GSK breach over Seretide TORCH mortality display: non-significant “16.5% risk reduction” presented too prominently (AUTH/2006/5/07)

📊

Key facts

⌄

Case number	AUTH/2006/5/07
Complainant	Pharmacist practitioner at a general practice
Company	GlaxoSmithKline UK Ltd
Medicine	Seretide (salmeterol/fluticasone)
Setting	GlaxoSmithKline meeting; exhibition display boards/panels
Issue	Mortality graph from TORCH highlighted “16.5% risk reduction” despite p=0.096 (non-significant), creating a misleading overall impression
Key disputed presentation	“Primary outcome - Seretide 500 Accuhaler survival result” with prominent “16.5% risk reduction… p=0.096” and downward arrow
Applicable Code year	2006
Breach clauses	Clauses 7.2 and 7.4
Sanctions	Undertaking received; Additional sanctions: Not stated
Complaint received	30 May 2007
Case completed	26 July 2007
Appeal	No appeal

Reviewed by Dr Anzal Qurbain (FFPM) — ABPI Final Signatory

🤖

Got a question about this case?

Ask one of our 13 specialist ABPI advisors — instant answers, 24/7.

Ask AskAnzal AI

🎬 Expert Video Walkthrough

🎬

Video walkthrough — coming for members

Subscribe now and get expert video analysis for every case as we publish them.

Subscribe — from £299/yr

📋

⌄

A pharmacist practitioner complained about GlaxoSmithKline’s promotion of Seretide (salmeterol/fluticasone) at a meeting using small exhibition display boards.
One panel showed TORCH study mortality data with a prominent claim: “16.5% risk reduction with Seretide 500 Accuhaler vs control p=0.096”, plus a downward arrow and a “Primary outcome - … survival result” heading.
Although the panel text stated the result was non-statistically significant, the complainant argued the overall presentation could be construed as a mortality benefit claim to a passing observer.
GSK argued the p-value and non-significance were shown, the data reflected a subgroup within the UK licence (FEV1 <50%), and HCPs could interpret the information in context.
The Panel assessed the panel as needing to stand alone and considered the overall impression created by the design and prominence of the “16.5%” claim.

⚖️

⌄

Breach found.
The mortality exhibition panel was ruled misleading because it did not make non-statistical significance sufficiently clear on a quick read; the graph’s impression could not be qualified by the smaller explanatory text.
Breaches of Clauses 7.2 and 7.4 were ruled.

🔒

Members get the complete breakdown — Clauses, Sanction, Signatory Lens, Audit checklist, and 3 Key Questions.

Best value

£249/year

Annual — save £99

£29/mo

Monthly

Join Now — Instant Access