Vaxart's oral COVID pill faces efficacy readout after 16-month completion date slip
The Phase 2b comparative efficacy data due in Q4 2026 will test whether VXA-CoV2-3.3 beats an mRNA booster, but enrollment fell 45% and the primary completion date moved to May 2027.
Executive Summary
- Vaxart expects comparative efficacy and safety topline data from about 5,000 participants in the KP.2 cohort of its Phase 2b oral COVID-19 vaccine trial in Q4 2026, testing VXA-CoV2-3.3 head-to-head against a marketed mRNA booster. That makes this the first disclosed readout of an oral pill vaccine against an injectable comparator at this scale.
- Enrollment fell from 10,400 to 5,485, a 45% cut, and the primary completion date moved 16 months, from January 2026 to May 2027. Protocol stability is rated Unstable, with five registry change events logged since the trial started. That instability is a reason to discount how clean the eventual readout will be, independent of the biological result.
- The registry lists seven primary endpoints split between sentinel-cohort safety measures and KP.2-cohort symptomatic PCR-positive COVID-19 rates at 14 days and 12 months. Neither the company's disclosures nor the dossier specify which endpoint anchors the Q4 topline, which limits how the model's 49.4% probability of success should be read.
- AppliedXL's endpoint-met probability sits at 49.4%, carrying a LOW confidence label and a wide uncertainty band, with no upper or lower bound reported. The read leans on operational scale (enrollment per arm) and phase-level endpoint base rates more than on endpoint quality or scientific plausibility, both of which the model scores below neutral for this trial.
- The setup turns on whether Vaxart can deliver a stable, interpretable readout after two completion-date amendments and an enrollment cut, not on a settled efficacy question.
The catalyst
Vaxart, Inc. expects the full comparative efficacy and safety topline readout from the roughly 5,000-participant KP.2 cohort of its Phase 2b COVID-19 trial in the fourth quarter of 2026 Press ReleasePress ReleaseMar 12, 2026. The trial, registered as NCT06672055, compares oral VXA-CoV2-3.3 against a commercially available mRNA COVID-19 booster in adults who previously received at least two mRNA vaccine doses NCT06672055A Study to Evaluate VXA-CoV2-3.3 COVID-19 Vaccine Against Currently Approved/Authorized mRNA COVID-19 Injectable Booster Vaccine in Adults Previously Immunized Against COVID-19 InfectionNCT06672055. Chief Executive Officer Steven Lo called the COVID-19 program a strategic priority tied to a worldwide collaboration with Dynavax, saying the company is "moving quickly toward two critical milestones: the release of 12-month data from our 400-person sentinel cohort expected early in the second quarter, followed by the full readout of comparative efficacy and safety from the main cohort of approximately 5,000 participants in our Phase 2b trial, which we expect in the fourth quarter of 2026" Press ReleasePress ReleaseMar 12, 2026. What is at stake is whether an oral pill vaccine can match an injectable mRNA booster on real-world efficacy, the first test of this delivery format against an active comparator at this enrollment scale NCT06672055A Study to Evaluate VXA-CoV2-3.3 COVID-19 Vaccine Against Currently Approved/Authorized mRNA COVID-19 Injectable Booster Vaccine in Adults Previously Immunized Against COVID-19 InfectionNCT06672055.
Probability of SuccessBased on the AppliedXL Probability of Success model. For more information about the methodology, read the research here.
Enrollment and timeline instability
Registry history shows enrollment moved from 10,000 to 400 (sentinel cohort only) in April 2025, back up to 10,400 in June 2025, then down to 5,485 as of 2026-06-25, a 45% cut from the fuller cohort figure. The primary completion date shifted from 2026-01-01 to 2026-11-01 in June 2025, then to 2027-05-01 in June 2026, a cumulative slip of 16 months from the original date. AppliedXL's protocol stability tool rates the trial Unstable, logging 4.2 change events per year against 5 total changes, and flags prim_comp_date_cumulative_delay and enrollment_significant_change as high-severity registry risk signals. Those changes do not by themselves say anything about the biological result, but they do mean the Q4 2026 window carries real execution risk beyond the science.
The endpoint bar is unclear
The trial registers seven primary endpoints: four are safety and reactogenicity measures scoped to the 400-person sentinel cohort (abnormal safety labs, unsolicited adverse events, treatment-emergent adverse events, and solicited local/systemic reactogenicity), and two are efficacy measures scoped to the KP.2 cohort, both tracking symptomatic PCR-positive COVID-19, one at 14 days post-vaccination and one at 12 months. Neither Vaxart's March 2026 business update nor the dossier specifies which of these anchors the Q4 2026 topline announcement or what effect size would count as a win Press ReleasePress ReleaseMar 12, 2026. Because several of these primary endpoints are safety measures, any model probability tied to "endpoint met" needs to be read against the specific endpoint it resolves, not treated as a single efficacy proxy.
Model read and its basis
AppliedXL's endpoint-met probability sits at 49.4%, carrying a LOW confidence label and a wide, unbounded uncertainty band [AXL-MODEL]. The driver mix behind that number leans on operational and structural features: enrollment per arm, phase-level endpoint base rates, and same-mechanism competitor counts carry the largest positive weights, while endpoint-type success rate carries the single largest negative weight [AXL-MODEL]. Scientific plausibility and endpoint quality scores pull against the read rather than support it. That mix means the 49.4% figure rests more on trial scale and phase base rates than on a mechanistic or endpoint-quality signal, a reason to treat the number as a rough prior rather than a sharp forecast.
Competitive and regulatory frame
No same-target, same-indication comparator trial exists in the competitive dataset; the nearest same-target study is a 20-person Phase 1 trial from Rokote Laboratories Finland, not a comparable efficacy trial. The competitive landscape data mark this pairing first-in-class on a no-peer-trials basis. The trial is explicitly non-registrational, and no Breakthrough, Fast Track, or Priority Review designation is on record, meaning a positive result would not by itself trigger a defined regulatory pathway. Landscape metrics also describe the broader SARS-CoV-2 Spike Protein field as decelerating, with recent trial activity down to 8% of its prior level.
This analysis was produced using AI-assisted reporting systems, AppliedXL data, and official public records. These systems undergo editorial review, quality checks, and regular audits by human experts. Errors may still occur, as with any automated system. Always consult the linked primary sources. Read our AI Editorial Policy.