None of the families have yet been notified of the findings from the Office for Human Research Protections, which safeguards people who participate in government-financed research. But the agency’s conclusions were listed in great detail in a letter last month to the University of Alabama at Birmingham, the lead site in the study. In all, 23 academic institutions took part, including Stanford, Duke and Yale.

The letter stated that the study did have an effect on which infants died and which developed blindness, and that those risks were not properly communicated to the parents, depriving them of information to make a decision about whether to participate.

The 1,300 infants who participated in the study, which took place between 2004 and 2009, and whose results were published in The New England Journal of Medicine in 2010, were born at just 24 to 27 weeks gestation, a very high-risk category that was already prone to death and eye disease.

Babies assigned to a high-oxygen group were more likely to develop eye disease and blindness, and babies assigned to a low-oxygen group were more likely to die. Ultimately, 130 babies out of 654 in the low-oxygen group died, and 91 babies out of 509 in the high-oxygen group developed an eye condition, which in serious cases can lead to blindness.

 The letter said that “the researchers had sufficient available information to know, before conducting the study, that participation might lead to differences in whether an infant survived, or developed blindness, in comparison to what might have happened to a child had that child not been enrolled in the study.”

The risk the consent form did mention was the far less significant one of abrasion of the infants’ skin by an oxygen monitoring device. It also said there was a possible benefit to participating — a decreased need for eye surgery depending on the group the infant was assigned to.

Richard B. Marchase, vice president for research at the University of Alabama at Birmingham, said in a telephone interview that the study kept the infants within the standard band of treatment for oxygen levels — 85 to 95 percent — and that its findings would help inform a discussion by the American Academy of Pediatrics this year about standards for oxygen levels.

But the federal office critical of the consent process said in its letter that the risks of participating in the trial “were not the same as those receiving standard care.” The office stated that the consent form should have explained that “there is significant evidence from past research indicating that the oxygen provided to an infant can have an important effect on many outcomes including whether the infant becomes blind, develops a serious brain injury or even possibly whether the infant dies.”

The consent forms were written by researchers at the University of California, San Diego, said a spokesman for the University of Alabama at Birmingham said, and were approved by the review boards of all 23 institutions in the study. It was financed by the National Institutes of Health.

Dr. Marchase said he had assured the office that in the future, “we will to the best of our ability let the subjects or their parents know as thoroughly as possible what previous studies suggest in terms of risk.” He added, “We are going to be very sensitive to that going forward as we look at these consent forms.”

He said that a similar group of infants born around the same time in the same hospitals who did not participate in the study, but were eligible, died at higher rates than those in the low-oxygen group. Those infants were not a control group in the study, but were similar to those in the study group, he said; they had a 24 percent mortality rate, compared with a 20 percent mortality rate for the infants in the low-oxygen group.

But Michael Carome, a physician who is the deputy director of the health research group at Public Citizen, a consumer advocacy organization, pointed out that a study in Pediatrics in 2012 that looked closely at the two groups, determined that the infants who did not participate, or were not chosen, were in worse health, and therefore more prone to die.

“To try to retrospectively make comparisons like this is very fraught because of the difference in the populations,” he said. “That’s not a valid way of doing science.”

Dr. Marchase replied that the conclusions of the Pediatrics study “are not universally accepted among those who know the field best.”

Medical ethics in the United States have improved drastically since the 1930s, when scientists conducted research on impoverished black men in Alabama in the Tuskegee syphilis experiment without explaining the objective of the study to the men or helping them get treatment once it became widely available. The experiment led to a public outcry and numerous patient protections, like the requirement of “informed consent.”

Omitting risks from a consent form these days is relatively rare, experts said, as institutions often want to reduce the chances that they will be sued. The result is voluminous forms that do more to protect the institution than to empower the potential subject.

“Informed consent has been captured by worries about legal liability, so risks tend to come billowing forward like a huge fog,” said Arthur Caplan, head of the division of medical ethics at New York University Langone Medical Center. “It’s a truth dump, so they are covered should something go wrong.”

Dr. Caplan said often in such emergency medical situations, parents often rely more on what doctors say in deciding whether to participate, than on the fine print of a consent form.

Clinical treatment of premature infants has a troubled history. Attempts to treat them with higher oxygen levels that were thought to improve their odds of survival led to many cases of blindness. Premature babies need oxygen because their lungs are underdeveloped and they often need help breathing.

Doctor generally make treatment decisions based on the needs of the infant, often in consultation with parents, experts said.

“It’s usually calibrated to how the baby is doing,” said Norman Fost, professor of pediatrics at the University of Wisconsin School of Medicine and Public Health. “It’s not a fixed formula.”

A spokesman for the University of Alabama at Birmingham said that the treatment for the babies in the study was flexible, based on decisions by doctors, and that the infants were not held in rigid bands of oxygen levels.

But Dr. Carome, who alerted news media outlets to the letter on Wednesday, said the purpose of the study was to determine optimal oxygen levels and, in order to gauge that, researchers grouped some infants in higher zones and some in lower zones.

“They were trying to confine them to a specific region in the curve,” he said. “Making these changes does have a clinical effect.”