Probiotics for the prevention of allergy: A systematic review and meta-analysis of randomized controlled trials _ Carlos A. Cuello-Garcia, MD,a,b Jan L. Brozek, MD, PhD,a,c Alessandro Fiocchi, MD,d Ruby Pawankar, MD,e a,f  Yepes-Nun ~ ez, MD, MSc(Epi), Luigi Terracciano, MD,g Shreyas Gandhi, BHSc,c,h Arnav Agarwal, BHSc,c,h Juan Jose a € nemann, MD, MSc, PhDa,c Yuan Zhang, MD, and Holger J. Schu Hamilton and Toronto, Ontario, Canada, Monterrey, Mexico, Vatican City, Tokyo, Japan, Medellin, Colombia, and Milan, Italy Background: Allergic diseases are considered a health burden because of their high and constantly increasing prevalence, high direct and indirect costs, and undesirable effects on quality of life. Probiotics have been suggested as an intervention to prevent allergic diseases. Objective: We sought to synthesize the evidence supporting use of probiotics for the prevention of allergies and inform World Allergy Organization guidelines on probiotic use. Methods: We performed a systematic review of randomized trials assessing the effects of any probiotic administered to pregnant women, breast-feeding mothers, and/or infants. Results: Of 2403 articles published until December 2014 identified in Cochrane Central Register of Controlled Trials, MEDLINE, and Embase, 29 studies fulfilled a priori specified inclusion criteria for the analyses. Probiotics reduced the risk of eczema when used by women during the last trimester of pregnancy (relative risk [RR], 0.71; 95% CI, 0.60-0.84), when used by breast-feeding mothers (RR, 0.57; 95% CI, 0.47-0.69), or when given to infants (RR, 0.80; 95% CI, 0.68-0.94). Evidence did not support an effect on other allergies, nutrition status, or incidence of adverse effects. The certainty in the evidence according to the Grading of Recommendation Assessment Development and Evaluation approach is low or very low because of the risk of bias, inconsistency and imprecision of results, and indirectness of available research.

From the Departments of aClinical Epidemiology and Biostatistics and cMedicine, McMaster University, Hamilton; bTecnologico de Monterrey School of Medicine, Monterrey; dPediatric Hospital Bambino Gesu, Vatican City; ethe Department of Pediatrics, Nippon Medical School, Tokyo; fUniversity of Antioquia, School of Medicine, Medellin; gthe Department of Child and Maternal Medicine, University of Milan Medical School at the Melloni Hospital, Milan; and hthe Faculty of Medicine, University of Toronto. Supported by the World Allergy Organization. Disclosure of potential conflict of interest: This study was supported by the World Allergy Organization (WAO). The institution of J. L. Bro_zek, A. Fiocchi, J. J. Yepes-Nu~ nez, L. Terracciano, S. Gandhi, C. A. Cuello-Garcia, Y. Zhang, and H. J. Sch€ unemann, has received consultancy fees from the WAO, as well as support for travel to meetings and other study-related purposes and fees for participating in review activities. A. Fiocchi has received consultancy fees from GlaxoSmithKline, as well as payment for delivering lectures from Danone. R. Pawankar received royalties from Springer and is employed by the Nippon Medical School in Tokyo, Japan, which has received or has grants pending from the Japanese Ministry of Education. L. Terracciano has received consultancy fees from Heinz-Plada. The rest of the authors declare that they have no relevant conflicts of interest. Received for publication January 13, 2015; revised April 17, 2015; accepted for publication April 20, 2015. Corresponding author: Jan L. Bro_zek, MD, PhD, Department of Clinical Epidemiology and Biostatistics, McMaster University Health Sciences Centre, Rm 2C19, 1200 Main St, West Hamilton, Ontario L8N 3Z5, Canada. E-mail: [email protected]. 0091-6749/$36.00 Ó 2015 American Academy of Allergy, Asthma & Immunology http://dx.doi.org/10.1016/j.jaci.2015.04.031

Conclusion: Probiotics used by pregnant women or breastfeeding mothers and/or given to infants reduced the risk of eczema in infants; however, the certainty in the evidence is low. No effect was observed for the prevention of other allergic conditions. (J Allergy Clin Immunol 2015;nnn:nnn-nnn.) Key words: Allergy, prevention, probiotics systematic reviews

Allergic diseases are on the increase, and they pose a considerable burden on health care because of potentially life-threatening allergic reactions, reduced quality of life, and associated direct and indirect costs.1-3 It is estimated that up to 20% of the population experiences an allergic condition, such as atopic dermatitis, food allergy, asthma, allergic rhinitis, and/or conjunctivitis.4 At the same time, a decrease in infectious diseases has been observed in developed countries and is frequently associated with the increased risk for allergies, giving rise to the so-called hygiene hypothesis of allergic diseases.3 The composition of the gastrointestinal microbiota has been postulated to play a role in the development of allergies because it promotes potentially antiallergenic processes: TH1-type immunity, generation of TGF (which has an essential role in suppression of TH2-induced allergic inflammation and induction of oral tolerance), and IgA production, an essential component of mucosal immune defence.5 The gut microbiota hypothesis suggests that alterations in those microbiota, the early and most massive source of microbial exposure, might underlie the allergic epidemic. From this perspective, the use of probiotic supplementation seems an attractive option for the prevention and treatment of allergic diseases. Probiotics are defined as ‘‘live microorganisms which, when administered in adequate amounts as part of food, confer a health benefit on the host.’’6 They can act as promoters of an adequate balance in the gut microbiota, which in turn could prevent the development of allergies. Although randomized controlled trials and systematic reviews have evaluated the use of probiotics for the prevention and treatment of allergic diseases, they have only assessed individual outcomes or are outdated. For example, the most recent reviews on prevention assessed the use of probiotics individually on the outcomes of sensitization,7 asthma or wheezing episodes,8 and atopic dermatitis.9-12 These studies have suggested a modest reduction in atopic dermatitis (eczema) and provided evidence of a mild effect in reducing sensitization. For other types of allergic conditions (ie, conjunctivitis, food allergy, and allergic rhinitis), there are insufficient data to provide conclusions. A Cochrane systematic review12 with relevant subgroups and outcomes was published more than 7 years ago. 1

2 CUELLO-GARCIA ET AL

J ALLERGY CLIN IMMUNOL nnn 2015

Abbreviations used GRADE: Grading of Recommendation Assessment Development and Evaluation RR: Relative risk

Probiotics have been used in the clinical setting in different ways: (1) directly given to infants as an oral preparation or with milk formula in those infants not being breast-fed; (2) indirectly given through breastmilk when given to mothers during breast-feeding; (3) transplacental when given to mothers during pregnancy; or (4) a combination of the above. Informing the recommendations in the World Allergy Organization clinical practice guideline for the prevention of allergies or Guideline for Allergic Disease Prevention,13 we conducted a systematic review to answer 3 clinical questions: 1. Should supplementation of probiotics versus no such supplementation be used in pregnant women to prevent development of allergy in their children? 2. Should supplementation of probiotics versus no such supplementation be used in breast-feeding mothers to prevent development of allergy in their children? 3. Should supplementation of probiotics versus no such supplementation be used in infants to prevent development of allergy?

METHODS Criteria for considering studies for this review Types of studies. We included randomized controlled trials with a minimum follow-up of 4 weeks that compared any type of probiotic with placebo, irrespective of their language or publication status. Types of participants. Studies must have included 1 or more of the following groups of participants: pregnant women, breast-feeding mothers, and infants and children. We included studies that assessed the use of probiotics in any age group, from newborn infants to preschool and school-age children (up to 9 years of age). Types of interventions. We included studies that used any probiotic supplementation, irrespective of formulation (capsules, oil droplets, suspension, and supplements in infant formulas or cereals), microorganism, supplement composition (single vs multiple strains), or dose. Types of outcome measures. The World Allergy Organization guideline panel members in a formal process determined the outcomes of interest. The following outcomes were deemed critical to the decision whether to use probiotics for prevention of allergies: eczema, asthma and/or wheezing, food allergy, allergic rhinitis, any adverse effects, and severe adverse effects.

Search methods We searched the Cochrane Central Register of Controlled Trials (from inception to December 2014), Ovid MEDLINE (from inception to December 2014), and Ovid Embase (from January 1980 to December 2014). We present all search strategies in Table E1 in this article’s Online Repository at www.jacionline.org. We used the Cochrane highly sensitive search strategy for randomized controlled trials for retrieving citations in MEDLINE and Embase. We also contacted authors of identified studies and experts in the field to find any additional studies we had not identified by database searches, reviewed abstracts and conference proceedings, and crosschecked references of included studies for additional sources. We also searched clinicaltrials.gov for ongoing studies.

FIG 1. Study flow diagram.

Data collection and analysis Selection of studies. Three review authors (C.A.C.-G., J.J.Y.-N., and S.G.) independently screened in duplicate the titles and abstracts identified in database searches and resolved differences by consensus. The same authors assessed full-text publication for inclusion using prepiloted screening forms. Any disagreements were resolved by consensus or by a fourth review author (J.L.B.). Data extraction. Four review authors (C.C.G., J.J.Y.-N., S.G., and A.A.) extracted the data using predefined and piloted data extraction forms. We recorded study sequence generation, concealment of allocation, masking, exclusions, patient dropout, loss to follow-up, and noncompliance; participants’ characteristics, such as country of origin and setting where participants were enrolled, number of patients in the study, number of patients randomized, age, age range, sex, and inclusion and exclusion criteria; intervention characteristics, such as type of probiotic, dosage in colonyforming units, placebo use, and duration of treatment; and individual outcomes, specifically the length of follow-up and the event rate in each of the outcomes mentioned above. We also recorded trial registration status and funding for each study. We performed data extraction in duplicate, and any discrepancies were resolved by consensus. Assessment of risk of bias in included studies. We assessed the risk of bias in the included studies using the Cochrane Collaboration risk of bias tool.14 Measures of treatment effect. We estimated relative risks (RRs) and 95% CIs for dichotomous outcomes and mean differences with associated 95% CIs for continuous outcomes. Dealing with missing data. We did not impute missing data. We contacted the study authors to obtain additional information if it was not available from the published report or reports from studies.

J ALLERGY CLIN IMMUNOL VOLUME nnn, NUMBER nn

CUELLO-GARCIA ET AL 3

FIG 2. Studies included in relation to the period/population when the intervention (probiotic) was administered (shaded areas). Direct evidence is displayed in boxes 1-3, when the probiotic was given to pregnant women, breast-feeding mothers, or infants. Indirect evidence is displayed in boxes 4-7, when probiotics were offered to any of the 3 possible populations.

Assessment of statistical heterogeneity and inconsistency. We assessed inconsistency between studies by means of visual inspection of forest plots, the Q statistic (with a P value of .05 or less as a suggestion of important statistical heterogeneity), and the I2 value.15 When inconsistency was found, we planned to explain it by differences in study populations, dosages, presentation, or strains of probiotics used. Assessment of publication bias. If possible, we investigated publication bias by means of visual assessment of funnel plot asymmetry. Data synthesis. We used RevMan version 5.316 to perform meta-analyses using the random-effects model according to the methods suggested in the Cochrane Handbook for Systematic Reviews of Interventions.14

Subgroup analysis and investigation of inconsistency. We considered 3 main groups (meta-analyses), each representing a clinical question. Because in many studies probiotics were given to more than one population of interest, we investigated both direct and indirect evidence (in other populations). We considered whether the population of interest (eg, infants) received the intervention as direct evidence and when not only the population of interest but also other populations received the intervention (eg, probiotics were administered to infants and their mothers during pregnancy or breast-feeding) as indirect evidence. We also planned a subgroup analysis of studies reporting eczema to investigate whether probiotics would have a different effect depending on the inclusion criteria of ‘‘any eczema’’ and ‘‘atopic eczema’’ (ie, eczema and an IgE-mediated sensitization to an allergen).

Sensitivity analysis. We performed sensitivity analyses to determine the effect of including or excluding the studies at higher risk of bias that we planned a priori. Certainty in the evidence (quality of evidence). We rated the overall quality of the body of evidence (also known as certainty in the evidence or confidence in the effect estimates) for the outcomes included in the review following the Grading of Recommendation Assessment Development and Evaluation (GRADE) approach (C.C.G., J.J.Y.-N., H.J.S., and J.L.B.).17 We used the GRADEpro Guideline Development Tool (www. guidelinedevelopment.org) to produce a summary of findings tables.

RESULTS Our search strategy identified 2403 unique bibliographic citations, of which 178 were selected for full-text assessment (Fig 1). We identified 29 randomized controlled trials comparing at least 1 probiotic with placebo18-46 and meeting the inclusion criteria. Fig 2 lists the included studies according to the different populations and periods when the probiotics were administered (eg, during pregnancy only or during pregnancy and breast-feeding). Table E2 in this article’s Online Repository at www.jacionline. org shows the characteristics of included studies and the probiotics used for supplementation. Table E3 in this article’s

4 CUELLO-GARCIA ET AL

J ALLERGY CLIN IMMUNOL nnn 2015

TABLE I. Probiotics versus no probiotics in pregnant women for prevention of allergy in their children (direct evidence): GRADE approach Anticipated absolute effects Outcomes

Eczema, follow-up: range, 12-24 mo

Atopic eczema, follow-up: range, 6-36 mo

Asthma/wheezing, follow-up: range, 12-24 mo

Food allergy, follow-up: range, 12-24 mo

Allergic rhinitis, follow-up: range, 12-24 mo

Any allergy, follow-up: range, 12-24 mo

Adverse events, follow-up: range, 12-24 mo

No. of participants (studies)

242 (1 RCT)

Quality of evidence (GRADE)

Relative effect, RR (95% CI)

CCBB

0.88 (0.63-1.22)

Risk with no probiotics

Risk difference with probiotics

Study population

Low*

208 (1 RCT)

CCBB

392 per 1000

47 fewer per 1000 (145 fewer to 86 more) Study population

188 per 1000

11 fewer per 1000 (88 fewer to 128 more) Study population

242 per 1000

19 fewer per 1000 (102 fewer to 109 more) Study population

0.94 (0.53-1.68)

Low*

242 (1 RCT)

CCBB

0.92 (0.58-1.45)

Low*

(0 studies)

(0 studies)

(0 studies)

(0 studies)

Not estimable 0 per 1000

0 fewer per 1000 (0 fewer to 0 fewer) Study population

0 per 1000

0 fewer per 1000 (0 fewer to 0 fewer) Study population

0 per 1000

0 fewer per 1000 (0 fewer to 0 fewer) Study population

0 per 1000

0 fewer per 1000 (0 fewer to 0 fewer)

Not estimable

Not estimable

Not estimable

GRADE Working Group grades of evidence: High quality, further research is very unlikely to change our confidence in the estimate of effect; moderate quality, further research is likely to have an important effect on our confidence in the estimate of effect and might change the estimate; low quality, further research is very likely to have an important effect on our confidence in the estimate of effect and is likely to change the estimate; very low quality, we are very uncertain about the estimate. The basis for the assumed risk (eg, the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). RCT, Randomized controlled trial. *The CI does not exclude appreciable benefit or harm.

Online Repository at www.jacionline.org lists the studies that did not fulfill the inclusion criteria with the reasons for the exclusion. Our search for ongoing studies in Clinical Trials Registry (ClinicalTrials.Gov) revealed 5 ongoing randomized controlled trials that would potentially meet our inclusion criteria but are expected to be completed no earlier than 2017.

Risk of bias in included studies There were moderate-to-serious concerns about the risk of bias in most studies (see Fig E1 in this article’s Online Repository at www.jacionline.org), especially about adequate concealment of allocation and the number of patients who were lost to follow-up. Table E4 in this article’s Online Repository at www.jacionline.org presents the risk of bias for individual outcomes. Effects of interventions Tables I-VI show the summary of findings for all 3 questions. Forest plots of meta-analyses are depicted in Figs E2 to E4 in this article’s Online Repository at www.jacionline.org.

Use of probiotics during pregnancy. Table I includes only studies with direct evidence of the effect of probiotics administered to pregnant women, whereas Table II shows the summary of findings from studies in which probiotics were used by pregnant women directly and otherwise. Only one study18 assessed the use of probiotics administered exclusively to pregnant women (Fig 2, box 1), whereas the remaining 20 studies19,20,22-26,28-35,37,39,40,45,46 evaluated the use of probiotics either in pregnant and then breast-feeding mothers,19,20,22,23 in pregnant women and their infants,24-26,33-35,37,39 or in all 3 groups (Fig 2, boxes 4, 5, and 7).28-32,40,45,46 Overall, the final estimate from the meta-analysis of direct and indirect evidence showed a reduced risk of eczema in infants, as shown in Fig E2 (RR, 0.72; 95% CI, 0.61-0.85; 3509 participants, 15 trials). The certainty in the evidence was judged as very low because of the risk of bias, inconsistency of results among studies, and indirectness of the whole body of evidence. Direct evidence from 1 trial in which probiotics were taken by pregnant women only showed a similar magnitude of effect on development of eczema, but the CI was very wide because of the small number of events (RR, 0.88; 95% CI, 0.63-1.22; 1 study, 242 participants). There was

CUELLO-GARCIA ET AL 5

J ALLERGY CLIN IMMUNOL VOLUME nnn, NUMBER nn

TABLE II. Probiotics versus no probiotics in pregnant women for prevention of allergy in their children (indirect evidence): GRADE approach Anticipated absolute effects Outcomes

Eczema, follow-up: range, 12-24 mo

Atopic eczema, follow-up: range, 12-36 mo

Asthma/wheezing, follow-up: range, 12-24 mo

Food allergy, follow-up: range, 12-24 mo

Allergic rhinitis, follow-up: range, 12-24 mo

Any allergy, follow-up: range, 12-24 mo

Adverse events, follow-up: range, 12-24 mo

No. of participants (studies)

3109 (14 RCTs)*

Quality of evidence (GRADE)

Relative effect, RR (95% CI)

CBBB

0.71 (0.59-0.84)

Risk with no probiotics

Risk difference with probiotics

Study population

Very low à§

2777 (10 RCTs)*

CBBB

309 per 1000

90 fewer per 1000 (127 fewer to 49 fewer) Study population

187 per 1000

65 fewer per 1000 (84 fewer to 41 fewer) Study population

137 per 1000

8 fewer per 1000 (38 fewer to 31 more) Study population

39 per 1000

3 more per 1000 (11 fewer to 23 more) Study population

123 per 1000

17 fewer per 1000 (69 fewer to 86 more) Study population

369 per 1000

26 fewer per 1000 (74 fewer to 30 more) Study population

0.65 (0.55-0.78)

Very lowà§k

2011 (7 RCTs)

CBBB

0.94 (0.72-1.23)

Very low§{#

355 (3 RCTs)

CBBB

1.08 (0.73-1.59)

Very low§#**

1743 (5 RCTs)

CBBB

0.86 (0.44-1.7)

Very low§#  

1152 (3 RCTs)

CBBB

0.93 (0.8-1.08)

Very low§#  

791 (3 RCTs)

CBBB

1.18 (0.78-1.78)

Low§# 222 per 1000

40 more per 1000 (49 fewer to 173 more)

The basis for the assumed risk (eg, the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). GRADE Working Group grades of evidence: High quality, further research is very unlikely to change our confidence in the estimate of effect; Moderate quality, further research is likely to have an important effect on our confidence in the estimate of effect and might change the estimate; Low quality, further research is very likely to have an important effect on our confidence in the estimate of effect and is likely to change the estimate; Very low quality, we are very uncertain about the estimate. RCT, Randomized controlled trial. *Although there are 14 studies, in the corresponding forest plot 15 reports appear as one study includes 2 intervention arms (Wickens 2008a and Wickens 2008b).46  Moderate risk of bias in 12 studies (unclear for allocation concealment) and one study with 6% weight with high risk of bias; only one study was free of risk of bias. àClinical heterogeneity as different probiotics were used among studies. §Probiotics were administered during pregnancy plus breast-feeding, pregnancy plus infancy, or during the 3 periods. kRisk of bias because of incomplete follow-up and concealment of allocation. {Only one study without risk of bias; the rest with moderate risk of bias for allocation concealment and adequate follow-up. #The CI does not exclude appreciable benefit or harm. **All studies with moderate risk of bias (inadequate follow-up).   Studies with moderate risk of bias (inadequate follow-up and allocation concealment).

no effect on other outcomes of interest or an increase in adverse effects (see forest plots in Fig E2). Use of probiotics by mothers during breast-feeding. Tables III and IV show a summary of findings for studies in which probiotic supplements were used by mothers during breast-feeding. Of 13 studies19-23,28-32,40,45,46 assessing the use of probiotics in breast-feeding mothers, only one21 evaluated their use exclusively during this period (Fig 2, box 2). All remaining studies used probiotics during pregnancy and breastfeeding19,20,22,23 and/or in pregnant and, later, breast-feeding mothers and their infants (Fig 2, boxes 5 and 7).28-32,40,45,46 Overall, probiotics reduced the risk of eczema (RR, 0.61; 95% CI, 0.50-0.74; 1595 participants, 10 trials); however, the confidence in this estimate is very low because of the indirectness and inconsistency of results among studies. Direct evidence from

1 trial in which probiotics were taken by breast-feeding mothers only showed a similar magnitude of effect on the development of eczema, but the CI was very wide because of the small number of events (RR, 0.57; 95% CI, 0.29-1.11; 1 study, 88 participants). Evidence from currently available trials on the use of probiotics by breast-feeding mothers shows no effect on the prevention of other allergies in their children (see forest plots in Fig E3). Use of probiotics in infants. Tables V and VI show a summary of findings for studies of probiotic supplementation in infants. Twenty-three studies evaluated the use of probiotics in infants,24-46 of which only 7 assessed the intervention administered only to infants (Fig 2, box 3).27,36,38,41-44 Probiotics were administered to pregnant women and then to their infants in 8 of the remaining studies,24-26,33-35,37,39 and in another 8 studies

6 CUELLO-GARCIA ET AL

J ALLERGY CLIN IMMUNOL nnn 2015

TABLE III. Probiotics versus no probiotics in breast-feeding women for prevention of allergy in their children (direct evidence): GRADE approach Anticipated absolute effects Outcomes

Eczema, follow-up: range, 12-24 mo

Atopic eczema, follow-up: range, 12-24 mo

Asthma, follow-up: range, 12-24 mo

Food allergy, follow-up: range, 12-24 mo

Allergic rhinitis, follow-up: range, 12-24 mo

Any allergy, follow-up: range, 12-24 mo

Adverse events, follow-up: range, 12-24 mo

No. of participants (studies)

88 (1 RCT)

Quality of evidence (GRADE)

Relative effect, RR (95% CI)

CCBB

0.57 (0.29-1.11)

Risk with no probiotics

Risk difference with probiotics

Study population

Low*

(0 studies)

163 fewer per 1000 (269 fewer to 42 more) Study population

0 per 1000

0 fewer per 1000 (0 fewer to 0 fewer) Study population

0 per 1000

0 fewer per 1000 (0 fewer to 0 fewer) Study population

0 per 1000

0 fewer per 1000 (0 fewer to 0 fewer) Study population

0 per 1000

0 fewer per 1000 (0 fewer to 0 fewer) Study population

54 per 1000

15 fewer per 1000 (48 fewer to 212 more) Study population

Not estimable

(0 studies)

Not estimable

(0 studies)

Not estimable

(0 studies)

88 (1 RCT)

378 per 1000

Not estimable

CCBB

0.73 (0.11-4.92)

Low*

(0 studies)

Not estimable 0 per 1000

0 fewer per 1000 (0 fewer to 0 fewer)

GRADE Working Group grades of evidence: High quality, further research is very unlikely to change our confidence in the estimate of effect; Moderate quality, further research is likely to have an important effect on our confidence in the estimate of effect and might change the estimate; Low quality, further research is very likely to have an important effect on our confidence in the estimate of effect and is likely to change the estimate; Very low quality, we are very uncertain about the estimate. The basis for the assumed risk (eg, the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). RCT, Randomized controlled trial. *The CI does not exclude appreciable benefit or harm.

probiotics were administered to mothers during pregnancy and breast-feeding and to their infants (Fig 2, boxes 4 and 7).28-32,40,45,46 Overall, probiotics reduced the risk of eczema in children (RR, 0.81; 95% CI, 0.70-0.94; 3447 participants, 15 trials). The confidence in this estimate is very low because of the risk of bias, indirectness of evidence, and inconsistency and imprecision of the results. Direct evidence from 5 trials showed a similar magnitude of effect (RR, 0.83; 95% CI, 0.58-1.19). We found evidence of no effect of using probiotics in infants on other outcomes, with very low certainty in the evidence (see forest plots in Fig E4).

DISCUSSION In this systematic review we found that supplementation with probiotics during pregnancy, in breast-feeding mothers, or in infants decreases the risk of eczema (including atopic eczema) in infants. However, the certainty in the evidence is very low because of the risk of bias across trials, indirectness of the evidence, inconsistency of results among studies, and imprecision of the pooled estimates. Currently available evidence suggests

that supplementation of probiotics does not reduce the risk of other allergies because of the imprecision of the estimates of effect of probiotics on these outcomes. Adverse effects were mild and similar across the intervention and control groups in every population receiving probiotic supplements. This systematic review has several strengths. Our review comprehensively summarizes the direct and indirect evidence from randomized trials about using probiotics in pregnant women, breast-feeding mothers, and their infants with an intention to prevent the development of allergy in infants. It also summarizes the evidence about the effects on all, rather than selected, patient-important outcomes. Our comprehensive search of several literature databases amended by reviewing the citations in identified studies and reviews, as well as seeking the information form study authors and experts in the field, makes it unlikely that any relevant study was missed. Therefore it presents the most comprehensive summary of all experimental data about the use of probiotics in the prevention of allergy available today. The limitations of our findings stem from the limitations of the available body of evidence on this topic. Our confidence that one would observe effects in real life is low to very low (low to very

CUELLO-GARCIA ET AL 7

J ALLERGY CLIN IMMUNOL VOLUME nnn, NUMBER nn

TABLE IV. Probiotics compared with no probiotics in breast-feeding women for prevention of allergies in their children (indirect): GRADE approach Anticipated absolute effects Outcomes

Eczema, follow-up: range, 12-24 mo

Atopic eczema, follow-up: range, 12-36 mo

Asthma/wheezing, follow-up: range, 12-24 mo

Food allergy, follow-up: range, 12-24 mo

Allergic rhinitis, follow-up: range, 12-24 mo

Any allergy, follow-up: range, 12-24 mo

Adverse events, follow-up: range, 12-24 mo

No. of participants (studies)

1507 (9 RCTs)*

Quality of evidence (GRADE)

Relative effect, RR (95% CI)

CBBB

0.61 (0.49-0.76)

Risk with no probiotics

Risk difference with probiotics

Study population

Very low à§

1225 (6 RCTs)*

CBBB

361 per 1000

141 fewer per 1000 (184 fewer to 87 fewer) Study population

198 per 1000

73 fewer per 1000 (101 fewer to 36 fewer) Study population

125 per 1000

6 more per 1000 (51 fewer to 108 more) Study population

59 per 1000

41 more per 1000 (25 fewer to 233 more) Study population

56 per 1000

8 fewer per 1000 (45 fewer to 139 more) Study population

469 per 1000

9 more per 1000 (136 fewer to 216 more) Study population

0.63 (0.49-0.82)

Very low à§

801 (4 RCTs)

CBBB

1.05 (0.59-1.87)

Very low §k

167 (2 RCTs)

CBBB

1.7 (0.58-4.96)

Very low§k{

664 (3 RCTs)

CCBB

0.86 (0.21-3.47)

Low§k

129 (1 RCT)

CBBB

1.02 (0.71-1.46)

Very low§k#

149 (1 RCT)

CCBB

1.52 (0.79-2.96)

Low§k 343 per 1000

178 more per 1000 (72 fewer to 672 more)

The basis for the assumed risk (eg, the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). GRADE Working Group grades of evidence: High quality: further research is very unlikely to change our confidence in the estimate of effect; Moderate quality: further research is likely to have an important effect on our confidence in the estimate of effect and might change the estimate; Low quality: further research is very likely to have an important effect on our confidence in the estimate of effect and is likely to change the estimate; Very low quality, we are very uncertain about the estimate. RCT, Randomized controlled trial. *Although there are 14 studies, in the corresponding forest plot 15 reports appear as one study including 2 intervention arms (Wickens 2008a and Wickens 2008b).46  Only one study with low risk of bias; the rest with higher concerns for risk of bias (concerns on allocation concealment and follow-up). àDifferent probiotics used among studies. §Probiotics were administered during pregnancy and breast-feeding periods or during pregnancy, breast-feeding, and to the infant. kThe CI does not exclude appreciable benefit or harm. {Both studies with risk of attrition bias. #Unclear risk of bias (allocation concealment and follow-up).

low certainty in the evidence). This is a result of the relative paucity of direct evidence in any of the 3 groups in whom probiotics could be used, the high likelihood of bias in primary studies, and the serious imprecision of the estimated pooled effects (Tables I-VI). The 3 initially posed clinical questions were addressed mainly through indirect bodies of evidence that largely overlapped; this could undermine the extent to which the results are applicable to a particular population. Furthermore, high variability of probiotics used in the studies makes it impossible to infer whether any specific probiotic could have different and possibly more beneficial effects compared with other probiotics. Our results extend the findings of previous reviews that assessed selected allergy outcomes7-10,12 showing that use of probiotics in any of these populations might reduce the risk of eczema and that there is no proved effect on development of other allergies. We have also found no excess in adverse effects, which

is in agreement with the health technology assessment report form the Agency for Healthcare Research and Quality that assessed the safety of probiotics used with various intents in a variety of populations.47 There is a surprising paucity of direct experimental evidence about the use of probiotics for the prevention of allergy, specifically about the use of probiotics by pregnant women and breast-feeding mothers. Development of some types of allergy (eg, food allergy or allergic rhinitis) has not been measured in studies assessing the effectiveness of probiotics when used by pregnant women and breast-feeding mothers. For all outcomes, except eczema, indirect evidence (ie, studies in which probiotics were given to >1 of the 3 groups that could receive them: pregnant women, breast-feeding mothers, and infants) also does not provide estimates precise enough to exclude an appreciable benefit and either no effect or even possible harm. A relatively large proportion of patients that were lost to follow-up in most studies make one even less confident in these effects. We also

8 CUELLO-GARCIA ET AL

J ALLERGY CLIN IMMUNOL nnn 2015

TABLE V. Probiotics compared with no probiotics in infants for the prevention of allergies (direct) Anticipated absolute effects Outcomes*

Eczema, follow-up: range, 6-24 mo

Atopic eczema, follow-up: range, 6-36 mo

Asthma/wheezing, follow-up: range, 6-24 mo

Food allergy, follow-up: range, 6-24 mo

Any allergy, follow-up: range, 6-24 mo

Adverse effects, follow-up: range, 6-24 mo

No. of participants (studies)

790 (5 RCTs)

Quality of evidence (GRADE)

Relative effect, RR (95% CI)

CBBB

0.83 (0.58-1.19)

Risk with no probiotics

Risk difference with probiotics

Study population

Very low à§

427 (2 RCTs)

CCBB

296 per 1000

50 fewer per 1000 (124 fewer to 56 more) Study population

90 per 1000

60 more per 1000 (2 fewer to 172 more) Study population

121 per 1000

5 more per 1000 (45 fewer to 85 more) Study population

167 per 1000

20 fewer per 1000 (75 fewer to 72 more) Study population

447 per 1000

85 more per 1000 (63 fewer to 291 more) Study population

1.67 (0.98-2.92)

Low§k

412 (3 RCTs)

CBBB

1.04 (0.63-1.7)

Very low à§

349 (3 RCTs)

CBBB

0.88 (0.55-1.43)

Very lowà§{

153 (1 RCT)

CCCB

1.19 (0.86-1.65)

Moderate§

187 (2 RCTs)

CBBB

1.27 (0.51-3.18)

Very lowà§# 532 per 1000

Nutritional status, assessed with growth and weight, follow-up: range, 3-36 mo

254 (2 RCTs)

CBBB

Very lowà§#

Not estimable

Mean nutritional status: 0 infants in the control group

144 more per 1000 (261 fewer to 1160 more) SMD, 0.06 lower (0.31 lower to 0.18 higher)

The basis for the assumed risk (eg, the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). GRADE Working Group grades of evidence: High quality, further research is very unlikely to change our confidence in the estimate of effect; Moderate quality, further research is likely to have an important effect on our confidence in the estimate of effect and might change the estimate; Low quality, further research is very likely to have an important effect on our confidence in the estimate of effect and is likely to change the estimate; Very low quality, we are very uncertain about the estimate. RCT, Randomized controlled trial; SMD, standardized mean difference. *There were no studies for allergic rhinitis.  Concerns with high risk of bias for allocation concealment and blinding; unclear risk of bias in random allocation and adequate follow-up. àClinical heterogeneity: different probiotics among studies. §The CI does not exclude appreciable benefit or harm. kBlinding of assessment was not adequately reported, and studies had incomplete outcome data. {One study with high risk of bias for concerns on the blinding process, allocation concealment, and random sequence generation. #One study with unclear description of the randomization process, allocation concealment, blinding, and follow-up.

found large variability in probiotics and their mixtures used in the studies, precluding any conclusions about the effects of specific strains. Our results are in agreement with those of previous systematic reviews that have assessed the effect of probiotics on selected allergy outcomes.7-10,12 The studies by Elazab et al,7 Azad et al,8 Pelucchi et al,9 and Doege et al10 did not evaluate the direct and indirect evidence separately, although one review10 included only direct evidence in pregnant women.

Suggestions for future research There is a need for more well-designed and executed randomized trials that would properly measure and report the effects of probiotics in the prevention of all allergic diseases, as well as potential adverse effects, and thus reduce the overall risk of bias. Proper assessment of any long-term adverse effects might require larger observational studies in infants and children. It might be beneficial if future trials focused on the most common

probiotics and perhaps also directly compares the effects of probiotics used in different populations, such as pregnant women versus breast-feeding mothers, as well as different dosages and specific probiotic strains. We also found no direct experimental evidence of using probiotics in formula, which might warrant further investigation.

Conclusions Supplementation of probiotics during pregnancy, in breastfeeding women and/or infants reduced the risk of eczema, including atopic eczema, in infants. However, our certainty in this evidence is low because of the risk of bias, inconsistency, and imprecision of estimated effect and mostly because of use of bodies of indirect evidence to answer each question. Currently available evidence demonstrates that supplementation of probiotics does not reduce the risk of other allergies in children but also does not exclude such a possibility. Further studies, if done, are likely to substantially contribute to this knowledge.

CUELLO-GARCIA ET AL 9

J ALLERGY CLIN IMMUNOL VOLUME nnn, NUMBER nn

TABLE VI. Probiotics compared with no probiotics in infants for the prevention of allergies (indirect): GRADE approach Anticipated absolute effects Outcomes

Eczema, follow-up: range, 6-24 mo

Atopic eczema, follow-up: range, 6-36 mo

Asthma/wheezing, follow-up: range, 6-36 mo

Food allergy, follow-up: range, 6-24 mo

Allergic rhinitis, follow-up: range, 12-36 mo

Any allergy, follow-up: range, 6-24 mo

Adverse events, follow-up: range, 6-36 mo

No. of participants (studies)

2657 (10 RCTs)*

Quality of evidence (GRADE)

Relative effect, RR (95% CI)

CBBB

0.80 (0.64-0.93)

Risk with no probiotics

Risk difference with probiotics

Study population

Very low à§k

2218 (7 RCTs)*

CBBB

289 per 1000

58 fewer per 1000 (104 fewer to 20 fewer) Study population

198 per 1000

73 fewer per 1000 (95 fewer to 46 fewer) Study population

145 per 1000

3 fewer per 1000 (39 fewer to 46 more) Study population

27 per 1000

0 fewer per 1000 (20 fewer to 79 more) Study population

145 per 1000

25 fewer per 1000 (89 fewer to 115 more) Study population

369 per 1000

26 fewer per 1000 (74 fewer to 30 more) Study population

0.62 (0.51-0.76)

Very lowà§{

1733 (6 RCTs)

CBBB

0.98 (0.73-1.32)

Very lowà§k{

295 (2 RCTs)

CBBB

1 (0.25-3.91)

Very low§k#

1465 (4 RCTs)

CBBB

0.83 (0.39-1.79)

Very lowà§k**

1152 (3 RCTs)

CBBB

0.93 (0.8-1.08)

Very lowà§k{

642 (2 RCTs)

CBBB

1.03 (0.7-1.52)

Very lowà§k   196 per 1000

Nutritional status assessed with growth and weight, follow-up: range, 3-36 mo

454 (1 RCT)

CBBB

Very low§k  

Not estimable

Mean nutritional status: 0 infants only in the control group

6 more per 1000 (59 fewer to 102 more) SMD, 0 higher (0.18 lower to 0.18 higher)

The basis for the assumed risk (eg, the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). GRADE Working Group grades of evidence: High quality, further research is very unlikely to change our confidence in the estimate of effect; Moderate quality, further research is likely to have an important effect on our confidence in the estimate of effect and might change the estimate; Low quality: further research is very likely to have an important effect on our confidence in the estimate of effect and is likely to change the estimate; Very low quality: we are very uncertain about the estimate. RCT, Randomized controlled trial; SMD, standardized mean difference. *Although there are 14 studies, in the corresponding forest plot 15 reports appear as one study including 2 intervention arms (Wickens 2008a and Wickens 2008b).46  Only one study with low risk of bias; the rest with concerns on allocation concealment and adequate follow-up. àDifferent probiotics among studies were used. §Probiotics were administered during infancy plus breast-feeding, during pregnancy plus infancy, or during infancy plus breast-feeding plus pregnancy. kThe CI does not exclude appreciable benefit or harm. {Most studies with concerns in allocation concealment and adequate follow-up. #High risk of bias for allocation concealment and unclear for adequate follow-up of outcome data. **Incomplete outcome data (attrition bias) in one study.   Concerns on allocation concealment and adequate follow-up.

Key messages d

Probiotics given to pregnant women, breast-feeding mothers, and/or infants reduced the risk of eczema in infants.

d

Currently available evidence does not indicate that probiotics prevent the development of other allergies.

d

Design of future studies should consider the weaknesses identified regarding the risk of bias and indirectness of the evidence.

REFERENCES 1. Holgate ST. The epidemic of allergy and asthma. Nature 1999;402:B2-4. 2. Asher MI, Stewart AW, Wong G, Strachan DP, Garcia-Marcos L, Anderson HR, et al. Changes over time in the relationship between symptoms of asthma, rhinoconjunctivitis and eczema: a global perspective from the International Study of Asthma and Allergies in Childhood (ISAAC). Allergol Immunopathol (Madr) 2012;40:267-74. 3. Bach JF. The effect of infections on susceptibility to autoimmune and allergic diseases. N Engl J Med 2002;347:911-20. 4. Ebert CS Jr, Pillsbury HC 3rd. Epidemiology of allergy. Otolaryngol Clin North Am 2011;44:537-48, vii. 5. Prokesova L, Lodinova-Zadnikova R, Zizka J, Kocourkova I, Novotna O, Petraskova P, et al. Cytokine levels in healthy and allergic mothers and their children during the first year of life. Pediatr Allergy Immunol 2006;17:175-83.

10 CUELLO-GARCIA ET AL

6. Food and Agriculture Organization of the United Nations, World Health Organization. Probiotics in food: health and nutritional properties and guidelines for evaluation. Rome: Food and Agriculture Organization of the United Nations, World Health Organization; 2006. 7. Elazab N, Mendy A, Gasana J, Vieira ER, Quizon A, Forno E. Probiotic administration in early life, atopy, and asthma: a meta-analysis of clinical trials. Pediatrics 2013;132:e666-76. 8. Azad MB, Coneys JG, Kozyrskyj AL, Field CJ, Ramsey CD, Becker AB, et al. Probiotic supplementation during pregnancy or infancy for the prevention of asthma and wheeze: systematic review and meta-analysis. BMJ 2013;347:f6471. 9. Pelucchi C, Chatenoud L, Turati F, Galeone C, Moja L, Bach JF, et al. Probiotics supplementation during pregnancy or infancy for the prevention of atopic dermatitis: a meta-analysis. Epidemiology 2012;23:402-14. 10. Doege K, Grajecki D, Zyriax BC, Detinkina E, Zu Eulenburg C, Buhling KJ. Impact of maternal supplementation with probiotics during pregnancy on atopic eczema in childhood—a meta-analysis. Br J Nutr 2012;107:1-6. 11. Tang ML, Lahtinen SJ, Boyle RJ. Probiotics and prebiotics: clinical effects in allergic disease. Curr Opin Pediatr 2010;22:626-34. 12. Osborn DA, Sinn JK. Probiotics in infants for prevention of allergic disease and food hypersensitivity. Cochrane Database Syst Rev 2007;(4):CD006475. 13. Fiocchi A, Pawankar R, Cuello-Garcia C, Ahn K, Al-Hammadi S, Agarwal A, et al. World Allergy Organization-McMaster University Guidelines for Allergic Disease Prevention (GLAD-P): probiotics. World Allergy Organ J 2015;8:4. 14. Cochrane Handbook for Systematic Reviews of Interventions, Version 5.1.0. The Cochrane Collaboration; 2011. Available at: http://www.cochrane-handbook.org. Accessed December 16, 2014. 15. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557-60. 16. Saengtawesin V, Tangpolkaiwalsak R, Kanjanapattankul W. Effect of oral probiotics supplementation in the prevention of necrotizing enterocolitis among very low birth weight preterm infants. J Med Assoc Thai 2014;97(suppl 6):S20-5. 17. Guyatt GH, Oxman AD, Kunz R, Atkins D, Brozek J, Vist G, et al. GRADE guidelines: 2. Framing the question and deciding on important outcomes. J Clin Epidemiol 2011;64:395-400. 18. Boyle RJ, Ismail IH, Kivivuori S, Licciardi PV, Robins-Browne RM, Mah LJ, et al. Lactobacillus GG treatment during pregnancy for the prevention of eczema: a randomized controlled trial. Allergy 2011;66:509-16. 19. Dotterud CK, Storro O, Johnsen R, Oien T. Probiotics in pregnant women to prevent allergic disease: a randomized, double-blind trial. Br J Dermatol 2010;163:616-23. 20. Huurre A, Laitinen K, Rautava S, Korkeamaki M, Isolauri E. Impact of maternal atopy and probiotic supplementation during pregnancy on infant sensitization: a double-blind placebo-controlled study. Clin Exp Allergy 2008;38:1342-8. 21. Ortiz-Andrellucchi A, Sanchez-Villegas A, Rodriguez-Gallego C, Lemes A, Molero T, Soria A, et al. Immunomodulatory effects of the intake of fermented milk with Lactobacillus casei DN114001 in lactating mothers and their children. Br J Nutr 2008;100:834-45. 22. Rautava S, Kainonen E, Salminen S, Isolauri E. Maternal probiotic supplementation during pregnancy and breast-feeding reduces the risk of eczema in the infant. J Allergy Clin Immunol 2012;130:1355-60. 23. Rautava S, Kalliomaki M, Isolauri E. Probiotics during pregnancy and breast-feeding might confer immunomodulatory protection against atopic disease in the infant. J Allergy Clin Immunol 2002;109:119-21. 24. Abrahamsson TR, Jakobsson T, Bottcher MF, Fredrikson M, Jenmalm MC, Bjorksten B, et al. Probiotics in prevention of IgE-associated eczema: a double-blind, randomized, placebo-controlled trial. J Allergy Clin Immunol 2007;119:1174-80. 25. Allen SJ, Jordan S, Storey M, Thornton CA, Gravenor MB, Garaiova I, et al. Probiotics and atopic eczema: a double-blind randomised controlled trial. Arch Dis Child 2012;97:A2-A. 26. Allen SJ, Jordan S, Storey M, Thornton CA, Gravenor M, Garaiova I, et al. Dietary supplementation with lactobacilli and bifidobacteria is well tolerated and not associated with adverse events during late pregnancy and early infancy. J Nutr 2010;140:483-8. 27. Hascoet JM, Hubert C, Rochat F, Legagneur H, Gaga S, Emady-Azar S, et al. Effect of formula composition on the development of infant gut microbiota. J Pediatr Gastroenterol Nutr 2011;52:756-62.

J ALLERGY CLIN IMMUNOL nnn 2015

28. Kalliomaki M, Salminen S, Arvilommi H, Kero P, Koskinen P, Isolauri E. Probiotics in primary prevention of atopic disease: a randomised placebocontrolled trial. Lancet 2001;357:1076-9. 29. Kalliomaki M, Salminen S, Poussa T, Arvilommi H, Isolauri E. Probiotics and prevention of atopic disease: 4-year follow-up of a randomised placebocontrolled trial. Lancet 2003;361:1869-71. 30. Kalliomaki M, Salminen S, Poussa T, Isolauri E. Probiotics during the first 7 years of life: a cumulative risk reduction of eczema in a randomized, placebo-controlled trial. J Allergy Clin Immunol 2007;119:1019-21. 31. Kim JY, Kwon JH, Ahn SH, Lee SI, Han YS, Choi YO, et al. Effect of probiotic mix (Bifidobacterium bifidum, Bifidobacterium lactis, Lactobacillus acidophilus) in the primary prevention of eczema: A double-blind, randomized, placebo-controlled trial. Pediatr Allergy Immunol 2010;21:e386-93. 32. Kopp MV, Hennemuth I, Heinzmann A, Urbanek R. Randomized, double-blind, placebo-controlled trial of probiotics for primary prevention: no clinical effects of Lactobacillus GG supplementation. Pediatrics 2008;121:e850-6. 33. Kuitunen M, Kukkonen K, Juntunen-Backman K, Korpela R, Poussa T, Tuure T, et al. Probiotics prevent IgE-associated allergy until age 5 years in cesareandelivered children but not in the total cohort. J Allergy Clin Immunol 2009;123: 335-41. 34. Kukkonen AK, Kuitunen M, Savilahti E, Pelkonen A, Malmberg P, Makela M. Airway inflammation in probiotic-treated children at 5 years. Pediatr Allergy Immunol 2011;22:249-51. 35. Kukkonen K, Savilahti E, Haahtela T, Juntunen-Backman K, Korpela R, Poussa T, et al. Probiotics and prebiotic galacto-oligosaccharides in the prevention of allergic diseases: a randomized, double-blind, placebo-controlled trial. J Allergy Clin Immunol 2007;119:192-8. 36. Lodinova-Zadnikova R, Prokesova L, Kocourkova I, Hrdy J, Zizka J. Prevention of allergy in infants of allergic mothers by probiotic Escherichia coli. Int Arch Allergy Immunol 2010;153:201-6. 37. Marschan E, Kuitunen M, Kukkonen K, Poussa T, Sarnesto A, Haahtela T, et al. Probiotics in infancy induce protective immune profiles that are characteristic for chronic low-grade inflammation. Clin Exp Allergy 2008;38:611-8. 38. Morisset M, Aubert-Jacquin C, Soulaines P, Moneret-Vautrin DA, Dupont C. A non-hydrolyzed, fermented milk formula reduces digestive and respiratory events in infants at high risk of allergy. Eur J Clin Nutr 2011;65:175-83. 39. Niers L, Martin R, Rijkers G, Sengers F, Timmerman H, van Uden N, et al. The effects of selected probiotic strains on the development of eczema (the PandA study). Allergy 2009;64:1349-58. 40. Ou CY, Kuo HC, Wang L, Hsu TY, Chuang H, Liu CA, et al. Prenatal and postnatal probiotics reduces maternal but not childhood allergic diseases: a randomized, double-blind, placebo-controlled trial. Clin Exp Allergy 2012;42: 1386-96. 41. Prescott SL, Wiltschut J, Taylor A, Westcott L, Jung W, Currie H, et al. Early markers of allergic disease in a primary prevention study using probiotics: 2.5-year follow-up phase. Allergy 2008;63:1481-90. 42. Soh SE, Aw M, Gerez I, Chong YS, Rauff M, Ng YP, et al. Probiotic supplementation in the first 6 months of life in at risk Asian infants—effects on eczema and atopic sensitization at the age of 1 year. Clin Exp Allergy 2009;39:571-8. 43. Taylor AL, Dunstan JA, Prescott SL. Probiotic supplementation for the first 6 months of life fails to reduce the risk of atopic dermatitis and increases the risk of allergen sensitization in high-risk children: a randomized controlled trial. J Allergy Clin Immunol 2007;119:184-91. 44. West CE, Hammarstr€om ML, Hernell O. Probiotics during weaning reduce the incidence of eczema. Pediatr Allergy Immunol 2009;20:430-7. 45. Wickens K, Black P, Stanley TV, Mitchell E, Barthow C, Fitzharris P, et al. A protective effect of Lactobacillus rhamnosus HN001 against eczema in the first 2 years of life persists to age 4 years. Clin Exp Allergy 2012;42:1071-9. 46. Wickens K, Black PN, Stanley TV, Mitchell E, Fitzharris P, Tannock GW, et al. A differential effect of 2 probiotics in the prevention of eczema and atopy: a double-blind, randomized, placebo-controlled trial. J Allergy Clin Immunol 2008;122:788-94. 47. Hempel S, Newberry S, Ruelaz A, Wang Z, Miles JN, Suttorp MJ, et al. Safety of probiotics used to reduce risk and prevent or treat disease. Evid Rep Technol Assess (Full Rep) 2011;200:1-645.