BALTIMORE, DEC. 14 -- Johns Hopkins Hospital officials today called on the federal government to promptly issue new recommendations on how hospitals should deal with doctors, nurses and other health care workers who are infected with the AIDS virus and who might pose a risk of transmitting the virus to patients.

The hospital has been embroiled in controversy since officials there revealed earlier this month that a prominent Hopkins breast surgeon, Rudolph Almaraz, had died of AIDS in November. Hospital officials also announced today that state and federal experts will cooperate with Hopkins in a study of 1,600 of Almaraz's patients to determine whether any became infected during surgery.

Hamilton Moses, the hospital's vice president for medical affairs, said at a news conference that lack of action by the Centers for Disease Control has created a "policy vacuum" that has left hospitals unable to require testing of employees and uncertain whether infected doctors or nurses should be barred from doing surgery or invasive procedures.

"We want help," he said. "We need policy and direction from those best able to provide" it.

Public concern about health care workers infected with the human immunodeficiency virus (HIV) has increased since last August, when the CDC reported the first case in which a health care worker with AIDS may have infected a patient. That case, in which it appeared that a Florida woman contracted the virus from her dentist during a tooth extraction, is still being investigated by the CDC and the Florida health department and has not been confirmed, according to James W. Curran, director of the CDC's AIDS division.

Curran said in an interview Thursday that the agency is working on new guidelines covering infected health care workers, but he did not say when they would be released.

"There's a need for more clear guidance," he said. He said the CDC is considering several issues, including whether to recommend that HIV-infected health care workers should be barred from doing surgery or other invasive procedures; whether certain operations pose a special risk of viral transmission; and who should be in charge of making such decisions when a worker is found to be infected.

Several studies in which patients of HIV-infected surgeons were tested have failed to find a single case in which a patient became infected. "At present the risk seems to be low -- not zero, but close to zero," Moses said.

Curran said the Florida dental case galvanized public fears, but has not helped AIDS experts to quantify the risk. "That case demonstrated that it can occur, without saying precisely how or how frequently," he said.

Hopkins officials said yesterday that so far about 400 of Almaraz's former patients have agreed to be tested for infection with the virus. Two patients have sued the hospital, charging emotional trauma, although there is no evidence that Almaraz infected anyone.

Moses said Almaraz, a breast cancer expert who limited his practice almost exclusively to breast biopsies and mastectomies, gave up surgery last March and resigned from the hospital staff in June. He said rumors that Almaraz had AIDS reached hospital officials early last summer. "We set about immediatley trying to confirm them," he said. He said neither Almaraz nor his family would confirm that the surgeon had AIDS.

Last fall, Hopkins officials informed the state health department about the rumors and suggested that the state begin a study of Almaraz's patients. State health officials declined to do so without confirmation that Almaraz was infected, said Timothy R. Townsend, the hospital epidemiologist.

Moses said it is likely that if all of Almaraz's patients are tested, a few will test positive, since the frequency of HIV infection in the U.S. population is estimated to be between 2 and 4 per 1,000 people. But hospital officials said they knew of no surgical accidents in which Almaraz might have bled during an operation.

"In Dr. Almaraz's mind, he was not putting his patients at risk," said Moses.

The Hopkins medical community first confronted the issue of HIV-infected workers in 1987, when a doctor there, Hacib Aoun, sued the hospital for breach of confidentiality for revealing that he was infected. Aoun had become infected when a glass tube of HIV-infected blood broke and cut his hand. The suit was settled out of court.

At the time, according to a source who asked not to be identified, a doctor who supervised surgeons-in-training warned a group of trainees that if they feared they had become infected, they should "be tested outside the hospital and make your own decisions. It's obvious that your confidentiality probably cannot be guaranteed."

Moses said Johns Hopkins Hospital adopted guidelines this fall under which HIV-infected workers are "encouraged" to inform a superior of their status. Officials may not legally ask an employee whether he or she is infected or require a worker to be tested, said Paul M. Rosenberg, the hospital's general counsel.

"If it is estimated that practicing medicine or performing surgery confers a risk on patients, we would . . . curtail {infected physicians'} privileges and also attempt to find something else they could do," Moses said.

Many hospitals around the country have similar guidelines, but most were written early in the epidemic and may not be adequate to deal with the new concerns about the risk of a patient's becoming infected, said Deborah Cotton, former clinical director for AIDS at Boston's Beth Israel Hospital.

"This issue has been around for many years, and no one wanted to face it proactively," she said. "Yet everyone, I think, knew that if a case occurred, we were going to be developing policy in a kind of crisis mode."

AIDS experts at Hopkins and elsewhere worry that their inability to measure the risk will make it difficult to strike a balance between protecting patients and respecting the rights of workers. "It's a no-win situation," said John G. Bartlett, director of infectious disease at Johns Hopkins Hospital. "There is no concrete evidence that a health care worker presents a risk to a patient. There is a suggested case that is yet to be confirmed. To try to design a strategy or policy in a vacuum of information would be dangerous."