BACKGROUND. The value of probiotics for primary prevention is controversial. Published trials vary considerably in study design and the applied probiotics, thereby limiting comparability of the results.
OBJECTIVE. The purpose of this trial was to study the preventive effect of the probiotic Lactobacillus GG on the development of atopic dermatitis.
METHODS. In a double-blind, placebo-controlled prospective trial, 105 pregnant women from families with ≥1 member (mother, father, or child) with an atopic disease were randomly assigned to receive either the probiotic Lactobacillus GG (American Type Culture Collection 53103; 5 × 109 colony-forming units of Lactobacillus GG twice daily) or placebo. Ninety-four families (89.5%) completed the trial. The supplementation period started 4 to 6 weeks before expected delivery, followed by a postnatal period of 6 months. The primary end point was the occurrence of atopic dermatitis at the age of 2 years. Secondary outcomes were severity of atopic dermatitis, recurrent episodes of wheezing bronchitis, and allergic sensitization at the age of 2 years.
RESULTS. Atopic dermatitis was diagnosed in 14 (28%) of 50 in the Lactobacillus GG group and in 12 (27.3%) of 44 in the placebo group. The risk of atopic dermatitis in children on probiotics relative to placebo was 0.96 (confidence interval 0.38–2.33). Severity of atopic dermatitis was comparable between the 2 groups. Notably, children with recurrent (≥5) episodes of wheezing bronchitis were more frequent in the Lactobacillus GG group (26%; n = 13), as compared with the placebo group (9.1%; n = 4). No difference was observed between both groups in total immunoglobulin E concentrations or numbers of specific sensitization to inhalant allergens.
CONCLUSIONS. Supplementation with Lactobacillus GG during pregnancy and early infancy neither reduced the incidence of atopic dermatitis nor altered the severity of atopic dermatitis in affected children but was associated with an increased rate of recurrent episodes of wheezing bronchitis. Therefore, Lactobacillus GG cannot be generally recommended for primary prevention.
Atopic dermatitis, allergic rhinitis, and bronchial asthma in childhood represent a significantly large segment of chronic disease in the Western world.1,2 To combat the rising prevalence of allergy, there is still a need to analyze the underlying mechanisms and to develop concepts that address primary prevention.
Improved hygiene, the increased use of antimicrobial medication, the consumption of sterile food, and reduced family size resulting in lower rates of infection during childhood have reduced early contact to microbes.3,4 This may interfere with the development of the child's immune system, which tends to be directed toward a T-helper (Th)2 phenotype in infants, whereas postnatal maturation is associated with gradual inhibition of Th2 and increasing Th1 affinity.5–7 Experimental and epidemiologic data show strong evidence for the hypothesis that early microbial exposure is a critical feature for Th1-skewed immune response in healthy children during the postnatal period.8,9 Less exposure to microbial agents at an early age would, therefore, result in reduced activation of the immune system and subsequent polarization toward a Th2 phenotype. In this context, the idea of modulating the intestinal flora has really caught on and is gaining in popularity. This concept is supported by epidemiologic data, which show that children with allergy have a different intestinal flora from healthy children with higher levels of clostridia and lower levels of bifidobacteria.10–12 Based on these data, “harmless” microbial agents, that is, probiotics, have been test administered for their efficacy in the treatment and prevention of allergy in infants. Probiotics are defined as products or preparations containing viable microorganisms in numbers thought to alter the host's microflora, thereby exerting beneficial health effects.13 One strain of probiotics, Lactobacillus rhamnosus (Lactobacillus GG [LGG]: American Type Culture Collection 53103), has proved safe at an early age.
So far there has been only 1 published double-blind, randomized, placebo-controlled trial that showed a preventive effect of LGG on atopic disease in a population at high risk. The frequency of atopic dermatitis (AD) was reduced by half in infants given probiotics compared with those on placebo at the ages of 2, 4, and 7 years.14–16 However, no difference was observed in the number of sensitized children between the groups. Furthermore, the mechanism of action is largely unexplained.
Recently, these results have also been questioned by a trial conducted by an Australian group, who reported no difference in the development of AD but observed increased sensitization to allergens in neonates supplemented with a different strain of Lactobacillus.17 Therefore, we used a prospective double-blind, placebo-controlled study to verify the importance of probiotics for primary prevention of atopic diseases. We hypothesized that LGG has a comparable clinical efficacy compared with the data reported by Kalliomäki et al.14–16
Participants and Study Design
The study design of this double-blind, randomized, placebo-controlled trial was similar to the concept of the Finnish study published by Kalliomäki et al.14 Two modifications of the protocol of Kalliomäki et al14 were made. First, we continued to supplement LGG for 3 months after birth exclusively to the breastfeeding mothers and the following 3 months only to the neonates. Second, pregnant women in our population commenced with LGG or placebo for 4 to 6 weeks, whereas Finnish mothers received supplementation only during the last 4 weeks of pregnancy. Pregnant women were recruited from outpatient departments of local gynecologists in Freiburg between July 2002 and June 2004. Inclusion criteria were as follows: (1) an uneventful pregnancy of the mother; (2) absence of underlying chronic disease of the mother (ie, diabetes mellitus, rheumatoid arthritis, or chronic infectious disease); and (3) a family history of atopic disease; that is, ≥1 family member (mother, father, or older sibling) with AD, allergic rhinitis, or asthma and a confirmed allergic sensitization against an inhalant allergen.
On the basis of our sample-size calculation before the study, where the efficacy of LGG supplementation was assessed based on the data published by Kalliomäki et al,14 50 families had to be researched in each group. At the beginning of the study, 105 mothers were randomly assigned to receive 2 capsules of placebo (microcrystalline cellulose) or L rhamnosus GG (American Type Culture Collection 53103) containing 5 × 109 colony-forming units of LGG (Infectopharm, Heppenheim, Germany) daily for 4 to 6 weeks before expected delivery. Randomization was performed in blocks of 4 according to a computerized randomization list. LGG and placebo capsules were matched for appearance, taste, smell, and packing.
Capsules were taken postnatally for 6 months. Directly after delivery, breastfeeding mothers took the capsules; otherwise, children received the agents (n = 2 in the LGG group; n = 3 in the placebo group). After 3 months, capsules were given only to the neonates. The capsule contents were mixed with water and then administered with a spoon.
Questionnaires were distributed to the parents during the neonatal period and when the children were 6, 12, and 24 months of age. All of the children were examined at the age of 24 months. In the case of suspected eczema, children were also seen at the age of 12 months. The primary study end point was the manifestation of an atopic disease during the first 2 years of life. Children were grouped into those with this disorder (children with AD) and those without (healthy children). Secondary end points were the severity of AD (Scoring Atopic Dermatitis Index), the cumulative incidence of AD, the incidence of recurrent (>5 episodes) wheezing bronchitis, total immunoglobulin E (IgE) and the number of children with sensitization to an inhalant allergen at the age of 2 years. The recruitment of pregnant women started on July 1, 2002, and ended on June 30, 2006. The last follow-up visits at the age of 2 years were performed during September 2006. The study was approved by the ethics committee of the University of Freiburg (local ethics committee of the University of Freiburg, code No. 46/02; 28.03.2003). Written informed consent was obtained from children's parents.
Clinical Examination and Questionnaires
The primary end point of our study was the manifestation of an AD at the age of 2 years. The physicians (Drs Kopp and Hennemuth) who conducted the physical examinations and diagnoses of atopic disease were unaware of the contents of the capsules until the end of the study in September 2006. To assure validity and comparability of the clinical assessment, the physicians underwent a standardized training program. Physical examination included inspection of skin, eyes, ears, and nose; auscultation of heart and lungs; palpation of abdomen; and assessment of neurologic development. Parents were asked about any signs and symptoms in their child that might be related to atopic disease: redness, dryness, oozing, and scratching (itching) of the skin; redness, discharge, sneezing, and rubbing (itching) of the eyes and nose; and coughing, wheezing, and shortness of breath. AD was confirmed by pruritus, facial or extensor involvement, or both, and chronic relapsing course by using United Kingdom Working Party criteria.18 The criterion “chronic relapsing course” was fulfilled if the child had had eczema for ≥1 month. The severity scoring of atopic dermatitis (Scoring Atopic Dermatitis Index) was used to assess eczema severity.19 Possible confounders that could influence the relationship of probiotics and allergic disease were obtained by means of a questionnaire. These data included information about other clinical diseases and common exposures (vaccination, infection, episodes of wheezing, period of breastfeeding, and use of antibiotics), the home environment (including carpeting, siblings, and pets), and recurrent obstructive bronchitis, which was diagnosed if the child had had ≥5 episodes of wheezing bronchitis during the first 2 years.
Treatment codes were kept by the supplier until data had been collected and analyzed, that is, until September 2006. Normally distributed data are expressed as means with 95% confidence interval (CI) and skewed data as geometric means with 95% CI. Values were compared between the groups by unpaired t test (Wilcoxon test). The χ2 test was used to compare proportions between the groups. A P value of <.05 was regarded as statistically significant. To determine the odds ratio (OR) of developing AD in the probiotic group compared with placebo, logistic regression was used. Finally, we calculated a linear regression model to describe a possible effect of LGG supplementation on the development of AD. Other than the LGG or placebo supplementation, the model includes mode of delivery; birth weight; gender; family history of mother, father, and siblings; use of antibiotics; and duration of breastfeeding as independent variables. The regression model gives parameter estimates for the differences between the groups as a description of possible LGG effects. Analyzing our data, information regarding the primary end point of our study was mandatory. Therefore, the analysis was performed in the “per protocol” group instead of the “intention-to-treat” population.
Characteristics of the Study Population
As shown in the trial profile (Fig 1), 108 pregnant women were recruited, and 105 were randomly assigned for the study. Of these, 11 withdrew from the study (Fig 1) during follow-up or were excluded from the analysis because of major protocol violation. A total of 94 mothers and their healthy term infants completed the study (89.5%), 50 in the LGG group and 44 in the placebo group.
There were no significant differences between the neonates of the 2 groups in gender, birth weight, gestational age, method of delivery, maternal or paternal age, or parity (Table 1). Furthermore, environmental variables, which might be a potential confounder for the development of atopic disease, did not differ between the groups. The number of older siblings, the lactation period, the time point of introduction of solids, passive smoke exposure, and the number of pets at home were equally distributed (Table 1). The atopic status of the family is detailed in Table 2. No difference in the number of atopic family members was noted between the groups (Table 2).
Effects of LGG on Clinical Outcomes
The primary aim of this study was to examine the effect of LGG supplementation on manifestations of AD at the age of 2. At that time, the rates of AD between the LGG group (28%; n = 14) and the placebo group (27.3%; n = 12; P = 0.93; Fig 2) were nearly identical. The risk of AD in children on probiotics relative to placebo was 0.96 (OR; CI: 0.38–2.33). Similarly, no significant difference was observed in the cumulative incidence of atopic symptoms during the first 2 years (Fig 3). In affected children, the difference of severity of AD between the LGG (median LGG: 18.5) and the placebo group (22.5; P = 0.8; Table 3) was not statistically significant. Factors that could account for possible confounding effects like mode of delivery, birth weight, gender, family history, use of antibiotics, or duration of breastfeeding were examined in a regression model. The parameter estimate for the differences between the groups was 1.18 (CI: 0.45–3.11; P = 0.73), indicating that possible confounding factors did not change the key study finding.
Secondary clinical end points were the occurrence of episodes with upper respiratory tract infections, fever, use of antibiotics, and number of wheezing bronchitis. There was no evidence that LGG provides protection from upper respiratory infections. Episodes of fever and the use of antibiotics during the first year of life were equally distributed between both groups (Table 3). No difference was observed between both groups in total IgE concentrations or numbers of specific sensitization to inhalant allergens (Table 3). The median IgE concentration was 17.9 kU/L (fifth to 95th percentile: 0.3–140.0 kU/L) in the LGG and 28.9 kU/L (fifth to 95th percentile: 1.3–279.0 kU/L) in the placebo group (Table 3). A specific sensitization against an inhalant allergen was observed in 4 (LGG) and 5 (placebo) children at the age of 2 years.
Notably, children with recurrent (≥5) episodes of wheezing bronchitis were more frequently in the LGG group (26%; n = 13), as compared with the placebo group (9.1%; n = 4; P = 0.03, Fig 4). The administration of LGG was well tolerated without any notable adverse effects attributable to the supplementation of probiotics (data not shown).
To confirm the hypothesis that LGG is effective in the prevention of early atopic disease, as based on the results of the clinical study of Kalliomäki et al,14–16 a second validating study in an independent population is deemed to be necessary. Therefore, we used a prospective double-blind, placebo-controlled trial to reevaluate the findings of the Finnish study and to characterize potential effects of LGG. In contrast to previous published findings, we have shown that the probiotic LGG has no preventive effect on the development or the severity of AD at the age of 2 years in a German population of infants at high risk. Instead, there was a significantly higher risk of ≥5 episodes with wheezing bronchitis during the first 2 years in the LGG group, as compared with placebo.
Over the last few decades, public reports have shown increasing enthusiasm for the potential health effects of probiotics. However, clinical studies of the efficacy of probiotics showed no therapeutic effect in allergic rhinitis20 and little or no benefit in the therapy of AD.21,22
Until now, 3 clinical studies have been published that have focused on the use of probiotics for primary prevention. The Finnish study of Kalliomäki et al14 was the first report to describe that the frequency of AD in the probiotic group was half that of the placebo. We adapted the protocol of Kalliomäki et al14 and continued to supplement LGG for 3 months after birth to the breastfeeding mothers and the following 3 months only to the neonates. This modification was made to achieve a more consistent probiotic delivery. Second, Finnish mothers received supplementation during the last 4 weeks of pregnancy, whereas pregnant women in our population commenced with LGG or placebo for 4 to 6 weeks. We extended the prenatal supplementation period, because a 4-week period is thought to possibly be too short for suspected in utero effects of LGG supplementation.
However, the discrepancy between the data of Kalliomäki et al14 and our study cannot be explained by minor deviations in the study design. We observed no preventive effect of LGG on the development of AD in our population nor any trend in this direction. There are some differences in confounding variables that probably contributed to the discrepancy in these clinical findings. First, the number of atopic family members per individual was 1.64 (LGG) and 1.75 (placebo) in our population as compared with 1.19 (LGG) and 1.11 (placebo) in the Finnish population. This may lead to our population being of higher risk compared with the Finnish population, which might account for the differing results. Second, we recruited more infants with older siblings (50% [LGG] and 68.2% [placebo] compared with 37% [LGG] and 34% [placebo] in the Finnish study), which might be also a potential confounder. Finally, the Finnish and German populations are of different genetic background. We recruited a fairly homogeneous ethnic group, with a proportion of >95% of mothers and fathers born and raised in Germany. No data were presented by Kalliomäki et al14 with respect to the ethnic background of the Finnish population. In a second preventive study in a Finnish population, a mixture of prebiotic and 4 probiotic bacterial strains failed to reduce the incidence of allergic disease by the age of 2 years but reduced the manifestation of AD eczema from 32% to 26%.23 Even if the effect in the intervention group was much smaller and the supplementation strategy was different compared with the study of Kalliomäki et al,14 these data support our hypothesis that the susceptibility to probiotics might differ between individuals because of different genetic backgrounds.
In support of our findings, a recently published, double-blind, placebo-controlled trial in a large Australian population failed to demonstrate a preventive effect of probiotics on atopic disease.17 Taylor et al17 supplemented L acidophilus directly postnatally to infants with an allergic mother for 6 months. Similar to our results, not even a trend of different rates of AD was observed between the groups. After 12 months, 34 (43.2%) of 88 children in the probiotic group and 34 (39.1%) of 87 children in the placebo group developed an AD. The higher rate of children with AD in both groups compared with our study might be because of the higher risk for atopic disease in Australian children with maternal atopy. It is striking that the proportion of children with AD and allergic sensitization in the study of Taylor et al17 was significantly higher in the probiotic group. Moreover, a significantly greater proportion of children in the probiotic group developed wheezing during the first 6 months of life. This finding also corresponds with data of our population, where we observed a significantly higher proportion of children with recurrent (≥5 episodes) wheezing bronchitis during the first 2 years in the LGG group. It is interesting that, in the study of Kalliomäki et al,15 a nonsignificant trend toward an increased rate of atopic disorders in the group treated with Lactobacillus, as compared with those given placebo at the age of 4 years, was also observed: 10 of 53 and 5 of 54 developed allergic rhinitis, and 3 of 53 and 1 of 54 developed asthma, respectively. Allergic rhinitis and asthma also tended to be more common in the probiotic group at the age of 7 years.16 However, one must be careful when concluding that there is a causal relationship between Lactobacillus supplementation and wheezing bronchitis. In our trial, episodes of wheezing were based on both diagnosis by a physician and symptoms recorded by the parents. It was not feasible to confirm every episode of wheezing by a doctor of the study center. Nevertheless, with respect to the lacking data showing efficacy of probiotics for primary prevention and higher incidence of wheezing in the Lactobacillus group observed in 2 independent studies, probiotics should not be recommended for allergy prevention.
Additional data from the Taylor et al17 study point in the same direction regarding allergic sensitization, also suggesting that the use of probiotics for primary prevention must be exercised with caution. At the age of 12 months, the rate of sensitization to common allergens was significantly higher in the probiotic group. In our population, no significant difference in total IgE levels or number of children sensitized against inhalant allergens was observed. However, we tested a panel of inhalant allergens and screened for food allergens only on individual request. In line with our data, Abrahamsson et al24 could not confirm a preventive effect of probiotics on infant eczema in a recently published study. However, he observed that the treated infants had less IgE-associated eczema at 2 years. Moreover, skin prick test reactivity was also less common in the treated group than in the placebo group, but this difference reached significance only for infants with allergic mothers. Again, there are differences in the administered probiotic (L reuterii) and the supplementation period (1 year), which might explain the discrepancy compared with the Taylor et al17 study and our study.
Based on the effects of LGG published by Kalliomäki et al,14 we calculated 50 families to be researched in each group. However, because 11 families from the originally randomly assigned 105 families were lost to follow-up, we ended up with 50 and 44 families in each group with complete data sets, so that we included fewer patients than planned. However, even if all of the randomly assigned families (LGG n = 54; placebo n = 51) had finished the study, and even if all the children who left the study had shown favorable results of LGG supplementation, this would not have changed our key end point (data not shown).
In the placebo group, we observed more families with older siblings, a slightly longer duration of breastfeeding, and less furry pets at home, factors that are discussed as allergy-protective determinants. On the other hand, children in the placebo group had slightly more siblings with a history of AD, but all of the differences between the groups were not statistically significant. Moreover, by using a regression model, all of these variables did not change the key study finding. Until now, many mechanisms have been proposed for the potential favorable effects of probiotics on allergy prevention. In addition to in vitro modulation of the cytokine response, probiotics have been observed to increase production of secretory immunoglobulin A,25 to improve the barrier function of the intestinal mucosa,26 and to enhance enzymatic degradation of dietary antigens by enzymes from probiotics.27 However, the exact modes of action are not yet known.
Supplementation with LGG during pregnancy and early infancy did not reduce the incidence of AD. In contrast, more children receiving LGG developed episodes of recurrent bronchitis. Thus far, an allergen-preventive effect of LGG could not be consistently established. Therefore, LGG cannot be generally recommended for primary prevention. Additional studies should clarify whether any susceptible subgroups exist, and how these subgroups benefit from a specific supplementation with probiotics can be defined.
The study was supported by the University of Freiburg and by Infectopharm (Heppenheim, Germany).
Other members of the study group were: Michael Goldstein, Alicija Dietschek, Michael Moseler, Gabriele Ihorst, Hannah Arends, Franziska Feindt, Julia Soffke, and Claudia Seydi.
We thank the children and their parents for their tireless cooperation. We are indebted to Maria Lickert for her reliable work in the laboratory, as well for her active support during the study.
- Accepted September 21, 2007.
- Address correspondence to Matthias Volkmar Kopp, MD, University Children's Hospital, Mathildenstrasse 1, D-79106 Freiburg, Germany. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
We affirm that the study has been registered at a qualified national registry, the registry for clinical studies at the University of Freiburg. Registration number is UKF000505.
What's Known on This Subject
Epidemiologic data show that children with allergy have a different intestinal flora from healthy children. Based on these data, “harmless” microbial agents, that is, probiotics, have been tested for their efficacy in the prevention of allergy in infants. However, the value of probiotics for primary prevention is controversial.
What This Study Adds
Supplementation with Lactobacillus GG during pregnancy and early infancy did not reduce the incidence of atopic dermatitis or alter the severity of atopic dermatitis in affected children but was associated with an increased rate of recurrent episodes of wheezing bronchitis. Therefore, Lactobacillus GG cannot be recommended generally for primary prevention.
- ↵Russell G. The childhood asthma epidemic. Thorax.2006;61 (4):276– 278
- ↵Strachan DP. Hay fever, hygiene, and household size. BMJ.1989;299 (6710):1259– 1260
- ↵Prescott SL, Macaubas C, Holt BJ, et al. Transplacental priming of the human immune system to environmental allergens: universal skewing of initial T cell responses toward the Th2 cytokine profile. J Immunol.1998;160 (10):4730– 4737
- ↵Sudo N, Sawamura S, Tanaka K, Aiba Y, Kubo C, Koga Y. The requirement of intestinal bacterial flora for the development of an IgE production system fully susceptible to oral tolerance induction. J Immunol.1997;159 (4):1739– 1745
- ↵Havenaar R, Huis in 't Veld JHJ. Probiotics: a general view. In: Wood, BJB, ed. Lactic Acid Bacteria in Health and Disease. Vol 1 of The Lactic Acid Bacteria. Amsterdam, Netherlands: Elsevier Applied Science Publishers; 1992
- ↵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 (1):184– 191
- Copyright © 2008 by the American Academy of Pediatrics