Rett Syndrome Data Already Out – Anybody Pays Attention?

This Blog Is Only For Educational Purposes. Do Not Trade On This Blog Alone. Do Your Due Diligence or Consult a Professional.

Not So Subtle Changes in RSBQ

RSBQ stands for Rett Syndrome Behavioral Questionnaire. It consists of 45 questions, each questions have 3 answers, valued 0, 1 and 2. The highest possible score is 90 and represents the strongest extend of disease symptoms. Value 0 on a question corresponds to “never”, 1 to “sometimes” and 2 to “always”. The first question I had was “what is the average score of Rett Syndrome patient”. I found the immediate answer in this paper https://jneurodevdisorders.biomedcentral.com/articles/10.1186/s11689-015-9104-y

The questionnaire divides the symptoms into 8 categories and these are having varied number of questions making some weight more than other on the final score.

Let’s see what RSBQ data was presented by $AVXL on the 6 open label patients in PK arm of the first Rett Trial.

From the outmost left graph we read the average combined score for the 6 patients was 50 on RSBQ scale of 90. 50 is above an average score as given by the Welsh survey. Now, let us discuss what that means for the patients and the care givers. There is a drop of 15 points, which in general makes these patient to have scores below average Rett Syndrom patient from a score above the average, from 120% of average to 83% of average. The 15 points for an average patient could have meant change from (2) “always” to “sometime” (1) or from “sometime”(1) to “never”(0) on 15 questions. Please, observe the p-value for just 6 patients which suggests besides a good distance between means. The downgrading of the symptoms on 15 questions out of 45 happens on low dose. Dr. Missling has revealed that was the case, in this pilot program aimed at dosing with potentially the least “harmful” dose since Blarcamesine has never been tried on Rett syndrome patients. The low dose is the reason we can only expect greater efficacy in the second Rett Syndrome trial. The confidence of Dr. Missling in Blarcamesine is corroborated by making the first 6 dosed patients in PharmaKinetics Open Label so that we have early reading of the trial. The Part B is to obtain better p-values on the test which by the Part A p-values is kind of redundent. Yet, the next 21 patients can be expected to lower the p-value score even further, unless the dose after the Open Label PK Arm has been increased, and then it becomes vey interesting, especially witht the 12 weeks open label extension. Just a remainder, the p-values below .05 are taken as a threshold to assume the outcomes are statistically valid.

The timing of the 3 Rett syndrome trials is such that the results of the previous feeds into the next, probably having written in their protocols some dose of flexibility. Ergo, FDA allowing such arramgement of trials exhibits high confidence in Blarcamesine and practically expedites the whole process without approving the drug outright on some contigency due to high efficacy, which though might not be excluded from possible scenarios once the trials confirm it be overwhelming. Anavex is basicly almost having Breakthrough Therapy protocol without calling one. I believe that the initiative to have such expedited development schedul was all due to Dr Missling. FDA just acquiesced to it, already recognizing the high efficacy in murine studies, and maybe Phase 2a Alzheimer’s. The Orphan Drug designation is more like a recognition that you play in the field of Orphan Diseases. The voucher after approval become a realized incentive. These are perks for playing in the field of Orphan Diseases and by themselves are part of the package. These facts should strengthen the case for FDA already recognizing the looming efficacy of Blarcamesnie.

The plot on the left represents the distribution for those 6 patients on measures of improvement. CGI-I is just 7 point scale, very subjective and dependent on the physician knowing the state of the patient before dosing.

Since we have distribution of RSBQ scores for all 6 patients and we make certain assumptions like that we have over 50 subjects in an arm with the same distribution of score like the 6 patients before dosing, we can calculate Effect Size under these assumptions. Also we assume that the SD about the mean before dosing is the same as for the average patient listed above then simplified Cohen D calculations for Effect Size give us the results seen below.

To qualitatively put it, it is a “very large” effect. If decent Effect Size to approve a drug is, the threshold of, 0.50 then indeed on the first Rett trial Blarcamesine earned its approval. Remember, I can tell you what Dr. Missling can not. I am looking into the future using clues from the present. I can be wrong but most likely I am only wrong in matters of details not the shape of things. The Effect Size might be different but the judgement on its quality should stay.