Relapsed kids are as severe as before

Here is an interesting observation from the follow-on paper of the Lidcombe trial. They note that those three children that had dysfluent speech recordings stuttered with a similar severity than before the treatment. First, is this observation just a statistical fluctuation or a real event? Second, can we observe the same effect in adult who relapsed after treatment? And thirdly, does it tell us something about the nature of stuttering?
I think it is a true effect for both children and adults: the relapsed have a tendency to go back to the same level of severity than before treatment. But I do not know of research that looked at this aspect. I did not consider this aspect when analyzing the outcome data of the Kassel Stuttering Therapy (KST). But I plan to re-visit the data.
Assuming it's true, can we learn a bit more about stuttering? My theory-biased interpretation is as follows: Stuttering at its core is a neurological dysfunction modulated by the response of the brain to this dysfunction. A stuttering treatment is most of all a behavioural therapy: you change some behaviours and you become fast more fluent. So going back to your old behaviours, you go back to your own severity. I cannot imagine that, especially in adults, treatment improves the underlying dysfunction and then they relapse due to a worsening of the dysfunction to before treatment. With children, new behaviours might be more easily ingrained in the brain or underlying weaknesses rectified or well compensated. But there is not a clear case for either interpretations.