Soy formula for prevention of allergy and food intolerance in infants

When babies are not exclusively breastfed, evidence suggests that using a soy formula instead of a cow's milk formula does not reduce allergies in infants and children. Infant formulas have been designed to try to lower the chances of developing allergy or food intolerance. These formulas include hydrolysed cow's milk and soy formulas. A review of trials found that in infants at high risk of allergy who are unable to completely breastfeed, there is no reduction in allergies in later infancy and childhood associated with feeding soy formula compared to a cow's milk formula. No eligible studies were found that compared a soy with a hydrolysed protein formula.

Authors' conclusions: 

Feeding with a soy formula cannot be recommended for prevention of allergy or food intolerance in infants at high risk of allergy or food intolerance. Further research may be warranted to determine the role of soy formulas for prevention of allergy or food intolerance in infants unable to be breast fed with a strong family history of allergy or cow's milk protein intolerance.

Read the full abstract...
Background: 

Allergies and food reactions in infants and children are common and may be associated with a variety of foods including adapted cow's milk formula. Soy based formulas have been used to treat infants with allergy or food intolerance. However, it is unclear whether they can help prevent allergy and food intolerance in infants without clinical evidence of allergy or food intolerance.

Objectives: 

To determine the effect of feeding adapted soy formula compared to human milk, cow's milk formula or a hydrolysed protein formula on preventing allergy or food intolerance in infants without clinical evidence of allergy or food intolerance.

Search strategy: 

The standard search strategy of the Cochrane Neonatal Review Group was used. Updated searches were performed of the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 1, 2006), MEDLINE (1966 - March 2006), EMBASE (1980 - March 2006), CINAHL (1982 - March 2006) and previous reviews including cross references.

Selection criteria: 

Randomised and quasi-randomised trials that compare the use of an adapted soy formula to human milk, an adapted cow's milk or a hydrolysed protein formula for feeding infants without clinical allergy or food intolerance in the first six months of life. Only trials with > 80% follow up of participants and reported in group of assignment were eligible for inclusion.

Data collection and analysis: 

Eligibility of studies for inclusion, methodological quality and data extraction were assessed independently by each review author. Primary outcomes included clinical allergy, specific allergies and food intolerance. Where no heterogeneity of treatment effect was found, the fixed effect model was used for meta-analysis. Where significant or apparent heterogeneity was found, results were reported using the random effects model and potential causes of the heterogeneity were sought.

Main results: 

Three eligible studies enrolling high risk infants with a history of allergy in a first degree relative were included. No eligible study enrolled infants fed human milk. No study examined the effect of early, short term soy formula feeding. All compared prolonged soy formula to cow's milk formula feeding. One study was of adequate methodology and without unbalanced allergy preventing co-interventions in treatment groups. One study with unclear allocation concealment and 19.5% losses reported a significant reduction in infant allergy, asthma and allergic rhinitis. However, no other study reported any significant benefits from the use of a soy formula. Meta-analysis found no significant difference in childhood allergy incidence (2 studies; typical RR 0.73, 95% CI 0.37, 1.44). No significant difference was reported in one study in infant asthma (RR 1.10, 95% CI 0.86, 1.40), infant eczema (RR 1.20, 95% CI 0.95, 1.52), childhood eczema prevalence (RR 1.10, 95% CI 0.73, 1.68), infant rhinitis (RR 0.94, 95% CI 0.76, 1.16) or childhood rhinitis prevalence (RR 1.20, 95% CI 0.73, 2.00). Meta-analysis found no significant difference in childhood asthma incidence (3 studies, 728 infants; typical RR 0.71, 95% CI 0.26, 1.92), childhood eczema incidence (2 studies, 283 infants; typical RR 1.57, 95% CI 0.90, 2.75) or childhood rhinitis incidence (2 studies, 283 infants; typical RR 0.69, 95% CI 0.06, 8.00). One study reported no significant difference in infant CMPI (RR 1.09, 95% CI 0.45, 2.62), infant CMA (RR 1.09, 95% CI 0.24, 4.86), childhood soy protein allergy incidence (RR 3.26, 95% CI 0.36, 29.17) and urticaria. No study compared soy formula to hydrolysed protein formula.