the virus clears before the patient arrives

Problem statement

Dengue causes an estimated 390 million infections a year (Bhatt et al. 2013; WHO's published range is 100–400 million) and has no specific antiviral treatment, but the reason drug development has stalled is not that candidate molecules lack activity — it is that clinical trials cannot measure viremia in patients by the time they enroll. DENV plasma viremia typically peaks within 24–48 hours of fever onset and declines rapidly over the ensuing days, such that the majority of patients have resolved their viremia and are afebrile 5–7 days after symptom onset; the standing trial-design recommendation is therefore that diagnosis, screening, enrollment, randomization, and first dose all occur within 24–48 hours of fever onset (Simmons et al. 2012). But "many patients with dengue currently do not present to health care providers until they are several days into their illness, by which time their adaptive immune response is already beginning to resolve their infection" (Simmons et al. 2012). Designing a trial around viremia as the primary endpoint — the most direct measure of antiviral effect — means measuring an outcome that is already over in a large share of enrolled patients. This is a trial design failure as much as a drug failure, and it has produced the same result: two discontinued small-molecule programmes and no specific treatment.

Why this matters

The phase 2 trial of AT-752 (Atea Pharmaceuticals) planned to enroll 60 dengue patients in 3 cohorts from 3 countries across 7 study sites; it was terminated in March 2023 after just 21 patients were recruited, which Atea attributed to logistical challenges and escalating costs associated with enrolling enough patients to account for the marked variability in viremia levels. The interim data show why that variability was fatal to the design: 23% of the treatment group and 43% of the placebo group already had viremia below the limit of detection at baseline, so the primary endpoint — change in dengue viremia from baseline — was unevaluable. Although the trial intended to enroll patients within 48 hours of fever onset, viremia was below the limit of detection in 30% of patients at recruitment and was in rapid decline by day 4 of illness (Kwek & Low 2025). The second discontinued small molecule, mosnodenvir (JNJ-1802, Janssen), was stopped for a different proximate reason — reprioritisation by Janssen ended its treatment and prophylaxis trials early (McBride et al. 2026) — though Kwek & Low describe it as "shelved for similar reasons." Meanwhile WHO recorded over 14.6 million dengue cases and more than 12,000 dengue-related deaths in 2024, against 505,430 reported cases in 2000, with about half the world's population now at risk. The absence of any specific treatment is not acceptable as a stable equilibrium, but the current trial infrastructure cannot generate the evidence needed to change it.

What’s been tried and why it hasn’t worked

Recruitment of trial patients has relied on the point-of-care DENV NS1 rapid test, but its sensitivity is low in the first two days after fever onset and in secondary dengue — precisely the window in which a viremia-based enrollment criterion would be meaningful (Kwek & Low 2025). Serology-based enrollment detects past or resolving infection, not current viremic state. Dengue PCR is more sensitive, but screening acutely febrile patients with it "would not be practical in most clinical settings" (Kwek & Low 2025), leaving a logistical gap between where trials are conducted and where rapid viremia confirmation is feasible. Nor is there a settled substitute endpoint: the Dengue Therapeutics Consortium's 2025 meeting report finds that dengue trials to date "have been mostly markedly underpowered to detect meaningful clinical outcomes," that "synthesis of trial data has not been possible due to the use of disparate and non-comparable outcomes," and that trials "have used surrogate outcomes, which may not have direct relevance to patient outcomes" (McBride et al. 2026). ISARIC's DEN-CORE Delphi consensus core outcome measurement set (Yacoub et al., Lancet Infect Dis, 2025, doi:10.1016/S1473-3099(25)00500-6) is the current attempt to harmonize those outcomes; it is an outcome-measurement standard, not a validated regulatory surrogate for antiviral activity. Trial activity continues around the unresolved endpoint problem rather than through it: a team at the Federal University of Minas Gerais is testing molnupiravir's effect on rate of viral clearance in early symptomatic dengue in Brazil, and DNDi announced in July 2026 that a phase III trial of a Serum Institute of India monoclonal antibody (formerly VIS513) is expected to begin in the first quarter of 2027 across Malaysia, Thailand, and Brazil with about 1,000 participants.

What would unlock progress

Two parallel workstreams are needed. First, a point-of-care viremia diagnostic with sensitivity sufficient to confirm active viremia at the time of enrollment — enabling reliable patient selection within the treatment window. Second, a validated clinical endpoint or composite endpoint that regulators will accept as evidence of antiviral efficacy when viremia measurement is unreliable at enrollment — which requires regulatory science investment, not additional drug discovery. Neither is a conventional drug development activity, which is why neither is being led by the pharmaceutical companies whose compounds are failing in trials.

Entry points for student teams

A diagnostic design team could specify the sensitivity and turnaround-time requirements for a point-of-care dengue viremia test that would solve the enrollment timing problem, and assess what sensing technologies are closest to meeting those requirements. A clinical trial design team could model alternative primary endpoint options — symptom duration, biomarker panels, severity composites — and evaluate their statistical power and regulatory precedent. A systems team could map the care-seeking delay problem: what structural features of dengue healthcare-seeking in endemic settings put most patients past the 24–48-hour post-fever-onset window before they reach a clinic, and what community-level interventions could pull presentation earlier into the viremic window?