Great questions
@Zugzug! As always, I appreciate your engagement in discussing the details...
Why everyone got a little worse, and some fell out of significance:
Frequency Therapeutics has shown that a single dose of FX-322 only penetrates to a certain depth in the cochlea, and therefore can only activate progenitors to a certain depth where the drug is concentrated enough. They actually measured the perilymph to provide additional evidence. It is shown here:
View attachment 44565
But more simply, here:
View attachment 44566
This is generally reinforced by subjects 1, 3, and 4 showing improvements of 10 dB+ on the audiogram at 8 kHz. So, we can conclude that the subjects, all 5 of them, had improvements detecting sounds in the highest frequency ranges. In the English language, the language used for the word score tests, this is where the fricative sounds are distinguished. IE: Consonant sounds created by, "f", "th", "h", "z", "v", "s", "sh"... As English speakers, looking at those consonant, many, many words are made up of fricatives that we use every day, that could easily be mistaken for one another on a 50-word C-N-C word list. As has been researched and shared here, as the IHC/OHC cells degrade over time, it becomes more difficult to distinguish the subtle differences between these sounds. So, the evidence points to the fact that restoring the hair cells with FX-322 made these consonants distinguishable to the listener.
So how did the word scores decline in 1-2 years?
As we are all aware, and as all research has shown for SNHL, damage is cumulative and damage is permanent to these fine hair cells. So, all of these participants have accumulated significant life-long damage. And, more recently, damage can be an accelerant to earlier hearing loss in life.
As we are aware, the entire cochlea wasn't regenerated. The 8 kHz+ area was restored with "new" cells, but below that, the cells continued to contain the damage absorbed from birth, and continued to wear at the normal expected rate for someone with considerable hearing loss damage. So, while they may have been able to distinguish fricatives at the follow-up, the set of hair cells below 8 kHz continued to decline in sensitivity from the original baseline, and therefore reduced sensitivity to distinguish affricates, for example sounds like: "ts", "ch", "j", and soft-"g", respectively.
Notes on Subject 2 and 5:
Their follow-up tests were the furthest out, so they absorbed more wear to the "original equipment cells" than subject 1, 3, and 4. It also stands to reason that their losses were more concentrated more deeply in the cochlea where no FX-322 could reach. So, while FX-322 did help provide some gains, the continued wear of already highly damaged cells outweighed the benefits of the high frequency cells that were regenerated.
On OHC + IHC wear:
It is believed that in a healthy cochlea, research shows that OHC
tend to wear out first, and then the IHC follow. The mechanisms for wear actually point to OHC actually creating a turbulence buffer in the cochlea, that "protects" the IHC. Also, research shows that OHC are more sensitive to the turbulence in the cochlea due to their mechanical nature, and can be damaged more easily. IHC though, are also protected from a certain amount of turbulence in the perilymph because OHC are in place. So, if OHC were never regenerated to begin with, the IHC would wear at a faster rate because they are subject to more turbulence in the perilymph.
Why are they all worse?
I actually think this shows promise. It's accepted that hair cells continue to wear from life to death. So, we're observing regular wear and tear over time from a single dose. It would actually be more suspicious, in my opinion, if a patient continued to improve significantly, while others continued to decline. What's interesting though, is 4/5 actually fell by a somewhat similar amount.