RELIABILITY OF PRACTICAL IRANIAN CRITERIA (PIC) FOR CLASSIFICATION OF BRAIN INFARCT

Background: Various classification criteria of brain infarct are used in clinical trials and stroke registries. The practical Iranian criteria (PIC) is designed for clinical practice. Methods: From March 2001 through March 2003, all consecutive stroke patients admitted to Vali-e- Asr Hospital, Birjand, Khorasan, Iran were included in this prospective observational study. Patients underwent a standard battery of diagnostic investigations by a stroke neurologist. Data on patients, demographics, clinical presentations, and diagnostic work-up were kept in a database. Two stroke neurologists and a general practitioner independently reviewed the data of 20 randomly selected patients and classified patients according to the PIC classification of stroke topography and etiology. The PIC is designed by stroke neurologists and approved in the University of Alberta, Canada. The degrees of interrater agreement were measured with unweighted k-statistics. Results: Among 302 stroke patients, 20 patients (11 females, 9 males) were randomly selected. The three interrater agreement for topographic subtyping of the patients was 0.95%, k = 0.915 (0.662 – 1), P < 0.0001 and for etiologic diagnosis was 0.90 %; k = 0.9022 (0.753 – 1), P < 0.0001. Stroke neurologists agreed in topographic diagnosis for 20 out of the 20 cases (100%; k= 1; 95% CI, 1.0 to 1.0; P < 0.0001). The general practitioner arrived at the same topographic diagnosis for 19 out of the 20 cases (0.95%; k = 0.875; 95% CI, 0.638 to 1.0; P < 0.0001). Stroke neurologists agreed in etiologic diagnosis for 18 out of the 20 cases (0.90%; k = 0.855; 95% CI, 0.66 to 1.0; P < 0.0001). The general practitioner arrived at the same etiologic diagnosis for 18 out of the 18 cases that the two stroke neurologists agreed (100%; k = 0.875; 95% CI, 1.0 to 1.0; P < 0.0001). Conclusion: The interrater agreement of PIC for topographic and etiologic classification of brain infarct is much higher than the other classification systems. The PIC could be useful in stroke practice.