Although nitric oxide can halve the risk of developmental problems in premature babies with undeveloped lungs, the gas could be harmful to newborns who are too small and too ill, according to two studies published on Thursday in the New England Journal of Medicine.

The findings mean doctors should avoid using the treatment during the first week of life in premature babies who are seriously ill and weigh less than 1,000 grams (2.2 pounds), Richard Martin and Michelle Walsh wrote in an editorial in the same publication where the studies appear.

“Short-term use of inhaled nitric oxide cannot be considered an effective rescue therapy for very preterm infants with profound respiratory failure,” said Martin and Walsh, both of Rainbow Babies and Children’s Hospital in Cleveland.

The good news for slightly larger, healthier premature infants is that the gas makes the lungs mature rapidly, cutting the risk of developmental problems in half. That could help about half of the 60,000 premature infants born each year with dangerously underdeveloped lungs.

“Babies born at two pounds (0.97 kg) have only a 50 percent chance at 2 years of age of being considered totally normal,” said pediatrician Michael Schreiber of the University of Chicago, whose team tracked 138 children.

They found that by the second birthday, 24 percent of the children given nitric oxide had problems such as blindness, hearing loss and cerebral palsy, against 46 percent for those 2-year-olds who did not receive the gas.

“Not only does nitric oxide extend life in a large group of premature infants, it also improves the quality of life for the children and their parents,” Schreiber said.

But findings in the second study qualified that conclusion.

The research team led by Krisa Van Meurs of the Stanford University School of Medicine found that the 210 premature babies with severe respiratory failure who got nitric oxide were just as likely to die or suffer lung damage as the 210 who were given a placebo.

The study, however, focused on babies who tended to be smaller and sicker than the ones in the Schreiber research.

The Van Meurs team also found that infants weighing over 1000 grams (35 ounces) benefited from nitric oxide, but those weighing less were more likely to die or develop severe brain bleeding than those who got standard care.

The Schreiber study was paid for in part by INO Therapeutics, which makes the nitric oxide treatment. INO is a division of Germany’s Linde AG . The Van Meurs study was financed largely by the National Institute of Child Health and Human Development.