINeMA (Confidence in Network Meta-Analysis) certainty of evidence and effect estimates for the comparisons of active interventions vs placebo. Bold numbers represent statistically significant results. B. animalis: Bifidobacterium animalis; L. casei: Lactobacillus casei; L. rhamnosus: Lactobacillus rhamnosus; L. helveticus: Lactobacillus helveticus; RTI: respiratory tract infection; RR: risk ratio; URTI: upper respiratory tract infection; SMD: standardized mean difference; PRO: probiotics; SYN: synbiotics; MultiProb: multiple-strain probiotics; VC-low: low-dose VC (average daily dose ≤ 500 mg); VC-moderate: moderate-dose VC (average daily dose of 501–1000 mg); VC-high: high-dose VC (average daily dose >1000 mg); VD-standard: standard-dose VD (average daily dose <2000 IU); VD-high: high-dose VD (average daily dose ≥2000 IU).

Fig. 4.

Network meta-analysis results for the primary and secondary outcomes. (A) RTI Incidence: The effect size of the network meta-analysis is shown in the lower triangle, and direct comparisons are shown in the upper triangle. (B) URTI/Common Cold: Lower triangle; COVID-19/Influenza: Upper triangle. (C) Symptom Duration: Lower triangle; Symptom Severity: Upper triangle. For each row and column comparison, a risk ratio less than 0 or 1 indicates that the intervention in the blue cell in the upper-left corner is more favorable. Bold numbers represent statistically significant results. The color of each result cell represents the certainty of the evidence in the pairwise comparison. B. animalis: Bifidobacterium animalis; L. casei: Lactobacillus casei; L. rhamnosus: Lactobacillus rhamnosus; L. helveticus: Lactobacillus helveticus; RTI: respiratory tract infection; RR: risk ratio; URTI: upper respiratory tract infection; SMD: standardized mean difference; SYN: synbiotics; MultiProb: multiple-strain probiotics; VC-low: low-dose VC (average daily dose ≤ 500 mg); VC-moderate: moderate-dose VC (average daily dose of 501–1000 mg); VC-high: high-dose VC (average daily dose >1000 mg); VD-standard: standard-dose VD (average daily dose <2000 IU); VD-high: high-dose VD (average daily dose ≥2000 IU).

Fig. 5.

Heatmap of SUCRA for nutrient supplementation and RTI outcomes. SUCRA values rank nutrients on a continuous scale from 0 to 1. The higher the value, the better the preventive effect. B. animalis: Bifidobacterium animalis; L. casei: Lactobacillus casei; L. rhamnosus: Lactobacillus rhamnosus; L. helveticus: Lactobacillus helveticus; SYN: synbiotics; MultiProb: multiple-strain probiotics; VC-low: low-dose VC (average daily dose ≤ 500 mg); VC-moderate: moderate-dose VC (average daily dose of 501–1000 mg); VC-high: high-dose VC (average daily dose >1000 mg); VD-standard: standard-dose VD (average daily dose <2000 IU); VD-high: high-dose VD (average daily dose ≥2000 IU).

URTI/common cold incidence

75 studies reporting on the incidence of URTI or common cold enrolled 51,454 adults and evaluated 15 preventive interventions (Fig. 2B). Of the 19 available direct comparisons, the I² of 4 comparisons was more than 50%. Inconsistency was only detected in comparisons between high-dose VD, standard-dose VD, and placebo (Table S11 in Supplement). Global inconsistency was not significant (P = 0.069).

Catechin (RR = 0.80, 95% CI: 0.67, 0.96; SUCRA = 73.0; moderate certainty) proved of best effectiveness in preventing URTI or common cold vs placebo. Low-certainty evidence suggests that SYN (RR = 0.69, 95% CI: 0.56, 0.87; SUCRA = 88.3) and zinc (RR = 0.69, 95% CI: 0.51, 0.94; SUCRA = 86.6) may be associated with a reduction in URTI or common cold incidence compared to placebo (Fig. 3, Fig. 4B, and Fig. 5). B. animalis (RR = 0.80, 95% CI: 0.64, 0.99; SUCRA = 72.9; moderate certainty), multi-strain probiotics (RR = 0.90, 95% CI: 0.82, 0.98; SUCRA = 55.2; high certainty), and Moderate-dose VC (RR = 0.93, 95% CI: 0.87, 1.00; SUCRA = 45.4; moderate certainty) demonstrated intermediate effectiveness (Fig. 3, Fig. 4B, and Fig. 5). Fig. 4B shows the comparative effectiveness and certainty for all pairwise comparisons. Figure S4 shows the number of participants and effect size of each study. Table S6 provides specific details of the CINeMA certainty assessment for the URTI/common cold incidence outcome.

COVID-19/influenza

The 32 studies (52,601 adults) reporting COVID-19 or influenza involved 11 interventions with 12 direct comparisons, among which High-dose VD vs placebo showed substantial heterogeneity (I² > 50%) (Fig. 2C). No global or local inconsistency was found (Table S12 in Supplement). High-dose VD (vs placebo, RR = 0.66, 95% CI: 0.51, 0.86; SUCRA = 55.4; moderate certainty) proved of best effectiveness in reducing COVID-19 or influenza incidence. Synbiotics (vs placebo, RR = 0.40, 95% CI: 0.25, 0.63; SUCRA = 82.3; low certainty) may be among the best effectiveness. Catechin (vs placebo, RR = 0.61, 95% CI: 0.37, 1.00; SUCRA = 60.3; moderate certainty) was of intermediate effectiveness (Fig. 3, Fig. 4B, and Fig. 5). Figure S7 shows the number of participants and effect size of each study. Table S7 shows specific details of the CINeMA certainty assessment for the COVID-19/influenza incidence outcome.

Duration of RTI symptoms

We identified 41 studies (13,191 RTI episodes) that reported duration of RTI symptoms involving 13 interventions (Fig. 2D). Of the 16 direct comparisons, 3 had significant heterogeneity (I² > 50%), and no global or local inconsistency was observed (Table S13 in Supplement).

Compared with placebo, Catechin (MD = −2.64 days/RTI, 95% CI: −4.92, −0.35; SUCRA = 94.3; high certainty) and multi-strain probiotics (MD = −0.97 days/RTI, 95% CI: −1.78, −0.16; SUCRA = 69.6; moderate certainty) were considered the best and most reliable interventions in shortening the duration of RTI symptoms. Moderate-dose VC (MD = −0.68 days/RTI, 95% CI: −1.13, −0.25; SUCRA = 57.3; moderate certainty) demonstrated intermediate effectiveness. Low-certainty evidence suggested L. casei (MD = −0.89 days/RTI, 95% CI: −1.50, −0.28; SUCRA = 68) may help reduce the duration of RTI symptoms (Fig. 3, Fig. 4C, and Fig. 5). Figure S10 shows the number of participants and effect size of each study. Table S8 provides specific details of the CINeMA certainty assessment for the duration of RTI symptoms outcome.

Severity of RTI symptoms

Based on 31 studies reporting on the severity of RTI symptoms, involving 5802 RTI episodes, 11 interventions were investigated (Fig. 2E). Of the 12 direct comparisons, two showing substantial heterogeneity (I² > 50%). We found no statistical evidence of global or local inconsistency (Table S14 in Supplement).

Multi-strain probiotics (SMD = −0.33, 95% CI: −0.51, −0.14; SUCRA = 77.8; moderate certainty) were among the best effective intervention in alleviating the severity of RTI symptoms, compared with placebo. Synbiotics (SMD = −0.33, 95% CI: −0.56, −0.10; SUCRA = 78.6; low certainty) maybe effective (Fig. 3, Fig. 4C, and Fig. 5). Figure S13 shows the number of participants and effect size of each study. Table S9 shows specific details of the CINeMA certainty assessment for the duration of RTI symptoms outcome.

Adverse events

30 trials including 68,462 adults investigated adverse events, involving 11 interventions (Fig. 2F). None of the interventions increased the risk of adverse events versus placebo, indicating that all interventions had a good safety profile (Figure S12). However, safety evaluation data for VC was lacking.

Subgroup and sensitive analysis

We conducted a subgroup and sensitivity analysis for the primary outcome, RTI incidence. Figures S19–S21 presented the results for adults aged <65 years. The effect of high-dose VD improved further (RR = 0.69, 95% CI: 0.59, 0.80; SUCRA = 0.85), while the effects of B. animalis, zinc, and moderate-dose VC were non-significant. In adults aged ≥65, only multi-strain probiotics showed statistical significance, but only 7 interventions were available for comparison (Figures S22–S24).

When excluding studies with a high risk of bias, the effects of moderate-dose VC and zinc became statistically non-significant. When excluding studies involving high-intensity-trained population as the study population, the effect of moderate-dose VC also lost statistical significance. Other results did not change significantly (Figures S25–S33). Meta-regression analysis indicated that age, proportion of females, follow-up duration, self-reported RTI, high-intensity-trained population, and study risk of bias had no significant impact on the results (Table S15). Additionally, a nonlinear dose-response relationship between VC supplementation and RTI incidence was observed (Figures S34 and S35).

Publication bias

Comparison-adjusted funnel plots across outcomes showed significant evidence of small-study effect in RTI, URTI/common cold and COVID-19/influenza (Figures S36–S42).

Discussion

In this systematic review and NMA, the benefits of micronutrient supplementation, probiotics, synbiotics, and bioflavonoids in preventing the incidence of RTIs and alleviating RTI symptoms in adults were compared. We found moderate to high-certainty evidence suggesting that catechins were among the best measures for preventing RTIs and URTI/common cold as well as shortening symptom duration, high-dose vitamin D (average daily dose ≥2000 IU) may be the relatively optimal choice for reducing the incidence of COVID-19/influenza, and multi-strain probiotics were most effective in alleviating symptom severity. We did not observe any significant modification effect for age, proportion of females, follow-up duration, self-reported RTI, high-intensity-trained population, and risk of bias.

Flavonoids are diverse, and their specific effects cannot be generalized. A meta-analysis of 20 RCTs found flavonoid supplements reduced RTI incidence (RR = 0.81, 95% CI: 0.74, 0.89) and duration (Weighted mean difference = −0.56 days, 95% CI: −1.04, −0.08), but the purity and potency of the supplements were unclear, which may introduce heterogeneity and bias.¹⁴ Our NMA found that compared to other nutrient interventions, catechin supplementation was the best choice for preventing RTIs, while quercetin, another flavonoid, showed no effect. Catechins are a complex of flavanols, including epigallocatechin gallate (EGCg), which exhibits antiviral effects in vitro.¹⁵¹ Studies show that EGCg and EGC target viral RNA polymerase, inhibiting RNA synthesis and blocking viral proliferation.¹⁵² Recent docking simulation studies on SARS-CoV-2 also indicate that EGCg, EGC, and other catechins have significant binding affinity to the main protease of the virus.¹⁵³ In mice, oral administration of an EGC/EGCg mixture increases IgA production in the intestinal mucosa, promoting mucosal immunity.¹⁵⁴ A previous meta-analysis showed that the use of tea and catechins is effective in preventing influenza and URTIs (RR = 0.79, 95% CI: 0.66, 0.94), which is consistent with our study findings.⁸

Two meta-analyses found that VD supplementation reduced the incidence of COVID-19 and influenza,^17,18 but neither explored the dose-dependency of VD effects. The relationship between VD dosage and RTIs remains controversial.^{48,64,141,155} A recent meta-analysis showed that ≥2000 IU/day VD (RR = 0.85, 95% CI: 0.75, 0.96) was more effective than <2000 IU/day (RR = 0.91, 95% CI: 0.94, 0.98) in preventing adult RTIs.⁴ In this NMA, high-dose VD was found to have a good effect in preventing COVID-19 or influenza. High-dose vitamin D exerts multiple immunomodulatory effects on innate and adaptive immune cells, enhances the antibacterial activity of macrophages, and improves immune health and defense against infectious diseases.¹⁵⁶

We found that moderate-dose VC (501–1000 mg/day) were also effective in preventing RTIs. No dose-response relationship between VC supplementation and adult RTIs has been established. This study showed an inverted U-shaped relationship between VC dosage and RTI risk. Vitamin C plays a critical role in both innate and adaptive immunity, affecting epithelial barriers, supporting phagocytosis, and enhancing lymphocyte differentiation.¹⁵⁷ However, studies indicate that daily supplementation exceeding the tolerable upper intake level (2000 mg) in adults may cause adverse reactions such as diarrhea, headaches, gastrointestinal discomfort, and an increased risk of other diseases, impacting patients' quality of life and potentially reducing compliance.¹⁵⁸ Interestingly, our dose-response meta-analysis showed that the effect of vitamin C in preventing RTIs begins to weaken at doses greater than 1500 mg/day, which may be related to the above. A previous meta-analysis found that moderate-dose vitamin C significantly reduces adult RTIs (RR = 0.95, 95% CI: 0.92, 0.99),⁴ consistent with our findings. However, a Cochrane review suggested this effect is only significant in participants under severe physical stress (RR = 0.48, 95% CI: 0.35, 0.64), not the general population (RR = 0.97, 95% CI: 0.94, 1.00).¹⁵ Sensitivity analysis in our study showed that when excluding studies involving high-intensity-trained populations, the effect of moderate-dose vitamin C lost statistical significance. Therefore, further research is needed to establish the preventive effect of vitamin C on RTIs in the general adult population.

Previous meta-analyses suggested probiotics reduce RTI risk by 9%–47%^12,159, 160, 161 and shorten RTI duration by 0.8–2.7 days.^159,160,162 In this NMA, B. animalis and multi-strain probiotics reduced adult RTI risk by 20% and 10% respectively, and multi-strain probiotics and L. casei shortened RTI duration by 0.97 and 0.89 days respectively, similar to the recent research findings.¹² In single-strain studies, B. animalis demonstrated better efficacy compared to strains within the Lactobacillus genus. Research indicates that B. animalis subsp. lactis (e.g., BB-12 strain) significantly enhances natural killer (NK) cell activity, reduces inflammatory cytokines (e.g., IL-6, TNF-α), and boosts antiviral responses (e.g., production of IFN-γ and IL-12).¹⁶³ Additionally, B. animalis has high cell wall hydrophobicity, which facilitates better adhesion to the intestinal epithelium, resulting in superior gut colonization compared to L. casei.¹⁶⁴ The ability of B. animalis subsp. lactis to produce short-chain fatty acids is significantly greater than that of Lactobacillus salivarius, Lactiplantibacillus plantarum, and others, benefiting respiratory defense by enhancing gut barrier integrity, reducing systemic inflammation, and supporting immune cell function.¹⁶⁵

For multi-strain probiotics, studies indicate that different strains can enhance bioactivity through synergistic effects.¹⁶⁶ However, more strains are not necessarily better. A systematic review evaluated the preventive effects of single-strain probiotics versus multi-strain mixtures on various diseases (including URTI), and in most cases, single strains were comparable to mixtures.¹⁶⁷ Therefore, selecting appropriate probiotics should be based on evidence from efficacy trials to identify specific single or multi-strain probiotic combinations. The multi-strain probiotics in our NMA primarily involve the genera Lactobacillus ssp., Bifidobacterium ssp., Streptococcus ssp., providing some guidance. In fact, the results of this NMA show that B. animalis (RR = 0.79, 95% CI: 0.63, 0.99; SUCRA = 70.0) outperforms multi-strain probiotics (RR = 0.90, 95% CI: 0.82, 0.98; SUCRA = 51.3) in efficacy ranking.

Although synbiotics and zinc showed significant beneficial effects in our study, their evidence level was downgraded due to the limited number of studies and low evidence quality. Recent two meta-analyses found that zinc was ineffective for RTIs, but they both included experimental challenge studies, which may differ from infections under natural conditions.^4,16 Additionally, the use of VA, VE, and other types of single-strain probiotics could not be pooled due to insufficient data points. More high-quality related studies are necessary in the future.

To elucidate our key findings, catechins exhibit superior efficacy in preventing respiratory infections due to their multifaceted mechanisms: potent antioxidant and anti-inflammatory effects, direct inhibition of viral replication, and enhancement of immune responses and mucosal barriers.¹⁶⁸ Probiotics rank second, exerting broader systemic effects via the gut-lung axis,¹⁶⁹ surpassing the localized or single-target actions of VC and VD, though their effects are slower and indirect. VC supports antioxidant and immune functions but is limited by its singular target and weak antiviral activity.¹⁵⁷ VD modulates immune balance via the vitamin D receptor pathway, with restricted antiviral and anti-inflammatory efficacy.¹⁵⁶

Considering cost-benefit is necessary. Studies show probiotics reduce RTI burden, saving Can$1.3–8.9 million in healthcare and Can$61.2–99.7 million in productivity in Canada by averting 0.57–2.3 million RTI days.¹⁷⁰ In France, savings are €14.6–37.7 million, cutting 2.4–6.6 million RTI days.¹⁷¹ In the US, savings are $4.6–373 million for healthcare and up to $1.4 billion with productivity, reducing 19–54.5 million RTI days.¹⁷² This NMA and current evidence suggest therapeutic supplementation with VD,¹⁷³ VC,^4,174 and zinc¹⁶ may outperform preventive supplementation in reducing RTI duration or symptom severity, indicating a potentially more cost-effective approach for symptom relief. Evidence for catechins and probiotics in adult RTI treatment is limited, requiring further research.

Investigating optimal supplementation dosage and intervention duration is crucial for effective practical implementation. A dose-response relationship may exist for catechins,⁸ suggesting that supplementation can be appropriately increased within safe limits (<704 mg EGCG/day).¹⁷⁵ Results from one meta-analyses failed to show a relationship between probiotic efficacy and dose administered¹⁷⁶; selecting appropriate strains (e.g., B. animalis) is more critical. Moderate-dose vitamin C (average daily dose of 501–1000 mg) and high-dose vitamin D (average daily dose ≥2000 IU) may represent optimal choices. Additionally, the median preventive intervention duration of trials included in this network meta-analysis was 12 weeks, with winter being the peak season for respiratory infectious diseases, indicating that supplementation during winter may be the most effective timing.¹⁷⁷

Our review has the following strengths: the first comprehensive comparative assessment of all relevant nutritional interventions for reducing the incidence of adult RTIs; the analysis of different dosage groups for VC and VD, as well as different types of flavonoids and probiotics, addressed the current controversies; and the use of the CINeMA method to semi-objectively assess the certainty of NMA effect estimates, as well as providing a simple and user-friendly representation of the relative performance of each intervention for each outcome (Fig. 3).

The limitations of this review: First, apart from direct comparisons with the control group, there is a lack of head-to-head direct comparisons between other interventions (especially those involving different major categories of interventions), making it difficult to assess inconsistency in many comparisons. Second, differences in administration methods (such as capsules, beverages, and mouthwashes) may lead to heterogeneity, which to some extent affects the overall estimation of the NMA. Finally, for many interventions, only low or very low-quality evidence was available (Fig. 3), and potential publication bias may further reduce the quality of the evidence.

In conclusion, this systematic review and NMA find that moderate to high-quality evidence suggest catechin, B. animalis, and multi-strain probiotics are among the best measures for preventing RTIs. High-dose vitamin D is optimal in reducing the incidence of COVID-19 or influenza. Catechin and multi-strain probiotics are associated with a shorter symptom duration. Multi-strain probiotics are most effective in alleviating symptom severity. Future high-quality studies are needed to provide direct evidence on the optimal intervention type, frequency, and form/dose of administration to guide clinical practice effectively.

Contributors

YC and ZXZ designed the study. ZXZ, XXZ, YRC, BZ, and YC conducted the literature search and literature screening. ZXZ and XXZ extracted the data. ZXZ performed the data analysis and statistical interpretation. YC supervised the statistical methodology. ZXZ and XXZ wrote the first draft of the manuscript. All authors contributed to data interpretation and critical revision of the manuscript. All authors read and approved the final version of the manuscript. ZXZ, XXZ, and YC had access to and verified the underlying data.

Data sharing statement

The data generated in this study are available from the corresponding authors upon reasonable request, for research purposes only. The data will be available starting from the date of publication.

Declaration of interests

All authors declare no competing interests.