The eligibility criteria were developed based on the population, interventions, comparator, outcomes, and study design (PICOS) elements.

This review included epidemiologic studies of all populations, excluding studies investigating persons with dairy sensitivity. There were no restrictions on geographical location, sex, age, or health status.

We included studies of exposure or dietary intervention involving whole dairy products, dairy proteins, or other components of dairy. Whole dairy products of interest included cow’s milk, yogurt, and cheese (both standard products and those fermented with additional probiotics). Yogurts using the traditional starter cultures Lactobacillus (L.) bulgaricus and Streptococcus (S.) thermophilus were referred to as traditional yogurt, while probiotic yogurts were those with additional probiotics added. Dairy proteins included whey (soluble milk protein) and casein (insoluble milk protein). Other dairy components of relevance were the fat components of milk (i.e., milk phospholipids and the milk fat globular membrane). Milk powders, milk peptides/proteins, and dairy products fermented with experimental/ non-traditional bacterial strains were considered relevant. Studies assessing dietary patterns, including prenatal and maternal exposures, were included. Studies where dairy products or components were administered through a feeding tube were included. Studies of bovine colostrum, non-bovine milks, hyperimmunized milk, and raw/unpasteurized milk were excluded. Studies that administered probiotics alone or in a vehicle other than a dairy product were also excluded. Studies that did not calculate an effect estimate or conduct any statistical comparisons were excluded.

Studies were required to have comparison group(s) of low or no dairy product/component consumption or pre- and post-intervention outcomes.

The registered protocol for the systematic literature review specified all outcomes related to immune function, excluding biomarkers of inflammation (which have been reviewed previously 814) and outcomes related to milk allergies/sensitivities. In this publication, we evaluated the following outcomes reported in the included studies: 1) immunoglobulin levels (IgA, IgD, IgE, IgG, and/or IgM) in the absence of vaccination; and/or 2) immunological responses to vaccines, specifically vaccine-specific immunoglobulins, vaccine-specific antibody titers, seroprotection rates, and/or seroconversion rates. Antibody titers are used to assess immunogenicity of various vaccines (e.g., the hemagglutinin inhibition ^HI titer for influenza) 17. The seroprotection rate refers to the proportion of individuals reaching an established protective antibody titer level (e.g., 1:40 for influenza, which is associated with a 50% reduction in the risk of acquiring laboratory-confirmed influenza), while the seroconversion rate describes the proportion of patients that reach a predefined increase in the HI titer that indicates a response (e.g., fourfold for influenza) 18.

The publication start date was not restricted. This review included peer-reviewed publications with the following study designs: prospective or retrospective cohort studies, case-control studies, cross-sectional studies, and clinical trials. Reviews, meta-analyses, case series, and case reports were excluded. Conference abstracts and articles for which neither the abstract nor the full text were available in English were excluded.

If more than one article from the same study population were published, data from the publication with the longest follow-up or most relevant population and/or outcomes were evaluated. For studies with overlapping data, data from the publication with the larger population size or most relevant population and/or outcomes were considered.

The pre-determined literature search strategy was followed at all stages of the review. Searches were conducted in the PubMed and Embase databases on May 19, 2022, with the human and English language filters applied (Supplemental Table 1). Standard software to conduct systematic literature reviews, i.e., DistillerSR (Version 2023.4) 19, was used to deduplicate the literature search results from PubMed and Embase and to track the identified publications at each stage of review.

One reviewer examined the titles and abstracts of the deduplicated articles for inclusion based on the eligibility criteria. The articles considered to be relevant at the title and abstract level were independently evaluated at the full-text level by 2 reviewers; all conflicts were resolved by a senior reviewer. Following the PRISMA guidelines, bibliographies of relevant reviews were also assessed to identify any additional citations of interest meeting the PICOS elements.

In DistillerSR, data abstraction was conducted for all included studies. The following information was abstracted for each included study: study design; geographical location(s); study period; dairy product(s)/component(s) under evaluation; dosing and duration; population size; health status; age range; immunoglobulin levels measured (IgA, IgD, IgE, IgG, and/or IgM); and the vaccine administered and concurrent or subsequent effects on immunoglobulins, vaccine-specific antibody titers, seroprotection rates, and/or seroconversion rates. Any relevant effect estimates, confidence intervals, and statistical testing for these outcomes were abstracted.

One reviewer conducted data abstraction, while a second reviewer independently reviewed the entries for complete quality control. All conflicts were resolved by a senior reviewer.

Risk of bias (RoB) assessment was also conducted for all included studies. To assess study validity and quality, we used the Academy of Nutrition and Dietetics Quality Criteria Checklist 20. This tool examined several domains pertaining to relevance, validity, and bias due to funding source. RoB assessment was conducted by one reviewer; the results were reviewed independently by a second reviewer (100% quality control). A senior reviewer resolved any conflicts and finalized the RoB results. Study quality was determined as positive, neutral, or negative, depending upon the scoring results from the domains (Supplemental Table 2).

For each outcome, we systematically summarized the data by study quality, dairy exposure/intervention, and publication year. Qualitative synthesis was done, as meta-analysis could not be performed due to the heterogeneous nature of the dairy exposures and reported outcomes.

(Table 1) presents the characteristics of the 61 studies: 27 (44.3%) were determined to have positive study quality, 33 (54.1%) were determined to have neutral quality, while one study (1.6%) 21 was determined to have negative quality. Except for one cross-sectional survey 22 and one cohort study 23, the remaining 59 studies were clinical trials. Four studies provided vaccine-specific immunological response data only 24252627, 50 studies reported the effects of dairy on immunoglobulin levels without vaccinations only, and 6 presented results on both outcome types 282930313233. The period of study enrollment and follow-up was not reported in 32 studies; of those providing enrollment and follow-up data (n=29), the years ranged from 1983 to 2017. Fifty-nine studies reported geographical location, with 80% (n=47) conducted in European nations.

Ten clinical trials evaluated dairy interventions in conjunction with vaccination and their effect on vaccine-specific IgG, IgA, IgM, and antibody titers, seroprotection rates, and/or seroconversion rates 24252627282930313233 (Table 1). Among 5 trials specifying the study enrollment and follow-up period, the years ranged from 2005 to 2011 2425293031 (Table 1). The vaccines studied included diphtheria 273033, tetanus 273033, pertussis 2730, polio 263032, influenza 24252931, hepatitis B 33, Haemophilus influenzae type B 3033, flagellin from Salmonella adelaide28, and Streptococcus pneumoniae or pneumococcus3032, with studies often administering combination or multiple vaccines (Table 2). Four trials were conducted in hospitalized elderly patients 24252729, while 3 were in adults (healthy: N=2; allergy: N=1) 263132 and 3 were in infants and children (healthy: N=2; stunted growth: N=1) 283033 (Table 2).

In 3 trials, differences between treatment arms were not observed in any of the reported analyses 303233 (Table 2). The remaining 7 trials reported differences between the treatment arms or between pre- and post-intervention periods for at least one outcome. These 7 studies are described in the upcoming sections by the dairy product/component intervention 24252627282931 (Table 2).

Overall, the evidence base indicated that whole dairy products enhanced vaccine-specific immune response to tetanus and Salmonella Adelaide, while probiotics added to whole dairy products amplified vaccine-specific immune response to influenza and polio.

Two trials evaluated changes in vaccine-specific antibody titers following milk powder consumption 2728; study years were not reported in either trial (Table 1 and Table 2).

Elderly patients in retirement centers and long-term care facilities in the United States were given 6 g milk powder (n=10) or isoflavone soy protein (n=11) twice daily for 8 weeks and administered the diphtheria, tetanus, and pertussis (DTaP) vaccine at week 4 27 (Table 2). Post-vaccine tetanus antibodies were higher in the milk powder intervention compared with the soy protein group at week 8 (p=0.034). The mean change in the tetanus antibody level was also higher in the milk powder group (p=0.029).

Prepubertal children in New Guinea with growth deficiencies were given skim milk powder (n=30) or no intervention (n=24) for 8 months and administered flagellin (i.e., protein) from Salmonella adelaide at month 7 28 (Table 2). Total antibody titers at 6-weeks post-vaccination were higher in the skim milk powder group compared with the untreated group (p=0.002).

Five trials evaluated whether probiotics added to dairy products altered the immune response to vaccination, including 4 studies administering an influenza vaccine 24252931 and one administering the polio vaccine 26 (Table 2).

During the 2010-2011 season, enterally-fed elderly patients in Japan were given a milk-based formula with added prebiotics and probiotics (L. delbrueckiisubsp. bulgaricus and S. thermophilus)(n=12)or a standard milk-based formula (n=12)for 14 weeks, with H1N1/H3N2/B influenza vaccination at week 4 29 (Table 2). The antibody titer to the influenza B antigen was lower in the intervention group compared to the control group at weeks 6 and 8 (p<0.05).

During the 2006/2007 influenza vaccine campaign in Spain, study participants aged 65-85 years received trivalent influenza vaccines 25 (Table 2). Probiotic consumption was started 3-4 months after vaccination. Nineteen elderly patients were randomized to high-dose skim milk powder with L. plantarum CECT 7315/7316, 14 were randomized to low-dose skim milk powder with L. plantarum CECT 7315/7316, and 15 were randomized to skim milk powder without the probiotic; the milk powders were administered for 3 months. For each treatment arm, the investigators compared immunoglobulin levels in the post- vs. pre-intervention. An increase in influenza-specific IgG was observed only for the high-dose intervention arm, comparing the post- vs. pre-intervention levels (p=0.023). Influenza-specific IgA was increased in both the high- and low-dose intervention arms, comparing post- vs. pre-intervention levels (p=0.008 and p=0.039, respectively).

During the 2008-2009 influenza season, healthy adults in Italy were randomized to 4 intervention arms. Two treatment groups were relevant for this review with 56 receiving an acidified dairy drink containing L. paracaseissp. paracasei (L. casei 431) and 54 receiving a placebo acidified dairy drink for 6 weeks 31 (Table 2). The trial participants were also administered the A/H1N1/A/H3N2/B influenza vaccine 2 weeks after starting the dairy intervention. Change between post-intervention and baseline plasma levels of influenza vaccine-specific total IgG, IgG1 and IgG3 were higher in the intervention group compared with the placebo (p=0.01, p<0.01, and p<0.001, respectively). The plasma IgG1 and IgG3 seroconversion rates were higher in the intervention group, compared to the placebo (p<0.001 and p<0.001, respectively).

A trial was conducted in France during the 2005-2006 (pilot study) and 2006-2007 (confirmatory study) influenza seasons 24 (Table 2). Differences between groups were observed in the confirmatory study only. In the confirmatory study, 113 participants were given a traditionally fermented dairy drink containing L. casei DN-114 001 (along with the traditional ferments of S. thermophilus and L.bulgaricus) for 13 weeks, and 109 were given non-fermented dairy drink. The influenza vaccine (A/H1N1, A/H3N2, and B) was administered 4 weeks after starting the consumption of the study products. The geometric mean antibody titers for the B strain were higher at 3 weeks (p=0.029), 6 weeks (p=0.027), and 9 weeks (p=0.025) after vaccination in the intervention arm, compared with the placebo. The seroconversion rate at 5 months after vaccination was also higher in the intervention arm, compared with the placebo for the B and A/H3N2 strains only (p=0.016 and p=0.031, respectively). The seroprotection rate at 3 weeks after vaccination in a subgroup of participants who were non-seroprotected at baseline was increased in the intervention group, compared to the placebo, for the A/H1N1 strain only (p=0.045).

In Germany, 22 healthy adults were given 100 g acidified milk product containing L. rhamnosusGG(LGG)(intervention 1) daily, 21 were given the same milk product with L. acidophilus paracaseisubspeciesparacasei(CRL431)(intervention 2) daily, and 20 received placebo acidified milk product 26 (Table 2). The treatment period spanned over 5 weeks for both intervention arms; the study year was not reported. Oral polio vaccination occurred at day 8. The poliovirus-1 IgA titer was increased in intervention arm 1 (p=0.036) and the poliovirus-2 IgM titer was increased in intervention arm 2 (p=0.040), compared to the placebo. Increased neutralizing antibodies of poliovirus-1 and -2 were also found with intervention arm 1, compared with the placebo (p=0.048 and p =0.014, respectively). Increased neutralizing antibodies were observed for poliovirus-3 (p=0.011), with intervention arm 2 compared to the placebo.

Fifty-six studies evaluated dairy’s effects on immunoglobulins without vaccinations. The results were heterogenous, with some studies reporting significant enhancement of immunoglobulins (IgA, IgE, or IgG), while others observed no differences between treatment groups. Supplemental Table 3 presents the immunoglobulin information reported in these 56 studies. Supplemental Materials provide detailed summaries of the evidence.