Given that the prevalence of multi-drug resistant (MDR) bacteria continues to rise, and new superbugs continue to emerge across various continents (see recent blog for details), developing novel antibiotics to treat these MDR pathogens has become an increasingly urgent public health imperative. In 2013, the FDA published a draft guidance outlining their thoughts on streamlining the development and designing clinical trials for novel drugs for patients with high unmet needs, including patients with infections caused by MDR pathogens. However, the agency emphasized that this draft guidance aimed to serve as a platform for continued discussions with key stakeholders on how best to design MDR pathogen-specific trials. As such, a concrete development pathway still remains to be established, and clarity on the design of pathogen-specific trials remains elusive. To address this problem, the FDA recently organized a workshop entitled “Facilitating Antibacterial Drug Development for Patients with Unmet Need”, to discuss difficulties in conducting pathogen-specific trial with all stakeholders including academics, clinicians, and representatives from biopharma companies and regulatory agencies. Notably, two companies actively engaged in antibiotic development, Achaogen and The Medicines Company, shared their experience with designing and conducting MDR pathogen-specific clinical trials for plazomicin and Carbavance™ (meropenem/vaborbactam), respectively.
Unlike traditional “single body site” infection clinical trials, which enroll patients with a specific body site infection like pneumonia or a urinary tract infection, a pathogen-specific clinical trial could help evaluate a novel antibiotic for the treatment of specific MDR pathogens. Compared with patients infected with drug-susceptible pathogens, patients infected with MDR pathogens, as Achaogen reported during the workshop, tend to have very different characteristics, including higher rates of polymicrobial infections, co-morbidities, and mortality. Patients with MDR infections are also more likely to need organ support, and their pharmacodynamics are less predictable and more variable as compared with patients with drug-susceptible infections. These differences generally prevent the inclusion of MDR pathogen-infected patients into traditional antibiotic clinical trials, which typically only include patients who do not have severe co-morbidities. Even if a proportion of enrolled subjects in a traditional antibiotic clinical trial were MDR pathogen-infected patients, the total number of enrollees with MDR infections would be too small to conduct a meaningful sub-group analysis. As a result, the safety, efficacy and dosing profile for antibiotics treating MDR infections can only be reliably evaluated through these pathogen-specific trials, rather than traditional single-body site infection clinical trials.
However, there are many challenges associated with designing and conducting an MDR pathogen-specific trial. The major challenge comes from the difficulties associated with enrolling enough patients infected with MDR pathogens to provide the required power for statistical analyses. Infections caused by MDR pathogens are rare, hence providing a very small patient-pool from which clinical trials can recruit. Furthermore, it is difficult to identify patients with MDR infections as these infections are sporadic and pathogen confirmation requires additional tests, which could take days to complete. In addition, infections caused by MDR bacteria can quickly become life-threating, making it difficult for hospitals to refer patients to a clinical trial testing center. As a result, a clinical trial testing center for a MDR pathogen-specific trial can only enroll patients admitted into their hospital, making the enrollment rate very slow.
Achaogen’s experience with its Combating Antibiotic-Resistant Enterobacteriaceae (CARE) trial illustrates how difficult it is to recruit MDR pathogen infected patients. Using colistin as the comparator, Achaogen’s CARE trial was designed to evaluate the efficacy of a novel aminoglycoside, plazomicin, in the treatment of carbapenem-resistant Enterobacteriaceae (CRE) infections, and was expected to enroll 360 patients within three years. However, after prescreening the first 659 patients, Achaogen was only able to enroll 14 patients. The main problem was that the percentage of prescreened patients with confirmed CRE infections was much lower than predicted. Further, among patients with a confirmed CRE infection, many had received empiric antibiotic treatments for longer than 72 hrs, violating the preset trial inclusion criterion. In addition, many of the excluded patients had polymicrobial infections that required antibiotics other than those outlined in the CARE trial protocol. Based on the low enrollment rate, Achaogen amended the trial protocol to include an additional treatment arm with more relaxed inclusion criteria that allow for the enrollment of patients with polymicrobial infections. Even with these relaxed criteria, however, Achaogen still lowered its target enrollment, and now expects to only enroll 100 patients by the end of the trial.
Another challenge with MDR pathogen-specific trials is the selection of an appropriate comparator for the clinical trial. The Medicines Company initiated an MDR pathogen-specific clinical trial to evaluate its novel antibiotic combination Carbavance™ in patients with CRE infections. Rather than specifying the active comparator, The Medicines Company’s trial was designed to use the best available therapy. So in order to design their pathogen-specific trial, the company conducted a retrospective study to understand what constituted “best available therapy” for infections caused by CRE. The study found that 69 different antimicrobial regimens, including both monotherapy and combination therapy, have been prescribed as “best available therapy” to treat 225 CRE patients across four different countries including Greece, Italy, the U.K., and the U.S. With such a large variability in choice for “best available therapy”, it becomes very difficult to accurately assess the efficacy of antibiotic candidates relative to a standard-of-care. Another complication arises when patients are ineligible for the selected comparator. During the FDA’s workshop, Achaogen reported that many patients with CRE infections had also developed resistance to the active comparator, colistin, further reducing the enrollable population for this trial. With these challenges, it is not surprising that MDR pathogen-specific trials can become extremely complicated, expensive, and time consuming.
In order to make an MDR pathogen-specific trial a viable option for developers, several solutions were proposed during the FDA’s workshop. The most promising strategy is the establishment of the Clinical Trial Network (CTN), which was proposed by Biomedical Advanced Research and Development Authority (BARDA) officials. Unlike the traditional “one drug, one trial” clinical trial design, the CTN would follow the innovative Platform Trial design, which aims to evaluate multiple antibiotic candidates simultaneously, hence reducing the cost and time commitment for each drug developer compared to developing these antibiotic candidates independently. In the CTN’s simplest scenario, two antibiotic candidates from different companies would be tested with a master protocol within the same network of clinical trial testing centers. As a result, both antibiotics could share the same control patient group(s), and therefore reduce the number of patients required for enrollment.
As a way to further alleviate the burden of patient recruitment, the CTN would continue enrolling MDR infection patients into the active comparator arm(s) even without any novel antibiotic candidates being actively tested. Moreover, the CTN could consolidate the total expenditure for the building and maintenance of an infrastructure and staff training, and would also preserve the knowledge of hard-learned lessons in order to improve trial-quality. The CTN could even provide a platform for the development of diagnostic tests for MDR pathogens, which is an additional area of unmet need. A similar Platform Trial design has already been implemented in oncology clinical studies, although these Platform Trials have patient populations that are much larger (over 1000 patients) than what is feasible for a MDR pathogen-specific trial (e.g., I-SPY2 breast cancer trial and STAMPEDE prostate cancer trial). Despite being able to provide long-term cost savings, however, CTN would require a large upfront investment, estimated by BARDA to be about $100 million, which could be the major hurdle for the implementation of the CTN.
Another proposed solution is to pool information from different sources of drug development data using Bayesian statistical method to compensate for the small sample size of MDR pathogen-specific trials. By using Bayesian method instead of traditional inferential statistical model, developers can integrate data from different body site infection together, or leverage prior evidence from early-phase clinical trials, animal studies or even in vitro studies. This solution would not require upfront investment like the CTN. In addition, this methodology could be applied to a range of clinical trial designs with minimal changes needed, and therefore offers a convenient option for interpreting the data from ongoing or even completed trials.
In summary, there are many challenges to designing and conducting MDR pathogen-specific trials. Based on the experiences from Achaogen and The Medicines Company, an MDR pathogen-specific trial would be more affordable if it could only target enrollment of less than 100 patients, allow tested antibiotic to be used as empiric treatment, apply more sensitive endpoints, utilize external or historical data as controls (e.g., through CTN), and implement non-standard statistical methods, such as the Bayesian method. However, even with a streamlined MDR pathogen-specific trials, a traditional Phase III “single body site” antibiotic clinical trial will likely still be required for any new antibiotic development, given that new drug’s approval must meet the statutory standard of effectiveness with data from at least one adequate and well-controlled clinical trial. Nevertheless, developers now have the opportunity to directly participate the conversation with the regulatory agencies and design their MDR pathogen-specific clinical trial together with government authorities, especially while regulators remain open to suggestions in shaping the path forward for the development of antibiotics against MDR pathogens.