Abstract Background and aims Glial fibrillary acidic protein (GFAP) is a leading biomarker candidate for prehospital differential diagnosis of acute stroke code patients. We explored the diagnostic capability of prehospital GFAP to rule-in and rule-out intracranial hemorrhage (ICrH) in a large cohort of acute stroke code patients, optimizing cut-offs to provide high diagnostic certainty. Methods Laboratory measurement of GFAP was performed with Alphalisa® in plasma samples collected 6 hours of symptom onset, with a median (IQR) last-known-well to sampling time of 60 (38-107) minutes. ICrH rule-in and rule-out cut-off values were selected separately for each sex and age group (male 60, 60-75, 75 years, female 60, 60-75, 75 years). Results We included 938 patients (122 ICrH, 615 acute cerebral ischemia ACI, 201 stroke mimic SM). With rule-out cut-offs, GFAP ruled out ICrH with high certainty (NPV 98.1%) in 479 patients (51.1% of the cohort, 57.6% of ACI/SM patients), with 9 false negative cases (patients with very small hematoma volume). With rule-in cut-offs, GFAP diagnosed ICrH with moderate certainty (PPV 88.9%) in 36 patients (3.8% of the cohort, 26.2% of ICrH patients), with 4 false positive cases. Overall, 423 (45.1%) were left in-between rule-out and rule-in cut-offs in a diagnostic gray area. Conclusions In the ultra-early differential diagnosis of stroke code patients, prehospital plasma GFAP and high-certainty cut-offs correctly diagnosed 53.5%, incorrectly diagnosed 1.4% and left 45.1% undiagnosed. While single-timepoint GFAP has potential to provide significant improvement to prehospital diagnostics currently lacking in regular EMS units, adjunct diagnostic methods will be needed. Conflict of interest All authors: nothing to disclose
Building similarity graph...
Analyzing shared references across papers
Loading...
Olli Mattila
University of Helsinki
Saana Pihlasviita
University of Helsinki
Tiina Nukarinen
University of Helsinki
European Stroke Journal
University of Helsinki
University of Gothenburg
Helsinki University Hospital
Building similarity graph...
Analyzing shared references across papers
Loading...
Mattila et al. (Fri,) studied this question.
synapsesocial.com/papers/69fd7ee0bfa21ec5bbf0721f — DOI: https://doi.org/10.1093/esj/aakag023.955