To find out the composite risk factors associated with early neurological deterioration (END) in acute ischemic stroke, to reveal the mechanism of END, and to provide the data base for the early prediction and prevention.

Five hundred and fifty-eight patients with cerebral infarction admitted to our hospital from October 2009 to December 2012, were screened. Among them, 107 patients met the diagnostic criteria of END in acute ischemic stroke, 451 patients met the diagnostic standard of early acute ischemic stroke without END. Neurological function scale and other variables included 58 related factors of 9 categories were selected. Association rule mining methods were used to analyze relations between END in acute ischemic stroke and risk factors sets.

The results of association rule mining discovered that there were 2 individual risk factors, 3 double-factor combinations, 7 triple-factor combinations and 15 four-factor combinations related with END in the early stage of acute ischemic stroke; the more the composite factors, the higher the probability of neurological deterioration; the composite factors were mostly the combination of variables of different categories, involving neurological function scale scores, infection condition, dysphagia, personal life history (smoking and drinking), infarction location, age, and levels of electrolyte, C reactive protein, and homocysteine. Besides some independent risk factors which had been reported in the literatures, the results of this study found that heart rate and time interval from onset to hospitalization also related with END in early acute ischemic stroke.

END in acute ischemic stroke may be attributed to the combination effect of variable factors; all risk factors should be considered and a variety of targeted measures should be taken to prevent and treat the patients with END in early acute ischemic stroke.

Ischemic stroke; Early neurological deterioration; Association rules mining method; Risk factor