That breast-feeding reduces the risk of various childhood ills, some serious, has been confirmed by a comprehensive review of previous studies conducted by the Agency for Healthcare Research and Quality.

The agency, in a little-publicized study released in April, found breast-feeding is strongly linked with lower risk of ear infection, gastroenteritis, severe lower respiratory tract infections, atopic dermatitis, asthma in children 10 and younger, obesity, diabetes, childhood leukemia, sudden infant death syndrome and necrotizing enterocolitis. Women who nurse their babies have a reduced risk of breast cancer and ovarian cancer.

The large-scale analysis looked at more than 9,000 abstracts, 86 studies of infant or maternal health outcomes and 29 meta-analyses that themselves reviewed 400 individual studies.

The research did not confirm all the claims made about breast-feeding benefits, however. The study found no link between breast-feeding and an infant's cognitive development. And any relationship between heart disease and breast-feeding was unclear, as was any link between breast-feeding and infant mortality in developed countries, the report says. Impact of breast-feeding on cholesterol levels as an adult was also not shown.

The report also documented benefits for mothers who breast-feed infants, including reduced risk of type 2 diabetes and breast and ovarian cancer. Not breast-feeding or stopping early was linked to increased risk of maternal postpartum depression, although researchers warned that a woman suffering such depression might be less likely to start or continue breast-feeding. They found no link between lactation and later risk of osteoporosis and no clear evidence of the effect on postpartum weight loss.

The report also came with a number of warnings about making too strong a cause-effect connection, although it said it focused on studies that had accounted for "confounders," other things that could account for or influence the findings. For instance, the analysis discounted studies that examined a link with improved intelligence, because maternal intelligence and home environment were not considered and those could be confounders that account for the results.

Most of the data came from observational studies, rather than blinded randomized assignment of women into breast-feeding and not breast-feeding groups. Because of known benefits of breast-feeding, asking women not to breast-feed, even to further science, raises ethical issues.