Novartis secretly ran its own Japan drug trials — the results turned out fake

Starting in 2000, Novartis ran clinical trials of its hypertension drug Diovan (valsartan) at five Japanese universities, aiming to show it prevented strokes and heart attacks beyond simply lowering blood pressure — a claim that would set it apart from cheaper rivals. A Novartis employee worked on the statistical analysis of every trial without his employer's affiliation being disclosed in the published papers.

Kyoto Prefectural University of Medicine and Tokyo's Jikei University later retracted their Diovan papers after outside statisticians found the data had been altered to favor the drug — some patient records were changed to show a prior stroke that had never happened. Novartis had also paid the two universities 570 million yen (about $5.4 million) for 'running classes' and Chiba University 91 million yen in scholarships during the trial period, funding research it was publicly presenting as independent.

Japan's health ministry filed a criminal complaint against Novartis and the employee on January 9, 2014 — the first time Japan's pharmaceutical affairs law, which bars exaggerated advertising, had been used against fabricated academic data rather than a marketing claim. The ministry also moved to suspend Novartis's Japan operations for roughly 15 days over the scandal and a separate failure to report more than 3,200 adverse-event cases involving its drugs.

Diovan had been earning Novartis roughly 100 billion yen a year in Japan and over 108 billion yen globally as recently as 2012. A Tokyo court acquitted both Novartis and the employee in 2017, ruling the advertising law did not cover academic papers — but the retracted trials, the criminal case and the funding disclosures had already done the commercial damage by then.

A criminal complaint and a roughly 15-day suspension of Japan operations were the price of fabricated trial data.

Novartis sought a claim — that Diovan cut strokes, not just blood pressure — strong enough to justify its price against generic-adjacent rivals.

It ran the trials that would make that claim through its own undisclosed employee, embedded in the universities' analysis without appearing on the papers.

It funded the same universities directly while presenting their findings as independent academic research.

The conflict surfaced only when outside statisticians reviewed the retracted papers — internal review had not caught it.

A trial the manufacturer secretly ran and funded is not independent evidence, whatever letterhead it carries — and disclosure gaps surface eventually, at a worse time than admitting them.

Both retracted trials stayed retracted even after the 2017 acquittal. Japan tightened its clinical-trial law in response, and the case remains the country's reference scandal for pharmaceutical research conflicts of interest.