Search This Blog

Wednesday, September 21, 2016

Sereptas Cargo Plane Has Landed

“there were serious methodological concerns identified by FDA,” according to the documents.

To measure the drug’s effect on muscle function, the company performed a six-minute walk test on the trial’s participants. The FDA reports there was “no nominally significant difference” between patients taking either the higher dose of eteplirsen, the lower dose or the placebo. The agency also griped about the fact that the company chose to compare the performance of the patients on the six-minute walk test against “historical controls,” or DMD patients who were in different trials in the past. - Forbes

You can read all about the "serious methodological concerns" here.

The simple description of Sereptas plan to demonstrate the efficacy of their product:

1) expression of an altered messenger RNA in muscle (pharmacodynamic)
2) production of dystrophin protein in muscle (pharmacodynamic)
3) improvement or preservation of muscle function (clinical).

Throughout the approval process, critics have expressed concerns with the small population — 12 patients — involved in clinical trials for the rare disease, along with flaws in how the clinical study was designed. These factors make “judgment on science difficult,” Califf (the FDA commissioner) said.

No it doesn't. More data points lead to better understanding of the accuracy of what is being measured. It is simple precision and accuracy. You may leave college with a poor grasp of the math behind statistics but you get the concept. The more data points the clearer the picture gets. Poor precision leads to greater doubts about accuracy. You aren't making a "judgment on science" that is "difficult". You are doing math.

Here is an example of data on the production of dystrophin protein in muscle.

Western blots and immunofluorescence were used to quantitate dystrophin.

Table 2: Applicant’s Quantification of Dystrophin by Western Blot and Immunofluorescence Analyses

Patient Western Blot % of normal Immunofluorescence % positive fibers
A 2.05 18.5
B 1.15 19.1
C 0.38 33.5
D 1.62 24.0
E 0.52 21.5
F 0.98 12.8
G 0 7.1
H 2.47 20.7
I 0.96 28.2
J 0 1.4
L 0.14 4.5

A quick glance at Figure 1: Correlation between Two Methods Used to Quantify Dystrophin in Skeletal Muscle: Patients from Study 201/202 tells you all you need to know. As the FDA put it, "Of note, the correlation between the two independent methods used to quantify dystrophin in muscle samples was weak." They also stated, "As discussed above, we believe that immunofluorescence analysis (percent positive fibers) is not a reliable method to quantify dystrophin content." What other criticisms of Sereptas methodology did the FDA panel mention?

Regarding the first pharmacodynamic goal, to demonstrate expression of an altered mRNA in muscle the FDA states;
Because even a minimal PCR signal is interpreted as “positive,” this biomarker provides little support of efficacy for eteplirsen; it does provide evidence that eteplirsen causes at least some degree of exon 51 skipping, as intended.

Regarding the second pharmacodynamic goal, the demonstrate production of dystrophin protein in muscle the FDA states;
FDA conducted an inspection of the facility where the data reported in the publication were generated. Significant methodological concerns were identified, which cast serious doubt on the reliability of assessments from the first three biopsies.
Thus, the review team does not consider “percent dystrophin‐positive fibers” to be a meaningful way to estimate dystrophin content, and we believe the results reported by the applicant on this measure do not establish that a significant increase in dystrophin occurred in response to eteplirsen treatment
In any case, the level of dystrophin was 0.9% of normal after 3.5 years, such that, in absolute terms, the increase from baseline would be, at most, 0.9%, assuming a “worst case” for untreated patients, i.e., zero dystrophin.

Regarding the clinical goal, improvement or preservation of muscle function, the FDA stated;
Two patients in the 30 mg/kg group became unable to ambulate soon after the study start. The applicant then pooled the six remaining eteplirsen patients and compared them to the four placebo patients, an unplanned post hoc analysis. No nominally significant difference between eteplirsen and placebo was identified in that post hoc analysis.
The applicant conducted a number of additional post hoc analyses, comparing the 6 patients who received eteplirsen in the 24‐week double‐blind phase of Study 201 and could still ambulate at the end of Study 201 (and continued on open‐label eteplirsen in Study 202) to those originally treated with placebo in the double‐blind phase of Study 201, and later switched to open‐label eteplirsen.
The applicant conducted a post hoc comparison of the patients in Study 201/202 to data from the “Italian DMD Registry” and the “Leuven Neuromuscular Reference Center” registry.
The problems of externally‐controlled studies are well recognized.

Dr. Ronald Farkas, who led the FDAs clinical team in the neurology products division, suddenly departed the agency just before the approval of Exondys 51. The arguments for the governments approval of Exondys 51 not come in a highly detailed document such as the one Dr. Farkas and his team presented.
Both FDA camps had “exercised reasonable scientific judgment,” Califf found, adding that it’s “exceedingly rare” to overrule a decision by the director of the Center for Drug Evaluation and Research. Without any additional technical expertise of his own, Califf said, he deferred to CDER Director Dr. Janet Woodcock.
An appeal to the authority of someone elses knowledge!

The jury is still out. Serepta still has to put up or shut up. The FDA has just bought them some time and money. How much money? $300,000 per year. Good grief. With BS artists like these, who needs scientists anymore?

Sunday, September 11, 2016

Risk of Cheating Barometer

Cheating in all areas of life is something that will always exist. We frown upon cheating. We throw people in jail if they cheat others out of their hard earned money. It is in general a bad thing for someone to get something through trickery or fraud. The problem is that we all cheat to some degree. We try to pay as little taxes as possible. We beef up our resumes. We tell little white lies when the need arises. That is why we have skepticism. We need a homeland-security-esque risk of cheating threat level system.

If Hillary Clinton debates Donald Trump, the threat level for cheating should be high for both candidates. They will cheat by exaggerating their accomplishments and lying about their failures. If you like Hillary you might think that Trump is the only one cheating up there on the stage, and vice versa. If you are a fact checker for the debate you operate under a different belief. Your job is to check what they say and compare that to the facts. That is the most honest person in the mix.

If you are a scientist and your job is to create new drugs, you are operating at a high level risk of cheating. The FDA will serve as the fact checker. Your boss is a different story. Your boss is also under a high level risk of cheating. He or she may have a lot of pressure to get the latest antibody drug to show the kind of data the CEO can show the board at the next meeting. Does this mean you are going to cheat? No. But you might end up on a path that will lead to the termination of your project/job. What does an honest person do under these circumstances?

The last post I put on the blog was about race. Black vs white in America is a high risk of cheating topic. "Hands up don't shoot" was the mantra after the Mike Brown incident in Ferguson MO. It is highly probable that that was an inaccurate description of what took place. Facts not supporting the narrative include a discharged weapon prior to the final lethal shots and the strong armed robbery committed by Brown prior to the incident. At his funeral a friend stated the Mr. Brown was out spreading the word of God prior to his death. That narrative would indeed make everyone skeptical about the police departments conclusion that the officer who shot Mike Brown, Darren Wilson, was defending his life. Who needs to shoot and kill a person out spreading the word of God? The problem is that the facts of the gunshot and robbery have been verified. The preaching has not.

Race is a topic where our Risk of Cheating threat level will be high. Was Mike Brown feeling invincible the day he died or was he out saving souls? That information is not proof positive of anything. It is information that only supports a narrative.

Back to BS - biotech science.

Our narrative is simple. We apply basic research as avenues to treating and/or curing disease. Science is the most honest and pure way of thinking. We set aside our biases and stick only to the facts. Narratives involving our motivations (money, fame, altruism...) are not factors in our scientific method. Those who have chosen to question underlying threats to our honesty are not understanding how science works. That is the big picture narrative behind the benevolence of BS - biotech science. We make drugs to help people because we have dedicated our lives to serve our fellow humans. Just don't bring up the history of that narrative.

Quick example: My favorite cargo cult of Seattle WA, Juno. August 31, 2016: Juno CEO Hans Bishop nets $1.28M from 42,673 shares of his stock. Sept. 8, 2016: Juno stock surges on positive clinical trial data.

The stock dives after four deaths in a trial of 129 patients are reported. It goes from $40-something per share to $20-something. The stock creeps back up to around $30 per share. The CEO socks away a cool million in spite of the narrative pointing to a bright future. Threat level HIGH!

Beware investors! Bullshit level high. Store your capital far from the shores of biotech.