Cancer Drug Research: Contrary to Allegation, NIH Hydrazine Sulfate
Studies Were Not Flawed (Letter Report, 09/13/95, GAO/HEHS-95-141).
Pursuant to a congressional request, GAO provided information on the
National Cancer Institute (NCI)-sponsored clinical trials of the
anticancer drug hydrazine sulfate, focusing on: (1) NCI protocol design
and data management procedures; (2) how NCI and the trials'
investigators dealt with the drug's potential incompatibility with
certain agents; (3) the extent to which patients received these
incompatible agents; and (4) how the investigators reported the issue.
GAO found that: (1) the three large NCI-sponsored clinical trials showed
that the drug did not prolong cancer patients' survival; (2) controversy
surrounding the trials focused on trial participants' use of
tranquilizers, barbiturates, and alcohol, which were allegedly
incompatible with the drug; (3) clinical trial records showed that
participants used tranquilizers under varying circumstances,
particularly for relief from vomiting; (4) the investigators believed
that it was unethical to withhold antiemetic medications from patients
undergoing chemotherapy; (5) subsequent analyses of patients' use of
concurrent medications did not invalidate NCI conclusions that the drug
was ineffective; (6) the Food and Drug Administration (FDA) may have
contributed to the confusion surrounding the trials due to its more
conservative position on how the drug should be administered to some
patients; (7) although FDA approved more than 70 applications permitting
the use of hydrazine sulfate, it cautioned physicians about their
patients' use of tranquilizing agents while on the drug; (8) there were
lapses in recordkeeping and reporting because NCI did not require
complete and accurate research records on the patients' use of
tranquilizing agents during the trials; and (9) NCI-sponsored
investigators only recently analyzed this issue, since published results
did not accurately describe the widespread use of tranquilizers during
the trials.
--------------------------- Indexing Terms -----------------------------
REPORTNUM: HEHS-95-141
TITLE: Cancer Drug Research: Contrary to Allegation, NIH Hydrazine
Sulfate Studies Were Not Flawed
DATE: 09/13/95
SUBJECT: Pharmacological research
Drugs
Documentation
Cancer research
Therapy
Information disclosure
Data collection operations
Research program management
IDENTIFIER: Russia
**************************************************************************
* This file contains an ASCII representation of the text of a GAO *
* report. Delineations within the text indicating chapter titles, *
* headings, and bullets are preserved. Major divisions and subdivisions *
* of the text, such as Chapters, Sections, and Appendixes, are *
* identified by double and single lines. The numbers on the right end *
* of these lines indicate the position of each of the subsections in the *
* document outline. These numbers do NOT correspond with the page *
* numbers of the printed product. *
* *
* No attempt has been made to display graphic images, although figure *
* captions are reproduced. Tables are included, but may not resemble *
* those in the printed version. *
* *
* A printed copy of this report may be obtained from the GAO Document *
* Distribution Facility by calling (202) 512-6000, by faxing your *
* request to (301) 258-4066, or by writing to P.O. Box 6015, *
* Gaithersburg, MD 20884-6015. We are unable to accept electronic orders *
* for printed documents at this time. *
**************************************************************************
Cover
================================================================ COVER
Report to the Chairman and Ranking Minority Member, Human Resources
and Intergovernmental Relations Subcommittee, House Committee on
Government Reform and Oversight
September 1995
CANCER DRUG RESEARCH - CONTRARY TO
ALLEGATION, NIH HYDRAZINE SULFATE
STUDIES WERE NOT FLAWED
GAO/HEHS-95-141
Hydrazine Sulfate
(108998)
Abbreviations
=============================================================== ABBREV
CALGB - Cancer and Leukemia Group B
FDA - Food and Drug Administration
IND - investigational new drug
NCCTG - North Central Cancer Treatment Group
NCI - National Cancer Institute
NIH - National Institutes of Health
UCLA - University of California Los Angeles
Letter
=============================================================== LETTER
B-261063
September 13, 1995
The Honorable Christopher Shays
Chairman
The Honorable Edolphus Towns
Ranking Minority Member
Human Resources and Intergovernmental
Relations Subcommittee
Committee on Government Reform and Oversight
House of Representatives
Cancer takes almost 500,000 lives annually and is the second leading
cause of death in the United States. About 400,000 of the 1 million
Americans who will develop cancer this year, however, will be
cured.\1 Despite advances in treating cancer, some forms of cancer
are often resistant to current therapies and have a very poor
outcome. Based on the findings of an earlier study that suggested
the drug hydrazine sulfate may improve survival for some patients
with advanced cancer, the National Cancer Institute (NCI) sponsored
three studies of hydrazine sulfate.
All three NCI-sponsored studies, however, failed to demonstrate any
benefit from hydrazine sulfate. The developer of hydrazine sulfate
therapy alleged in documents furnished to the Subcommittee that NCI
compromised its studies on hydrazine sulfate. He asserted that the
NCI-sponsored clinical trials\2 did not demonstrate any benefit for
hydrazine sulfate because NCI's clinical investigators permitted the
concurrent use of tranquilizers, barbiturates, and alcohol. He and
other proponents of hydrazine sulfate therapy have interpreted animal
studies and other data to suggest that hydrazine sulfate is
incompatible with these agents. This allegation was echoed in the
popular media.
This report responds to your request that we provide further
information about the NCI-sponsored clinical trials of hydrazine
sulfate. It provides background on the trials, including information
on protocol design and data management procedures. It also discusses
how NCI and the trials' investigators dealt with the issue of
potential incompatibility with certain agents, the extent to which
patients may have received these agents, and how this issue was
reported.
To obtain information for this report, we reviewed NCI's policy
guidance for conducting clinical trials of investigational agents.
We also reviewed agency memorandums pertaining to protocol
development for the hydrazine sulfate clinical trials and related
correspondence. For each of the hydrazine sulfate clinical trials,
we visited the research facilities and reviewed a sample of the
research records documenting patient medications, including
concurrent antiemetic (antinauseant and antivomiting) and barbiturate
medications.\3 We discussed the conduct of these trials with NCI
officials and investigators, Food and Drug Administration (FDA)
officials, and proponents of hydrazine sulfate therapy to obtain
their perspectives on the issues involved. A more detailed
description of our scope and methodology is in appendix I.
--------------------
\1 Cured is defined as being free of any evidence of cancer for 5
years or more. These cured individuals will have the same life
expectancy as others of the same age and sex who have never had
cancer.
\2 Clinical drug trials involve testing a new drug in humans to
determine whether it has therapeutic benefit.
\3 Information was not available for us to document possible alcohol
use by study patients.
RESULTS IN BRIEF
------------------------------------------------------------ Letter :1
In three large clinical trials, NCI found that hydrazine sulfate did
not prolong survival for cancer patients. Nevertheless, controversy
and confusion developed, in part, because some researchers have
suggested that hydrazine sulfate is incompatible with tranquilizers,
barbiturates, and alcohol. In testing hydrazine sulfate, NCI
permitted study patients to use tranquilizing agents, barbiturates,
and alcohol in one NCI-sponsored clinical trial. In the other two
trials, NCI prohibited the use of barbiturates and alcohol, but
patients were permitted to use tranquilizing agents as antiemetics to
control nausea and vomiting. However, subsequent analyses of the use
of concurrent medications found no evidence to invalidate NCI's
conclusion that hydrazine sulfate is ineffective.
Nonetheless, there were lapses in record-keeping and reporting in
these clinical trials. NCI did not require that complete and
accurate research records be kept during one clinical trial
documenting the use of tranquilizing agents, barbiturates, and
alcohol by study patients. Also, NCI-sponsored investigators did not
analyze this issue until recently, and the published results did not
accurately describe the use of tranquilizing agents during one of
these clinical trials.
FDA may have contributed to the confusion surrounding these trials of
hydrazine sulfate with its more conservative position on how the drug
should be administered to some patients. While accepting NCI's study
designs and therapy plans, FDA apparently had concerns over possible
incompatibility. About the same time these three trials were
occurring, FDA approved more than 70 applications permitting the
compassionate use of hydrazine sulfate. These approvals were
accompanied by an FDA caution to physicians that their patients
should avoid taking tranquilizers, barbiturates, or alcohol while
taking hydrazine sulfate. Other FDA documents reflect this same
concern.
BACKGROUND
------------------------------------------------------------ Letter :2
NCI, within the National Institutes of Health (NIH), Public Health
Service, Department of Health and Human Services, is the world's
largest sponsor of clinical trials in cancer treatment research. NCI
spends about 20 percent of its research budget on clinical trials.
Using its clinical trials network that includes cooperative groups,\4
NCI funds therapeutic research that includes evaluating the safety
and efficacy of investigational drugs in large multicenter clinical
trials. NCI also sponsors therapeutic drug development through the
submission of investigational new drug (IND) applications to FDA.
FDA is responsible for ensuring the safety of the public in matters
related to clinical research with investigational drugs. FDA
regulations define the terms under which clinical research may
proceed. Through the INDs, FDA reviews the experimental rationale
for conducting clinical drug trials, including results of animal
toxicology studies, manufacturing data, purity and stability
information, and an initial plan for clinical investigation. The
responsibility for monitoring the trials rests with the sponsor.
--------------------
\4 Clinical cooperative groups are composed of multiple institutions
and investigators who collaborate to develop and implement treatment
research in large numbers of patients. There are 11 cancer clinical
cooperative groups, and all involve highly organized collaborations
among geographically dispersed institutions with central data
management offices and large statistical centers that support the
administrative requirements of the research and perform data
collection and analysis.
EARLY STUDIES OF HYDRAZINE
SULFATE FOUND MIXED AND
INCONCLUSIVE RESULTS
---------------------------------------------------------- Letter :2.1
Unexplained weight loss and physical deterioration of the body
(cachexia) commonly accompany advanced cancer. Moreover, cachexia is
associated with decreased survival. For example, data have shown
that in patients with lung cancer, weight loss is associated with a
50-percent reduction in survival time.
Joseph Gold, M.D., director of the Syracuse Cancer Research Institute
in New York, proposed a theory to explain why cachexia commonly
accompanies advanced cancer. After extensive research, Dr. Gold
proposed the use of hydrazine sulfate, a chemical that interrupts the
abnormal sugar metabolism associated with weight loss, to arrest and
reverse cancer cachexia.
In 1973, Dr. Gold reported on results of animal tests indicating
that hydrazine sulfate inhibited the growth of various rodent tumors
and enhanced the antitumor action of some chemotherapeutic drugs. In
1975, Dr. Gold reported the results of hydrazine sulfate's use in
cancer patients. Using reports from physicians whose advanced cancer
patients were taking hydrazine sulfate, Dr. Gold noted several cases
of tumor regression and subjective improvement in cancer patients
treated with hydrazine sulfate. Additionally, Russian investigators
have claimed some successes with hydrazine sulfate for more than 20
years.\5 Although early clinical studies conducted in the United
States found mixed results, later studies evaluating hydrazine
sulfate as an anticachexia agent suggested that the drug benefited
some cancer patients.
In the 1980s, studies at the Harbor-University of California Los
Angeles (UCLA) Medical Center indicated that adding hydrazine sulfate
to a standard chemotherapy regimen improved the nutritional status
and survival time of some cancer patients. Of particular interest
was a randomized clinical trial--involving 65 patients with advanced,
inoperable non-small-cell lung cancer\6 --that compared chemotherapy
and hydrazine sulfate with chemotherapy and placebo. Data from the
study suggested that hydrazine sulfate may benefit some cancer
patients. While overall survival differences between the two
treatment groups were not significant, researchers found that
hydrazine sulfate improved survival in a subset of study patients who
began the trial in better overall condition. Given the
inconclusiveness of the study, UCLA investigators believed that
further trials of hydrazine sulfate were warranted to determine its
effectiveness in improving survival.
--------------------
\5 The Russian studies, however, were not randomized or blind and
have involved various types and stages of cancer. Randomized
clinical studies are the "gold standard" for evaluating medical
interventions. In a randomized study, each patient is randomly
assigned to one of a number of treatments in order to ensure unbiased
comparison of outcomes. Random assignment is the best method for
ensuring that the patients in each group have equivalent prognoses.
Because the groups should not differ on variables related to cancer
survival, their outcomes can be directly compared, and any difference
in survival can be attributed to the difference in treatment.
\6 Non-small-cell lung cancer is commonly used as a catchall phrase
to include all of the primary malignant lung cancers except small
cell and accounts for approximately 70 to 80 percent of lung cancers.
Almost all primary malignant cancers of the lung are one of four
types: squamous cell carcinoma (cancer originating in the glandular,
skin, or linings of internal organs), adenocarcinoma, large cell
carcinoma, and small cell carcinoma.
NCI-SPONSORED TRIALS FOUND
HYDRAZINE SULFATE
INEFFECTIVE
---------------------------------------------------------- Letter :2.2
NCI sponsored three clinical trials that were designed to assess the
effect of hydrazine sulfate on survival, weight gain, and quality of
life. Two trials in patients with advanced lung cancer assessed the
efficacy of hydrazine sulfate as an adjunct to chemotherapy. One of
these trials, in patients with advanced lung cancer, was conducted by
the Cancer and Leukemia Group B (CALGB) and led by a principal
investigator at the Scripps Clinic and Research Foundation in San
Diego, California.\7 The other trial in advanced lung cancer
patients, was conducted by the North Central Cancer Treatment Group
(NCCTG) and led by a principal investigator at the Mayo Clinic in
Rochester, Minnesota. The third trial assessed the efficacy of
hydrazine sulfate as the sole medical intervention in patients with
advanced colon cancer. This trial was also conducted by NCCTG and
led by the same principal investigator at the Mayo Clinic. Figure 1
shows highlights of activities surrounding NCCTG's and CALGB's
clinical trials of hydrazine sulfate.
Figure 1: Time Line of the
UCLA and NCI-Sponsored Clinical
Trials of Hydrazine Sulfate
(See figure in printed
edition.)
The NCI-sponsored clinical trials did not find the survival advantage
observed in the earlier UCLA study. Results from the three clinical
trials were published in June 1994.\8
The data showed that hydrazine sulfate therapy does not result in any
significant benefit. Specifically, in two trials involving over 500
patients with inoperable non-small-cell lung cancer, the addition of
hydrazine sulfate to a standard chemotherapy regimen resulted in
somewhat worse quality of life, no effect on weight gain or loss, and
a suggestion of decreased survival when compared with placebo. In
the trial evaluating the use of hydrazine sulfate as the sole
therapeutic intervention in 127 patients with metastatic colon
cancer, survival time for patients receiving hydrazine sulfate was
decreased compared with patients given placebo.
--------------------
\7 We are using the term principal investigator to describe the role
of the scientific coordinator (protocol chair) of a multicenter
clinical trial. The principal investigator is responsible for
developing and monitoring the study as well as analyzing, reporting,
and publishing its results.
\8 Results of the three studies of hydrazine sulfate supported by NCI
were published in the Journal of Clinical Oncology, Vol. 12 (June
1994), pp. 1113-29. The articles were "Cisplatin, Vinblastine, and
Hydrazine Sulfate in Advanced, Non-Small-Cell Lung Cancer: A
Randomized Placebo-Controlled, Double-Blind Phase III Study of the
Cancer and Leukemia Group B," by Michael P. Kosty and others;
"Randomized Placebo-Controlled Evaluation of Hydrazine Sulfate in
Patients With Advanced Colorectal Cancer," by Charles L. Loprinzi
and others; and "Placebo-Controlled Trial of Hydrazine Sulfate in
Patients With Newly Diagnosed Non-Small-Cell Lung Cancer," by Charles
L. Loprinzi and others.
NEGATIVE TRIAL RESULTS
CONTESTED
---------------------------------------------------------- Letter :2.3
Criticisms regarding the design of the three clinical trials arose in
the media. Proponents of hydrazine sulfate therapy alleged that NCI
compromised the trials by permitting study patients to ingest agents
that they believe are incompatible with hydrazine sulfate. Some
proponents believed that the concurrent use of tranquilizers,
barbiturates, or alcohol with hydrazine sulfate would nullify the
therapeutic effect of hydrazine sulfate and cause toxicity in
patients. They based their beliefs on Russian and unpublished animal
studies as well as some pharmacological data that they said suggested
that hydrazine sulfate interacts with tranquilizing agents
(particularly tranquilizing agents classified as benzodiazepines),
barbiturates, and alcohol.\9
--------------------
\9 Dr. Mikhail Gershanovich, one of Russia's leading cancer
specialists and a professor at the Petrov Research Institute of
Oncology, told us that he had no evidence of incompatibility between
hydrazine sulfate and tranquilizers. He confirmed, however, that his
studies with hydrazine sulfate have demonstrated an incompatibility
with barbiturates and alcohol.
NCI CONCLUDED CONCERNS THAT
HYDRAZINE SULFATE WAS
INCOMPATIBLE WITH SOME AGENTS
WERE UNFOUNDED
------------------------------------------------------------ Letter :3
NCI rejected concerns that the concurrent use of hydrazine sulfate
with tranquilizers, barbiturates, or alcohol would nullify the
therapeutic effect of hydrazine sulfate. NCI concluded that there
was no objective evidence or published studies of humans addressing
interactions between hydrazine sulfate and these alleged incompatible
agents to support the concerns. NCI concluded that, at most, Russian
animal data suggested that large doses of alcohol or barbiturate
medications consumed with hydrazine sulfate can increase the total
overall toxicity. NCI also concluded that unpublished animal data
did not support the hypothesis that the short-term use of
tranquilizing agents with hydrazine sulfate would increase toxicity
or prevent clinical benefit. These conclusions were based on the
assessment of NCI scientists in consultation with CALGB and NCCTG
researchers. In addition, NCI officials told us that because the
UCLA study did not specifically prohibit patients from taking
barbiturates and tranquilizers or consuming alcohol, the
NCI-sponsored confirmatory trials did not have to be any different in
that regard.\10
Nevertheless, in response to the issue of incompatibility, NCCTG
investigators (at the Mayo Clinic) prohibited patients from taking
barbiturates or consuming alcohol. Furthermore, patients were
prohibited from taking tranquilizing agents, except for antiemetic
purposes. The NCCTG principal investigator told us that he felt it
would have been unethical to perform either of the Mayo clinical
trials without allowing the use of antiemetic drugs, including drugs
otherwise considered to be tranquilizers, by patients experiencing
nausea or vomiting.
CALGB investigators, however, decided that it was more important to
replicate the UCLA trial than attempt to address concerns of
incompatibility by prohibiting the use of tranquilizers,
barbiturates, and alcohol. Because the NCI-sponsored trials were
designed to confirm the improved survival observed in the UCLA trial,
CALGB investigators believed they should use essentially similar
methods to those used in the earlier UCLA trial.
--------------------
\10 Initially, there was some confusion surrounding whether these
alleged incompatible agents, particularly tranquilizers, were
permitted in the UCLA trial. Reports in the media were apparently
inaccurate in suggesting that the UCLA researchers did not allow
patients to take tranquilizers. Some of the confusion may have
developed because in research documents describing protocols for
other trials of hydrazine sulfate, UCLA researchers previously
expressed concerns over potential incompatibility with hydrazine
sulfate and alcohol, tranquilizers, and barbiturates. As noted on p.
10, however, the principal researcher said he did not have such
concerns in the more recent UCLA trial.
NCI-SPONSORED TRIALS INCLUDED
SOME USE OF TRANQUILIZING
AGENTS
------------------------------------------------------------ Letter :4
Published reports of the three trials did not disclose the extent of
tranquilizer use among study patients. In examining research
records, however, we found that patients in all three NCI-sponsored
clinical trials of hydrazine sulfate were prescribed tranquilizers
under varying conditions. In the NCCTG and CALGB clinical trials in
non-small-cell lung cancer, virtually all patients received a variety
of antiemetic drugs, particularly tranquilizing agents, for the
short-term relief of chemotherapy-induced vomiting.
Our review of over 50 percent of CALGB's standard research forms
revealed that patients with advanced lung cancer routinely received
benzodiazepine and phenothiazine tranquilizing agents to relieve the
nausea and vomiting associated with chemotherapy. Although CALGB
investigators decided not to collect data on concurrent medications
on their standardized research forms, some research associates
voluntarily provided information on antiemetic usage. Data on the
use of antiemetic medications for about half of the patients in our
review were recorded on the research forms. Our analysis of research
forms listing antiemetic medications revealed that 88 percent of the
patients received benzodiazepine and 71 percent received
phenothiazine tranquilizing agents.\11 \12 Generally, it appeared
that most patients were prescribed tranquilizing medications for
short-term emetic relief. In several instances, however, patients
were prescribed tranquilizing agents on an "as needed" continual
basis. We also found one instance where a patient was prescribed a
barbiturate.
At our request, NCCTG and CALGB research associates reviewed research
forms, medical records, or both for patients enrolled in their lung
cancer trials to collect data on the concurrent use of antiemetic and
barbiturate medications. Table 1 shows the number of patients
receiving hydrazine sulfate and various antiemetic medications.
Table 1
Number of Patients Receiving Specific
Antiemetic Medications Along With
Hydrazine Sulfate in the Lung Cancer
Clinical Trials
CALGB NCCTG
trial trial
(n=135 (n=116
Medication ) )
------------------------------------------------------ ------ ------
Benzodiazepines 119 43
Lorazepam 113 43
Other 6 0
Phenothiazines 108 43
Prochlorperazine 98 37
Other 10 6
Serotonin antagonists 0 91
Ondansetron 0 90
Other 0 1
----------------------------------------------------------------------
Investigators said it was necessary to prescribe antiemetic
medications, including tranquilizing agents, to patients in all three
clinical trials. In the two clinical trials involving patients with
advanced lung cancer, patients received chemotherapy in addition to
hydrazine sulfate or placebo. Because the chemotherapeutic regimen
used to treat advanced lung cancer induces severe nausea and vomiting
in almost all patients, NCCTG and CALGB investigators did not deem it
feasible or ethical to administer chemotherapy without the concurrent
use of antiemetic drugs.
During the time when NCCTG and CALGB were conducting their trials,
the most effective antiemetic regimens available involved the use of
tranquilizing agents. Accordingly, many study patients in both
clinical trials were prescribed phenothiazines and
benzodiazepines.\13 Although patients in NCCTG's colon cancer trial
were not undergoing chemotherapy, some patients with advanced colon
cancer experience nausea and vomiting associated with their disease.
NCCTG and CALGB investigators told us that they would not deny
standard medical care to control nausea and vomiting in patients who
were dying from cancer.
Also, patients enrolled in the UCLA clinical trial reportedly
received tranquilizing agents while taking hydrazine sulfate.
Medical records for 40 study patients treated at Harbor-UCLA Medical
Center indicated that 22 patients received hydrazine sulfate and
chemotherapy.\14 In addition, these patients received tranquilizing
agents to control their chemotherapy-induced vomiting. Specifically,
patients who received hydrazine sulfate also received a total of 16
doses of benzodiazepines and 20 doses of phenothiazines. Other
possible uses of tranquilizing agents and barbiturates outside of
chemotherapy treatment as well as possible alcohol use are not known.
--------------------
\11 The primary benzodiazepine and phenothiazine drugs prescribed
were lorazepam (Ativan) and prochlorperazine (Compazine),
respectively.
\12 Our results closely paralleled the results obtained in a
comprehensive subsequent review done by CALGB and presented in table
1.
\13 In the late 1980s, tranquilizing agents were used extensively in
oncology to treat chemotherapy-induced vomiting. In addition to
sedation, phenothiazines and benzodiazepines have some antinauseant
and antiemetic effects. Furthermore, benzodiazepines were useful
additions to antiemetic regimens because they relieve anxiety about
nausea or vomiting surrounding chemotherapy treatment.
\14 Because data on concurrent medications were not collected on
standard research forms, we asked the principal investigator to
review the medical records for study patients maintained at
Harbor-UCLA Medical Center. We did not attempt to verify these data.
RETROSPECTIVE ANALYSES FOUND NO
EVIDENCE OF ADVERSE EFFECTS
FROM THE USE OF ALLEGEDLY
INCOMPATIBLE DRUGS
------------------------------------------------------------ Letter :5
Analyses of data from NCI-sponsored clinical trials found no evidence
of adverse effects on survival associated with hydrazine sulfate and
the use of tranquilizing agents as antiemetics and barbiturates.
Researchers at the Mayo and Scripps clinics retrospectively analyzed
clinical trial data in an attempt to address the issue of
incompatibility raised by hydrazine sulfate proponents. Their
analyses suggested that the concurrent use of hydrazine sulfate with
tranquilizing agents or barbiturates did not adversely affect the
survival of lung cancer patients enrolled in the hydrazine sulfate
trials. Also, their post-trial analyses did not change the
conclusions originally drawn from the clinical trials: There was no
benefit for patients who received hydrazine sulfate compared with
those who received placebo.
Because patients who entered later in NCCTG's lung cancer trial did
not receive benzodiazepine tranquilizing agents as antiemetics, NCCTG
investigators were able to retrospectively compare the clinical
outcomes of patients who received benzodiazepines with those of
patients who did not.\15 The data showed no statistically significant
differences in survival time between patients who received hydrazine
sulfate and a benzodiazepine tranquilizer as an antiemetic and
patients who received hydrazine sulfate and new non-benzodiazepine
antiemetics. Furthermore, analyses showed no statistically
significant differences in terms of time to disease progression for
patients who received hydrazine sulfate and a benzodiazepine
tranquilizer compared with those who did not.
CALGB researchers also looked retrospectively at this incompatibility
issue. Beginning in January 1995, CALGB conducted a retrospective
review of primary medical records and documented the medications that
were used by patients enrolled in its clinical trial of hydrazine
sulfate. On June 5, 1995, we received the results of CALGB's
examination of the effect of benzodiazepines, barbiturates, or
phenothiazines on patient survival. The data showed no statistically
significant differences in survival between patients who received
hydrazine sulfate and barbiturates or benzodiazepines or
phenothiazines and patients who received hydrazine sulfate but none
of these allegedly incompatible agents. Furthermore, the data also
showed no statistically significant differences in survival between
patients who received hydrazine sulfate and barbiturates or
benzodiazepines or phenothiazines and patients who received placebo
and any of these agents.
--------------------
\15 Patients entered later in NCCTG's clinical trial were treated
with new antiemetic agents, primarily ondansetron (Zofran).
Ondansetron was the first of a whole new class of
antiemetics--serotonin receptor antagonists--to become widely
available for the prevention of nausea and vomiting associated with
emetic chemotherapy. The development of serotonin receptor
antagonists represented a vast improvement in controlling vomiting.
For example, ondansetron works well even against chemotherapy
regimens most notorious for causing severe vomiting in most people.
FDA HANDLED THE ISSUE OF
POSSIBLE INCOMPATIBILITY
SOMEWHAT DIFFERENTLY FROM NCI
------------------------------------------------------------ Letter :6
FDA handled the issue of possible incompatibility differently in
approving the use of hydrazine sulfate by individual physicians than
it did in approving NCI's sponsored clinical trials. FDA recommended
that NCI-sponsored investigators monitor study patients to detect
possible interactions between hydrazine sulfate and possible
incompatible agents. However, while NCI was conducting its clinical
trials, FDA was cautioning other physicians to avoid possible
incompatible agents when administering hydrazine sulfate.
In reviewing NCI's IND applications to conduct clinical trials of
hydrazine sulfate, FDA raised safety concerns to NCI regarding
hydrazine sulfate's interactions with other drugs, including
tranquilizing agents. In his review of NCI's IND, the FDA medical
officer stated, "The following drugs are interdicted, due to known
interactions: ethanol [alcohol], barbiturates, and tranquilizers."
This was followed by a recommendation that NCI outline all
precautions to be taken by study investigators "to fully explore the
neurotoxic potential of hydrazine." NCI complied.
FDA took a more conservative view of the use of possible incompatible
agents with hydrazine sulfate under its compassionate use program.
Before completion of NCI's sponsored clinical trials, FDA approved
more than 70 applications permitting the compassionate use of
hydrazine sulfate. Because of publicity given to hydrazine sulfate,
FDA received many requests from individual physicians for approval to
use hydrazine sulfate on a case-by-case "compassionate" basis on the
chance that patients with no other available effective therapy might
benefit.\16
A central nervous system depressant effect associated with hydrazine
sulfate consistently prompted FDA to caution patients regarding the
use of hydrazine sulfate with any potential sedative agent.\17 In its
approvals, FDA staff requested that physicians caution their patients
to avoid tranquilizers, barbiturates, and alcohol while taking
hydrazine sulfate. FDA officials told us that the reason for this
instruction was that these physicians were not trained clinical
investigators and, under the circumstances, would be less likely to
recognize adverse reactions from interactions between hydrazine
sulfate and possible incompatible agents.
--------------------
\16 In the case of a serious disease, a drug that is not approved for
marketing but is under clinical investigation may be made available
for a serious or immediately life-threatening disease in patients for
whom no comparable or satisfactory alternative drug or other therapy
is available.
\17 FDA may also have been influenced by positions taken by other
researchers in trials of hydrazine sulfate. In earlier INDs
submitted to FDA, a few researchers had developed study protocols or
therapy plans that said patients should not receive alcohol,
barbiturates, or tranquilizers.
NCI DID NOT ENSURE COLLECTION
OF DATA ON THE USE OF ALLEGED
INCOMPATIBLE AGENTS
------------------------------------------------------------ Letter :7
NCI contributed to the subsequent controversy surrounding these
trials by not requiring better data collection and analysis of this
issue. Although NCI officials were aware of the concerns surrounding
the use of allegedly incompatible agents with hydrazine sulfate, they
did not believe it was necessary to maintain research records during
its trial regarding concurrent medications and possible alcohol
use.\18 NCI and CALGB documents, however, stated that all data,
including concurrent medications taken by study patients, would be
recorded on standardized research forms.
The Chairman of CALGB told us that the cooperative group decided
before starting the trial that they would not require such data to be
recorded on standardized research forms because the evaluation of
concurrent medications with hydrazine sulfate was not a study
objective. Even after the trial began, however, the CALGB principal
investigator and an NCI official assured proponents of hydrazine
sulfate that CALGB would collect and analyze data on antiemetic
medications to determine the possible effects of benzodiazepines and
phenothiazines on patients taking hydrazine sulfate. A statement
submitted by NCI staff to media representatives pointed to the
purpose for this:
"[A]ll concurrent medications were well documented in the Cancer
and Leukemia Group B (CALGB) study (a routine component of
clinical trials data collection) so that any differences in
study outcomes could be reviewed from the perspective of these
potential `incompatibles'."
Despite these assurances, CALGB did not uniformly collect data on the
use of concurrent medications, including tranquilizing agents and
barbiturates, and possible alcohol use. Furthermore, in a published
article describing the results of the clinical trial, CALGB
investigators incorrectly reported that data on the use of concurrent
medications were recorded on standardized research forms. CALGB
investigators should have accurately reported their data collection
efforts. In addition, NCI should have ensured that CALGB
investigators prospectively collected data on concurrent medications
and alcohol use on research forms to permit investigators to analyze
trial data to determine the possible effects of these agents on
patients taking hydrazine sulfate.
--------------------
\18 As stated above, data on the actual antiemetics as well as any
concurrent medications used by patients in CALGB's clinical trial
were available from patients' primary medical records.
CALGB'S PUBLISHED TRIAL RESULTS
DID NOT ACCURATELY DESCRIBE
TRANQUILIZER USE
------------------------------------------------------------ Letter :8
A paper presenting the final results of the CALGB clinical trial did
not clearly describe the use of tranquilizing agents by study
patients. Authored by the principal investigator for the trial, this
scientific paper did not accurately reflect the widespread use of
tranquilizing agents in the CALGB lung cancer trial.
In the published paper, the investigator wrote that "no patients
received barbiturates and virtually no patients received
phenothiazine-type tranquilizers, with the exception of
prochlorperazine . . ., which was used as a short-term antiemetic."
Data from the medical records, however, showed that phenothiazines,
including prochlorperazine, were prescribed to 80 percent of study
patients. In addition, over 88 percent of study patients were
prescribed benzodiazepines. Medical records also showed that
approximately 5 percent of study patients were treated with
barbiturates. The principal investigator told us that he used data
submitted by some research associates to form his "impressions" of
concurrent medication usage. Because CALGB did not routinely collect
data on concurrent medications, however, the data used to support his
impressions are not an accurate and complete reflection of
information contained in the medical records.
In a letter to us dated February 27, 1995, the Chairman of the CALGB
cooperative group said the principal investigator would prepare a
letter to the Journal of Clinical Oncology correcting his statement
regarding study patients' use of barbiturates. The Chairman told us,
however, that he believed the description of tranquilizer use was
accurate. He based his assessment on, first, the fact that most
medical records did not indicate that phenothiazines were prescribed
for long-term use as tranquilizers. Second, the tranquilizing
agents, phenothiazines and benzodiazepines, were interchangeable in
the investigator's description of their use as short-term
antiemetics. Accordingly, he concluded that the principal
investigator was justified in stating that "virtually no patients
received phenothiazine-type tranquilizers." We disagree with the
Chairman in this regard. We believe the investigator erred in not
reporting the widespread use of benzodiazepine tranquilizing agents.
In June 1995, the Journal of Clinical Oncology published a letter to
the editor from CALGB correcting and clarifying CALGB's published
results.\19 The letter corrected information on the use of
barbiturates during CALGB's clinical trial. The letter also
clarified that in addition to the use of a phenothiazine
tranquilizing agent as an antiemetic, many patients received a
benzodiazepine antiemetic.
--------------------
\19 Vol. 13 (June 1995), pp. 1529-30.
CONCLUSIONS
------------------------------------------------------------ Letter :9
In three large, randomized, placebo-controlled clinical trials
sponsored by NCI, hydrazine sulfate was ineffective in extending the
survival time for certain cancer patients. The developer of
hydrazine sulfate therapy has suggested that the trials were
compromised because investigators permitted some study patients to
take agents that are possibly incompatible with hydrazine sulfate.
We confirmed that all three trials permitted some use of
tranquilizing agents to varying degrees and one trial permitted the
use of barbiturates and alcohol. Specifically, many patients
received short-term dosages of tranquilizers for antiemetic purposes.
Retrospective analyses, however, found no evidence that the use of
allegedly incompatible agents adversely affected NCI's clinical trial
results. Although our work did not support the allegation that the
studies were flawed, NCI should have made sure that complete and
accurate records were kept during CALGB's clinical trial regarding
concurrent medications and possible alcohol use. Furthermore, this
issue should have been analyzed on a more timely basis in the NCCTG
and CALGB clinical trials, and the published results of CALGB's trial
should have been accurate with regard to tranquilizer use.
AGENCY COMMENTS
----------------------------------------------------------- Letter :10
In commenting on a draft of this report, the Public Health Service
agreed with the report's main conclusion that there is no evidence to
support the allegation that the three trials sponsored by NCI were
flawed. In addition, retrospective analyses suggested that the use
of tranquilizers as antiemetic agents, barbiturates, or alcohol by
patients receiving hydrazine sulfate did not produce greater
toxicities or interfere with hydrazine sulfate's alleged benefits.
(See app. II for a copy of the Public Health Service's comments.)
The Public Health Service did not agree, however, that either NCI or
the clinical investigators were remiss for not ensuring that
concurrent medications were recorded on research forms. NCI and
CALGB documents provided that data on concurrent medications would be
recorded on research forms. As noted previously, NCI staff wrote to
media representatives that all concurrent medications were well
documented as a routine part of the trial's data collection so that
any differences in outcomes could be analyzed in terms of the
allegedly incompatible agents.
Although CALGB informed NCI that the use of concurrent medications
would be captured on the patient research forms in accordance with
the research plan, CALGB investigators did not uniformly record this
information on the forms as originally intended. Under the terms of
the cooperative agreement that provided funding for CALGB's clinical
trial, it was the responsibility of CALGB to record such data. NCI
officials told us that the agency has specific expectations with
respect to cooperative group performance and it is the grantee's
responsibility to successfully accomplish these. We believe,
however, that NCI, as the funding agency, has the oversight
responsibility for ensuring that their expectations are met.
Furthermore, CALGB should have complied more completely with its
proposed plan for data recording.
The Public Health Service also does not believe that NCCTG and CALGB
clinical investigators should be criticized for not having analyzed
data on concurrent medications promptly. The issue of
incompatibility was consistently part of the public controversy
surrounding the NCI-sponsored clinical trials of hydrazine sulfate.
Therefore, we believe that NCI was remiss as were NCCTG and CALGB
investigators for not settling the controversy by promptly analyzing
data on the impact of specific medications on the effects of
hydrazine sulfate.
The Public Health Service agreed that the initial published article
describing the findings of CALGB's study was not accurate with
respect to the use of tranquilizing agents as antiemetics and
barbiturates. NCI criticized this lapse and ensured that a letter
from the CALGB investigator was published that provided more complete
and accurate information.
The Public Health Service also provided technical comments which have
been incorporated where appropriate, in our report.
--------------------------------------------------------- Letter :10.1
We are sending copies of this report to the Secretary of Health and
Human Services and the Director of NCI; the Commissioner of Food and
Drugs; and interested congressional committees. Copies will also be
made available to others upon request.
Please call me at (202) 512-7119 if you or your staff have any
questions. Other major contributors to this report include Barry D.
Tice, Assistant Director, (202) 512-4552, and Gloria E. Taylor.
Mark V. Nadel
Associate Director, National
and Public Health Issues
SCOPE AND METHODOLOGY
=========================================================== Appendix I
To obtain information for this report, we reviewed NCI's policy
guidance for conducting clinical trials of investigational agents,
agency memorandums documenting protocol development for the hydrazine
sulfate clinical trials, and related correspondence. We also
discussed the conduct of these trials with NCI officials, cooperative
group representatives and investigators, FDA officials, officials in
the Office of Research Integrity, and proponents of hydrazine sulfate
to obtain their perspectives on the issues involved. We performed an
extensive literature search on hydrazine sulfate as well as topics
related to cancer research, the conduct of clinical trials,
approaches to chemotherapy treatment, and drugs to control
chemotherapy-induced vomiting. In addition, we discussed the issue
of incompatibility with a leading Russian researcher and viewed
several hours of a taped interview with senior Russian oncologists.
We also discussed the interpretation of animal data with experts in
pharmacology.
In our examination of the extent to which barbiturates and
tranquilizers were used during the clinical trials, we reviewed
research records maintained by the data management and statistical
centers for each cooperative group. For the CALGB clinical trial, we
visited the cooperative group's Data Management Center located at
Duke University. We randomly selected research records for 137 of
291 study patients for review. For the NCCTG clinical trial, we
visited the Data Management Center for the cooperative group at the
Mayo Clinic. Before our arrival, NCCTG research associates had
compiled a list of antiemetic medications administered to each study
patient. We randomly selected 15 percent of 116 study patients'
research records to verify the accuracy of NCCTG's data collection
efforts.
We conducted our work from July 1994 to April 1995 in accordance with
generally accepted government auditing standards.
(See figure in printed edition.)Appendix II
COMMENTS FROM THE PUBLIC HEALTH
SERVICE
=========================================================== Appendix I
(See figure in printed edition.)
(See figure in printed edition.)
(See figure in printed edition.)