Baby formula which claims to reduce the risk of allergies doesn't work, according to new research.
The study, published in The BMJ, reviewed figures from 37 different trials into hydrolysed baby formula - a type of formula treated with heat to break down the milk proteins.
The idea is that giving the formula to children at risk of conditions such as milk allergy and eczema, instead of standard formula, can reduce the chance of infants developing the conditions.
The US Food and Drug Administration has previously approved a manufacturer's claim that a partially hydrolysed formula may reduce risk of eczema in some infants.
A review in 2006 found limited evidence that feeding a baby hydrolysed formula may reduce the risk of milk allergy.
Current baby feeding guidelines, including those in North America, Australasia, and Europe, recommend hydrolysed cows' milk formula, in place of standard infant formula, to prevent such diseases in infants during the first months of life.
But the new research showed there was no statistically significant reduction in risk of these conditions among babies using hydrolysed formula.
The paper also revealed conflicts of interest in many of the studies - due to financial links with baby formula manufacturers.
Senior author Dr Robert Boyle, of Imperial College London, said: "Despite parents being advised these hydrolysed milk formulas may reduce the risk of conditions such as milk allergy and eczema, we found no evidence to support these claims."
In the paper, which was funded by the UK Food Standards Agency, the team analysed studies that included more than 19,000 participants.
Dr Boyle said the new study is the most complete and robust assessment of all the evidence to date.
The study found there was no significant reduction in risk of developing eczema, wheezing, or food allergy.
Some of the trials also investigated the risk of developing type 1 diabetes, but no link was found.
Dr Boyle said: "Not only did we find no evidence of reduced risk from hydrolysed formula, but we found very few studies which were methodologically sound and without a conflict of interest.
"For instance, in some of the studies all babies were started on the formula at birth, or a few days after. This raises questions about whether enough was done to promote breastfeeding to the mothers in those studies."
The study's senior statistician Professor Jo Leonardi-Bee, of Nottingham University, said the results may not be representative of all studies conducted in the area.
She said: "Our research suggests that there was evidence of publication bias, where some studies that showed formula milk didn't actually reduce allergies may not have been published."
She said further studies into infant milk formula must be more robust, adding: "We need to ensure future studies are carefully designed, so that our guidance to parents is based on reliable evidence."