To combat antimicrobial resistance, the FDA needs a reset

Antimicrobial drug development

• Antibiotics are not working as well as they used to – and some don’t work at all for diseases they used to cure. As pathogens evolve to evade formerly effective anti-infectives, clinicians are having to reach toward products, or combinations of products, that may be less effective, expensive, potentially toxic, and which themselves may drive resistance. In the case of extremely drug-resistant tuberculosis, up to four concomitantly administered antibacterials may not be enough to eradicate infection. Antimicrobial resistance (AMR) is a global problem that is acutely felt in the U.S.
• Meanwhile, a 2022 report from the pharmaceutical trade association BIO termed the antibacterial innovation pipeline “fragile and failing,” with just 64 new chemical entities in clinical trials for antibacterial indications at the time of the report, with a third of these centered on combating tuberculosis and bacterioides species – a common source of healthcare-acquired infection (HAI). By comparison, the group pointed out, 158 new products were in development for breast cancer. Expense, small patient populations, and lack of incentives have all been cited as reasons for the lagging pace of discovery – along with slowed scientific advances.
• New and more targeted anti-infectives are one approach to tackle the AMR crisis; another critical strategy, known as “antimicrobial stewardship,” has clinicians avoiding antimicrobial use when not indicated by guidelines. When antibiotic or antifungal treatment is needed, stewardship principles dictate avoiding broad-spectrum antibiotics when necessary, although for serious illness when the causative agent is not known, broad-spectrum empiric treatment may be needed.
• More on the scope of the AMR problem came from a 2022 joint CDC-FDA meeting looking at specific strategies to fight AMR, including “decolonizing” hospitalized patients who carry pathogens associated with HAIs. At the meeting, MICHAEL CRAIG, the CDC’s senior advisor for antibiotic resistance, cited CDC data that 48,000 deaths annually can be attributed to resistant infections, with about one in 31 hospitalized patients having a healthcare-acquired infection (HAI). AMR infections, and deaths from these infections were up “at least 15%” in 2020, said Craig. These figures reverse an earlier trend that had been seen through 2019 toward reduction of deaths from antibiotic resistance overall and in hospitals, with significant decreases in some of the most common resistant pathogens.
• Pressure from inappropriate antibiotic prescribing, including use of broad-spectrum antibiotics where more targeted ones would suffice, has continued to drive antimicrobial resistance (AMR). Craig presented a sobering timeline that mapped the introduction of new antibiotics against the seemingly inevitable rise of organisms resistant to these drugs. Multi-drug resistant organisms, and even pan-resistant organisms, may elude every single tool in the antimicrobial toolbox. “If we do not seriously invest in preventing infections, we will lose modern health care,” said Craig.

Case study: A recent FDA meeting addressing febrile neutropenia highlighted many of the regulatory gaps and hindrances for developers of new therapies

• On April 23, the FDA convened a workshop entitled Drug Development Considerations for Empiric Antibacterial Therapy in Febrile Neutropenic Patients. U.S. regulators were joined by colleagues from the European Medicines Agency (EMA) and Japan’s Pharmaceuticals and Medical Devices Agency (PMDA), as well as clinical and academic infectious disease experts and drug development experts.
• Febrile neutropenia, also known as neutropenic fever, is “a condition marked by fever and a lower-than-normal number of neutrophils in the blood,” according to the National Cancer Institute (NCI). Although a variety of etiologies can ultimately lead to a case of febrile neutropenia, where infection-fighting neutrophils are at a dangerously low level in a patient with fever, this situation occurs most frequently in patients receiving chemotherapy. This condition is not a diagnosis so much as a constellation of symptoms, and is thus sometimes referred to as a syndrome. The presence of a fever in a patient with neutropenia is a medical emergency, but in the vast majority of cases, the actual cause for the fever is unclear initially.
• Since the actual cause of the fever is typically unknown, febrile neutropenia is initially managed empirically with drugs that target the most likely causes of the fever in that patient. This involves the use of broad-spectrum antibiotics, begun within one hour of presentation.
• Diagnostic testing will help physicians sort patients into one of four categories: 1) Microbiologically defined, or the source of the fever has been identified down to the causative microorganism; 2) Clinically defined, or the patient’s clinical presentation indicates an infection, but without confirmatory microbiological data); 3) Unexplained fever, which could possibly be due to an infection; 4) Non-infectious fever. Most patients fall fairly evenly into one of the first three categories.
• Therefore, diagnostic testing is critically important and must be initiated immediately. Blood cultures must be obtained before antibiotics are started and are critical to determining whether there is an infection in a patient’s bloodstream and, if so, identifying the causative microorganism.
• At the point that a cause for the fever is identified, the patient’s treatment plan will change dramatically, with the diagnosis shifting to the verified infection, caused by a certain organism and sourced from a certain location. The patient’s treatment from this point forward will be guided by the best practices for that specific type of infection, albeit in the context of a serious underlying illness (i.e., cancer) and poor immune function.
• In general, the appropriateness of empiric treatment for febrile neutropenia will be determined within 48-72 hours. At this point in time, about 30-50% of patients will have an identified source of infection, and treatment will be modified accordingly.
• Thanks to scientific advances, survival rates are now high. For example, patients deemed at 20% or higher risk of febrile neutropenia from chemotherapy now receive preemptive treatment to stimulate the bone marrow to make more white blood cells and restore immune function. In 1987, mortality in patients with febrile neutropenia was documented to be around 26%; in recent years, mortality has dropped to about 2%.

Despite these advancements in outcomes, the FDA’s views on the topic seemed to be stuck in decades past

• PETER KIM, FDA’s director of its Division of Anti-Infectives (DAI), spoke at the febrile neutropenia workshop, noting that only two drugs have ever received FDA approval for “empiric therapy for febrile neutropenic patients” and that both therapies were approved in 1997, over 25 years ago. Additionally, both drugs are administered intravenously. He also noted the difficulty in determining an ideal study design and primary endpoint.
• Multiple FDA presenters provided a brief overview of drug development from a regulatory perspective, with a handful of targeted considerations. In particular, the presenters homed in on the need for at least one adequate and well-controlled investigation, which could be supported by results from previous clinical investigations for a similar but closely related indication (i.e., an antibiotic already approved for the treatment of a specific infection). Additionally, presenters discussed the pros and cons of non-inferiority versus superiority trial designs, and the importance of patient selection within each of these contexts. For example, the inclusion of patients with “unexplained fever” could hamper a study’s ability to demonstrate superiority; conversely, the use of additional diagnostics which identify patients most likely to have an active infection could be used to enrich a study.
• Notably, RAMA KAPOOR, a senior medical officer within FDA’s DAI, stated the regulator’s preference for either all-cause or infection-related mortality as a primary efficacy endpoint, due to the objective and clinically relevant nature of these outcomes. However, with the significant advances in managing febrile neutropenia (which, as noted above, have reduced the mortality rate to about 2%), the use of these endpoints would require very large clinical trials. This point was further emphasized by FDA statistician DANIEL RUBIN, who provided a table of trial sample sizes using mortality as a primary endpoint. These trial sizes ranged from about 400-11,000 patients depending on the success rates of drug versus placebo, the rate of patient evaluability, and the use of a superiority or non-inferiority design.

Clinicians and drug developers, on the other hand, felt that the advances in care have completely reshaped the landscape

• “Mortality is simply not an endpoint as it was 40 to 50 years ago,” said ALLISON FREIFELD, infectious disease specialist at University of Nebraska and primary author on a key febrile neutropenia guideline. She put the FDA on notice that the use of mortality as an endpoint for this indication no longer makes reasonable sense, nor is it clinically relevant to this patient population. DOUG GIRGENTI, head of drug development at Melinta Therapeutics, laid out a similar argument, pointing out that response times have been greatly trimmed. “Given the anticipation that we will have a bug in 2-3 days, or [know] whether the patient is improving or not,” he advocated for a clinical endpoint that occurs very early, with a primary efficacy evaluation falling within the first 72 hours.
• “We’ve got to determine what we’re trying to do,” said KIEREN MARR, chief medical officer at Elion Therapeutics, speaking to the heterogeneity of this patient population. Considering that up to 70% of patients will never have an identified microbiological source of infection, she pointed out that an infection-based endpoint (i.e., infection-related mortality) is also not appropriate in this population.
• From here, clinicians and developers posed a question that diverged further from the regulator perspective: Is febrile neutropenia even a distinct indication anymore? Marr made a forceful argument away from any clinical investigations into febrile neutropenia specifically, with Freifeld also supporting this approach.
• Girgenti, coming from industry, proposed a novel idea for regulatory consideration: Could the use of antibiotics during neutropenia actually be evaluated by modeling alone? If a drug has shown clinical activity against a certain bacterial spectrum, he proposed that perhaps any potential differences in efficacy in patients with neutropenia could be identified via model-informed drug development (MIDD) techniques. “From the sponsor perspective, I would love to be able to extrapolate and use modeling and simulation” to identify any potential differences – if any exist – between these patients and “other patients with, say, a hospital-acquired infection.”

The meeting also highlighted the promise of advances in diagnostics

• In a situation such as febrile neutropenia, where an infection constitutes a medical emergency but the ability to confirm an infection is limited, speed of diagnosis is paramount. As reiterated multiple times during the workshop, diagnostic tools for this population remain limited, with a microbiological source confirmed in just about 30-50% of patients, after final cultures are available at 48-72 hours.
• During the workshop, pathologist KIMBERLY HANSON of the University of Utah discussed some promising molecular diagnostics currently gaining traction. She reviewed the timeline of diagnosis offered by molecular tests that can provide diagnostic information within hours, rather than days for a culture.
• When evaluated clinically, these molecular diagnostics have dramatically shortened the time to microbiological identification, resulting in clinical benefit. The study cited by Hanson ultimately found that the time to microorganism identification dropped from 22.3 hours in the standard care arm to 1.3 hours in the intervention group; this resulted in fewer doses of broad-spectrum antibiotics in the intervention group. Another promising exploratory technology – metagenomic sequencing of circulating microbial cell-free DNA (mcfDNA) – has the potential to identify organisms within and outside of the blood, within 24 hours.
• Advances in molecular diagnostics – which already exist and are already being actively validated – may thus narrow the window of febrile neutropenia even further, perhaps to 24 hours or less, calling into question whether there’s even any benefit in evaluating therapies specifically for this condition.

Looking back, how was the FDA so off-base in the febrile neutropenia meeting?

• In some ways, the FDA seemed fully aware of the altered landscape – but was simply too rigid to do anything with that information. The dramatic improvement in mortality rate was clearly discussed and portrayed within an FDA presentation, immediately after the presenter reiterated the importance of the use of mortality as an endpoint in clinical trials. And this reduced mortality rate was top of mind during the FDA statistician’s review of barriers to clinical trials, which included a discussion of the large trial sizes that would now be necessary to evaluate this outcome. Still, the FDA chose to propose clinical trial designs from the 1990s, when the last approval was granted for this indication and mortality rates were much higher.
• No recent antibiotic approvals have been granted on the basis of a mortality-related outcome. It has been generally accepted over the past decade that, in the case of antibiotics, nearer-term and more focused endpoints are valid and appropriate. These tend to include proof of activity against the target organism within the relevant portion of the body, targeting symptom resolution (clinical response) and/or the eradication of bacteria (culture conversion). In fact, every antibiotic approval between 2010-2015 and 2016-2019 was completed on the basis of these nearer-term and more evaluable endpoints. If mortality was not considered necessary in the context of these other serious infections, why is it considered relevant here?
• Perhaps one explanation for the regulator’s apparent rigidity in this case is the segmented, siloed and bureaucratic nature of the FDA’s review departments. The Office of New Drugs is divided into multiple sub-offices, which are then divided into multiple divisions. For example, the Office of Infectious Disease, which houses the DAI, also houses the Division of Antivirals and the Division of Pharm/Tox for Infectious Diseases. The other review divisions are structured around both disease type and body system, resulting in a hodgepodge of divisions that work in parallel but not together, sometimes reviewing drugs for complex conditions that – in clinical practice – would involve the input of multiple different specialties. Instead, due to the FDA’s outdated administrative structure, these drugs are reviewed within one specific and highly specialized division. In the case of febrile neutropenia, for example, most clinicians would likely argue that input from oncology and hematology experts, not just infectious disease experts, would be relevant.

Looking beyond the siloed review of drug products, it was clear that the FDA’s bifurcated approach to the review of drugs and diagnostics also severely limited the utility of the febrile neutropenia workshop.

• Particularly in the realm of infectious disease, for which treatment is very regularly guided by specific diagnostic results, the approval of drugs in isolation from diagnostics can, at times, be almost counterproductive. This workshop presents a clear case in which any drug development must be considered in the context of the available (and emerging) diagnostics, which can dramatically alter treatment decisions over a relatively short span of time.
• In the U.S., diagnostic products are regulated as medical devices, and held to regulatory standards of safety and performance under the medical device requirements. Notably, navigating the device regulations is notoriously difficult for diagnostic developers – and this has been highlighted as a key concern for antibiotics stakeholders. At a December 2023 meeting hosting by the Duke-Margolis Center for Health Policy that examined barriers to antimicrobial innovation, multiple experts called out the dearth of “good diagnostics” as a serious problem for developing novel antibiotic products. Attendees saw the FDA’s expectations for those developing diagnostics as stifling development of new diagnostic products; for example, the volume of samples the FDA expects developers to retain can pose a practical burden. Others at the meeting expressed concern that the FDA’s device regulators weren’t taking the clinical context into consideration, since prescribers employ clinical judgment when deciding how to treat the patient in front of them.
• The FDA has taken some recent action to be more agile in this space. In September 2023, the agency updated its guidance on antimicrobial susceptibility testing (AST) to allow test makers to respond more quickly shifting patterns of susceptibility of a given pathogen to a given antimicrobial, in response to evolving resistance patterns. Under the new policy, manufacturers of these types of tests can ask the agency for a type of authorization that allows them to update their test over time, using what’s known as a Pre-Determined Change Control Plan (PCCP). The policy lets AST manufacturers update their tests to reflect updated breakpoints (the thresholds for a pathogen’s susceptibility), need for increased exposure, or resistance to an antimicrobial. This newly permissible agility can help AST test makers push out updates more quickly, supporting evidence-based prescribing of a targeted antimicrobial (at an appropriate dose), a framework that gives clinicians confidence they can practice safely while avoiding broad-spectrum antibiotics.

Regulatory and communication gaps don’t stop at diagnostics

• Although febrile neutropenia represents just one indication for antimicrobial therapy, that recent meeting showed the FDA’s lack of agility in its support of clinical development programs by adhering to an endpoint – overall survival – that, the clinicians at the recent febrile neutropenia meeting all agreed, has little relevance for the patient population at hand.
• At the Duke-Margolis meeting on antibiotic innovation, stakeholders pointed out that non-traditional antibacterial agents have additional regulatory considerations. For example, the CARB-X portfolio consists primarily of non-traditional antibacterial agents, used as adjuncts to standard of care. Examples of non-traditional antibacterial agents include products that can chelate metals needed for bacterial enzyme activity, nucleic acids that can interfere with bacterial DNA, or even materials that can absorb bacterial toxins, among many others. However, considering the wide range of potential mechanisms of action and even routes of administration, a major challenge in this space is finding the right endpoints to demonstrate the benefit of these drugs. While CARB-X has convened stakeholders and discussions, this is an area that could benefit from FDA guidance.
• Further silos between medicine and public health regulators: Although the 2022 workshop on AMR and HAI decolonization strategies was co-sponsored by the FDA and the CDC, discussions at that event made clear that there’s been little cross-pollination between the medico-pharmaceutical world and the population-based approaches that dominate the public health perspective. FDA officials were, for the most part, circumspect in any observations they made about the challenges described, though the CDC’s Craig called for regulators across agencies to carve out development and regulatory pathways that have a “prevention mindset.”
• However, Kim, the FDA’s head of anti-infectives, acknowledged the need to address existing gaps. As a meeting moderator at the FDA-CDC meeting, Kim addressed a specific question regarding developing patient-reported outcomes for HAI therapeutics. “There’s a reason why the system exists the way that we have it,” he said, though he acknowledged “the gray spaces that exist.”
• Nontraditional approaches to combat bacterial infections hold promise to combat AMR, yet they often fall into these regulatory “gray spaces.” From phage therapies to chelators to novel diagnostic strategies, an armamentarium of resources are in development – yet the regulators shepherding these products through to approval may be isolated from each other, and from other advances in the field.
• “I think that if we can all agree that there is a new world of conversations that needs to be furthered, and how we get there, and how it might change what we have held fast to as the dogma of how we conduct studies or approve products… I think what we’re saying through all of this is that these things need to shift in some way,” said Kim. “None of us are certain exactly how, but we need to come closer together. Because there’s a lot of people who need these products, and they need them in a way that can’t be decades in the making.”

Featuring previous research by Amanda Conti, Laura DiAngelo, and Kari Oakes
To contact the author of this item, please email Chelsey McIntyre ( cmcintyre@agencyiq.com).
To contact the editor of this item, please email Kari Oakes ( koakes@agencyiq.com).

Key Documents and Dates

• • Drug Development Considerations for Empiric Antibacterial Therapy in Febrile Neutropenic Patients, April 23, 2024