Surgeon William C. DeVries said today there was a "95 percent chance" that the mechanical heart he implanted in William J. Schroeder almost three weeks ago was responsible for the stroke Schroeder suffered Thursday.

DeVries said Schroeder, the second person to receive an artificial-heart implant, was "quite stable and doing very, very well" today but cautioned that his recovery in the days to come would be "like being on a roller coaster," with "some times that are very good and some times that are very bad."

DeVries said Schroeder's stroke most likely was caused by a small blood clot traveling to the brain from the mechanical device or by the altered blood flow it produced in the remaining portion of his natural heart.

Based on animal work and experience with the mechanical-heart valves in other patients, DeVries said stroke was a "known complication" but one that had appeared to be "rare." He emphasized that Schroeder and his family were told of the possibility of stroke before the Nov. 25 surgery that took nearly seven hours.

Because the mechanical heart is made of "foreign substances" -- metal and plastic -- there is a greater chance of blood accumulating on the heart's surfaces and causing clots, but the device was designed to minimize this possibility, DeVries said. In addition, Schroeder has been kept on an anticoagulant or blood-thinning drug to reduce the chance of clots.

Doctors at the Humana Hospital-Audubon will be watching and waiting to see if the stroke was a rare occurrence or something that may be more likely in heart implants than they had expected.

DeVries said it is something he will "worry about as long as [Schroeder's] alive."

The stroke "is a complication of the artificial heart . . . , " DeVries said. "I would say it was probably about a 95 percent chance that it came from the actual device, the valves, or the natural atria which are behind the heart." The atria are the upper chambers of Schroeder's natural heart that were attached surgically to the larger lower chambers of the mechanical heart.

Later in the briefing, DeVries said, "I said the possibility was 95 percent . . . . I'm not sure of anything . . . . If you were a betting man, you would have to say it came from the valves or the heart."

DeVries' comments today were more pessimistic than those of Dr. Allan M. Lansing, the head of Humana Heart Institute, which is treating Schroeder. Lansing has been in charge of the regular news briefings. He said Friday that Schroeder was recovering "brilliantly."

But DeVries said, "I'm still optimistic. I think he can get out of here and get out of the hospital and live a good life."

The two physicians today also had somewhat differing interpretations of another problem Schroeder is suffering: edema, or swelling from water retention in his tissues, particularly in his legs.

DeVries appeared to be less concerned and attributed the swelling to the artificial-heart surgery. He said that although the swelling had shifted more to his legs from his back, Schroeder's weight had dropped. He weighed about 205 pounds before surgery, 229 pounds shortly afterward and is now down to 202 pounds, DeVries said.

Like many stroke patients, DeVries said, Schroeder today fluctuated between long periods of sleep and drowsiness and shorter periods of alertness in which he watched television and talked slowly but understandably with a slurred voice. He also was moving his arms and legs on his stroke-weakened right side.

The drowsiness is considered a result of the physical and emotional fatigue of a major setback, as well as a temporary swelling in his brain from the stroke.

Even though Schroeder seems to be recovering well from it, the stroke is a reminder of how quickly things can change when doctors move into the frontiers of medicine, where the unknowns are greater than the knowns.

Schroeder, like Barney B. Clark, the first mechanical-heart implant patient, is a pioneer participating in an experiment and may not realize a long-term benefit from the artificial heart beating in his chest.

This morning, Lansing said the edema seemed to be a result of chemical imbalances due to nutritional deficiencies, particularly a lack of protein. Schroeder is now receiving supplemental protein through a feeding tube.

Like DeVries, Lansing reflected more today on the experimental nature of their heart work. Only two implants have been attempted, and "even if you lost the first two games you wouldn't pack up and go home," Lansing said.

"There are seven games in the World Series. There are seven implants planned. We may have had an error in the first game and an error in the second game. We haven't lost the second game yet, but we're still in only the second game. The purpose of this whole thing is to find out answers to questions . . . "

"We would like to win the series," he said, but cautioned, "it's still early on."

Asked what would be considered a win, Lansing said, "It's the patient going home, enjoying life with his family, fishing and working in the garden."

Asked whether the fact that Clark and Schroeder both had suffered neurological or nervous-system problems suggested a larger problem with the artificial heart, Lansing said the origins of the problems appeared to be "entirely different."

Schroeder's fast-paced progress in the first 18 days after receiving the heart, allowing him to move around the hospital and consider leaving within a few weeks, may have been a misleading indicator of his prospects.

While Schroeder's stroke appeared to result from a clot coming from the artificial heart, Clark suffered seizures within a week after his implant in 1982. The seizures later were attributed to a chemical imbalance in his deteriorated body and perhaps to overly rapid pumping of the artificial heart in the beginning. After that, his mental status and alertness were impaired until his death in March 1983, 112 days after the implant.

Although Schroeder has had a major setback, he is still faring better than Clark.