|
Guidance for Industry
Catheter-Related Bloodstream Infections - Developing Antimicrobial Drugs for Treatment
DRAFT GUIDANCE
[Acrobat version of this document]
This guidance document is being distributed for comment purposes only.
Comments and suggestions regarding this draft document
should be submitted within 90 days of publication of the Federal Register notice
announcing the availability of the draft guidance. Submit comments to Dockets Management
Branch (HFA-305), Food and Drug Administration, 5630 Fishers Lane, rm. 1061, Rockville, MD
20852. All comments should be identified with the docket number listed in the notice of
availability that publishes in the Federal Register.
For questions on the content of the draft document contact
Renata Albrecht, 301-827-2336.
U.S. Department of Health and Human
Services
Food and Drug Administration
Center for Drug Evaluation and Research (CDER)
October 1999
Clinical Antimicrobial
Guidance for Industry
Catheter-Related Bloodstream
Infections - Developing Antimicrobial Drugs for Treatment
For additional copies contact the:
Drug Information Branch
Division of Communications Management, HFD-210
5600 Fishers Lane, Rockville, MD 20857
(Tel) 301-827-4573
http://www.fda.gov/cder/guidance/index.htm.
Table of Contents
I. INTRODUCTION
II. BACKGROUND
III. CATHETER-RELATED BLOODSTREAM
INFECTIONS
GUIDANCE FOR INDUSTRY_
Catheter-Related Bloodstream
Infections - DevelopingAntimicrobial Drugs for Treatment
I. INTRODUCTION
This is one in a series of guidances intended to assist
pharmaceutical manufacturers in developing antimicrobial drug products to treat
infections. The information presented in this document will provide most if not all of the
information that should be used to plan the necessary clinical studies, design the
clinical protocols, implement and appropriately monitor the clinical studies, collect
relevant data needed for analysis, and perform the appropriate types and numbers of
analyses of the study data. The results of studies planned and conducted in accordance
with this guidance are expected to yield information that the Agency can use to determine
whether the antimicrobial under study is safe and effective in the treatment of the
specific infection. For general information on related topics, the reader is referred to a
draft guidance entitled Development of Antimicrobial Drug
Products - General Considerations (July 1998), which currently is being finalized.
This draft guidance focuses on developing antimicrobials
for the treatment of catheter-related bloodstream infections. For purposes of this draft
guidance, bibliographic references are provided in endnote format.
II. BACKGROUND
Over the years, the Agency has issued guidance to the
pharmaceutical industry on how to design, carry out, and analyze the results of clinical
trials for the development of antimicrobials for the treatment of infections in a variety
of forms. This draft guidance is the result of efforts to collect all pertinent
information on one type of infection and present it in one location. Where appropriate,
this guidance contains relevant information from several sources, including Clinical Evaluation of Anti-Infective Drugs (Systemic) (1977); IDSA's "Guidelines for the Evaluation of
Anti-Infective Drug Products" (1992) (IDSA guidance); Points to Consider: Clinical Development and Labeling of
Anti-Infective Drug Products (1992) (Points to Consider),
an FDA guidance on issues related to evaluating new drug applications for anti-infective
drug products; and Evaluating
Clinical Studies of Antimicrobials in the Division of Anti-Infective Drug Products (February 1997), a draft guidance discussed at a March 1997
advisory committee meeting on anti-infective drug products.
III. CATHETER-RELATED
BLOODSTREAM INFECTIONS
A. Disease Definition
For the purpose of this guidance, catheter-related
bloodstream infections are defined as bloodstream infections resulting from an
infected vascular access device or contaminated infusate, including central venous
catheters (tunneled [e.g., Hickman], subcutaneously implanted [e.g., Porta-cath], and
nontunneled), peripherally inserted central venous catheters (PICC lines), midline
catheters, vascular dialysis catheters (e.g., Quinton catheters), pulmonary artery
catheters, peripheral arterial catheters, and peripheral venous catheters. Not included in
this guidance are infections related to or associated with permanent intravascular devices
(such as vascular grafts or implantable pacemakers or defibrillators), intravascular
transplants (such as porcine cardiac valves), or nonintravascular devices (such as
peritoneal dialysis catheters or neurosurgical devices such as ventriculoperitoneal
shunts, ICP monitors or epidural catheters).
The most common bacterial pathogens in catheter-related
bloodstream infections are also common skin colonizers (with the suspected portal of entry
being the actual catheter insertion site in most cases) with staphylococcal species
accounting for one-half to two-thirds. Of these, coagulase-negative species predominate,
but Staphylococcus aureus remains a common cause of these infections.10
Enterococci, particularly vancomycin-resistant strains, account for 8 percent of all
catheter-related bloodstream infections.1,2 Candida albicans and other fungal
pathogens have become increasingly important causes of catheter-related bloodstream
infections in recent years, accounting for roughly 10 percent of nosocomial bloodstream
infections.3 Gram-negative enterics
account for the majority of the remainder, with pathogens such as Klebsiella spp., Enterobacter
spp., and Serratia marcescens most commonly seen in patients with such risk
factors as recent gastrointestinal or genitourinary tract surgery and/or manipulations.4 Among neutropenic patients, Pseudomonas
aeruginosa is a common pathogen.
B. Regulatory Synonyms
These infections are sometimes also referred to as catheter-related
bacteremia. However, the term catheter-related bloodstream infection is
preferable, since the latter term emphasizes the need for a diagnosis to be based on both
clinical and microbiologic criteria. Terms such as line sepsis, catheter-related
septicemia, primary bacteremia, and bacteremia of unknown origin are not
synonymous with the term catheter-related bloodstream infection.
C. Study Considerations
1. General Study Characteristics
Two statistically adequate and well-controlled trials are
recommended establishing safety and effectiveness (i.e., similar or superior effectiveness
to an approved product). Generally, superiority trials should be performed when there is
no approved comparator, as is the case with this indication at present. In these trials,
an evaluable patient should be both clinically and microbiologically evaluable. A single
superiority trial of the test drug may be sufficient under the circumstances outlined in
the FDA guidance for industry, Providing Clinical Evidence of Effectiveness for Human
Drug and Biological Products (May 1998). Two equivalence trials might be sufficient to
support approval under certain circumstances, as discussed in section III.1. Trials should
be double-blind whenever possible.
2. Scope
The purpose of this guidance is to propose consistent
methodologies in the design of clinical trials in which catheter-related bloodstream
infections are being studied. More specifically, bloodstream infections resulting from
either an infected vascular access device or contaminated infusate will be discussed.
This guidance focuses on bacterial infections, though many
of the concepts that will be proposed could apply to fungal bloodstream infections related
to intravascular access devices. The guidance focuses on bloodstream infections that have
been shown to be directly related to one of the intravascular devices listed. Thus, this
guidance is not intended for the study of patients with bacteremia of unknown origin
or with bacteremia due to a focus of infection other than the intravascular device. Entry
of patients into clinical trials evaluating catheter-related bloodstream infections
should, in part, depend on excluding another sources of the bacteremia.
This guidance is intended for use in studies in adult
patients, but as the clinical experience with catheter-related bacterial bloodstream
infections in pediatric patients (including neonates) expands, it is envisioned that this
guidance will be expanded to include this age group.
3. Diagnosis
The diagnosis of catheter-related bacterial bloodstream
infections is difficult for the following reasons:
a. Lack of pathognomonic clinical signs and/or symptoms
Although these infections are usually associated with the
presence of fever, a study of intensive-care unit patients with new onset of fever found
that 80 to 90 percent of these fevers were not associated with a documented catheter
infection.5 It
has been estimated that 75 to 85 percent of catheters are removed unnecessarily during
evaluation of new fever.5 In one study over 70 percent of documented central
venous catheter-related bloodstream infections were not associated with signs or symptoms
of local inflamation at the catheter entry site.6
The absence of specific clinical signs and symptoms associated with catheter-related
infection makes the diagnosis and evaluation of such infections difficult.
b. Difficulties with culturable material
When no obvious signs of inflammation at the catheter
entry site are seen, the diagnosis of a catheter-related infection depends on either blood
cultures drawn through the catheter or cultures of the catheter itself. A diagnosis of
catheter-related infection on the basis of blood culture alone (without cultures of
catheter hardware) can be made on the basis of quantitative differences between colony
counts of a pathogen isolated from a blood culture obtained through the catheter and
colony counts from a simultaneously obtained peripheral blood culture. Due to the cost and
relative unavailability of quantitative blood cultures, this technique has not been widely
used. The most accepted methods of diagnosing a catheter-related infection have involved
either quantitative or semi-quantitative cultures of the catheter tip.5 Thus,
removal of the catheter is often necessary to diagnose these infections.
c. Lack of consistency in diagnostic techniques
A recent meta-analysis surveyed the English-language
medical literature for the years 1966 to 1994 for studies evaluating techniques in
diagnosing catheter-related bloodstream infections.5 Sixteen different
diagnostic methods with 17 variations were described. Few studies have examined methods in
similar patient populations, but in those studies that have, large differences were noted
in both sensitivity and specificity. Due to such wide discrepancies in the ability of
various techniques to accurately diagnose a catheter-related bloodstream infection, it is
difficult to pool data from different studies.
Therefore, several standards exist that have been adopted
and used by investigators. Enrollment of patients into studies of these infections has
depended on microbiologic criteria and on the presence of fever, with secondary emphasis
placed on other clinical signs and symptoms. The following criteria have been most
commonly adopted:
· All other potential foci must be ruled out.
· Patients without another potential focus who have
inflammation and other signs of infection at the catheter insertion site or tunnel and a
concomitant positive blood culture are classified as having a true catheter-related
bloodstream infection.
· In patients without local signs/symptoms, diagnosis of
catheter-related bloodstream infections depends on the material available for culture. A
quantitative or semi-quantitative tip culture with growth of a pathogen identical to that
in a concomitant blood culture fulfills microbiologic criteria for catheter-related
infection.
· In situations where the catheter is not available for
culture, paired quantitative blood cultures obtained peripherally and from the catheter
have been compared. A 3:1 or 5:1 ratio between colony counts for a pathogen from the
catheter-drawn culture and a peripheral culture indicates a catheter-related bloodstream
infection.7,8 New methods, such
as comparing times to growth in automated blood culture systems or the use of staining
techniques (such as acridine orange) have been proposed as well.
4. Epidemiology
More than 150 million intravascular catheters are
purchased annually by clinics and hospitals in the United States, including more than five
million central-venous and pulmonary-artery catheters.7 However, due to the
differences in disease definition discussed above, the true incidence of catheter-related
bloodstream infections remains unknown. Estimates range from 25,000 up to 400,000 per
year.7,9 Based on bloodstream infection
rates reported in large Centers for Disease Control and Prevention studies, the estimate
of 400,000 may be closer to the true incidence. Catheter-related bloodstream infections,
because of the medical conditions with which they are associated, increase the risk of
morbidity (such as prolonged hospital stays)5 and death. Mortality rates
associated with catheter-related bloodstream infections range from 10 to 20 percent. The
estimated percentage of all bacterial bloodstream infections in the adult
population that are related to a catheter ranges from 5 to 15 percent, though experts in
the field believe the incidence to be higher.10
5. Therapy
As with diagnosis, the therapy of catheter-related
bloodstream infections has involved a wide variety of considerations.
· Catheter removal
When the source of a bacteremic infection is suspected to
be a peripheral intravenous catheter, the standard of care has been to remove the line and
establish access at a new site.7,11 For long-term catheters such as PICC lines,
central venous lines, and arterial lines recent literature strongly suggests that with
certain pathogens, particularly Staphylococcus aureus, Pseudomonas aeruginosa,
Gram-negative enterics, and Enterococcus faecium, catheter removal should be the
first step in the treatment of the related bloodstream infection. However, for the most
common group of pathogens, the coagulase-negative staphylococci, there continues to be
debate as to whether catheter removal is necessary. When this group of pathogens is
involved, the decision to remove the catheter is highly dependent on individual patient
factors. Pathogen factors, such as biofilm production or colony-size variants, may also be
important.
· Site of new catheter
When the catheter needs to be removed, the next issue to
consider is whether a new catheter insertion site needs to be established or whether a new
catheter can be placed into the former insertion site (i.e., changing a catheter over a
guidewire). Guidance concerning this matter has not been established. A recent
meta-analysis of all published articles dealing with this issue suggests that changing a
catheter over a guidewire carries a higher risk of reinfection than if a new site is
established.11 Of note, the increased risk was small and the authors suggested
that very large studies would be needed to establish whether this is a significant
difference.
· Whether to treat with antimicrobials
Another controversial issue is whether systemic
antimicrobial therapy is always needed, and for how long, after a potentially infected
catheter is removed, or whether only removal of the focus is needed to clear a
catheter-related bloodstream infection. Virulent pathogens and/or those known to readily
cause metastatic infections (such as Staphylococcus aureus) are treated with
antimicrobial therapy after catheter removal. The length of therapy depends on the
individual patient's clinical status, co-morbidities and the pathogen. However, with
coagulase-negative staphylococci, especially if the focus of infection is a peripheral
intravenous catheter, the importance of antimicrobial therapy relative to catheter removal
is less clear.
· Follow-up
With certain pathogens, notably Staphylococcus aureus,
a bloodstream infection due to an infected catheter may lead to distant infections that
may not manifest until weeks to months have elapsed (such as osteomyelitis). While such
infections can occur after a prolonged time, the literature is unclear about what
percentage of patients are expected to have such long-term sequelae and at what point the
initiation of antimicrobial therapy for the initial catheter-related bloodstream infection
will prevent these late infections.
6. Incorporating Guidance into the Design of Clinical
Trials
a. Primary Enrollment and Efficacy Endpoints
Enrollment and efficacy determinations will be driven by
microbiologic criteria. However, basic clinical signs and/or symptoms are proposed that
would be needed for enrollment and that would be used in the final efficacy analysis. The
clinical criteria chosen represent a compromise, recognizing that some patients with
catheter-related bloodstream infections may not meet the definitions proposed here.12 On the one hand, given the controversy as to
whether antimicrobial therapy is needed in certain situations, the criteria are strict
enough so that only patients who unequivocally require antimicrobial therapy would be
enrolled. On the other hand, due to the wide variability in clinical presentations of
catheter-related bloodstream infection, the criteria are flexible enough so as not to make
enrollment prohibitively difficult.
b. Microbiologic Criteria
Evaluability and efficacy decisions will be based
primarily on microbiologic criteria; therefore, the criteria proposed are intentionally
strict.
c. Line Removal
The criteria for line removal should be defined
prospectively and applied uniformly for all patients within a randomization stratum. If
line removal is not required at enrollment, patients requiring line removal more than 72
hours after initiation of therapy because of clinical failure or bacteriologic persistence
or relapse should be considered treatment failures.
Changing lines over a guidewire as a substitute for line
removal may cause a discrepancy in efficacy rates and is discouraged. If performed as part
of the study, criteria for this practice should be specified prospectively and applied
uniformly. When this approach is used, a separate subset analysis should be performed for
patients whose lines were changed over a guidewire.
d. Inclusion/Exclusion Versus Evaluability Criteria
Due to difficulties in diagnosing catheter-related
bloodstream infections, a large proportion of patients enrolled into a study may
ultimately be found not to have this infection. On the other hand, strict entry criteria
that are based on the presence of a proven catheter-related infection will not allow for
the enrollment of patients in whom empiric therapy must be started. Because a major
emphasis in the final approval decision will be on the results in the subset of patients
with a proven catheter-related bloodstream infection, sponsors are encouraged to enroll
enough patients in whom this infection is proven or strongly suspected.
e. Randomization
The sponsor should decide, prior to study initiation,
between a prospective stratification of randomization versus planned, poststudy subgroup
analyses. The former approach would be more valuable in a clinical trial when the study
population has either proven or strongly suspected catheter-related bloodstream infections
at the time of enrollment, so that the evaluability rates are high. The latter approach
would be more valuable in a clinical trial in which more severely ill patients are
enrolled in whom empiric therapy is started in a large percentage before a
catheter-related infection is proven. In such a study, large numbers of patients could be
found to be unevaluable, so that subgroup analyses would be more heavily relied on for
efficacy analysis. Potential strata to use in either analysis approach include presence or
absence of neutropenia, age, and severity of illness (such as stratification by APACHE II
scores). Other possible strata that would need to be discussed with the FDA in advance
could include type of device (e.g., arterial catheters, PICC lines), use of
antimicrobial-impregnated catheters, and pathogen(s) of interest.
D. Inclusion Criteria
To be enrolled, patients should have at least one of the
two clinical criteria listed below and at least one of the microbiologic
criteria listed below. However, there will be clinical trials where empiric therapy will
be started before microbiologic confirmation. In such situations, at least one clinical
criterion should be met for the patient to be enrolled, and the microbiologic criteria
should be used as part of the evaluability criteria.
Clinical criteria:
Temperature ³ 38.0°C or < 36°C, with one of the
following:
· WBC count >12,000 or <4,000, or with a
differential
count showing ³ 10% band forms
· Tachycardia: Pulse rate > 100 bpm
· Tachypnea: Respiratory rate > 20 breaths/minute
· Hypotension: Systolic blood pressure < 90 mm Hg
or
Signs and symptoms of localized catheter-related infection
(tenderness and/or pain, erythema, swelling, purulent exudate within 2 cm of entry site)
Microbiologic criteria:
The concordant growth of the same organism from peripheral
blood and one of the following:
· A blood culture aspirated from a catheter, as
shown by quantitative cultures of catheter-drawn and peripherally drawn blood cultures
with a catheter to peripheral blood culture organism ratio of 3:1 to 5:1, regardless of
pathogen.1,4,14,15
· A culture of a catheter segment, as shown by
quantitative cultures of the catheter segment where the number of organisms is ³ 103
CFU/segment, regardless of pathogen13;
or semiquantitative cultures of a catheter segment (i.e., Maki technique) where the number
of colonies of an organism cultured from the catheter tip is > 5 CFU/segment,
regardless of pathogen.14,15
· A culture of the interior surface of a catheter hub,
as shown by quantitative cultures of the catheter hub where the number of organisms is ³
103 per segment of catheter.5 This criterion applies to pathogens that are
common skin colonizers, such as coagulase-negative staphylococci. For pathogens that are
not common skin colonizers (e.g., Pseudomonas aeruginosa), concurrent cultures of
the interior surface of the catheter hub, regardless of colony count.16
· A culture of a catheter entry site exudate, as
shown by concurrent cultures of the catheter entry site, regardless of pathogen and
regardless of colony count.5, 17,18
· A culture of infusate, as shown by concurrent
cultures of the infusate, regardless of pathogen and regardless of colony count.
Definition of concordant
For all pathogens, the peripheral blood culture and the
catheter-related culture (as outlined above) should have growth of the same species. These
species should have either the same pulsed field gel electrophoresis (PFGE) profile or the
same antibiogram.19,20,21,22 For cases in which the pathogen is a common
colonizer for which different strains may have identical antibiograms (e.g.,
coagulase-negative staphylococci),23
use of PFGE is strongly recommended. Use of a particular method to demonstrate concordance
should be supported by data showing that the method is capable of distinguishing between
different strains of the same organism, and of distinguishing between contamination and
true infection.
E. Exclusion Criteria
The exclusion criteria have been divided into three
categories.
1. Exclusion of other endovascular infections:
· Patients with clinical and/or echocardiographic
evidence of endocarditis
· Patients with prosthetic cardiac valves
· Patients with vascular grafts
· Patients with septic thrombophlebitis
· Patients without a pre-existing vascular access device
with community-acquired bacteremia
2. Exclusion of other infections resulting in
bacteremia
· Patients with clinical or radiographic evidence of
osteomyelitis
· Patients with skin/skin structure infection, pneumonia,
urinary tract infection, joint infection, intra-abdominal infection, or other infection
known to be due to the organism cultured from the blood
3. Other exclusion criteria
· Administration of >24 hours of potentially effective
anti-microbial therapy within 72 hours of enrollment
· High probability that line removal alone will cure the
infection
· High probability of death from an unrelated underlying
disease within 14 days
· Hypersensitivity to the study drugs
· Renal or hepatic dysfunction, except as specifically
provided for in the protocol
F. Drugs and Dosing Regimens
1. Investigational Agent
Data should be submitted demonstrating that the pathogens
to be studied are susceptible in vitro to the study drug, including information from
animal models. Because some of the pathogens implicated in catheter-related bloodstream
infections can metastasize to various body sites (as seen with Staphylococcus aureus),
an investigational agent should be shown to achieve adequate concentrations in both serum
and various tissues and fluids. Preferably, the investigational agent should be
bactericidal against the pathogen(s) of concern.
Studies should be designed to demonstrate that, at the
dosing regimen to be studied, the investigational agent achieves and maintains
concentrations predicted to inhibit 90 percent of clinical strains of the pathogens of
concern (i.e., MIC90); for patients with impaired immunity (e.g., neutropenic
patients), achievement of bactericidal concentrations may be recommended. The
concentrations that need to be achieved will depend on the pharmacodynamic parameter most
related to the investigational drug's activity (e.g., concentration-dependent versus
time-dependent activity).
2. Comparator Agent
The sponsor should clearly specify the comparator to be
used in the clinical trial(s). At this time there are no approved agents for this
indication and, thus, the sponsor should choose the most appropriate standard of care as
the comparative agent(s). This choice should be discussed with the Agency prior to study
initiation. The sponsor can consider a dose-response study design. This approach may be
problematic when trying to show a dose/efficacy response, given the high efficacy rates
seen in clinical studies in which patients with mild-to-moderate severity of illness were
treated. A dose-response study design may be most feasible when studying a population of
patients with high severity of illness scores.
3. Adjunctive Therapy
With seriously ill patients, adjunctive and concomitant
therapies are commonly used, such as vasoactive drugs and anti-fungal agents. The sponsor
should make sure that the same standard of care is used in both the study drug and
comparator drug arms. In addition, the sponsor should consider any potential antagonistic
or synergistic effects due to drug-drug interactions. Such factors may affect not only
efficacy rates, but the adverse event profile as well.
4. Duration of Therapy
The duration and timing of therapy should be specified
prospectively in the protocol and may be pathogen-dependent. For example, a 14-day course
of therapy may be appropriate for more virulent pathogens while a shorter duration of
therapy may suffice for infections due to less virulent pathogens. The duration of therapy
will also depend on the nature of the study population enrolled, with longer courses
anticipated for neutropenic patients, as an example. For evaluation of a therapeutic
response the patient should receive at least 80 percent of the intended regimen for at
least 72 hours.
5. Switch in Therapy
Depending on the patient population to be studied, oral
therapy may be considered, either as the initial therapy or as the relay therapy after
several days of intravenous antimicrobial therapy. Criteria for switching from intravenous
to oral therapy should be prospectively defined in the study protocol.
G. Evaluation Visits
The following evaluations are recommended. At each of
these visits, two sets of peripheral blood cultures should be obtained; in situations
where the catheter is not removed, blood cultures through the catheter should be obtained
as well. In situations where the initially infected catheter is removed, cultures from the
new catheter are not needed unless there is evidence for infection of the new catheter.
These visits are:
1. Entry
At the initial evaluation, the following information
should be obtained and recorded: vital signs, clinical signs and symptoms, particularly
those suggesting local inflammation at a catheter site, type and site of catheter, and
laboratory results. Clinical and laboratory data regarding other potential foci of
infection should also be obtained and recorded. As described above, peripheral blood
cultures and either cultures of the catheter itself or blood cultures drawn through the
catheter should be obtained. In addition, cultures of the catheter hub or infusate should
be considered, since these represent potential sources of catheter-related bloodstream
infection.
2. On-Therapy
At 48 to 72 hours, a formal evaluation should be conducted
by the investigator, and a decision should be made whether the drug is showing
effectiveness. This decision should be based on results of blood cultures (i.e., whether
clearance of the pathogen from the bloodstream has been achieved) and evaluation of the
patient's clinical status. Patients who have a change in therapy due to poor effectiveness
of the initial regimen should be considered therapeutic failures. In addition, patients
who do not have their catheter removed initially, but have their catheter removed at this
visit (unless this removal is a pre-planned change), should be considered therapeutic
failures.
3. End-of-Therapy
This is an optional visit at which an investigator can
decide whether additional therapy is needed or not. If prolongation of therapy is
warranted, the protocol should prospectively define how these patients will be analyzed.
If an alternative therapy is initiated, these patients should be considered therapeutic
failures.
4. Early Follow-up (test-of-cure visit)
This visit should be at least 5 days post-completion of
therapy, with a longer period of time planned for study drugs with a long half-life. At
this visit, the investigator should also look for clinical signs or symptoms consistent
with possible metastatic phenomena (such as joint inflammation, bone pain, or signs of
endocarditis). This visit should occur at a uniform time from baseline for all study
groups (an issue when dealing with "short" versus "long" therapy
comparisons).
5. Late Follow-Up Visit
The primary purpose of this visit is evaluation for
possible metastatic infections.
This visit should be considered mandatory for patients in
whom a pathogen known for causing late-onset metastatic infections (e.g., Staphylococcus
aureus) is isolated in the entry cultures. Because the literature is unclear about the
appropriate timing of such a visit, a 4-week postcompletion of therapy visit is proposed.
H. Outcome
As noted previously, the major emphasis in the evaluation
of efficacy will be on the population of patients who have a proven catheter-related
bacterial bloodstream infection. A composite endpoint (i.e., clinical and microbiologic
response) at the test-of-cure visit will be the primary endpoint in the final regulatory
decision, with differences in all-cause and/or infection-related mortality rates also
considered. Clinical and microbiologic outcomes should also be examined separately. In
situations where the clinical and microbiologic outcomes differ, possible causes for the
discrepancy should be explored in the study report. Secondary endpoints that could be
considered include time to clearance of bacteremia, percentage of patients with documented
late metastatic sequelae, and development of resistance during therapy.
Analysis of the following populations is suggested:
· Modified Intent-to-Treat
All randomized patients who meet required clinical and
microbiologic inclusion criteria at randomization. In addition, subgroup analyses as
described in section III.C are suggested.
· Evaluable
All patients who meet required clinical and microbiologic
inclusion criteria at randomization; have none of the exclusion criteria; receive at least
80 percent of the study regimen for at least 48 hours; do not receive concomitant
antimicrobial therapy for reasons other than treatment failure; do not have
discontinuation of assigned therapy solely for adverse events; and have all follow-up
evaluations.
The following outcome categories are suggested:
· Cure
Patient shows complete resolution of entry signs and
symptoms and negative blood cultures at test-of-cure visit. Patients at risk for late
metastatic sequelae (e.g., S. aureus osteomyelitis) do not show such sequelae at
late follow-up.
· Failure
Patient shows any of the following:
- Incomplete resolution of entry signs and symptoms at
test-of-cure
- Clinical deterioration or relapse while on therapy
requiring change to alternative therapy
- Persistent or relapsing bacteremia while on therapy
- Death from infection
- Late metastatic infectious sequelae (e.g.,
osteomyelitis)
Separate reporting of clinical and microbiological
outcomes is also recommended.
I. Statistical Considerations
At present, there is no approved drug for this indication
for use as a comparator. In such a situation, evaluation of a new drug generally proceeds
using one of two approaches. If a drug exists that is a widely accepted standard of care
for the indication, the sponsor can use an equivalence trial, provided sufficient activity
can be documented in the comparator drug for the given indication. If there is no widely
accepted standard of care, or if the efficacy of the standard of care is difficult to
document, a superiority design will probably be the best approach.
A superiority trial could take any of a number of forms,
including:
· Test drug vs. comparator drug
· Dose response of test drug (e.g., high dose vs. vs. mid
dose vs. low dose)
· High dose of test drug vs. low dose of test drug vs.
comparator drug
Discussion of the choice of comparator drug and
considerations involved in the use of a dose-response design are discussed above in
section III.F.
Alternatively, two equivalence trials might
be sufficient to support approval, if the following conditions can be satisfied:
· The sponsor provides an analysis based on a
comprehensive review of historical data.
· The analysis supplies convincing evidence about the
level of activity that the comparator drug provides in this population. Specifically, this
analysis should address how much cure rates would differ between the following groups in a
hypothetical clinical trial:
Group 1: Comparator drug(s) + line removal (where
indicated) in a population such as that studied in the trial, receiving all background
therapy.
Group 2: Line removal (where indicated) in a
population such as that studied in the trial, receiving all background therapy.
The analysis should establish a defensible estimate of
difference in cure rates between Group 1 and Group 2. Let this difference be denoted _.
The delta used in the sponsor's equivalence trial should be smaller than this value _, and
also be sufficiently small to exclude clinically important differences. Delta should not
be greater than the smallest effect size that the active drug would be reliably expected
to have compared with placebo in the setting of the planned trial, but may be smaller
based on clinical judgment._
The analysis should consider the relative distribution of
the pathogens found in the trials, as well as other baseline characteristics.
A line-removal policy will be in effect in both arms of
the sponsor's trials; thus, historical data about patients in whom line removal practice
is not similar to what will be done in both groups of sponsor's trials is not pertinent to
this analysis.
Even when delta is appropriately selected prior to a
trial, circumstances of a particular trial, such as poor compliance or the characteristics
of the study population, could invalidate the suitability of this delta. Thus, the sponsor
should also document that its trial has assay sensitivity (also known as difference
detecting ability)._
J. Review Considerations
(Reserved)
K. Labeling Considerations
(Reserved)
FDA/Center for Drug Evaluation and Research
Last Updated: March 08, 2001<
Originator: OTCOM/DML
HTML by PKS |