Considering the importance of clinical data to a drug developer - and its investors - having to correct previously released information is a scenario always best avoided. Unfortunately for Vertex, the clarification it had to issue today referred to some of the highest profile data seen so far this year, on its promising cystic fibrosis candidates that in the last few weeks have been labelled future blockbuster products.
It remains to be seen whether the 16% plunge in Vertex’s share price in early trade today is justified - the new data might be less impressive than the previous read out but it is still firmly in positive territory. However the fact remains that this is still interim, phase II data. Despite the promise seen, this was always a risky product. The $2.5bn wiped from Vertex’s valuation this morning suggests expectations were already running ahead of themselves.
Vertex this morning released both corrected and new information from the trial of its combination regimen, VX-809 and Kalydeco.
What was previously described as the absolute change in lung function was actually the relative measure, the company said, a mistake it blamed on a “misinterpretation by a third party”. Absolute changes look at the improvement difference between baseline and end point, while relative is calculated as the absolute change over baseline.
Absolute data is the “gold standard” in measuring change in lung function and correlated with survival, and is what regulators want to see, according to Mark Schoenebaum, an analyst with ISI Group. Unfortunately, the real absolute data was not as good as previously thought.
As it stands now, 35% of patients had an absolute improvement in lung function of at least 5 percentage points while 19% had at least a 10 percentage point improvement. Previously the company stated that 46% of patients had the 5 percentage point improvement while 30% had the 10 percentage point improvement – these figures actually describe the relative improvements.
New data disclosed today revealed that the mean absolute improvement in lung function was 8.5 percentage points for patients on the drug combination, compared to placebo. However the gloss was taken off this impressive figure by the disclosure that lung function declined in the placebo group by 4.6 percentage points.
This is a surprisingly poor performance by a group that should be expected to show little change in lung function and is likely to inflate that improvement seen in patients on the drug.
It is this disclosure that likely contributed the most to Vertex’s share price plunge today, as it raises concerns about the reproducibility of results in phase III.
On a conference call today Vertex executives suggested the decline in the placebo arm could be due to antibiotic withdrawal. Finding the cause of that decline, along with further analyses of the new and corrected data, will no doubt continue in the coming months.
Vertex executives also stressed that their conclusions and confidence in the combination’s clinical activity has not changed, with the data showing a meaningful benefit in a substantial number of patients. Other drugs, such as TOBI, have been approved on the basis of a 4% improvement in absolute lung function, they pointed out.
As for the necessary data correction, the company can do little but eat humble pie and promise not to do it again. “This mistake is disappointing. It's not how we do business. Mid-year data will trump all of this, and we are confident that the effect we are seeing is real,” Chris Wright, Vertex's senior vice president of global medicines development and medical affairs, told investors.
Those investors will need more than fine words to bid Vertex shares back to $64.85, the price they closed at Friday. The stock was trading at $54 this morning, still well above the $37 seen prior to the original, albeit inaccurate, data released at the beginning of the month.
The full data from this trial - expected mid-year - was always needed to more accurately gauge this drug's potential, and that fact has not changed. What has changed is perception of this product's chance of success, which arguably now better reflects the risks associated with a mid-stage, novel candidate.