Long COVID Phenotypes: Symptom and Performance Profiles at 2 Years

Key Takeaways

• Among people assessed in the Netherlands approximately two years after SARS-CoV-2 infection, a combined post-exertional/ME/CFS-like symptom, sleep, depressive-symptom and coping component was significantly associated with symptom-defined PASC.
• Separate processing-speed, executive-functioning, physical-performance and comorbidity/initial-hospitalization components were not significantly associated with that classification.
• Persistent complaints were more prevalent in exploratory physical-impairment and lower-cognitive-performance clusters (77% and 84%) than in a relatively mild cluster (42%).

Ongoing fatigue and difficulty concentrating after coronavirus disease 2019 (COVID-19) can coexist with post-exertional symptoms, disturbed sleep, and varied physical or cognitive functioning. For people still reporting complaints long after infection, how symptom burden relates to measured performance remains uncertain. Investigators selected participants with and without persistent complaints and assessed symptoms alongside cognitive and physical function.

Post-acute sequelae of COVID-19 (PASC) was classified operationally by fatigue and concentration complaints in the Dutch VeCosCO case-control analysis. Participants in Verfaillie and colleagues' PLOS One study of long-term PASC phenotypes came from existing COVID-19 cohorts. The analysis included 153 people with complete measures: 88 met the study definition and 65 did not.

Assessment occurred approximately two years after the first confirmed severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection; eligibility began at least three months after diagnosis. The Checklist Individual Strength defined the groups by severe fatigue and concentration complaints, not a universal PASC diagnostic criterion. Investigators applied principal component analysis (PCA) to 19 tests and questionnaires spanning self-reported post-exertional malaise (PEM), myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS)-like symptoms, sleep, depressive symptoms and coping, alongside cognitive and physical performance. Measures included the Fatigue and Energy Scale and DePaul Symptom Questionnaire for exertional or fatigue-linked complaints, Stroop and Trail Making tests for cognition, and the Short Physical Performance Battery for physical function. Standardized component scores were modeled with adjustment for age, sex, education, nationality and time since infection; K-means clustering was explored separately.

PCA separated a combined self-reported PEM, ME/CFS-like symptom, sleep, depressive-symptom and coping dimension from processing-speed, executive-functioning, physical-performance and comorbidity/initial-hospitalization dimensions. Only the combined dimension was significantly associated with the fatigue/concentration-defined PASC group after adjustment and multiple-comparison correction (adjusted odds ratio, 97.10; 95% confidence interval, 16.52–570.34; p < 0.001). The other components were not significantly associated with that classification. The wide confidence interval limits precision, and the association does not establish an individual's risk of biologically distinct disease.

Exploratory clustering identified predominant physical impairment (n=31), below-average cognitive performance (n=33) and relatively mild or no substantial problems (n=89). Persistent complaints were more common in the physical and cognitive profiles than in the relatively mild profile. These profiles were distinct from the adjusted component association and did not themselves establish PASC diagnoses.

Missing measures excluded 77 people from the primary complete-case analysis, although the authors found no differences in the tested demographics between included and incomplete cases. Cross-sectional assessment cannot establish causal mechanisms or symptom trajectories. The symptom-based group definition overlaps the associated self-report component, and differences between questionnaire and performance measures may influence component separation. Published cluster-specific complaint counts do not reconcile with the stated overall PASC group total; the exploratory profiles are not validated diagnostic or treatment categories, and no phenotype-specific treatment benefit was established. The authors concluded that interrelated self-reported symptoms coexisted with physical and cognitive profiles and that longitudinal investigation is needed to establish causal relationships.

Clinician Questions

What fatigue and concentration thresholds defined PASC in the VeCosCO comparison?

Among previously SARS-CoV-2-infected VeCosCO participants, the group labeled PASC required both a Checklist Individual Strength fatigue severity score ≥35 and a concentration problems score ≥18. These were operational cutoffs for this comparison, not a universal PASC diagnosis.

How was post-exertional malaise assessed in the long-term COVID-19 phenotype analysis?

Among previously SARS-CoV-2-infected VeCosCO participants, investigators used the self-reported Fatigue and Energy Scale after neuropsychological assessment to gauge worsening of fatigue after cognitive exertion; they did not establish PEM through experimental exercise stress testing.