PLEASE NOTE: The Viewpoints on our website are to be read and freely shared by all. If they are republished, the following text should be used: “This Viewpoint was originally published on the REVIVE website revive.gardp.org, an activity of the Global Antibiotic Research & Development Partnership (GARDP).”
The views and opinions expressed in this article are solely those of the original author(s) and do not necessarily represent those of GARDP, their donors and partners, or other collaborators and contributors. GARDP is not responsible for the content of external sites.
Clinical trials of treatments for certain serious infections and for highly drug-resistant pathogens of low prevalence are difficult to conduct
Clinical trials of novel, urgently needed antibacterial drugs are especially difficult to conduct for the treatment of certain serious infections such as ventilator-associated bacterial pneumonia (VABP) and for highly drug-resistant pathogens of low prevalence (e.g. extensively drug-resistant tuberculosis (TB), Acinetobacter baumanii). The target patient population can be challenging to enroll, given complicated medical histories, comorbidities and strict limits on disease severity. The time window for study participant enrollment is narrow and may expire before a microbial diagnosis is made. Furthermore, treatment with prior and concomitant antibacterial therapies is often restricted to prevent confounding of the study results. Most registrational clinical trials utilize a non-inferiority study design where only patients with infections that are susceptible to the comparator may be included. As a consequence, drug developers may opt to run initial phase 3 registrational trials in patient populations that are easier to enroll with less severe diseases (e.g. complicated urinary tract infections or complicated intra-abdominal infections) with infections that are susceptible to current therapies. Healthcare providers are left with limited data for clinical decision-making. While some data, such as in vitro microbiology breakpoints, may be available, it is often uncertain how a drug would perform in critically ill patients, much less in patients receiving combinations of antibacterial drug therapy for highly resistant infections.
A single phase 2/3 trial in an at-risk population with confirmatory nonclinical evidence can support a finding of effectiveness against multi-drug resistant pathogens.
Accumulating evidence supports the feasibility of smaller clinical data packages
Accumulating evidence in individual programs demonstrates that it is possible to plan an antibiotic development program for an unmet need with a smaller clinical data package. A single phase 2/3 trial in an at-risk population with confirmatory nonclinical evidence can support a finding of effectiveness against multi-drug resistant (MDR) pathogens. The most striking case example is the 2019 US Food and Drug Administration (FDA) approval of pretomanid tablets in combination with previously approved bedaquiline and linezolid (BPaL) as part of a six-month, all-oral regimen for the treatment of people with extensively drug-resistant TB (XDR-TB) or MDR-TB who are treatment-intolerant or non-responsive (collectively “highly drug-resistant TB”).1 Prior to the approval of the BPaL combination, the only treatment options for the XDR and MDR TB patients were at least 18 months of daily injections of drug cocktails with high toxicity and low tolerability.2 Evidence for the contribution of each BPaL component drug to effectiveness in the treatment of MDR and XDR TB was obtained in a mouse model of chronic TB infection.3 It was not feasible nor acceptable to perform a factorial clinical trial design to assess the contribution of each drug in the regimen in this patient population. The safety and effectiveness of BPaL were demonstrated in a study of 109 patients with XDR, or MDR pulmonary TB.4 Treatments of 98 patients in the intention-to-treat analysis had a favorable outcome after six months.
The use of animal models to support regulatory decision-making appears to be evolving. Following the passage of the 21st Century Cures Act in 2016, the FDA published a guidance on the ‘Limited Population Pathway for Antibacterial and Antifungal Drugs’5 in August 2020, and a follow-up guidance ‘Antibacterial Therapies for Patients With an Unmet Medical Need for the Treatment of Serious Bacterial Diseases – Questions and Answers’ in May, 2022.6 The second publication addresses some of the Agency’s expectations regarding generating PK data using animal models. Some uncertainties remain, such as guidance on the information that would be acceptable for inclusion within the microbiology section of product labeling.
These examples demonstrate that robust evidence packages can be built with a combination of data from a well-designed clinical trial(s) supported with animal data for PK/PD evaluations at the target body site/organ system.
A second example of nonclinical data supporting approval of an antibiotic for unmet need is the development program for sulbactam-durlobactam (SUL-DUR), approved by the FDA in 2023 for the treatment of healthcare-associated bacterial pneumonia (HABP)/VABP caused by A. baumannii-calcoaceticus complex in patients 18 years of age and older. Infections due to carbapenem-resistant A. baumannii (CRAB) are associated with mortality rates ranging from 16% to 76%.7 A single Phase 3, randomized, assessor-blinded, active-controlled non-inferiority (NI) study in 177 hospitalized adults, together with confirmatory evidence provided by in vitro and animal model data, provided evidence of effectiveness. The study met the primary efficacy endpoint with a day 28 all-cause mortality rate within the 20% NI margin; the rates were 19.0% in the SUL-DUR group and 32.3% in the comparator (colistin) group.8 Animal models of infection included a neutropenic mouse thigh and a mouse model of pneumonia with sulbactam-resistant strains of CRAB. Both in vitro hollow-fiber and animal model data provided PK/PD targets for both plasma and epithelial lining fluid (ELF) exposures and demonstrated restoration of sulbactam activity in the presence of durlobactam.9 This conclusion appears in the microbiology section of the US prescribing information.10
Robust evidence packages can be built with a combination of clinical trials and animal data
These examples demonstrate that robust evidence packages can be built with a combination of data from (a) well-designed clinical trial(s) supported with animal data for PK/PD evaluations at the target body site/organ system.11-13 If activity is demonstrated in multiple in vitro and animal models with numerous parameters (“humanized” dosing, multiple bacterial strains, inoculation sites and sizes), along with appropriate experimental controls, then confidence could be increased. The nonclinical evidence must be compelling and some precaution is warranted. Animal models of disease pathophysiology often do not reflect human disease – multiple models may need to be evaluated to increase confidence; some models may be better established with some pathogens than others. Animal models of acute severe disease may not represent clinical disease states accurately (e.g. sepsis-like models). In addition, it is difficult to obtain clinical samples in pneumonia patients due to invasive sampling; for example, bronchoscopy for ELF samples is rarely performed. Public workshops14 and publications13-15 provided post-mortem analyses of antibacterial drug development failures, all focused on the need for adequate clinical pharmacometrics analysis to predict the probability of effectiveness based on nonclinical results.
An alternative development pathway is emerging to assess the potential benefits of novel antibacterial drugs in small, clinically relevant populations. Streamlined clinical data packages with supporting in vitro and animal model data could provide an approach to demonstrate the benefit of a drug where large clinical trials are not feasible and a great unmet medical need exists.