Federal scientists are abandoning efforts to develop a completely artificial human heart after spending more than a quarter-billion dollars, because every use of such devices has resulted in the eventual death of the patient.

From now on, government research support will concentrate instead on developing a smaller pump that can assist, rather than replace, a failing heart.There are still a lot of problems that cannot soon be solved with a total heart replacement device, Dr. Claude Lenfant, director of the National Heart, Lung and Blood Institute, said Thursday.

Every attempt at total heart replacement has resulted in a patient dying of either infection or stroke caused by blood clots that develop within the mechanical heart, said Lenfant.

"Those issues have not been resolved," he said in an interview.

The National Institutes of Health have given a total of $2.7 million in research grants on the total-heart replacement this year, but Lenfant said that will be the last. Since 1964, the agency has spent $239.31 million on various elements of artificial heart development technology.

Now, Lenfant said, the agency will support only four grants for the development of a self-contained, long-term "single ventricular" device. This would be inserted into the left ventricle of the heart, the chamber that endures most of the work stress in the four-chambered human heart.

"We believe there can be an extremely reliable single ventricle system developed for clinical application," said Lenfant.

The single ventricle device already has been used in more than 170 cases as a "bridge to transplant," but it has not been implanted for long-term use, he said.

John W. Wood Jr., president of Thermedics Inc. of Woburn, Mass., said his company now has such a long-term device being tested in experiments with animals.

He said the device, which is an electrically driven, diaphragm-like pump called the Heartmate, is inserted into the left ventricle of the heart and powered by a battery.

Powering the heart device has been solved with an electrical conduction system.

The Heartmate is equipped with an electrical coil, Wood said. A similar coil is placed on the outside of the body and electricity is conducted through the skin to recharge the battery. In this way, the patient will have no wires or tubes leading to the artificial pump, lowering the risk of infection.

Wood said the device has been tested successfully on calves and an application is pending for clinical trials.

A second assist device, designed to be used only temporarily, is now on the market. This device is inserted to assist the heart only until the heart has recovered from trauma or until a donor heart can be found.

Lenfant said the total artificial heart replacement has been attempted five times worldwide, but none was considered successful.

The Jarvik-7 artificial heart was the most widely used in the United States. Barney Clark received the first artificial heart on Dec. 2, 1982. His University of Utah surgeon was Dr. William DeVries. Clark died in March 1983.

William Schroeder, who received his artificial heart at the Humana-Audubon Hospital in Louisville, Ky., where DeVries practiced later, lived longest with the Jarvik device - almost two years.

Lenfant said the Jarvik device was successful mechanically, but "the interface with the body was never successful."

He said tubes leading to the outside may have contributed to the infections that killed some of the patients and the pump itself led to blood clots that cause strokes.

At the University of Utah, which had received one of the NIH artificial heart research grants, Dr. Donald Olsen told the Atlanta Constitution that cancellation of the research contracts "is a tremendous setback for scientific advancement in the field."

Four companies, however, will receive grants for continued study of the heart assist device, said Lenfant. In addition to Thermedics of Boston, they are Abiomed of Danvers, Mass., Nimbus Inc. of Sacramento, Calif., and Novacor of Oakland, Calif.