WASHINGTON — A government-sponsored study of two measles vaccines, begun in 1989 during a major U.S. epidemic and conducted on nearly 1,500 minority infants in Los Angeles, failed to disclose to parents that one of the vaccines was experimental, federal health officials said Sunday.

"A mistake was made," said Dr. David Satcher, director of the Atlanta-based federal Centers for Disease Control and Prevention, one of the study sponsors. "It shocked me."

Satcher said in an interview that the CDC plans to contact all the families involved. He said he was very concerned that the events not fuel suspicion in the minority community of government-sponsored medical research.

"Every little mistake like that seeds the concerns of people," he said. "We need to move to a new level of assurance so people can really trust what we're doing."

None of the Los Angeles children, most of whom are now 5, was injured by the unlicensed vaccine, the CDC said. However, similar clinical trials conducted in Africa and Haiti with the vaccine raised questions about its relationship to an increased death rate among female infants who received the more potent of two dosages being studied. Those children died within two years after the vaccination. In light of the questions, the Los Angeles study was halted in 1991.

The inquiry into the measles research was conducted after a physician connected to a public-interest vaccine safety group raised questions. Satcher, who was not CDC director at the time the study began, said he concluded from his review that "there was no ill intent" on the part of the agency in not telling parents that the vaccine had not been licensed for use in the United States, which is why it is termed "experimental" in this country.

"But things sometimes fall through the cracks," he said.

CDC officials acknowledged that the omission was serious and attributed it to researchers' knowledge that the lesser doses of the unlicensed vaccine, known as Edmonston-Zagreb, or E-Z, had been used safely for decades outside the United States and that it had been recommended by the World Health Organization.

"Our doctors just didn't think of it as being 'experimental,' " said Barbara Reynolds, an agency spokeswoman.

The study was co-sponsored by Kaiser Permanente of California and was begun during a national measles epidemic, with California among the most stricken states.

The purpose was to compare E-Z with the Moraten vaccine, the standard vaccine used in this country. Researchers were trying to determine whether immunity could be obtained by vaccinating children younger than a year old and which of the dosage strengths should be used.

Both vaccines are made of live but weakened measles virus.

The trial involved children in communities hardest hit by the disease, including East and West Los Angeles and Inglewood. The majority of the children were African American and Latino.

These neighborhoods were recruited because "you have to study the area where the disease is occurring," Reynolds said.

Worldwide, measles has been a devastating disease. Global health officials believe that 2.5 million lives are saved annually through vaccination. In the United States, before the era of vaccinations, an estimated 500,000 cases and 500 deaths occurred annually. From 1989 to 1991, 55,000 cases occurred across the nation, hospitalizing 11,000 people and killing more than 130 children and adults.

During the epidemic, 2,476 infants younger than 1 were hospitalized, and 35 died. At that time, the routine age for receiving the first dose of measles vaccine was 15 months, and 12 months in areas experiencing outbreaks. Today, the vaccine is routinely given 12-month-old children.

The "informed consent" papers signed by parents, which were reviewed and approved by institutional review boards at the CDC and Kaiser, said children would receive one of two vaccines and that two different doses of the E-Z vaccine would be studied.

The form did not say that E-Z was experimental or unlicensed. But a brochure that accompanied the form said: "This vaccine has been shown to be effective in younger children. Over 200 million children around the world have received this vaccine, but Los Angeles County is the first place in the United States where it is being offered."

Researchers became concerned in 1990 and 1991, when data from studies in Senegal and Guinea-Bissau, and later from Haiti, suggested an increased death rate among female infants receiving the stronger dose of the E-Z vaccine.

Dr. Stephen Hadler, director of the epidemiology and surveillance division of the CDC's national immunization program, said researchers have not confirmed a causal association between the more potent dose of E-Z vaccine and the deaths.

However, "it was enough to make [the World Health Organization] say that high [doses of the] vaccine should no longer be considered for use in kids," he said.