PubMedDate not supplied
INTRODUCTION: Cerebral infarction (CI) can be classified aetiologically in several different ways using explicit diagnostic criteria. However, the extent to which these diagnostic criteria are actually implemented in clinical practice is unknown. Aims. The aim of this study was to analyse the management and use of diagnostic tests in the aetiological diagnosis of CI in two county hospitals and to compare this with the most common recommendations. We also sought to analyse the clinical and demographic variables that may help to explain why these guidelines are not followed. PATIENTS AND METHODS: We reviewed the discharge abstracts of 307 cases of CI attended in two county hospitals between 1999 and 2000 and we analysed the clinical data, diagnostic tests and the final diagnosis. The diagnoses were reorganised using the TOAST, Laussane, NINDS and SEN 98 classifications and we analysed the frequency with which the diagnostic tests were employed in each aetiological subtype. RESULTS: Average age: 71.3 years; 59.3% were males. CAT scans were performed in 97.1% of cases, neurosonology was used in 40.1% and echocardiography was performed in 8.5%. The aetiological diagnosis was: atherothrombotic 22.4%, cardioembolic 10.7%, lacunar 26%, unusual causes 0.3% and unknown causes 1.6%. In 37.4% of cases the diagnosis was given as unspecified CI. On reclassifying the diagnoses according to SEN 98 criteria, we obtained the following: atherothrombotic 19.5%, cardioembolic 2.8%, lacunar 13.7% and of unknown origin 63.5%. 0.6% of the cases were unclassifiable. Factors that exerted an influence on the fact that diagnostic tests were less frequently carried out included age, level of awareness and mortality. The most frequent cause of incomplete studies was the absence of carotid Doppler. CONCLUSIONS: The guidelines for aetiological diagnosis of CI are not often followed. Systematic performance of a neurosonological study would improve aetiological diagnosis of CI.