Skip to main content

Table 3 Performance at different sampling times for studies with multiple sampling intervals (Area under receiver operator curve)†

From: Blood biomarkers for mild traumatic brain injury: a selective review of unresolved issues

Biomarker Study Year Use 0-1 0-6 6-24 24-48 48-144 >144  
GFAP Asken [79] 2018 concussion   0.63 0.71     
  McCrea [31] 2020 concussion   0.68   0.57    
  Meier [80] 2017 concussion   0.53   0.56    
  Papa [17] 2019 concussion 0.76 0.76 0.8 0.87 0.89   
  Posti [81] 2017 concussion    0.63 0.37    
  Yue [82] 2019 CT-/MRI+   0.72 0.85     
  Welch [65] 2017 CT+/CT-   0.84 0.94     
  McDonald [68] 2021 concussion     0.60 0.52 0.54  
  Giza [83] 2021 concussion   0.75   0.63    
NF-L McCrea [31] 2020 concussion   0.56   0.49    
  Shahim [84] 2017 PCS   0.82   0.73 0.83 0.66  
  Shahim [72] 2018 RTS   0.82   0.72 0.73 0.73  
  McDonald [68] 2021 concussion     0.73 0.85 0.79  
  Giza [83] 2021 concussion   0.75   0.63    
S100B Asken [79] 2018 concussion   0.75 0.69     
  Meier [80] 2017 concussion   0.72   0.58    
  Welch [65] 2017 CT+/CT-   0.78 0.75     
  Shahim [85] 2014 PCS   0.68    0.55   
  Rogatzki [86] 2021 concussion 0.48 0.24      
tau-T Asken [79] 2018 concussion   0.74 0.62     
  McCrea [31] 2020 concussion   0.55   0.31    
  Shahim [85] 2014 PCS   0.90    0.76   
  Gill [87] 2017 RTS   0.80    0.74   
  Shahim [72] 2018 RTS   0.67   0.57 0.52   
  McDonald [68] 2021 concussion     0.51 0.55 0.72  
  Giza [83] 2021 concussion   0.50   0.46    
UCH-L1 Asken [79] 2018 concussion   0.62 0.60     
  McCrea [31] 2020 concussion   0.66     0.51  
  Meier [80] 2017 concussion   0.74     0.55  
  Papa [17] 2019 concussion 0.65 0.65 0.77 0.74 0.50   
  Posti [81] 2017 concussion   0.52 0.46     
  McDonald [68] 2021 concussion     0.61 0.56 0.57  
  Giza [83] 2021 concussion   0.73   0.62    
  1. Use: concussion versus controls, mTBI with CT+ versus CT-, delayed return to sport versus no delay, CT- and MRI+ versus CT- and MRI-, post-concussion syndrome versus no post-concussion syndrome.
  2. Time intervals are in hours and are approximate across studies.
  3. Bolded AUROC is the highest row value.
  4. Values are rounded to 2 decimals. No tests of significance were applied.