Poor diagnostic accuracy and applicability of Siriraj stroke score, Allen score and their combination in differentiating acute haemorrhagic and thrombotic stroke.

OBJECTIVE: To evaluate the relevance of bed side clinical diagnostic scoring systems--Siriraj stroke score (SSS), Allen score and their combined use for differentiating acute haemorrhagic and thrombotic stroke. MATERIAL AND METHODS: The study was conducted on 240 admitted patients of stroke over a period of two years. SSS was calculated immediately and Allen score, 24 hours after admission. CT scan was done immediately and 48 hours after admission if required. The sensitivity, specificity, positive predictive value, negative predictive value and diagnostic gain were calculated for both the scores. Comparability between the scores and CT scan finding was determined with the help of kappa statistic programme. Receiver operating characteristic curves (ROC) were plotted to assess the diagnostic accuracy of both scores over a range of cut-off points. RESULTS: One hundred and thirty four patients (55.83%) had infarction and 106 patients (44.17%) had haemorrhage. SSS was applicable in 66.25% patients (159 out of 240) while Allen score was applicable in only 61.25% patients (147 our of 240). The sensitivity, specificity, positive predictive value, negative predictive value and diagnostic gain for SSS were 73%, 85%, 85%, 71% and 30% for infarction and 85%, 73%, 71%, 85% and 27% for haemorrhage respectively, whereas the corresponding figures for Allen score were 91%, 60%, 77%, 82% and 18% for infarction and 60%, 91%, 82%, 77% and 41% for haemorrhage respectively. There was overall moderate comparability between SSS and Allen score for diagnosing supratentorial stroke (k = 0.396). The comparability of these scores in terms of certain results was worse (k = 0.143). However when the results that were within the diagnostic range with both the scores were considered, the agreement in diagnosing infarction and haemorrhage was almost perfect (k = 0.874). While considering CT scan finding as gold standard for differentiation of infarction and haemorrhage, the overall accuracy of SSS and Allen score was seventy eight percent. CONCLUSION: (a) Applicability of SSS only in 66.25% patients and wrong diagnosis in 22.01% patients does not reflect its usefulness because adequate management of stroke requires a gold standard diagnosis which is only possible by immediate CT scan. (b) Allen score is not useful because it can be assessed only after 24 hours of onset of stroke. This deprives the management to all thrombotic patients in speculated time window of modern management.