Clinical Trial Design in Acute Bacterial Sinusitis Considerations for Future Guidance

1
Clinical Trial Design in Acute Bacterial
SinusitisConsiderations for Future Guidance

• John H. Powers, MD
• Lead Medical Officer
• Antimicrobial Drug Development and Resistance
  Initiatives
• Office of Drug Evaluation IV
• Center for Drug Evaluation and Research
• U.S. Food and Drug Administration

2
Introduction

• Re-addressing 1998 Guidance
• Defining the research question in ABS trials
• Considerations in clinical trials of ABS
• Defining the disease and Patient population
• Types of studies
• Endpoints and Evaluation
• Proposals for future guidance

3
Re-addressing previous Guidance

• Discussions at previous AIDAC and workshops and
  international meetings on selection of
  non-inferiority (NI) margins in clinical trials
• February 2002
• November 2002
• ICH-E10 document
• Lack of adequate selection of NI margin means
  cannot ensure adequacy of any drug over placebo
  in setting of NI trial
• ICH-E10 suggests other study designs if NI margin
  unknown

4
Re-addressing previous Guidance

• Agreement to examine previous placebo-controlled
  trials (PCTs) in each disease state to select
  appropriate margin
• review all pertinent studies not just those which
  show benefit
• Review of PCTs in acute bacterial sinusitis
• Assessment of 14 PCTs in ABS in literature showed
  several issues with selection of margin

5
Re-addressing previous Guidance

• Prior PCTs provide clues that antimicrobials may
  be effective in shortening duration of symptoms
  in ABS
• Accurate assessment of magnitude of benefit of
  antimicrobials in ABS remains unknown
• Need to discuss other study designs
• Other issues with ABS trials became apparent as
  part of review

6
Research Question

• What are we trying to measure?
• true benefit of antimicrobials as sole therapy
  in bacterial disease
• OR
• added benefit of antimicrobials above and beyond
  effect of standard non-antimicrobial therapies in
  patients with bacterial disease
• no need for antimicrobials if decongestants,
  saline and anti-inflammatory agents are effective
  alone
• other therapies do not result in antimicrobial
  resistance
• ABS fifth most common reason for prescribing
  antimicrobials in ambulatory setting
• McCaig L et al. JAMA 1995273214-9.

7
Defining the Disease

• Antibacterials would logically have greatest
  effect in patients with bacterial disease
• recent FDA labeling rule
• Others have hypothesized no proof that
  antibacterials have no effect in viral disease
• previous PCTs would seem to contradict this
• use anti-inflammatory drugs if anti-inflammatory
  effects of antimicrobials predominate give issue
  of resistance
• use in non-bacterial disease would seem to
  contradict current appropriate use guidelines

8
Defining the Disease

• Effects of inclusion of patients likely to
  recover spontaneously
• in NI trial, biases conclusion toward
  non-inferiority when there may be true
  differences between drugs
• in PCT, biases conclusion toward no difference
  from placebo when there may be important
  differences
• inclusion of higher proportions of patients with
  spontaneously resolving disease will result in
  less measured treatment effect of antimicrobials
• may explain results of many of previous PCTs

9
Defining the Disease

• Review of literature on correlation of clinical
  signs and symptoms or radiography and sinus
  puncture
• more rigorous criteria seem to select for higher
  proportion of patients with bacterial disease
• most rigorous criteria still would allow
  inclusion of 20 to 40 of patients without
  bacterial disease
• no adequately-sized prospective, reproduced
  studies to allow adequate selection criteria
• 7 days of symptoms in clinical practice but
  evaluation in clinical trial may be different
• antibacterials may exert greatest effect early in
  disease

10
Defining the Disease

• Sinus puncture remains gold standard
• procedure considered unpalatable by many
• newer procedures may obviate some of this
  discomfort and similar in performance to sinus
  endoscopy (through nose instead of oral approach)
• lack of accuracy with sinus endoscopy especially
  for certain organisms e.g. Staphylococcus aureus
• previous studies on nasopharyngeal and throat
  cultures show high level of discordance
• Wald E et al. Pediatrics 1986 77795-800.
• middle meatus not normally sterile site
• Klossek JM et al. J Laryngol Otol 1996110847-9.

11
Defining Disease

• Sinus puncture may be therapeutic by itself
• drainage of closed space infection
• Effect would be evenly distributed across arms of
  a trial
• Effect of puncture should be small relative to
  effect of antimicrobials
• otherwise one could question benefit of
  antimicrobials in this setting
• Antimicrobials should address residual
  inflammation caused by tissue invasion of
  pathogens

12
Defining Disease

• Sinus punctures are not done in clinical practice
• Prior PCTs with clinical entry criteria show
  minimal if any benefit for antimicrobials in ABS
• Clinical trials versus clinical practice
• many procedures done in setting of clinical trial
  that are not done in clinical practice
• drug efficacy trial versus strategy trial
• FDA mission is to approve effective drugs for
  disease under study
• Two reasons why strategy trial may fail to show
  a difference
• drug is not effective in disease under study
• strategy for selecting patients is not specific

13
Defining Disease and Sample Size

• Clinical enrollment criteria requires very large
  sample size or very large treatment effect to
  show a difference between drug and placebo
• Example
• assume 35 of patients in trial have bacterial
  disease
• assume cure rate in viral disease is 80 in both
  arms and cure rate with bacterial disease is 80
  with antibacterial agent and 65 in placebo
• THEN
• Cure rate with clinical enrollment criteria 80
  vs. 75 with sample size of gt2900 patients
• Cure rate with sinus puncture defined disease 80
  vs. 65 with sample size of 370 patients

14
Types of Trials

• Current guidance (two studies)
• Microbiological based non-comparative trial with
  presumed eradication based on clinical outcome
  (micro study)
• Clinical based NI study with clinical inclusion
  and outcome criteria (clinical only study)

15
Types of Trials

• Issues with previous Guidance studies
• Micro study
• presumes correlation of microbiological and
  clinical outcome that has never been shown in ABS
• examples of other diseases with less than optimal
  correlation with micro and clinical outcomes (e.g
  AOM)
• Johann-Liang R et al. Pediatr Infect Dis J
  200322936-7.
• Pre and post therapy microbiological data very
  helpful in ascertaining contribution of drug to
  treatment effect
• less helpful when used as sole measure of
  efficacy
• Clinical only NI study
• may include significant proportion of patients
  with non-bacterial disease
• timing of measurement of endpoints may not be
  optimal
• selecting NI margin in NI trials

16
Prior Placebo Controlled Trials

• Medline search for PCTs plus references of
  relevant trials
• Evaluated 14 PCTs from medical literature
• Varying quality of studies
• blinding in
• randomization
• sinus puncture with culture in only two but
  analyzed patients with and without positive
  culture together in primary anaylsis
• numbers of patients with positive cultures small
• Wide variations in methods precludes meta-analysis

17
Prior Placebo Controlled Trials

• Widely varying methods of assessment of outcomes
  (patients versus sinuses)
• clinicians assessment of symptomatic cure
• non-validated scoring systems for symptomatic
  cure
• radiological scoring systems
• ostial patency by pressure measurements
• changes in nasal cytology
• Timing of assessment of cure varies widely
• various fixed time points from days to weeks
  after EOT
• time to resolution of symptoms
• two used Kaplan-Meier curve but differing
  measures of outcome (one used pain only and other
  used non-validated symptom scale)
• several used modified time analysis of early
  fixed time points

18
1
14
2
11
3
4
10
7
5
12
6
8
9
16
13
15
Antimicrobial Efficacy ( 95 CI)
Study
19
Prior Placebo Controlled Trials

• Only two show benefit over placebo defined as
  lower bound of 95 CI above zero
• Two with sinus puncture data
• one shows much better outcome in subgroup with
  bacterial disease (25 vs. 3)
• other trial shows worse outcome in subset with
  bacterial disease ( -12.0 vs. 2.3)
• both subsets with small numbers of patients
• Point estimates in majority of studies show small
  benefit, on the order of approximately 4

20
Types of Trials

• ICH-E10 suggests choosing trial design other than
  NI trial if NI margin unknown
• cannot ensure benefit of any drug over placebo
• cannot scientifically justify picking margin
  without data
• Other types of trial designs
• dose-response trial
• placebo controlled trial
• more accurate to refer to as superiority trial
  vs. other symptomatic therapies
• give other appropriate medications other than
  antimicrobials for symptom relief

21
Types of Trials

• Antimicrobials plus other symptomatic therapies
  vs. other symptomatic therapies
• no issue of selection of appropriate NI margin in
  superiority trial
• trial has own internal validity since direct
  comparison rather than indirect comparison with
  no antimicrobial therapy
• PCTs suggested by independent reviews
• Given the small number of trials with
  heterogeneous results, additional
  placebo-controlled trials are needed to evaluate
  the efficacy of antibiotics.
• Williams JW Jr et al. Cochrane Database Syst Rev.
  2003(2)CD000243

22
Types of Trials

• Ethical consideration given rare but serious
  adverse events associated with ABS
• Balances risk of adverse outcomes of ABS with
  risks of adverse outcome with antimicrobials
• serious adverse events e.g. anaphylaxis
• spread of antimicrobial resistance
• No data on antimicrobials decreasing risk of
  complications
• would require very large sample size given rare
  events
• complications may be due to altered anatomy or
  host factors

23
Types of Trials

• Other infectious diseases with higher mortality
  and complications studied as PCTs
• new influenza drugs studied as PCTs despite
  availability of older drugs
• mortality of influenza in outbreak setting ranges
  from 10 to 600 per 100,000 persons in healthy and
  chronically ill respectively
• Barker WH et al Arch Intern Med 198214285.
• Selection of exclusion criteria to minimize risk

24
Types of Trials

• Exclusion criteria to minimize risk
• exclude frontal and sphenoidal disease
• ? exclude patients with certain organisms
• brain abscess associated with microaerophilic
  streptococci (S. anginosus group) in 70
• S. pneumoniae and H. influenzae in lt1
• Chun CH et al Medicine 198665415-31
• Kao PT et al. J Microbiol Immunol Infect
  200336129-36.
• excluding severe disease
• but has several associated issues

25
Types of Trials

• Severe disease
• severity defined as clinical characteristics
  predicting worse outcome (regardless of therapy)
  e.g PORT criteria for community- acquired
  pneumonia
• Fine MJ et al N Engl J Med 1997336243-50.
• no criteria exist for ABS
• criteria used in previous trials (fever) may
  select for bacterial disease
• patients with facial swelling may have
  periorbital cellulitis which may be considered a
  separate disease
• patients with severe disease may be most likely
  to benefit from antimicrobials

26
Trial Design and Sample Size

• More conservative assumptions
• Non-inferiority (D.05) 2700
• Placebo w/ 10 Day Endpoint (.80 v. .70) 780
• Placebo with time-to-resolution 520
• More aggressive assumptions
• Non-inferiority (D.10) 670
• Placebo w/ 10 Day Endpoint (.80 v. .65) 370
• Placebo with time-to-resolution 250
• Current sample sizes of ABS databases from NDAs
• 680 patients/drug in clinical only trials
  (range 142 to 1965)
• 580 patients/drug in micro trials (range 192 to
  1103)

27
Types of Trials

• Implications for pharmaceutical industry
• current trial design affords low risk of failure
  as difficult to distinguish any drug from placebo
• ABS is large market and proof that ABS does not
  need treatment in some portion of patients would
  decrease market size
• opportunity for smaller trials and more
  streamlined development program
• use of supportive data from CAP trials allows
  opportunity for one rigorous ABS trial
• less data but higher quality data

28
Endpoints and Timing

• Most appropriate measurement would examine
  resolution of signs and symptoms of disease
• several PCTs evaluated time to return to normal
  daily activities which may be patient dependent
• radiographic scores not validated and not
  correlated with clinical signs and symptoms
• correlation with microbiolgical endpoints and
  clinical outcomes not demonstrated to date
• Time to resolution of disease may be most
  appropriate in self-resolving diseases
• influenza and travelers diarrhea trials

29
Endpoints and Timing
Nicholson KG et al. Lancet 20003551845-1850.
30
Proposal for Future Guidance

• 1) Define population with bacterial disease by
  sinus puncture
• would not necessarily require 7 days of symptoms
• perhaps improve selection criteria for clinical
  practice
• 2) Superiority trial design of symptomatic
  therapy vs. symptomatic therapy plus
  antimicrobial
• discussion of appropriate exclusion criteria
• role of antimicrobial resistance
• 3) Endpoint of time to resolution of symptoms
• similar to previous trials of influenza
• requires validation of patient diaries

31
(No Transcript)