The only exclusionary testing available is Lyme testing, which is notoriously unreliable. This seems to be the whole point of the study, if you can´t distinguish Lyme from ME by using Lyme tests, can you distinguish them by using metabolomics? The problem then is whether they can make sure there aren´t any Lyme cases in the ME group, as this will affect the results. I guess if 100% of Lyme patients who test positive on the two-tier test have a particular biomarker, and most people with ME don´t, then that either suggests that they are different diseases or that seronegative Lyme is a different beast to seropositive Lyme (the first being a more likely hypothesis, obviously). They might then be able to go back over the results and exclude those ME cases that look a bit Lymey based on their metabolic profile, and see what effect that has on the sensitivity and specificity of the biomarkers.