The Food and Drug Administration said yesterday it will investigate the circumstances surrounding the mysterious death of a Catholic University nursing student who died in her sleep Saturday while taking part in a National Institute of Mental Health experimental drug study.

"We're concerned," said Heinz Wilms, deputy director of FDA's division of scientific investigations. "A death in a drug study is alway alarming."

Bernadette Gillcrist, a 23-year old senior from Bethesda, was one of 20 paid volunteers being used to test the effects of lithium, an antidepressant, combined with AMPT (alphamethylparatyrosine), and experimental drug, on mod and sleep patterns.

National Institutes of Health officials said yesterday that the drug study was discontinued after Gillcrist's death was discovered.

"What we're trying to do now is make sure it's not going on anywhere else in the country," said Dr. Charles McCathy, director of NIH's Office for Protection from Research Risks. "So far our people haven't turned up any similar program."

A spokesman for the FDA said yesterday that AMPT, manufactured by Merck and Co. in West Point, Pa., currently is beng used by only six physicians in the country, including the sleep study's director, Dr. Chris Gillin.

The drug was licensed in September 1979 under the trade name "Demser" and is used to reduce blood pressure in patient suffering from a rare form of cancer. The FDA gave NIH a test license last year to use the drug in conjunction with lithium to monitor sleep disorders. According to the FDA, the NIH is the only research center combining AMPT with another drug.

On the consent form signed by Gillcrist and the 19 other volunteers, the hospital states that the two drugs had not been used in combination previously. h"There is no pharmaceutical evidence," the form states, "to suspect that the combination will be harmful."

The study was considered a "low risk" experiment, according to hospital officials. The form signed by the volunteers did not contain a provision releasing the hospital from an liability.

The consent form does state what the risks to the volunteer are. "The risks associated with the AMPT study were side effects, such as tiredness and crystals in the urine," said Dr. Charles McCarthy. "There are risks involved in all research.Is one of the risks death? It can be."

Gillcrist's body was discovered by a hospital technician at 7 a.m. Saturday.The young volunteer had been given doses of lithium during the prior week. On Friday, she had received three doses of the AMPT.

Gillcrist had arrived at the hospital around 10 p.m. and had gone to sleep monitored by a brain scanning device. At 5 a.m., the machine "went flat," according to one hospital official. But the technician, believing that the machine was not operating properly, did not try to arouse Gillcrist.

Dr. Mortimer Lipsett, director of the NIH clinical center where the experiment took place, said Gillcrist had been given doses of AMPT in January at the start of the program. However, he said yesterday, Friday was the first time she had been given both AMPT and lithium.

Although NIMH volunteers are required to sign a form saying they had had no history of medical or psychological problems, Gillcrist manged to conceal the fact that she had been hospitalized four years ago for cardiac arrest and also had suffered from anorexia nervosa.

The disease, a form of emotional disorder, is typically found among young women. Most often, those who have the disease eat virtually nothing, although some may eat normally and then induce vomiting after meals.

According to her sister, Bernadette Gillcrist, who was 5 feet 8, weighed about 130 pounds when she died. She had lost 20 or 25 pounds since January, her sister said.

At the time of her death Gillcrist also was under the care of a private psychiatrist, according to the family. Sharon Gillcrist, 22, said Bernadette did not tell her family that she was participating in the experimental drug study. She also had been under the care of a Bethesda cardiologist.

A lawyer retained by the family said yesterday he would try to ascertain the cause of Gillcrist's death. "She did not have any heart disease," said Gordon Forrester Jr. "I believe the medical examiner will confirm she had a healthy body."

Asked if Gillcrist needed the $800 volunteer fee paid by MIMH, Forrester said, "She needed money just like any of them (the family's children) needed it. She was on scholarship, and there were four children in college."

Currently, there are approximately 60 volunteers at NIMH participating in the inpatient program. The volunteers, mostly college students, reside at the hospital for four to six months.

The outpatient volunteers are paid by the hour, a spokesman said yesterday.

Gillcrist was the second member of her family to take part in experiments at the National Institutes of Health. According to one family member, her older sister Marcia had volunteered to participate in unrelated studies involing eye drops.