Although preliminary, it's seems a little difficult to not find the results of the study significant, absent some gross errors in performing the testing.
A 5 or even 10% false positive rate in healthy controls seems relatively typical for ACPA testing. So 2 of 25 controls testing positive might be expected.
The Mayo Clinic is indicating that the false positive rate for ACPA for individuals with autoimmune diseases other than rheumatoid arthritis (RA), including for example systemic lupus erythematosus (SLE), may be about double the false positive rate for healthy persons.
Allowing for 1 or 2 (non-autoimmune) false positives among the patients, that leaves 5 or 6 true positive and/or non-RA autoimmune false positive patients - which is quite a large number out of 25 and which seems to require that either a significant number of the patients actually have RA, which seems unlikely, or that a large number have other autoimmune diseases - or that the ACPA may indeed be specifically positive for a subset of ME/CFS patients.
So, absent error, either way the results could be suggesting that the ACPA test may be useful in ME/CFS, if not as a true test for a subset of ME/CFS patients, then to indicate the possibility of autoimmune diseases other than RA. It'll be interesting to see what comes of this.