Copyright 1993 The Washington Post
The Washington Post

September 7, 1993, Tuesday, Final Edition

SECTION: FIRST SECTION; PAGE A1

LENGTH: 4442 words

HEADLINE: And Then the Patients Suddenly Started Dying;
How NIH Missed Warning Signs in Drug Test

SERIES: Occasional

BYLINE: John Schwartz, Washington Post Staff Writer

BODY:
For months, Jay Hoofnagle and his staff at the National
Institutes of Health had been discounting the complaints of
some of their patients. Several had told the doctors that the
experimental drug they were taking for chronic hepatitis B
was causing side effects, from nausea and weight loss to a
painful tingling in the feet.

The doctors had conducted extra tests, but found nothing
unusual. They told the patients that their symptoms were not
threatening and probably were not caused by the drug,
fialuridine (FIAU). After all, FIAU had seemed relatively safe
and wonderfully effective for some patients in two previous
tests.

Then the patients started to die.

On June 25, Howard Tichenor, a 44-year-old Arlington man,
checked into an emergency room with failure of his liver and
other organs. He had a rare but ruinous symptom: lactic acidosis
-- a potentially fatal buildup of waste materials in the cells --
which had afflicted another patient in a previous trial of the
drug.

"The word 'acidosis' kind of struck like a lightning bolt," said
Hoofnagle, who headed the study. "We realized that we had
been mistaken" in disregarding earlier, less serious problems in
other patients.

Hoofnagle stopped the trial the next day and called his patients
back to NIH for tests. Some were vacationing. Hoofnagle
persuaded police to break into one patient's house to find out
where he had gone, and found another by calling the bar where
his girlfriend worked.

Another patient, Carlton Lee, 35, had gone to Charleston, S.C.,
to visit family members. He had quit taking the drug on his own
at least two weeks before because he felt ill, but he still looked
awful, having lost at least 25 pounds from his previously chubby
build. His sister, Susan, 33, pressed him about his health,
obviously worried. An older sister, Darlene Kriewald, recalled,
"He snickered, and said, 'Susan, I'm not going to die.' "
Hoofnagle reached Lee by telephone, and told him to get back to
Bethesda right away.

Lee died on July 30, poisoned by FIAU. The drug cleared Lee's
body of the hepatitis virus, but it destroyed his liver, pancreas
and other major organs in the process. Of the 15 patients in the
NIH trial, 10 had taken FIAU for more than a month. Eventually
seven of those had to be sent to transplant centers for new livers.
Five, including Lee, have since died, one is still hospitalized in
fair condition following a liver transplant, and one has been
released from the hospital following a successful transplant.

Because they were so unexpected, the deaths have drawn
national attention even though drug trials are conducted to
determine exactly these kinds of risks. The researchers felt they
had found a safe treatment for a dangerous, drug-resistant
disease, and the tests had been reviewed and approved by all the
appropriate authorities. Only after patients began "crashing" one
by one did the warning signs become clear. But by then it was too
late.

The FIAU trial has been called "a catastrophe" and "a
nightmare" by a shaken Hoofnagle, the 50-year-old lead
researcher on FIAU and director of the division of digestive
diseases and nutrition for NIH's National Institute of Diabetes
and Digestive and Kidney Diseases. NIH is reviewing the case, as
is the Food and Drug Administration, which regulates clinical
trials and will hold hearings on the matter Sept. 21. In Congress,
Sen. Edward M. Kennedy (D-Mass.), whose Labor and Human
Resources Committee has oversight authority regarding NIH
and FDA, is monitoring the reviews. Through a spokeswoman,
Kennedy said: "The FDA and NIH have a heavy responsibility to
reassure Congress and the public that adequate procedures are
being followed in clinical trials of experimental drugs. Medical
research is essential, but it also has to be safe."

Should Hoofnagle and his team have seen the signs that stood
out so clearly after the fact? Or were FIAU's lethal ways so
deceptive that they would have blindsided even the most careful
clinicians at the NIH? It may take months to answer these
questions.

Giving FIAU 'One Last Try'

Fialuridine was born in the 1970s at the institution now
known as the Memorial Sloan-Kettering Cancer Center in New
York. Chemically, it belongs to the family of "nucleoside
analogues" -- that is, its structure is chemically similar to one of
the major components of DNA.

When a virus invades a cell, it commandeers the cell's nuclear
equipment to make copies of its own DNA, using whatever
natural nucleosides and other materials are at hand to assemble
a replica. One way to stop that process, researchers reasoned,
would be to introduce chemicals like FIAU that were close
enough to natural nucleosides to fool the virus into using them
-- but different enough so that, when used, they would make the
replica defective. Meanwhile, it was assumed, the more robust
mechanisms that protect human-cell DNA from that sort of
assembly error would keep the FIAU from causing damage to
the patient's body. That approach had proven somewhat
successful in the use of AZT to combat the AIDS virus, and with
various chemotherapy drugs used to fight leukemia and other
cancers.

FIAU, however, languished at Sloan-Kettering until the early
1980s, when Bristol-Myers (now Bristol-Myers Squibb) paid for
the right to conduct animal studies of FIAC, a drug that turns
into FIAU in the body. It dropped the license in 1988.

Oclassen Pharmaceuticals picked up FIAU on the rebound. The
company, based in San Rafael, Calif., with $ 16 million in sales a
year, is best known for medicated skin creams. Company
founder Glenn Oclassen said his researchers at first thought
FIAU might be used as a topical ointment to treat herpes sores.
But after plowing through the mountains of research material,
Oclassen said, "we began to feel this was a drug that could treat
patients systemically for serious viral infections." Oclassen tested
the drug against cytomegalovirus, a relatively rare illness that
has become more prevalent in the age of AIDS. The drug did
not perform well, and the company was considering dropping it
-- but decided to give it "one last try" against chronic hepatitis B.

Hepatitis B had not excited much interest from drug
companies in the past. People contract the disease a number of
ways, including by living in poor sanitary conditions, using
intravenous drugs or engaging in some sexual activities; it can
lead to cirrhosis of the liver, liver cancer and death. But it is not
always a killer: With or without drug treatment, patients who
show no cirrhosis have an 80 percent survival rate, and 10
percent of patients eventually recover on their own. "Ten years
ago I couldn't get a drug company interested in hepatitis B," said
Hoofnagle, who recalled "knocking on doors" at drug firms in
vain.

Still, the market was large. About 300 million people
worldwide suffer from the disease -- 50,000 to 100,000 in the
United States. Moreover, the rise of AIDS had caused an
explosion in anti-viral research, and several companies were
exploring hepatitis.

Suddenly Oclassen's gamble started to look good. The company
worked up a liquid form of the drug, taking pains to give it a
pleasant flavor. (Patients would later exclaim that it tasted like
Grand Marnier; one former Oclassen official explained, "Well, it
was Marin County.") In October 1990, Oclassen conducted a 10-
day trial of FIAU on 43 patients who had both HIV (the AIDS
virus) and hepatitis B. Glenn Oclassen recalled the results were
"unlike anything that had been seen before," with significant
reductions in hepatitis.

Hoofnagle wanted to see how it would perform in non-HIV
patients, and obtained approval from the FDA in March 1992 to
give the drug to 24 hepatitis B patients at four dosages. That test
began in April 1992. Patients with the lowest dose did not see
much in the way of results, and patients at the higher doses
tended to get nauseated. But patients on the middle two doses
made remarkable progress. Half cleared the disease from their
system entirely, and others showed a significant drop in the
level of virus, though they relapsed later. The new drug had
done as good a job against the disease in one month as a
four-month round of alpha interferon, without that drug's side
disadvantages. (Alpha interferon cannot be taken orally, but
must be injected; it also can have side effects, including flu-like
symptoms and a lowered white cell and platelet count, and has
triggered depression in patients with a history of mental
illness.) By comparison, FIAU "looked terrific," Hoofnagle said.

Oclassen saw big things on the way -- and knew that it would
need a big partner to help pay for broader drug trials with
hundreds of patients. On Sept. 1, 1992, the company announced
that pharmaceutical giant Eli Lilly & Co. would pay $ 7.5 million
to share the development and marketing rights.

'I Came to Feel Like a Number'

Hoofnagle quickly made arrangements for the next test -- a six-
month trial to begin in mid-March of this year. Chronic diseases
commonly require long cures. FIAU had reduced the amount of
virus in the patients' systems, but didn't always eradicate it.
Hoofnagle thought longer exposure to FIAU would wipe out the
hepatitis completely.

But even as he was planning that longer trial, problems from
the 1992 pilot study were beginning to appear. Four months after
that trial, a 60-year-old participant from Kansas, Patient 4D, died
of liver failure. (Drug trials are confidential. Unless patients or
their families have consented to have the names made public,
their patient numbers are used here.)

Some of Patient 4D's symptoms paralleled the deaths to come.
He had suffered abdominal pain and nausea, which doctors
attributed to a "flare" of hepatitis symptoms that normally
occurs when the body flushes out the disease. But he also
developed lactic acidosis and a high concentration of fat in his
liver, which are not commonly associated with hepatitis B. NIH
doctors attributed the death to complications from cirrhosis of
the liver and the man's recent gall bladder surgery, which NIH
doctors had advised against.

A second FIAU patient from the 1992 trial suffered side effects
that, while not fatal, also might have indicated some of the
dangers ahead. Paul Melstrom of Phoenix, known as Patient 1A,
developed a painful tingling in his feet 90 days after he stopped
taking the drug in May 1992. Such pains, known as "peripheral
neuropathy," have been linked to nucleoside analogues and
were cited as a possible side effect in the "informed consent"
forms that patients signed before beginning the drug trial. The
sensations usually go away; Melstrom's continued to get worse.

Melstrom, a recovering alcoholic who provides AIDS
information through computerized on-line services, said in a
recent interview that NIH doctors told him they did not believe
FIAU had caused his problems, in part because he had
experienced a mild version of the syndrome years before when
he was drinking heavily. The doctors, Melstrom said, "kept on
discounting the obvious. They were willing to blame everything
except FIAU for the side effects." The doctors arranged for
Melstrom to get treatment at the NIH's Bethesda-based pain
clinic. "We weren't thinking much of seeing toxicity at that
point," Hoofnagle said, because Melstrom and others had been
off the drug for so long.

Melstrom, an acerbic 53-year-old, wasn't satisfied, so he began
educating himself, tapping into the databases he had discovered
in his AIDS research and conferring with doctors and medical
students he had met on-line. "It was my layman's med school,"
Melstrom said. Fredrick Kenny, Melstrom's gastroenterologist in
Phoenix, was impressed with the depth of knowledge his
patient had developed: "He dug into this more than anybody --
certainly more than I did."

NIH doctors did take some action in response to Melstrom's
complaints, however. They reworked their plans for medical
tests on subjects in the long-term trial to include nerve tests that
would check for the neuropathy, and strengthened warnings
about it in informed consent forms and interviews. In a
letter to Melstrom dated June 10, Hoofnagle mentioned that
those warnings had been so strong that one patient decided not
to enter the study.

Melstrom kept up a steady stream of hectoring correspondence
with NIH officials, complaining that they still were not paying
sufficient heed to the dangers. In an angry letter dated June 20,
he wrote Hoofnagle that while the constant pain of his
neuropathy troubled him, "what equally irritates me was the
insensitivity and denial from NIH. . . . I came to feel like a
number, a mere casualty to research objectives."

Meanwhile, new patients were coming into the long-term
FIAU trial. One was a familiar face to the researchers. Carlton
Lee already had participated in four NIH drug trials for his
condition, including the 28-day pilot FIAU study in 1992. His
work in the Peace Corps, as a forestry manager in Sierra Leone,
West Africa, had left him with a mild case of the disease.

Lee was the kind of person people naturally take to. He came to
Washington in 1979 on the staff of newly elected Rep. Charles
W. Stenholm (D-Tex). In nearly 14 years in and out of the capital,
Lee rose to become congressional liaison for the National
Commission on AIDS, where he has been called "the driving
force" behind the nation's first federal AIDS initiative in 1987.
He was all energy, constantly working and networking, fueled by
soft drinks and candy bars and the pleasure that came from
doing work that he loved.

Lee told friends that even though hepatitis had never troubled
him much medically, he had been impressed with the feistiness
of the AIDS sufferers whom he worked to help. They took an
active role in fighting for a cure, and he would too.

Within a few weeks of starting on the new drug, Lee began to
hurt. He had suffered disagreeable side effects from the 1992 test,
but this time Lee began to complain of painful tingling that
made it hard to sleep at night. He had wrenching abdominal
pain and severe weight loss. According to his roommates, he
developed bizarre allergies, breaking out into blisters after eating
an orange.

Lee's friends said he told his doctors that he wanted to be taken
off the drug, but that several times in June 1992 Michael Fried,
the NIH researcher most directly involved with the patients,
encouraged him to remain in the trial. At that point, Hoofnagle
was not involved in the trial on a day-to-day basis. In mid-June
Lee ran into Hoofnagle on the NIH campus; when the doctor
asked him how he was doing, Lee relied, "As a matter of fact, not
well at all." Fried said in a recent interview that he ordered extra
blood tests to look for signs of toxicity, but that nothing showed
up in the "standard investigational techniques" NIH researchers
followed. Eventually Lee stopped taking the drug on his own,
but his condition worsened even as he continued to make
regular weekly visits to the NIH doctors.

By the end of June, another patient was worrying the research
team. A woman from North Carolina, Patient 6, had suffered
from nausea and vomiting early in the trial. The doctors took
her off the drug. She continued to have nausea, however, so she
was brought back to NIH for evaluation on June 17. At first
the doctors thought her problems might be related to "flaring,"
but tests showed that there was no virus left in her body. She
became jaundiced.

'The Nightmare Was Deepening'

Then, on June 25, Patient 2 -- Tichenor, the 44-year-old
Arlington man -- showed up at an emergency room in
Fredericksburg, Va., with multiple organ failure and lactic
acidocis. Hoofnagle saw the warnings behind him, one by one,
in sharp relief. He stopped the study on June 26 and called every
patient back to Bethesda for testing. Of the 10 people who had
taken the drug for longer than a month, all but three were
showing some sign of toxicity. "The nightmare was deepening,"
he said.

(Five other patients at NIH and five others in a separate trial
had begun taking the drug later, and had been exposed for less
than a month. To date, none has shown signs of toxicity.)

Hoofnagle and his medical team immediately began combing
the medical literature for clues, and quickly latched on to a
theory suggested by researcher Yung-Chi Cheng of Yale
University that seemed to fit the epidemic of lactic acidosis:
somehow, the FIAU was damaging the mitochondria, sub-units
of cells that provide energy, in organs such as the liver, pancreas
and kidneys. Could it be that the mitochondria, which have
their own DNA but lack the protective mechanisms of the
cellular nucleus, had taken up the FIAU and been damaged?

On July 1, Hoofnagle took a moment away from the crisis to
write a letter to his most persistent critic. "Dear Mr. Melstrom,"
it said. " . . . Recent events in our current study of FIAU have
caused us a terrible shock and show that you were indeed right
about the toxicity of FIAU."

When Melstrom received the letter, he turned to a stack of
unread medical reports on FIAU. One caught his eye: a May 1991
article from the journal Molecular Pharmacology by Yung-Chi
Cheng, the Yale researcher. The piece warned that some
nucleoside analogues showed evidence of giving delayed toxicity
-- a toxicity that seemed to present itself in damage to
mitochondria.

Hoofnagle had not seen Cheng's work until the weekend the
crisis began. He said he immediately contacted Cheng, who
became part of the team trying to make sense of the growing
disaster. Cheng's lab showed the first evidence of mitochondrial
damage in a biopsy sample from one of the NIH patients.

Hoofnagle convened emergency seminars with experts from
around the country who flew to NIH to consult and find possible
treatment. So many patients had so many things wrong with
them. "There was liver failure," Hoofnagle said, counting down
the list. "Abnormalities of the kidney. There was pancreatitis.
There were muscle problems. There were nerve problems."

Working feverish 16-hour days, the doctors devised a plan to
keep the patients alive and repair their damaged cells. They gave
the patients an intravenous solution rich in two natural
nucleosides: thymidine and uridine. They also tried to get each
patient to compensate for the loss of energy-producing
mitochondria: Patients were told to eat a box of cornstarch a day,
and the sickest patients were given a viscous sugar solution in a
line running directly into a neck vein.

On July 4, Hoofnagle's two sickest patients -- including
Tichenor -- received liver transplants. Tichenor died the next
day at the University of Pittsburgh hospital. The second patient
died July 6 after receiving a transplant at the University of
Virginia at Charlottesville. On July 16, a third patient died,
also in Charlottesville, while awaiting a transplant.

On July 28, Hoofnagle decided to refer Carlton Lee to the
University of Pittsburgh for a transplant. When Lee arrived in
Pittsburgh on the evening of the 29th, doctors there initially said
he seemed too healthy to qualify for an immediate transplant.

But the next morning when Lee's mother and a sister arrived
at the hospital, they couldn't get into his room -- it was crowded
with doctors trying to resuscitate him. A doctor told family and
friends that they were only keeping Lee alive through "brute
force medicine," extreme doses of their most powerful
drugs. A blood vessel in Lee's pancreas had ruptured during the
night, sending the rest of his weakened body systems cascading
into ruin. He died that afternoon.

On Aug. 31, Patient 6 (the woman from North Carolina)
became the fifth of the trial subjects to die, at the University of
Virginia hospital at Charlottesville. She had undergone two
liver transplants. Hospital officials said her new liver was
working well, but other major organs failed.

The FIAU deaths leave many questions -- questions that may
be answered only in a congressional hearing room or a
courtroom, if ever.

Why didn't the doctors recognize the signs of trouble earlier?
Several doctors in government and academia who have
reviewed the facts of the FIAU case say it would have taken an
exceptionally perceptive clinician to recognize the warnings in
the varied reactions patients had to FIAU. The gall bladder
patient's woes seemed to be related to his earlier cirrhosis, and
Melstrom had had peripheral neuropathy in the past. As for Lee,
Hoofnagle said the symptoms were not obviously related to the
toxicity of FIAU.

No More 'Jumping to Conclusions'

The risks from FIAU seemed slight in those early, short-term
trials. Nausea is a common reaction to many drugs, including
many vital medications such as those used in chemotherapy.
One of the difficult tasks in clinical trials is separating such
"nonspecific" symptoms from the "specific" ones, which indicate
a real problem. The doctors decided Lee's symptoms fell into the
"nonspecific" group, in part because Lee had suffered from an
irritable colon in the past.

Today Hoofnagle believes the pain might have been a sign of
autonomic neuropathy -- pains emanating from the nerves
along Lee's internal organs. But "I'm not jumping to
conclusions anymore," Hoofnagle said; "In retrospect, I wish
we had stopped everybody when they first burped."

Fried, who advised Lee to stick with the trial, said that while he
might have urged Lee to stay on, the patient was free to drop out
without any consequences from the NIH -- as Lee had done in a
trial of the drug interferon a few years before. "Ultimately the
care of the patient wins out," said Fried, who added that he
considered Lee a friend because of their long acquaintance
through NIH.

Why weren't the dangers of nucleoside analogues taken more
seriously? Drugs that are chemically similar can have widely
different effects, which is why clinicians test so many of them.
Ethanol and methanol are chemically close, but drinking the
first makes people drunk while drinking the other can lead to
permanent blindness.

In the case of FIAU, the doctors felt they had found a
nucleoside analogue with lower toxicity than others in its
family. In fact, it might be more dangerous, said Yale researcher
Cheng. Other nucleoside analogues join the DNA chain at the
end and drop off again relatively easily after the patient stops
taking the drug. Cheng believes that FIAU is taken up in DNA
well within the strand, so that "it's much tougher to repair."

While reluctant to criticize other researchers, Cheng said the
NIH doctors should have been mindful of the dangers he
described in papers from 1991 on. "Maybe they didn't pay
attention to it," Cheng said, because his work is so much
better known in the field of AIDS research than liver disease.
"It's a sad story."

Through an NIH spokeswoman, Hoofnagle said his team did
do a literature search on the "Medline" computer database for
toxicity of nucleoside analogues, stressing their link to
peripheral neuropathy. Hoonagle said, however, that he did not
discover Cheng's work until after the disaster struck. A recent
Medline search for articles containing the words "nucleoside
analogue" and "toxicity" by The Washington Post turned up
Cheng's May 1991 article linking nucleoside analogues to
mitochondrial damage in less than two minutes.

Stephen Straus, a co-investigator with Hoofnagle and chief of
the Laboratory of Clinical Investigation at the NIH's National
Institute for Allergy and Infectious Diseases, said that as a
virologist he knew of the potential for mitochondrial damage in
anti-retroviral drugs such as AZT -- but that FIAU had not been
commonly classified as an anti-retroviral drug. "Antiviral drug
development is still a young science, and it's going to have false
leads," Straus said. He suggested that FIAU's toxicity might be
limited to patients whose systems had been weakened by
hepatitis B, making it even harder to spot in preclinical tests.

Hindsight is a marvelous thing, and warning signs that seem
sharp and bright in retrospect may have been lost in the
mountains of research articles and symptoms that every drug
trial deals with. In the case of FIAU, the late onset of toxicity and
the hepatitis-like symptoms helped to camouflage the dangers
further.

One View: 'The System Worked'

The Food and Drug Administration is under constant pressure
from industry and patient support groups to speed its drug
approval process so that promising medications get to market
more quickly. While it does not appear that FDA rules regarding
clinical trials were violated, NIH officials said the drug
companies involved with the trials had pressed researchers to
shorten the length of time between trials in order to speed up
the process. "Pauses between studies were much more rapid
than we're used to," Hoofnagle said. "There was a bit of a rush
and push to get this evaluated quickly."

Had the six-month trial gone well, a large-scale trial with
hundreds of patients would have begun immediately. Patients
already were being recruited. "I didn't want [the drug companies]
to go to a multicenter trial so quickly," Hoofnagle says today. In
the case of a drug whose toxicity might show up only after many
months, haste meant that problems with the drug showed up as
the next trial was commencing.

A Lilly representative has made the case that the deaths, while
tragic, show the strength of the clinical trial system, which is
supposed to expose the smallest number of people to a drug in
the shortest amount of time to establish its safety and
effectiveness. "In view of that philosophy, the system worked,"
Allan Weinstein, vice president of Lilly Research Laboratories,
said in a July interview. Without a system of clinical trials in
place, far more deaths might have occurred, one government
official said; in other countries, FIAU might already be on the
market.

Could FIAU resurface some day, despite this disaster? Glenn
Oclassen says FIAU's strong showing against hepatitis B in short-
term treatment is hard to leave alone: "That leaves a continuing
incentive to try to understand what happened . . . to try to decide
if there's any way this program can go forward."

That sort of answer enrages Paul Melstrom, who is preparing
for one of the many lawsuits likely to result from the FIAU
tragedy. "If I ever hear that FIAU is going back into clinical trials
for any purpose," he said, "I will go into federal district court for
a restraining order to see that it is not used."

Hoofnagle was more succinct. The soft-spoken researcher,
when informed of Oclassen's statement, shot a withering look
toward the interviewer and deadpanned, "Yeah, right."

GRAPHIC: PHOTO, JAY HOOFNAGLE; PHOTO, VICTORIA
BUCKNER FOR TWP

TYPE: NATIONAL NEWS

SUBJECT: MEDICAL RESEARCH; DRUGS AND MEDICINES;
DRUG SAFETY; PATIENTS; MEDICAL TREATMENT;
PHARMACEUTICAL INDUSTRY; DEATH AND DYING; LIVER

ORGANIZATION: NATIONAL INSTITUTES OF HEALTH;
FOOD AND DRUG ADMINISTRATION

NAMED-PERSONS: JAY HOOFNAGLE; EDWARD M. KENNEDY

ENHANCEMENT: HEPATITIS