Chronic fatigue syndrome (CFS/ME) symptom-based phenotypes and 1-year treatment outcomes in two clinical cohorts of adult patients in the UK and The Netherlands. — CFSMEATLAS
Chronic fatigue syndrome (CFS/ME) symptom-based phenotypes and 1-year treatment outcomes in two clinical cohorts of adult patients in the UK and The Netherlands.
Collin, Simon M, Heron, Jon, Nikolaus, Stephanie et al. · Journal of psychosomatic research · 2018 · DOI
Quick Summary
This study looked at different patterns of ME/CFS symptoms in nearly 2,300 patients across the UK and Netherlands to see if the types of symptoms people have when they first seek treatment predict how well they'll improve. Researchers found three main symptom patterns: patients with few extra symptoms, patients with multiple symptoms, and patients with mainly pain symptoms. After one year of treatment, patients who started with fewer symptoms were significantly more likely to feel much better, while those with many symptoms or pain-focused symptoms were less likely to improve.
Why It Matters
This research suggests that symptom presentation at baseline may help predict treatment response in ME/CFS, potentially allowing clinicians to identify patients who may need different or more intensive treatment approaches. Understanding whether symptom complexity reflects disease severity, distinct biological pathways, or other factors could improve personalized treatment strategies and patient counseling about realistic recovery expectations.
Observed Findings
Three symptom-based phenotypes (oligosymptomatic, polysymptomatic, pain-only) identified in previous UK cohort were successfully replicated in new UK cohort of 918 patients.
UK patients with polysymptomatic phenotype were 57% less likely to report substantial health improvement at 1 year (MOR 0.43, 95% CI 0.19-0.94).
UK patients with pain-only phenotype were 67% less likely to report substantial improvement (MOR 0.33, 95% CI 0.13-0.84).
Dutch cohort (N=1,392) showed similar patterns: polysymptomatic patients 72% less likely to improve (MOR 0.28) and pain-only patients 55% less likely (MOR 0.45) compared to oligosymptomatic patients.
Inferred Conclusions
ME/CFS patients presenting with multiple symptoms or pain-dominant symptom profiles have substantially lower odds of reporting favorable treatment outcomes compared to those with minimal extra symptoms.
Symptom-based phenotypes are stable, replicable clinical features that may have prognostic value for predicting 1-year treatment response.
The consistency of findings across two geographically distinct cohorts suggests phenotype-outcome associations are robust and potentially generalizable.
Remaining Questions
Do these symptom phenotypes reflect distinct biological disease subtypes, or do they simply represent different points on a severity spectrum?
What mechanisms explain why polysymptomatic and pain-dominant presentations predict poorer outcomes—is it disease biology, treatment responsiveness, or unmeasured confounding factors?
What This Study Does Not Prove
This observational study cannot establish whether the symptom patterns cause poorer outcomes or merely reflect underlying disease severity that independently predicts poor response. The association between phenotype and outcome does not prove that reducing pain or extra symptoms would improve treatment response, nor does it clarify the biological mechanisms driving these associations. Reverse causality is also possible—patients with inherently poor prognosis may develop more symptoms rather than symptoms preventing recovery.
Would targeted interventions addressing specific symptom patterns (e.g., pain management for pain-only phenotype) improve outcomes in patients with complex phenotypes?
Do phenotypes change over the disease course, and if so, does change in phenotype correlate with change in treatment outcomes?