Should long-term prophylactic use of probiotics for infants and young children give cause for concern?

Abstract

In the past decade, several probiotic strains have been investigated for health benefit effects and Lactobacillus rhamnosus GG (LGG) has been one of the most studied probiotic strains. Short-term beneficial effects on viral infectious diarrhoea have been reported in a number of studies, but a prophylactic effect on diarrhoea has not been documented. Furthermore, while some studies have reported a prophylactic, or even a curative, effect of LGG on atopic eczema in young children, more recent studies have not confirmed these findings. Quite the contrary, an increased incidence of allergic sensitization has been reported among children receiving LGG-supplemented infant formula feed at an early age. To date, there is thus no certain documented prophylactic effect of LGG on any medical condition in children. When used as a supplement in infant formula and baby foods, LGG is intended for long-term daily use in children as opposed to a short-term, specific, treatment. Furthermore, the targeted consumer group includes children below the age of 12 months. These two aspects demand particular consideration with regard to the lack of knowledge of the effects of long-term treatment with large doses of live bacteria both on the ecology of the microbiota of the gastrointestinal tract and on its immunomodulating properties. Neither of these systems is fully matured in infants and small children, and may therefore be particularly vulnerable. Probiotics may also alter intestinal metabolism due to their enzymatic activities, transfer genes coding for resistance to antimicrobials and have other adjuvant effects of which we yet know very little. In conclusion, we consider that there is insufficient scientific knowledge to support manipulation of the immune system in a predictable manner by administration of probiotics to young children. Manipulation at an early age is of particular concern, as the effects may be non-reversible.

• probiotics
• diarrhoea
• infant formula
• baby foods
• Lactobacillus rhamnosus GG
• microbiota
• atopic eczema
• allergic sensitization

Introduction

The term probiotic derives from Greek, meaning ‘for life’. The microorganisms commonly used as probiotics for humans belong to Bifidobacterium and the lactic acid bacteria group, Lactobacillaceae, which have the ability to metabolize carbohydrates to lactic acid, thereby lowering the micro-environmental pH. A normal intestinal microbiota has been shown to be beneficial to the functioning of the normal intestinal mucosa and mucosal defence systems, and can inhibit the colonization of pathogenic microorganisms [1]. Thus, the possibility to control the process of newborn intestinal colonization is an appealing idea. Studies on the neonatal intestinal flora and differences in colonization patterns have influenced feeding practices and, more recently, formula development.

During the past 10 years, certain specific strains of lactobacilli, including Lactobacillus rhamnosus GG (LGG), have been included in a number of probiotic products [2]. A number of manufacturers are interested in establishing the use of probiotics for infants and children. This review article is based on two reports requested by the Norwegian Food Safety Authority. The reports were carried out by the Norwegian Scientific Committee for Food Safety in 2005 and 2007 and are available at www.vkm.no.

Intestinal microbiota in infants

The intestinal microbiota comprises a dynamic mixture of microorganisms. It develops over time, determined by interaction between genetic factors, contact with the environment, diet and disease. The abilities of the bacteria to modulate immune function, metabolize potential carcinogens and provide a direct barrier to invasion of the gut by pathogenic microorganisms are examples of the diverse functionality of the gut microbiota. Host-related factors have a major impact on the bacterial flora in the human gastrointestinal (GI) tract. Every individual has a unique microbiota and even homozygotic twins differ in the composition of their microbiota [3].

The colonic population of breast-fed babies is dominated by lactic acid bacteria and bifidobacteria, with comparatively few Bacteroides, Clostridium or coliforms. The low faecal pH of infants who are breast-fed may promote bacteria such as Lactobacillus and Bifidobacterium, which are more acid-tolerant. In contrast, greater variability in the gut microbiota is found in formula-fed babies, tending to contain larger numbers of Bacteroides, Clostridium, lactobacilli, gram-positive cocci, coliforms and other groups, all present in fairly equal proportions [4]. Bifidobacteria are also present in over half of formula-fed babies, but the number of these bacteria is much lower (about one-tenth) compared with breast-fed infants [5]. The mechanisms responsible for the differences between the microbiota of infants fed human milk and those fed with modern formula are probably numerous and are not fully understood. Immunological factors such as secretory IgA (sIgA) and lysozyme present in human milk may modulate the growth of some bacteria.

Many factors are involved in the development of a healthy intestinal flora, which is subsequently of importance for the development of normal immune responses to antigens and oral tolerance. The intestinal microbiota comprises a highly complex dynamic ecosystem in which modulation of one aspect may have many consequences downstream of the initial events. In infancy, this ecosystem may be particularly vulnerable to destabilization. Major factors that influence intestinal microbiota include genetic factors, mode of delivery at birth, early exposure to antimicrobial agents and diet. The long-term consequences of changing one single factor that affects the complex environment of the intestine are not known and have not been adequately studied. This area requires further research.

Role of LGG in disease prevention and treatment

One of the primary research areas for the use of probiotics in children has been in the treatment and prevention of diarrhoea. Regarding treatment, a shortening of the duration of acute viral diarrhoea, e.g. rotavirus-associated diarrhoea, has been demonstrated [6–11]. However, the preventative effect of probiotics on diarrhoea has not been adequately studied.

The use of probiotics in the treatment of atopic eczema in children has been investigated in several studies [12–17]. The effect of LGG alone [13] or the combination of LGG with L. rhamnosus 19070-2 and L. reuteri DSM 122460 [14], or LGG and Bifidobacterium lactis Bb-12 [12], has been studied. However, these studies were flawed by the small number of children in each group, and their positive findings were not confirmed subsequently in two larger studies carried out by other research groups [18], [19]. No overall improvement in eczema as assessed by scoring atopic dermatitis (SCORAD) [20] was noted in these studies, but in a subgroup of children with IgE sensitization a significant, but only moderate, effect was observed [13]. Infants who had developed atopic eczema while being exclusively breast-fed were weaned to a probiotic-supplemented hydrolysed formula [12]. After 2 months of treatment, SCORAD was significantly decreased in the infants who received formula supplemented with B. lactis Bb-12 or LGG, compared with infants who received unsupplemented hydrolysed formula. LGG and B. lactis Bb-12 were equally effective.

Conflicting results have been obtained in studies on treatment of atopic eczema by LGG administration. Studies by Kalliomaki et al. [21], Viljanen et al. [13] and Weston et al. [22], which indicated a beneficial effect, have been contradicted by data obtained in more recent studies [18], [19], [23]. In a randomized intervention trial, Folster-Holst et al. [19] studied the effect of LGG supplementation for 8 weeks in 54 infants aged 1–55 months with moderate to severe atopic eczema. No significant difference in the severity of the eczema was detected between the placebo and LGG treatment groups during the treatment period, as evaluated by SCORAD, pruritus, sleeplessness, use of corticosteroids and ECP (eosinophil cationic protein) levels in faeces. Among those children with a positive test for IgE sensitization, there was a SCORAD reduction of 6.9 in the LGG treatment group and 8.6 in the placebo group. However, randomization may have been unfortunate in this trial, as more allergic children were placed in the placebo group, particularly subjects with food allergy and rhinitis. This was a very small study, and may therefore have lacked the power to show significant differences. However, none of the non-significant trends in the data provide support for a positive effect of LGG; in actual fact, rather the opposite. Another study [18] investigated high-risk children receiving a Lactobacillus strain (other than LGG) or a placebo for the first 6 months of life. No effect of the probiotic on the presence of atopic disease (AD) was noted at either 6 or 12 months. However, the rate of allergic sensitization was significantly higher at 12 months in the probiotic group than in the placebo group. The authors’ conclusion was that no effect of probiotic administration on the presence of AD was noted. Thus the possible benefit of probiotics on AD is unclear. The higher allergic sensitization rate, especially to cow's milk in the probiotic group, must be seriously considered. A high degree of allergic sensitization was also apparent from the previous studies by Kalliomaki et al. [21]. No side effects of the probiotic supplementation were noted, apart from the sensitization, but this study was not designed to investigate safety.

Brouwer et al. [23] conducted a 3-month randomized, double-blind, placebo-controlled study on infants less than 5 months old with AD. They administered a hydrolysed whey-based formula either as a placebo or supplemented with L. rhamnosus or LGG. No statistically significant difference between the groups that was attributable to probiotic supplementation was found by SCORAD, sensitization, inflammatory parameters or cytokine production. There is also some evidence from the above-cited studies that administration of probiotic bacteria to pregnant women, and to children via infant formula, may increase the degree of allergic sensitization in children with atopic eczema [18], [21], [24].

The effect of LGG in the prevention of early AD in high-risk children (i.e. with a family history of atopy) was investigated [21]. LGG was given daily in capsules to mothers for 2–4 weeks before expected delivery and after delivery to the breast-feeding mothers and directly to the infants (capsule content mixed with water and given with a spoon) for 6 months. In this prospective study from birth, atopic eczema was diagnosed at 2 years of age in 46 of the 132 children. The frequency of atopic eczema in the group receiving LGG was half that of the placebo group (15/64; 23% vs 31/68; 46%). No differences in allergic sensitization were observed between the placebo and probiotic groups as studied by skin prick tests, and both total and specific IgE levels were similar. The study did not address cell-mediated immunity. The children were re-evaluated at 4 years of age [24] and atopic eczema was diagnosed on the basis of a questionnaire and a clinical examination. Fourteen of 53 children who had received LGG showed clinical symptoms of atopic eczema compared with 25 of 54 children in the placebo group [24]. There was, however, a non-significant tendency for an increased incidence of asthma and pollen allergy among the children in the treated group, commented on by Niers et al. [25]. Furthermore, the incidence of diagnosed cow's milk allergy was doubled in the probiotic group [24], although these differences were no longer apparent at the age of 4 years. Clearly, there is a need for further studies to clarify whether these were chance findings or whether the risk of other allergic diseases increases coincidently with a decrease in atopic eczema. Previous studies suggesting that LGG might have a moderately beneficial effect in the prevention of atopic eczema in some subpopulations of sensitized children are not supported by data from a new study, although in this study the probiotic supplement used was not LGG [26].

Risk and safety aspects of LGG

Historically, the lactobacilli that are naturally associated with food have been considered safe with no pathogenic potential and are only very occasionally isolated as opportunist pathogens. However, probiotics may, on rare occasions, be responsible for systemic infections, altered intestinal metabolism, gene transfer and immunomodulation and adjuvant effects. Each of these aspects is discussed below [27].

Bacteraemia and sepsis attributable to LGG have been reported in two patients who received Lactobacillus LGG [28]. DNA fingerprinting revealed that the LGG strain was similar to the probiotic strain ingested by the patients. Both patients were children (one aged 6 weeks, the other 6 years). The first patient had been admitted for scheduled repair of a double-outlet right ventricle and pulmonal stenosis. The other patient, who had cerebral palsy, microcephaly, mental retardation and a seizure disorder that required feeding through a gastrojejunostomy tube, had been admitted for treatment of a urinary tract infection.

Currently, knowledge of the in vivo enzymatic functions of LGG in infants and babies is limited. The intestinal microbiota is considered to have a metabolic capacity equalling the liver and, in addition to its metabolism of dietary constituents, is involved in the synthesis of vitamins, the conversion of cholesterol, bile acids and both the formation and elimination of toxic and procarcinogenic products. Microbially derived enzymes can participate in reversing detoxification processes that have taken place in the liver [29]. In newborns, the detoxification capacity of the liver is low. It is thus essential to gain knowledge of the in vivo enzymatic properties of probiotic microbes, especially when they are intended for use in infants.

Bacteria belonging to the genus Lactobacillus are intrinsically resistant to vancomycin, and therefore vancomycin-susceptible strains of these species do not exist [30]. Based on the published genomic information, vancomycin-resistant lactobacilli, including LGG, have not been shown to contain van genes, which encode for resistance against vancomycin in enterococci and staphylococci. To date, there is no indication that intrinsically vancomycin-resistant lactobacilli can transfer vancomycin resistance genes to other species [30]. It has not been elucidated whether intrinsic vancomycin resistance in Lactobacillus is associated with any other resistance genes.

In common with other microorganisms and microbial products, probiotics have the potential to initiate and modulate immune responses. They interact with the Toll-like receptors present on cells belonging to the immune system. Toll-like receptors belong to a family of receptors involved in the recognition of a wide range of microbial molecules, such as lipopolysaccharides (LPS) of gram-negative bacteria and peptidoglycans of gram-positive bacteria. Different microbes, or combinations of different microbes or microbial products, act through different types of Toll-like and other receptors, resulting in the release of different cytokines. In accordance with such a model, it has been reported that different cytokine patterns were observed depending on whether LGG was given alone or in combination with other probiotics. Different species within a genus may also cause varying effects with regard to cytokines. This has been illustrated in studies on different species of Bifidobacterium [31]. Knowledge of the signals by which microbes and microbial products exert their effects is relatively recent and incomplete.

Several studies show apparently favourable effects on some of the cytokines and immune markers studied [32], [33]. However, considering the complexity of the immune system, studies showing a favourable shift in one or two markers fail to convince us of the overall benefit of the supplementation. The immune response towards Th2 in allergic diseases is documented as being skewed, and this was the basis for the Th1/Th2 paradigm, in which it was believed that treatment or prophylaxis of allergic diseases could be obtained if the balance could be shifted towards an increase in Th1 responses. However, as autoimmune diseases are based on Th1 responses, this balance is clearly very delicate. Furthermore, a multitude of different disease phenotypes exist, phenotypes that are associated with different complex patterns of cytokines; not even the allergic phenotypes are limited to Th2 types. Thus, an increase in specific cytokines that are believed to be advantageous may be associated with an increase in less favourable cytokines. Also of note is the parallel increase in food allergy [24] and allergic sensitization [18] concurrent with improvement in atopic eczema among children given LGG as part of clinical trials.

Increased insight into the fine-tuning of the immune system has recently been gained, and the crucial role of regulatory T cells has been established. Rather than shifting the balance from Th2 to Th1, an independent down-regulation of the Th2 responses may be needed to prevent allergy, and at the same time a down-regulation in Th1 responses may be needed to prevent autoimmune diseases [34].

One major concern regarding supplementation of infant food with LGG, or indeed any other probiotic strain, is the absence of knowledge on possible long-term effects of any factors that have an effect on the establishment of intestinal microbiota in early infancy. The microbiota of the intestine is not fully established before 2 years of age. Factors that influence the composition of the microbiota at an early stage may therefore permanently affect further development of the ecosystem. This has been demonstrated by studies on mode of delivery and intestinal microbiota that are outlined below. Caesarean delivery influences the development of early intestinal microbiota, since infants delivered by caesarean section do not receive microbes from the birth canal, and differences in intestinal microbiota after birth have been reported in children delivered by caesarean section compared with children delivered vaginally. Of even greater interest is the finding that the altered microbiota among caesarean-delivered children is long-standing [35] and may still be evident at 6 months [36], and even at 7 years of age [37]. This could indicate that an early exposure/non-exposure of infants to probiotic microorganisms may also have long-term effects, as it may influence the subsequent selection of colonizing microbes. While there is some limited information regarding the effects of single-species bacterial load on the short-term composition of the intestinal microbiota of the newborn infant, no publications have addressed the effects on the long-term composition of the intestinal microbiota of infants and small children following supplementation of their diet with LGG over a prolonged period of time. Such information is crucial.

Recent studies have shown that microbes have the ability to alter gene expression in enterocytes, although specific knowledge on LGG is not available. This may affect the diversity of microbiota in the intestine. In vitro data presented by Yan et al. [38] showing that LGG may regulate intestinal epithelial cell survival and growth strengthened our concerns about possible long-term effects of a dietary mono-bacterial supplement given to the age groups in question.

It is therefore clear that we do not at present have the scientific knowledge necessary to start manipulating the immune system in a predictable manner by administration of probiotics, and in fact any manipulation at an early age is of particular concern as the effects may be non-reversible.

In the case of small children (< 6 months), infant formula may represent their only food and may be a substantial part of their diet up to 1 year. When infant formula is supplemented with a probiotic bacterial culture this culture will be present in large numbers in all the food consumed. This situation is entirely different when compared with the consumption by adults of a portion of yoghurt supplemented with probiotic bacteria, an amount that would constitute only a small part of their total diet. The challenge to the resident intestinal flora is therefore likely to be much greater in small children than in adults and the counter-challenge from the resident flora is likely to be considerably less. This could allow greater proliferation in vivo of the administered probiotic strain.

In nutritional as well as pharmaceutical studies, the value of any claim is strengthened if the compound under study is administered in a format similar to that intended for the market. With the exception of one study [39], a general weakness in the new clinical trials publications is that they have not used infant formula or baby foods in which LGG is an ingredient. The LGG preparations used in these studies differ in composition from infant formula with added LGG, and therefore any effect of the food matrix (infant formula and baby foods without LGG) is unknown.

Concluding remarks

We conclude that LGG has been widely studied and characterized in short-term trials and that no immediate deleterious effects of LGG have been found. However, although LGG seems to have a foreshortening effect on infectious diarrhoea in infants and young children, there is no proof that it has any prophylactic effect on diarrhoea. Furthermore, the data on the possible beneficial effect of LGG on atopic disease in infants and small children are conflicting – the most recent studies concluding that it has no prophylactic effect on atopic eczema. Some studies even indicate that early inclusion of LGG into the diet may provoke allergic sensitization. Long-term effects on immune function in general, or on the gut in particular, when LGG is consumed on a daily basis are not known and warrant caution.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.