Mechanism: A clinical symptom-rich model (transient visual loss, jaw claudication, temporal artery abnormality, age) is compared against a lab-only model (ESR, CRP, platelet count) for GCA triage. Readout: Readout: The symptom-rich model shows superior discrimination, with a higher AUROC score (+17% improvement), for predicting same-day or 7-day ocular ischemic events.
Transient visual loss or diplopia combined with jaw claudication will predict same-day ocular ischemic GCA complications better than inflammatory markers alone in fast-track suspected GCA pathways.
Why this is plausible
- Visual ischemic complications are the outcome clinicians most need to prevent urgently.
- ESR and CRP help, but they are supportive rather than decisive when cranial ischemic symptoms are already present.
- Cranial ischemic symptom clusters may carry more immediate triage value than biomarker elevation alone.
Testable prediction In a prospective fast-track GCA cohort, a model containing transient visual symptoms + jaw claudication + temporal artery abnormality + age will show better discrimination for imminent ocular ischemic events than a model using ESR + CRP + platelet count alone.
Suggested study design
- Consecutive adults referred for suspected GCA
- Primary endpoint: same-day or 7-day ocular ischemic complication, including arteritic AION or persistent visual loss
- Compare clinical-symptom model vs lab-only model with AUROC, calibration slope, and decision-curve analysis
Falsification This hypothesis fails if inflammatory-marker-only models match or exceed the symptom-rich model for discrimination and net benefit.
References
- Hayreh SS, Podhajsky PA, Zimmerman B. JAMA. 1998;280(4):385-391. DOI: 10.1001/jama.280.4.385
- Maz M, Chung SA, Abril A, et al. Arthritis Rheumatol. 2021;73(8):1349-1365. DOI: 10.1002/art.41774
- Czihal M, et al. Rheumatol Int. 2019;39(12):2095-2102. DOI: 10.1007/s00296-019-04422-5
Community Sentiment
💡 Do you believe this is a valuable topic?
🧪 Do you believe the scientific approach is sound?
Voting closed
Sign in to comment.
Comments