*3.3.3 Phase II trials*

The first test of therapeutic efficacy where both antimicrobial and clinical effects can be measured is traditionally a phase II study (also termed a "Proof -of- Concept" or PoC) in patients with well-defined disease characteristics. The dosing regimen chosen in these studies is derived from the preclinical animal data, PK/PD modeling, and results of phase I studies. A dose close to the maximum tolerated dose seen in phase I studies may be chosen with provision for modification if significant tolerability or safety issues occur. The initial PoC study population is usually adults because only limited amount of human safety information is available. Pediatric patients are typically included later after safety has been documented in juvenile preclinical toxicology studies and in adult clinical trials.

Sample sizes in PoC studies are usually limited though efficient designs with multiple parallel arms for regimen comparison and combination therapy have become more common.

Two pairs of published phase II trials that merit further discussion illustrate some of the methodological issues for PoC and dose finding. Features common across these four studies are a multicenter, multiarm, randomized, parallel group design; enrollment of RT-PCR-positive patients with chronic (mainly indeterminate) disease; use of serial short- and long-term biomarker measurements; and PK sampling for drug exposure.

The CHAGAZOL trial [22] was an open-label, multicenter randomized trial of 79 adults conducted in a Spain comparing high and low doses of posaconazole (POZ) to 60d of standard BZN. The subject screen failure rate for study entry was 57%, and most patients (> 90%) originated from Bolivia. Sixty-five percent of subjects

had chronic indeterminate disease (CID), and the remainder had some evidence of end-organ damage, primarily cardiac (22%). Serial RT-PCR measures showed a high PCR reversion in all arms during the treatment period, but by 10 months post treatment, more than 80% of POZ-treated patients relapsed versus 38% of BZN patients. The transient effectiveness of POZ in this study prompted exploration in a combination study with BZN. The STOP-CHAGAS trial [23] was a multinational study of 120 adults with CID. This partially blinded study randomized patients across four arms comprising POZ alone, BZN plus placebo, BZN plus POZ, and placebo. The higher screen failure rate of 70% may reflect the narrower patient population and broader site participation. The primary outcome measure was RT-PCR negativity at 180d. High response rates (>90%) were seen in the three active arms during treatment but was not sustained for POZ monotherapy with only 13% response at d180 compared with >80% for the BZN arms. Notably, the treatment discontinuation rate is 32% almost halving the per protocol population in the BZN arms.

The second pair of randomized studies are the trials of fosravuconazole (E1224). The single-blind, five-arm trial from the E1224 study group [24] allocated 231 adults with CID to low, high, or short doses of E1224 or to BZN or placebo. This study was conducted in two sites in Bolivia and had screen failure rate of 59% and the antiparasitic effect was assessed through serial RT-PCR and serology. The study was powered to show superiority over placebo for a primary endpoint of sustained PCR response at 12 months after treatment start. The high initial response observed for both E1224 and BZN during treatment was only sustained in BZN-treated group. The follow-on BENDITA trial [17] was a seven-arm, double-blind, double dummy study in 210 adults with CID conducted in three regions of Bolivia. Of the 518 patients screened, 210 were eligible and randomized to one of three BZN treatment arms (2,4, or 8 weeks), one of two BZN plus E1224 combination arms, or to placebo. The primary efficacy endpoint was sustained parasite clearance at 6 months based on RT-PCR. This trial showed that 2 weeks of BZN therapy was almost as good as standard therapy in achieving sustained response with far better tolerability. Whether a 2-week regimen can translate into longer-term benefit remains to be seen. The addition of E1224 to BZN did not improve efficacy but increased the incidence severe adverse events. The low rates of protocol violation and treatment discontinuation in these two trials attest to the quality of the design and execution of the study investigators.

These well-designed and well-executed phase II designs using RT-PCR in patients with CID provided a robust evaluation of drug effect and assist dose selection. This same design approach has also been used in the studies of fexinidazole (NCT02498782, NCT03587766).

### *3.3.4 Phase III trials*

The enormous heterogeneity in published studies conducted in Chagas discussed above emphasizes the need to have robust and consistent approaches to clinical testing. Typically, data from one or more adequately conducted randomized trial would be expected by regulatory authorities to support license approval. The clinical research gold standard for generating evidence of efficacy and safety is a prospective (ideally double blind) randomized clinical trial against the current standard-ofcare (SoC) or placebo. In the majority of Chagas*'* settings, either benznidazole or nifurtimox would be the active comparative control in registration enabling clinical studies. In certain circumstances, a historical or concurrent placebo may be acceptable. The evolution of potentially shorter BZN regimens from four ongoing trials

*New Therapeutics for Chagas Disease: Charting a Course to Drug Approval DOI: http://dx.doi.org/10.5772/intechopen.102891*

(NCT03191162, NCT03981523, NCT04897516, NCT03672487) may significantly impact both a new drug TPP and the design of future comparative trials.

Beyond the fundamental requirements for minimizing bias and variability through randomization, blinding, and stratification, the numerous possibilities for comparative phase III and adaptive study designs will not be covered here as they must be tailored to answer the precise clinical research question of interest. The logistic challenges around designing and executing such trials to meet current regulatory standards are significant and a major investment in research time, finances, and resources. It behooves drug developers to have compelling phase II data before deciding to embark on such studies and to engage in early discussions with regulatory authorities.

A clinically meaningful endpoint (or an acceptable surrogate marker) should have been chosen and agreed with Health Regulatory Authorities; a dose adequately determined; and a minimum follow-up period of 12 months (or more) planned. Ideally, a trial should be prospective, multicenter, multiregion randomized, doubleblind (double dummy), and powered at a minimum for noninferiority against the SoC (BZN/NFX) and superiority against placebo (where used). Factors known to influence variability in response such as age and geographical distribution required careful balancing. Preplanned interim analyses with early stopping rules may be useful to manage risk, and an independent data safety monitoring board is strongly recommended. Adjudication committees for some endpoints may be added as needed. Centralized laboratory assessment for serology and biomarkers is recommended.

The operational challenges of mounting large phase III interventional studies in Chagas disease include effective community engagement; cooperation with local centers of expertise; engagement of government health screening and awareness programs; rapid diagnostics and facilitated access to medical facilities capable of conducting regulatory standard studies. Rigorous trial execution and patient retention are needed to minimize protocol deviations that could compromise the noninferiority designs. Post-approval requirements may include further clinical trials or extended follow-up (3+ years) of study patients.

Where more than one drug candidate exists, prioritization decisions must be made as the resource requirements to run several simultaneous phase III programs may well be prohibitive. Clinical trial designs (master protocols) to improve efficiency of testing of multiple drugs of the same therapeutic class are well accepted by Health Authorities and have been used in oncology and more recently COVID-19 settings [58, 59]. An alternative strategic option would be a combination approach with two new drugs (antiparasitic or antiparasitic plus host response modulator) in a single phase II/III program.

#### *3.3.5 Alternative development strategies*

Discussion of drug development for Chagas disease would not be complete without considering the many outstanding strategic questions concerning the ultimate goal of halting disease progression.

After four decades of use, is it realistic to expect any more gains from BZN (or NFX) alone or is the problem simply related to poor tissue targeting [60]? Is chronic indeterminate disease already too late a setting to influence the course of the disease for any drug and if so, should efforts be redirected to early detection and acute disease where the risk of cardiomyopathy is higher [52]? Finally, would combinations of antiparasitic drugs with or without anti-inflammatory/immune-modifying drugs in

both ACD and CID be more effective than monotherapy? Combination antimicrobial therapy is well established in some infectious diseases (e.g., HIV, HCV, Tuberculosis) but relatively uncommon in Chagas disease. A phase I/II trial (ICTRP REBEC-RBR-5n4htp) studying the combination of BZN and disulfiram antitrypanosomal agents is currently underway in Brazil [61]. While immunomodulatory approaches to Chagas cardiomyopathy have met with some success in preclinical models [41], there are few human trials of antiparasitic and immunomodulatory combinations. Simvastatin combined with BZN was shown to reduce cardiac inflammation in a murine model despite parasite persistence [62], and the ongoing ATOCHA trial (NCT04984616) is evaluating the effects of atorvastatin combined with BZN or NFX with chronic indeterminate disease. In a murine model, attenuated cardiac dysfunction and tissue parasite clearance were seen with the combination of fenofibrate (a peroxisome proliferator-activated receptor ligand) and BZN [63], but no human study of this combination has been published. The poor progress with monotherapy in many settings warrants preclinical and clinical exploration of combination approaches despite the potential for increased costs, more side effects, drug interactions, and increased risk of medication nonadherence.
