Under a new American Medical Association policy, a doctor can ethically withdraw "all means of life-prolonging medical treatment," including food and water, from a patient in an irreversible coma.

The new policy goes beyond previous AMA statements in two respects: It specifically mentions nutrition and hydration -- food and water -- as a form of artificial life support, and it applies not only to terminally ill patients but also to those in an irreversible coma.

For example, it could have applied to Karen Ann Quinlan, who died last June after more than 10 years in a coma caused by a drug overdose. Although Quinlan was in an irreversible coma, she was not considered terminally ill. She survived disconnection of her respirator, but would have died had her parents requested, and the courts permitted, the withholding of food.

The four-paragraph statement, issued last month, is the product of nearly two years of deliberation by the AMA's Council on Ethical and Judicial Affairs.

"There is nothing in this statement that encourages physicians to go out and turn off food and water," said Dr. Nancy Dickey, a Houston family practitioner and chairman of the AMA council. "But there are times, even outside terminal illness, when physicians can ethically withdraw life-supporting measures, including food and water."

Such a decision, Dickey said, "will never be easily arrived at by any physician. But this [statement] will make the extreme decision of withdrawing treatment slightly easier for some physicians."

"In treating a terminally ill or irreversibly comatose patient," the AMA statement says, "the physician should determine whether the benefits of treatment outweigh its burdens." With the concurrence of the patient or those responsible for the patient, "it is not unethical" to withdraw or withhold all life-prolonging medical treatment.

Dickey said the AMA council took up the issue in an effort to clarify the physician's role in a dilemma confronting a growing number of patients and their families: When does prolonging life amount to merely postponing death? Besides the thousands of terminally ill patients, an estimated 10,000 irreversibly comatose patients are being kept alive in American hospitals by artificial means, including respirators and feeding tubes.

The council, whose seven members are physicians, interprets principles of medical ethics for the AMA. Its policies are not legally binding, but Dickey said she expected they would be cited by lawyers in several pending court cases involving whether life support may be withdrawn from irreversibly comatose patients.

A previous AMA statement, issued in 1982, said that "where a terminally ill patient's coma is beyond doubt irreversible and there are adequate safeguards to confirm the accuracy of the diagnosis, all means of life support may be discontinued."

But the 1982 statement, Dickey said, needed clarification because it did not spell out that "life-supporting measures" include nutrition as well as artificial respiration and did not say whether the policy could apply to non-terminal cases, such as when a patient is in an irreversible coma.

For some doctors and medical ethicists, the new statement "simply makes explicit something they've been reading between the lines," Dickey said. "But for many physicians, this is a big step."

While some practicing physicians look at the new policy "with a raised eyebrow," she said, public response to the new policy has been "by and large positive. Frankly, I've been a wee bit surprised there's been so little opposition."

But some experts have expressed concern about it.

"I see this as a classic case of head versus heart," said Daniel Callahan, director of the Hastings Center, a research institute that focuses on issues involving medical ethics. "One can make a rational case for withholding food and water from such a patient, but I find it basically repugnant to allow someone to starve to death or die of lack of water.

"I don't think I could do it myself and I don't think many health professionals could. It goes against an awful lot of our basic instincts."

A 20-member task force from the Hastings Center is drawing up its own detailed guidelines for the care of dying patients. The guidelines, which aren't expected to be finished until next year, will cover about 10 major issues, including the definition of death, economic considerations of treatment and withdrawal of life support.

"I think this is an issue that most people have yet to think about very much," Callahan said.

The new AMA statement "would count as a clarification rather than a change," said Robert M. Veatch, senior research scholar at the Kennedy Institute of Ethics at Georgetown University. "It makes explicit what was understood before."

But Veatch said he had "real problems" with the AMA position on several grounds, particularly the issue of informed consent.

"I'm very concerned about the possibility that physicians will decide that it's in the patient's best interest not to be fed, when in fact the patient regards it in his best interest to be fed."

The AMA policy states that where the doctor's duties "to sustain life and relieve suffering" become contradictory, "the choice of the patient, or his family or legal representative if the patient is incompetent to act in his own behalf, should prevail." In the absence of the patient's choice or an authorized proxy, "the physician must act in the best interest of the patient."

"But," Veatch said, "that leaves us in a very difficult spot in the case of an incompetent patient without family."

Acknowledging that it's a "very emotional issue," Veatch also said he doesn't see "any logical difference" between nutrition and hydration, on the one hand, and other medical treatments.

"To me, the key issue is whether the patient finds the treatment of proportional value," Veatch said. "If the burden exceeds the benefit, or if there's no benefit at all, the patient -- or his surrogate -- is within his rights to refuse the treatment."

In the most famous legal case involving the withholding of medical treatment, Karen Ann Quinlan's family sued for the right to have her disconnected from an artificial respirator in 1975, but when the respirator was turned off, she continued to breathe on her own.

Although respirators and feeding tubes are both means of life support for comatose patients, many people make an emotional distinction between the two. For example, although Quinlan's parents went to court for permission to disconnect her respirator, they chose not to withdraw nutritional support from her.

"There's an emotionality about food and water," Dickey said. Also, an intravenous line may seem more familiar and less intrusive to many people than an artificial respirator does.

"It seems less of an imposition on their loved one than a 4-by-8-foot piece of machinery that sets off alarms periodically and makes noise. It takes longer for a family to reach the point of saying that this, too, is a a technology and an intrusion."

Although some critics suggest that withdrawing nutritional support from a dying or comatose patient is dangerously close to murder, the AMA statement declares that "the physician should not intentionally cause death."

"I see a rather large jump between letting someone die and killing someone," said Dr. Russel H. Patterson Jr., chairman of neurosurgery at New York University-Cornell Medical Center and a member of the AMA judicial council.

Withdrawing extraordinary technological support from patients who have no hope of regaining consciousness can be the most humane treatment, said Patterson, who is president of the American Association of Neurological Surgeons.

"After a while -- maybe weeks or months of seeing the patient with no concept of the present, no memory of the past and no hope for the future -- a lot of families say, 'Why does this have to go on?

" 'What's the purpose?'"