An obituary on DAV-132—authors’ viewpoint on the current limits of pivotal trials in clinical microbiome research

There is an urgent need for novel approaches to reduce the spread and clinical impact of infections with MDR bacteria. Da Volterra (Paris, France) is a biopharmaceutical company specializing in the development of microbiota-protective therapies. In 2010, they introduced DAV-132, an active charcoal preparation designed to protect the gut microbiota from antibiotic selection pressure, without affecting antibiotic pharmacokinetics, at the Superbugs and Superdrugs Conference in London, UK.¹ In 2016, Da Volterra joined the EU Innovative Medicines Initiative (IMI) to further develop this promising product in close cooperation with academic partners. Phase I and II trials were completed successfully, confirming the protective effect of DAV-132 on the human gut microbiota.^2,3 Consequently, a Phase III trial—MICROCARE—was planned and initiated. On 2 December 2022, however, the company Da Volterra was declared bankrupt. Given the significant clinical potential of DAV-132, this turn of events can be considered tragic. Could it have been avoided?

The authors of this letter were closely involved in the design and conduct of the Phase III trial necessary for registration of DAV-132, which was classified as a medical device by the competent authorities. Many hurdles had to be overcome to design and implement a trial that would be feasible to execute. To meet EMA and FDA registration criteria, it was necessary to show that DAV-132 could confer a health benefit. Protecting the microbiota is a very useful property, but is not considered a health benefit per se and is not recognized as a surrogate endpoint. Therefore, an alternative clinical endpoint had to be identified. The evidence on potential health benefits associated with a balanced human microbiota seems to be growing exponentially.⁴ At the same time, very little of that evidence is quantifiable and directly relatable to endpoints recognized by regulators. For the Phase III trial, we needed patients with considerable antibiotic exposure. These would have to experience a measurable clinical outcome associated with their antibiotic exposure. In addition, the frequency of this outcome to occur would need to be high enough to result in a feasible sample size. The only indication that seemed to fit these criteria was the prevention of Clostridioides difficile infections (CDI). Even though CDI is considered the most frequent healthcare-associated infectious cause of diarrhoea, its overall incidence is declining. Predicting which individual patient might develop CDI remains a major challenge, despite the identification of age and intense antibiotic exposure as key risk factors for CDI.^5,6 In patients undergoing induction chemotherapy for AML or high-risk myelodysplastic syndrome the frequent co-occurrence of the latter risk factors results in a relatively high cumulative incidence of CDI, ranging from 9.4% to 23.8% within the first 120 days after initiation of induction chemotherapy.^7–10 Even in this selected population, a sample size of 900 was deemed necessary to show the prophylactic efficacy of DAV-132, while assuming a clinical efficacy of DAV-132 of 50% reduction in CDI cases. At the same time, selecting such a seriously ill patient population involves risks. Chemotherapy-associated side effects can significantly impair patient compliance with trial protocols, and an increased number of serious adverse events is to be expected in this population. However, given that the sample sizes calculated for all other populations were not within a realistic range, we were left with the choice of no trial versus a trial with a considerable risk of failure.

The first patient was enrolled in the trial on 19 July 2021. One year later, by 13 July 2022—despite numerous discussions to evaluate alternative solutions—the decision was made to terminate the MICROCARE trial prematurely, based on its slow recruitment rate, as well as compliance and retention issues. As a result of this decision, the last patient visit and database lock occurred on 13 October and 31 October 2022, respectively. During the first year, only 73 patient inclusions could be achieved, with 828 further inclusions pending. Of these, over 50% withdrew from study treatment prior to Day 120. Half of these treatment dropouts occurred within the first 2 weeks after randomization. As authors being closely involved in the trial design, we have discussed these shortcomings in depth. We conclude that the failure of clinical development of DAV-132 was the consequence of a lack of alternative endpoint options, which forced us to focus on an exceptionally difficult trial population. If we had had more options concerning microbiota-related trial endpoints, we might have had a higher chance of success and could potentially have made DAV-132 available to a broad range of patients, reducing antibiotic-related microbiota dysbiosis and subsequent colonization by and spread of drug-resistant bacteria. Clearly, there is a huge gap in the field of microbiome research quantitatively linking gut microbiota measures to different types of health benefits that are relevant to patients, acceptable for physicians and regulatory authorities, and important for public health. To better identify and characterize such cause–effect relationships, we should consider the use of novel analytical techniques outside the usual clinical trial toolbox, e.g. machine learning or structural equation modelling.^11,12 This type of evidence is urgently needed to avoid future bankruptcies of small- or medium-sized enterprises with promising products that could have contributed to the fight against antimicrobial resistance.

Funding

All authors were supported by the IMI Joint Undertaking (JU) (grant 115523), Combatting Bacterial Resistance in Europe, with resources including financial contribution from the EU’s Seventh Framework Programme and in-kind contributions from companies in the European Federation of Pharmaceutical Industries and Associations (EFPIA).

Transparency declarations

M.J.G.T.V. has received research grants from 3M, Astellas Pharma, BioNTech, Da Volterra, Evonik, Gilead Sciences, Glycom, Immunic, MaaT Pharma, Merck/MSD, OrganoBalance, Seres Therapeutics and Takeda Pharmaceutical, as well as speaker fees and/or consulting fees from Alb Fils Kliniken GmbH, Ardeypharm, Astellas Pharma, Basilea, bioMérieux, Da Volterra, Farmak International Holding GmbH, Ferring, Gilead Sciences, Immunic AG, MaaT Pharma, Merck/MSD, Pfizer, Roche, OrganoBalance, SocraTec R&D GmbH and Tillots Pharma. A.D. is a paid consultant for the pharmaceutical and biotechnology industry, including for many companies developing products for infectious diseases, including Da Volterra. C.H.v.W., L.M.B., L.T. and M.E.A.d.K. have no conflicts of interest.

References