Complications after acute stroke.

Abstract

BACKGROUND AND PURPOSE: We sought to observe the type, timing, and frequency of complications occurring in hospitalized patients after an acute stroke. METHODS: In a single hospital, we prospectively identified a consecutive cohort of patients who were either admitted after an acute stroke or who suffered a stroke while already an inpatient (n=613). We retrieved the case notes for 607 (99%) of these strokes, and a single observer, using predefined diagnostic criteria, reviewed the notes and recorded the type, timing, and frequency of complications that occurred during the inpatient period. We also measured the reliability of complication identification from case note review by comparing two observers on a sample of records. RESULTS: Complications were recorded after 360 strokes (59%); the most common individual complications were falls (complicating 22% of all strokes), skin breaks (18%), and urinary tract (16%) or chest (12%) infections. Miscellaneous "other" complications complicated 32% of strokes. Seizures and chest infections occurred early, whereas depression and painful shoulder were later problems. Complications were more common in older patients, who were more disabled before their stroke and had suffered more severe strokes. We demonstrated moderate to good agreement between the two observers for most complications. CONCLUSIONS: Complications after acute stroke are common, confirming that stroke rehabilitation requires active and knowledgeable medical input. Knowing the nature and timing of complications, together with the identification of high-risk patients, may be useful to those planning stroke services. The differences in our results and those previously reported, most notably for skin breaks, are probably due to the different methods used, in particular patient selection and diagnostic criteria for complications. Although complications may be useful as a measure of outcome in comparative studies (eg, therapeutic trials and audit), the methodological difficulties in accurately and reliably measuring them must be addressed.