Updated from Sunday with Clovis stock price.
CHICAGO (TheStreet) -- Clovis Oncology's (CLVS) lung cancer drug CO-1686 is turning patients into diabetics.
"Three or four" lung cancer patients in the ongoing phase I/II study of '1686 are now taking insulin to control hyperglycemia (high blood sugar) caused by the experimental drug, Clovis CEO Pat Mahaffay told a roomful of investors and analysts on Saturday night.
The new disclosure ratchets up the seriousness of the link between '1686 and hyperglycemia and should raise more investor concerns about the drug's tolerability. The safety issue is magnified in light of the tight, competitive race between Clovis and AstraZeneca (AZN - Get Report), which is developing a similar lung cancer drug known as AZD9291.
Clovis shares are dow 11% to $45.61 in Monday trading.
Both Clovis' '1686 and AstraZeneca's '9291 are designed to be effective against non-small cell lung cancer containing a genetic mutation known as T790M, which renders tumors resistant to treatment with drugs like Roche's (RHHBY) Tarceva. Updated, positive study results for both drugs were presented Saturday morning at the American Society of Clinical Oncology (ASCO) annual meeting. But Wall Street demands drug development be a zero-sum game, so healthcare investors spent much of the weekend debating whose drug -- Clovis' or AstraZeneca's -- was superior.
In the phase I/II study of Clovis' '1686, hyperglycemia was observed in 55% of lung cancer patients, with 22% of these patients reporting hyperglycemia severe enough to require treatment with metformin, a blood-sugar lowering pill commonly prescribed to Type 2 diabetics.
On Saturday night at a Clovis-sponsored event, investors and analysts learned for the first time about a small number of '1686-treated lung cancer patients unable keep their blood sugar levels under control with metformin, requiring doctors to start them on insulin injections.
Mahaffey is very savvy Wall Street guy, so he deftly pivoted the ominous conversation about the insulin disclosure to an uplifting and positive anecdote about a lung cancer patient with pre-existing diabetes who was responding remarkably well to '1686. Mahaffay's message: Sure, turning some of our '1686 patients into insuling-injecting diabetics sucks, but lung cancer sucks worse.
Undoubtedly true, but it's also true that the ramped-up diabetes concern weakens Clovis' assertion that '1686 is superior to AstraZeneca's '9291. An objective read of the data presented so far reveals no winner between Clovis and AstraZeneca. Lung cancer patients with T790 mutated tumors win because both '1686 and '9291 appear to be very effective with their own, unique side effect issues. Which drug works best or is better tolerated cannot be determined at this time.
Yale lung cancer expert Dr. Thomas Lynch took on the winner vs. loser issue head on during a discussion of the '1686 and '9291 data on Saturday morning.
Lynch's verdict: Both the Clovis and AstraZeneca drugs are active in the T790M-positive lung cancer patients and neither has an efficacy edge based on the data available.
Clovis, perhaps self-conscious about the numerically lower response rate compared to AstraZeneca, chose to play up the current estimate for median progression-free survival (PFS) in '1686-treated patients exceeding 12 months. "PFS favors us substantially," said Clovis Chief Medical Officer Andrew Allen, at the Saturday night investor event.
Here's the '1686 PFS curve to date:
It does look impressive, with median PFS not yet reached with follow-up going out beyond one year. But look closer and you'll notice the "flatness" seen in the latter half of the PFS curve is heavily dependent on a relatively few patients plus patients censored, meaning they either dropped out the trial or information about tumor progression status isn't known. Simply put, Clovis' claim that '1686 has superior PFS data compared to '9291 based on this graph is flimsy.
Here's the PFS curve from AstraZeneca's '9291 presentation:
Yale's Lynch also tackled a safety comparison between the Clovis and AstraZeneca drugs:
$AZN $CLVS toxicity comparisons from Discussant Tom Lynch #ASCO14 pic.twitter.com/PnE4FrZGI7 Adam Feuerstein (@adamfeuerstein) May 31, 2014My tweeted photo is hard to read, but for Clovis' '1686, the significant side effects include the aforementioned hyperglycemia and Qt prolongation (heart arrhythmia) seen in 15% of patients, with 7% of those reporting more serious grade 3 events.
For AstraZeneca's '9291, 27% of patients reported rash and 3% of patients experienced "interstitial lung disease-like" events. The '9291 presentation also disclosed some "grade 5" adverse events at the lower dose, which means patients died. But it's unclear if these are deaths due to tumor progression or related to '9291 treatment.
If Clovis can't differentiate the activity profile of '1686 from '9291, the smaller company may try to use it speed to get to market quicker. Clovis has started new lung cancer studies of '1686 with the goal of compiling enough clinical data to seek FDA approval in the middle of 2015.
AstraZeneca's current guidance is to seek approval of '9291 by the end of 2015 but a lot of investors here at ASCO believe the company will accelerate that timeline, especially now that Pfizer's takeover effort has been rebuffed. AstraZeneca is under pressure to prove to investors that it can make good on its revenue and earnings growth estimates. Getting '9291 approved quickly and turning the drug into a lung cancer blockbuster is a big part of that effort.