AstraZeneca

Opinion · Devil's advocate

Fact-check-o-meter

AstraZeneca's releases match its filings, but lead with core figures and give few trial numbers

AstraZeneca's half-year report is candid about the failed Wainua trial and the missed PFS goal for sone-ve, so we found no contradiction between releases and filings. The gaps are in framing: core EPS leads the key performance bullets while reported EPS grew 3% in the half year and fell 2% in the second quarter, both at constant rates, and two phase 3 releases call results clinically meaningful without giving hazard ratios. The legal note lists no securities class action.

1. Press release vs filing · weight medium

   Core EPS leads, reported EPS grew more slowly

   The company said: "Core Operating profit and Core EPS increased 11%" source

   The record says: The bullet is for the first half. The same table shows H1 Reported EPS of $3.60, up 4% as reported and 3% at constant exchange rates, against Core EPS of $5.21, up 12% (11% at constant rates). For the second quarter, Reported EPS was $1.61, up 2% as reported and down 2% at constant exchange rates, against Core EPS up 21% (18% at constant rates). It says the gap is "primarily due to costs relating to the amortisation of intangibles, impairments, legal settlements and restructuring charges." The footnote names legal settlements among the items excluded from Core measures. Separately, the legal note records a May 2026 settlement with the University of Sheffield for a payment by AstraZeneca of $220m. source

   Our view: The release gives both measures, so this is a matter of emphasis, not omission. Reported EPS grew less than Core EPS in both periods, and in the second quarter it fell at constant rates, which the bullets do not show.

2. Press release vs filing · weight medium

   Wainua miss is stated plainly, with a subgroup beside it

   The company said: "We remain confident in the strength of our pipeline and have more than twenty high-value readouts due over the next 18 months." source

   The record says: The report says Wainua in ATTR-CM "did not meet the primary efficacy endpoint of the composite outcome of CV mortality and recurrent CV clinical events up to 140 weeks compared with placebo." It adds that a prespecified monotherapy subgroup result "was nominally significant" while "In patients who were on stabiliser therapy at baseline, no treatment effect was observed." AstraZeneca's separate 9 July 2026 update on the trial also reported the primary miss. source

   Our view: AstraZeneca did not bury the failure, and the CEO calls it a disappointment. Because the subgroup was not the primary analysis, a nominal result there is a reason to look further, not evidence of benefit.

3. Science · weight high

   Sone-ve: the broad survival claim rests on a key secondary endpoint, and no numbers yet

   The company said: "Sonesitatug vedotin demonstrated a statistically significant and highly clinically meaningful improvement in overall survival in 2nd and later-line CLDN18.2-positive advanced gastric/GEJ cancers" source

   The record says: The release says the trial "met the dual primary endpoint of OS in 3rd and later-line treatment, and a key secondary endpoint of OS in the overall trial population" and that for PFS in the second-line-and-later population "results showed a trend toward improved PFS ... but did not reach statistical significance." It describes the trial as "randomised, open-label, sponsor-blinded". ClinicalTrials.gov lists 594 participants and two primary outcomes, PFS in all randomized participants and OS in 3L+ participants, with OS in all randomized participants as a secondary outcome. The H1 report's trial table labels CLARITY-Gastric01 as "Primary endpoint met", although its text discloses the PFS miss. source

   Our view: The headline claim covers the broader 2nd and later-line population, where OS was a key secondary endpoint, and the other primary endpoint, PFS, did not reach statistical significance. The release discloses both, but gives no survival gain or hazard ratio, so the size of the benefit is unknown until the data are presented at a medical meeting.

4. Science · weight medium

   SAFFRON: positive on PFS and OS, but open label against chemotherapy

   The company said: "First global Phase III trial to show significant progression-free and overall survival benefits in this setting" source

   The record says: The release describes SAFFRON as "a randomised, open-label, multi-centre, global Phase III trial" versus platinum chemotherapy in 338 patients, with "The primary endpoint is PFS and key secondary endpoints include OS and objective response rate." ClinicalTrials.gov lists one primary outcome, progression-free survival, and OS as a secondary outcome. source

   Our view: The trial is open label against chemotherapy, which means patients and investigators know the treatment, although ClinicalTrials.gov describes PFS as assessed by blinded independent central review. The release gives no hazard ratios, and OS was a key secondary endpoint, so the size of the survival result is not yet known.

5. Legal and disclosure record · weight low

   No securities litigation or enforcement disclosed in the latest filing

   The record says: The legal proceedings note in the H1 2026 report lists patent cases, commercial disputes and government matters. It says of the Texas qui tam case "In June 2026, the case settled for an immaterial amount" and of the University of Sheffield dispute "in exchange for payment by AstraZeneca of $220m." A putative class action over patient genetic information (Barone Privacy Litigation, Illinois federal court, filed March 2026) was followed by a consolidated complaint in June 2026 "not naming AstraZeneca as a defendant." The government section also lists a False Claims Act qui tam brought by a relator on behalf of the United States and several states in the US District Court for the Central District of California (340B Qui Tam), where the Ninth Circuit reversed a dismissal in March 2026 and the note describes it as a contingent liability. The note lists no securities class action, and it does not say the United States has intervened in that case. source

   Our view: The disclosed matters are mostly contract, patent and pricing disputes, several now concluded. The 340B qui tam is described as a contingent liability, and the Texas qui tam, Sheffield and Barone matters are marked concluded in the note.

This is opinion. We compare the company's own announcements with its filings and the scientific record, and link every source. It is not an allegation of wrongdoing and not investment or legal advice. Legal items list only matters disclosed in filings or by courts and regulators. Checked Oct 23, 2026. Companies can respond through our right of reply.