WEDNESDAY, April 27 (HealthDay News) -- Babies born prematurely need supplemental oxygen to survive, but just how much has been a matter of debate.

A new report finds higher oxygen concentrations improve survival, but not without risks.

There are dangers of giving too much oxygen, severe eye damage -- even blindness -- being the primary one, as well as lung and brain damage. But there are also dangers in giving too little oxygen, including death, a neonatal expert said.

"The problem is we are dealing with the risks associated with high oxygen, which everybody has been trying to avoid up until now," said Dr. Eduardo H. Bancalari, director of the University of Miami Miller School of Medicine's Division of Neonatology.

"I think that in an effort to avoid the complications of high oxygen, the pendulum had gone too far to the other side," said Bancalari, who was not involved with the new research. "Babies have been receiving concentrations that are too low that are also associated with complications -- in these cases increased death."

Bancalari noted that getting the right concentration of oxygen isn't easy. "You are walking a tightrope," he said. "The problem is to find the right balance, which is difficult because it is a continuously moving target."

In the April 28 issue of the New England Journal of Medicine, Dr. Ben Stenson, of the University of Edinburgh in Scotland, and colleagues report on two studies that included 3,631 infants born before 28 weeks of gestation. Oxygen concentrations of 91 percent to 95 percent improved survival, compared with concentrations of 85 percent to 89 percent, the researchers found.

After 36 weeks, 14.4 percent of the infants on higher concentrations of oxygen had died, compared with 17.3 percent of those on the lower concentrations, the results showed.

Looking only at 1,055 infants in a trial done in Australia and the United Kingdom, the so-called BOOST II trial, 13.3 percent of the babies on high concentrations of oxygen died, compared with 21.8 percent of the children on the lower concentration, the researchers reported.

Based on these findings, both the BOOST II trial and a U.S.-based trial called the SUPPORT trial stopped recruiting new patients, the researchers noted.

Stenson's team concluded: "These data allow no inferences about risks and benefits of other targets. Until longer-term data on survival and morbidity are available, we consider it prudent not to target an [oxygen saturation] of 85 to 89 percent in infants born earlier than 28 weeks of gestation."

Bancalari thinks the message for doctors is one of balance.

"We cannot go from one extreme to another; we have to stay in a range that achieves both avoidance of too much or too little oxygen. That's easy to say, but in the clinical setting is very difficult to accomplish," he said.

Bancalari said his neonatal unit never went to the low end of oxygen concentrations, but rather uses a range of 88 percent to 93 percent.

More information

For more on premature birth, visit the March of Dimes.