Where our evidence is thin: a tour of the weak spots

A paper is only as honest as the places it admits are soft. Here are ours.

It is easy to produce a confident-sounding document. It is harder — and more useful — to mark exactly where the confidence is not earned. Our paper tags speculative content with certainty values and flags weak-evidence areas; here is a plain-language tour.

The soft spots:

• Single-study threads. The norepinephrine story, the VMAT2/dopamine-circuit thread, and several others rest on a handful of small studies, sometimes from one group. Replication is pending.
• Animal and cross-disease evidence. Many mechanistic arrows (norepinephrine and clearance, norepinephrine and microglia, neuromodulators and cortical gain) are established in animals or other diseases, not measured in ME/CFS.
• Composite measures. New indices (like the norepinephrine “pathway index”) are useful but unvalidated; a striking result can partly be an artefact of the measure.
• Confounds that are not removed. Deconditioning, medication, and activity level can all move the biomarkers we discuss, and most studies cannot rule them out.
• The null and the negative. Some of our strongest statements are about what does not work — amantadine for MS fatigue (a clean null), clonidine in adolescent CFS (a clean null), and graded exercise therapy (contested, and a case where the endpoint — symptom worsening — is what matters, not absence of any effect). Those deserve as much prominence as the positive ones.

Why we flag them. Because the alternative — a smooth narrative with the uncertainty sanded off — is how patients get misled, and how research priorities get set on air. A flag is not a weakness; hiding one would be.