I don't infer this from the abstract - it states "Antibodies against β2, M3 and M4 receptors were significantly elevated in CFS patients compared to controls." This implies 'everyone' has these antibodies, but they are higher in CFS patients.
Then those who responded to RTX had their antibodies against β2 and M4 receptors (not M3) drop.
I am now going to read the entire paper - being someone of statistics background, I am interested in the results section
I agree that what I suggested does not seem to add up. However, having spent years trawling through studies like this - including my own data - I have come to see that there are all sorts of ways it can eem there is a useful result when there is not.
Patients with ME in the study had, on average, rather higher antibody levels. However, there was huge overlap so one would not be able to conclude anything much from an individual patient's result. Even for the highest antibody levels there were people in the control group at that level.
But the crucial point is that amongst the patients those with high antibody levels showed no statistically significant increase in likelihood of responding to rituximab. That is counterintuitive if the antibodies are causing disease in a subset of patients but it is what they found.
What
was found was that amongst those who responded to rituximab there was a greater fall in their antibodies. Now, this does not give us any evidence for those particular antibodies being relevant, because it might just be a sign that in these patients B cells were depleted more efficiently. And any benefit might be due to reducing some other antibodies that have not been measured and we know nothing about. If anything this piece of information supports the Norwegian idea that rituximab can benefit ME by reducing antibodies. However, since responses did not seem to be more common in patients with the specific antibodies tested it rather suggests that these are the wrong antibodies to look at. But there are further arguments about antibody populations and how tests reflect them that could counter that. It all becomes too speculative.
The bottom line remains, though, that if my memory is right the study indicates that antibody levels pre-treatment tell you nothing about likely response.