• 
• 

Three Part Question

In [adults with suspected stroke], can [blood biomarkers] [accurately diagnose ischemic stroke rapidly]?

Clinical Scenario

A 63-year-old gentleman with a history of hypertension comes to the Emergency Department complaining of sudden onset of right-sided weakness and slurring of speech 2 hours ago. You request an immediate non-contrast CT brain. However, you wonder if blood biomarkers could be used instead to diagnose a stroke, or more accurately, an ischaemic stroke.

Search Strategy

CCTR, CDSR, EMBASE, Ovid MEDLINE® 2011 – 2016 were searched with the following search criteria:
(biomarker[Title] OR marker[Title]) AND (cerebrovascular[Title] OR ischaemic stroke[Title] OR ischemic stroke[Title])) AND (("diagnosis"[Subheading] OR "diagnosis"[All Fields] OR "diagnosis"[MeSH Terms]) OR "sensitivity and specificity"[All Fields] OR ("probability"[MeSH Terms] OR "probability"[All Fields]))) NOT fibrillation[All Fields]) NOT (("infant, newborn"[MeSH Terms] OR ("infant"[All Fields] AND "newborn"[All Fields]) OR "newborn infant"[All Fields] OR "neonatal"[All Fields]) OR ("pediatrics"[MeSH Terms] OR "pediatrics"[All Fields] OR "pediatric"[All Fields]) OR ("pediatrics"[MeSH Terms] OR "pediatrics"[All Fields] OR "paediatric"[All Fields]) OR ("infant"[MeSH Terms] OR "infant"[All Fields] OR "infants"[All Fields]))) NOT FLAIR[All Fields] AND (("2011/01/01"[PDAT] : "2016/12/31"[PDAT]) AND "humans"[MeSH Terms] AND English[lang])

Search Outcome

Of 72 papers found, 21 duplicates were removed and 9 papers were chosen by relevant title and abstract.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Li et al 2016, China	206 patients admitted with first-ever AIS defined according to WHO criteria with symptoms within 72h had their blood collected at 6AM after admission to measure interleukin-33 (IL-33) levels.	Prospective cohort study	Difference in AIS serum IL-33 compared to control: 57.68 vs 45.00 ng/L (p<0.001)	Sensitivity, specificity and AUC not included	Circulating IL-33 levels were only measured once i.e. no serial measurements CSF IL-33 levels not measured
			Optimal cutoff value of serum IL-33 levels as an index to diagnose AIS: 47.92 ng/L	Sensitivity 69.9% Specificity 44.4% AUC 0.706
Ormstad et al, 2016, Norway	45 patients admitted with first-ever AIS defined according to ICD-10 had their blood collected at different time points to measure serum phenylalanine (PHE) and tyrosine (TYR) levels.	Prospective cohort study	Difference in AIS serum TYR compared to control: 74.3 vs 83.8 μm (p=0.035)	Sensitivity 76% Specificity not included AUC 0.82	Sample size calculation not included Small sample size with significant inclusion bias Did not include hemorrhage or stroke-mimics CSF TYR and PHE levels not measured
			Difference in AIS serum PHE compared with control: 124 vs 101 μm (p<0.001)	Sensitivity 85% Specificity N/A AUC 0.88
Walsh et al, 2016, USA	14 patients admitted with AIS diagnosed clinically and by neuroimaging had blood collected within 12h of symptom onset to measure apolipoprotein (Apo) A-I and paraoxonase-1 (Po-1) levels.	Prospective cohort study	Difference in AIS serum Apo A-I compared with control: 140 vs 175 mg/dL (no p-value)	Sensitivity, specificity and AUC not included	Small sample size Only recruited patients at a single site Did not draw samples within an even shorter time range (i.e. within 12h vs within 3h of symptom onset) Serial measurements were not taken CSF Apo A-I and Po-1 levels not measured
			Difference in AIS serum Po-1 compared with control: 250,500 vs 345,500 mg/dL (no p-value)	Sensitivity, specificity and AUC not included
Ranga et al, 2016, India	60 male patients admitted with AIS diagnosed by CT brain or MRI brain had blood collected within 24h of stroke onset to measure carcinoembryonic antigen (CEA).	Cross-sectional study	Difference in AIS serum CEA compared with control: 5.26 vs 1.38 ng/ml (p=0.001)	Sensitivity, specificity and AUC not included	Small sample size Only men were included Inability to determine presence of a causal relationship between rise in CEA and AIS due to the nature of the study CSF CEA levels not measured
Qi et al, 2015, China	312 patients with first-ever AIS defined according to WHO criteria and with symptom onset within 24h had blood collected a day after admission to measure serum thioredoxin levels.	Prospective cohort study	Difference in median AIS serum TRX when compared with controls: 15.03 vs 8.95 ng/ml (p<0.0001)	Sensitivity 80.3% Specificity 73.7% AUC 0.87	Immunoglobulins in serum might have affected results of immunoassays by binding to reagent antibodies CSF TRX levels not measured Serial measurements of TRX levels not done TRX measurements may not accurately reflect pre-stroke exposure
Peng et al, 2015, China	72 patients admitted with clinical history and MRI results supporting diagnosis of first-ever AIS had blood and CSF samples collected at different time points after stroke onset to measure let-7e levels.	Prospective cohort study	Difference in AIS serum let-7e compared with controls 1.5 vs 0.88 (p=0.0001)	Sensitivity and specificity not included AUC 0.86	Sample size calculation not included Small sample size Let-7e target proteins not measured Did not match cardiovascular comorbidities of the patients Method used required time
Tian et al, 2015 China	442 patients with first-ever AIS defined according to WHO criteria and with symptom onset within 24h had fasting venous blood collected on day of admission to measure serum levels of procalcitonin (PCT).	Prospective cohort study	Difference in AIS serum PCT when compared with controls: 1.04 vs 0.25 ng/ml (p<0.0001)	Sensitivity 79.6% Specificity 72.1% AUC 0.801	Did not record data on when and for how long PCT levels were elevated No follow-up data collected CSF PCT levels in CSF not measured
Long et al 2013, China	197 patients with first-ever AIS defined by ICD-9 had blood samples collected at different time points after stroke onset to measure levels of miR-30a, miR-126 and let-7b.	Cross-sectional study	Difference in large-vessel atherosclerosis AIS plasma let-7b when compared with controls: 70-75% (p<0.05)	Sensitivity 79.6% Specificity 72.1% AUC 0.801	Sample size calculation not included Small sample size Reasons behind varying expressions of let-7b among various types of AIS were not investigated CSF miRNAs levels not measured
			Difference in other types of AIS plasma let-7b when compared with controls: 3.51–14.42 fold increase (p<0.05)	Sensitivity 92% Specificity 84% AUC 0.93
Steigleder et al, 2012, Germany	42 patients admitted with AIS defined clinically had blood collected within 6h after symptom onset and different time points afterwards to measure glycogenphosphorylase isoenzyme BB (GPBB).	Prospective cohort study	Difference in AIS plasma GPBB when compared with controls: 68.040.5 vs <10 ng/ml	Sensitivity, specificity and AUC not included	Conference paper Sample size calculation not included Small sample size CSF GPBB levels not measured

Comment(s)

This BestBET review updates previous work by Jie (2011) to explore the diagnostic possibility of any new-found biomarkers for ischaemic stroke.

Jie commented that biomarkers for AIS should be “brain specific, plasmatic (not CSF), early expressed and detected after symptom onset”. This would suggest that while using these biomarkers, the focus should be on shortening the time (possibly using point-of-care testing) required for accurate diagnosis and improving cost effectiveness.

Despite some relatively promising sensitivities, specificities and AUCs observed in the 9 studies reviewed here, none of the reported biomarkers are reliable enough as a single diagnostic tool due to the small sample sizes and lack of power. In future, adequately powered investigations are needed to validate the accuracy of these and other biomarkers.

Editor Comment

RB

Clinical Bottom Line

Blood biomarkers are not currently suitable for the rapid accurate diagnosis of AIS.

References

1. Jie K.E. Diagnostic use of blood biomarkers for discerning ischemic from hemorrhagic stroke. [Online] 2011 BestBET database http://www.bestbets.org/bets/bet.php?id=2251, accessed 4 October 2017.
2. Li, Q.,Lin, Y., Huang, W., Zhou, Y et al. Serum IL-33 is a novel diagnostic and prognostic biomarker in acute ischemic stroke. Aging and Disease 2016 :7(5), pp.614-622.
3. Ormstad, H., Verkerk, R. and Sandvik, L. Serum phenylalanine, tyrosine, and their Ratio in acute ischemic stroke: on the trail of a biomarker? Journal of Molecular Neuroscience 2016:58(1), pp.102-108.
4. Walsh, K.B., Hart, K., Roll, S et al. Apolipoprotein A-I and paraoxonase-1 are potential blood biomarkers for ischemic stroke diagnosis. Journal of Stroke and Cerebrovascular Diseases, 25(6), pp.1360-1365. 2016: 25(6), pp.1360-1365.
5. Ranga, G.S., Bansal, N., Sharma, S.B et al. Human carcinoembryonic antigen is a useful biomarker for diagnosis of acute ischemic stroke – a pilot study. Neurology Asia 2016: 21(1), pp.1-6.
6. Qi, A.Q., Li, Y., Liu, Q et al. Thioredoxin is a novel diagnostic and prognostic marker in patients with ischemic stroke. Free Radical Biology and Medicine 2015: 80, pp.129-135.
7. Peng, G., Yuan, Y., Wu, S et al. MicroRNA let-7e is a potential circulating biomarker of acute stage ischemic stroke. Translational stroke research 2015: 6(6), pp.437-445.
8. Tian, D., Zhang, S., He, X et al. Serum procalcitonin as a diagnostic marker in acute ischemic stroke. Neuroreport 2015: 26(1), pp.33-37.
9. Long, G., Wang, F., Li, H et al. Circulating miR-30a, miR-126 and let-7b as biomarker for ischemic stroke in humans. BMC Neurology 2013:13(1), p.1.
10. Steigleder, T., Kugler, J., Hoffman, V et al. Glycogenphosphorylase isoenzyme BB (GPBB) – novel marker of ischemic lesion of the brain. Cerebrovascular Diseases. Conference: 21st European Stroke Conference 2012