Post-stroke recrudescence (PSR) usually occurs in the setting of infection, hypotension, hyponatremia, insomnia or stress, and benzodiazepine use. Animal studies have suggested an infection-related immunologic mechanism for PSR. This retrospective study was designed to assess whether infection-triggered PSR is related to a prior infection during the index stroke. We identified 95 patients admitted to Massachusetts General Hospital from 2000 to 2015 with post-stroke recrudescence who had adequate medical record information concerning the index stroke. The frequency of infections, as well as other triggers such as hypotension, hyponatremia, insomnia/stress, and benzodiazepine use, was compared between the index stroke and the PSR episode. Independent predictors of infection-related PSR were identified using a logistic regression model. The mean age was 66 ± 17 years (53% female); 29 (31%) had infections during the index stroke as compared to 40 (42%) during the PSR episode. The frequency of PSR triggered by infection was higher in patients with infections during the index stroke (65% vs 32%, p = 0.003). The same relationship occurred with benzodiazepine-triggered PSR (41% vs 12%, p = 0.008). The frequencies of other triggers such as hypotension, hyponatremia and insomnia/stress were not significantly different between the index stroke and the PSR episode. In a logistic regression model, infection during the index stroke was an independent predictor of infection-triggered PSR (odds ratio 4.85, 95% C.I. 1.7, 13.7). The association between infection during index stroke and infection-triggered PSR supports the immunologic mechanism postulated in animal models.
Keywords: Immune response; Infection; Recrudescence; Stroke.