Elena Zapata-Arriaza, MD

Boehme AK, Kulick ER, Canning M, Alvord T, Khaksari B, Omran S, et al. Infections Increase the Risk of 30-Day Readmissions Among Stroke Survivors: Analysis of the National Readmission Database. Stroke. 2018;49:2999–3005.

Infection is a most frequent medical complication after stroke. It is well recognized that acute stroke induces both the central nervous system and peripheral inflammatory responses, and, infection during stroke admission increases acute and longer-term inflammatory responses, complicating stroke outcomes. However, up to now, clinical trials with prophylactic aim in infection after stroke have failed. Therefore, stroke and infection develop a negative feedback between them.

Based on the known association between infection and stroke outcomes, the authors performed a weighted analysis of the federally managed 2013 National Readmission Database to assess the relationship between infection during a stroke hospitalization and 30-day readmission (30dRA) among ischemic stroke survivors. The main goal of this paper was to confirm the relation between stroke infection and 30dRA. The novelty of the paper is the identification of stroke associated infection as a predictor of readmission, not only as the most frequent cause of a new hospitalization. The authors employed the International Classification of Disease Ninth Revision [ICD-9] codes to identify ischemic stroke patients with ICD-9 codes present in the first diagnostic position ischemic stroke (referred as primary ischemic stroke patients) and patients with ischemic stroke code at any diagnostic position (referred as all ischemic stroke patients). The primary outcome, 30dRA, was classified as any hospitalization occurring within the 30-day postdischarge window and classified into planned or unplanned readmissions using previously validated ICD-9 codes. Secondary outcomes, 7dRA and 60dRA, were also assessed and were classified as any hospitalization occurring within the 7 days, or 60 days, postdischarge window.

Of the primary ischemic stroke admissions (319 317), 12.1% were readmitted within 30 days. Approximately 29% of the patients with a primary ischemic stroke code had an infection during their stroke stay, and 32% of patients with a stroke code in any position had an infection during their stroke stay. Primary ischemic stroke code patients who had an infection during their stroke stay had a higher odds of readmission (OR, 1.58; 95% CI, 1.53–1.64), and this relationship remained after adjusting for confounders. Urinary tract infection (UTI) was associated with 30dRA (OR, 1.10; 95% CI, 1.04–1.16) and with unplanned 30dRA (OR, 1.12; 95% CI, 1.06–1.18) in adjusted models among primary ischemic stroke patients. In patients with a stroke diagnosis in any diagnostic position, pneumonia (OR, 1.08; 95% CI, 1.01–1.17) and UTI (OR, 1.11; 95% CI, 1.06–1.17) increased the odds of readmission even after adjusting for confounders. All this information is summarized in Table 3. Among the primary ischemic stroke patients and all ischemic stroke patients, infection was associated with 7dRA. The relationship between infection during a stroke stay in primary ischemic stroke patients and all ischemic patients and readmission was stronger when the readmission period was extended to 60 days. The leading cause of 30dRA was acute cerebral infarction, which was the primary diagnostic code for 20.8% of the readmissions, and infections were the second leading cause of readmissions, accounting for 15.2% of readmissions.

Unadjusted and Adjusted ORs for Readmission by All-Cause Infection and by Infection Subtype

The main performance of this paper is that experiencing an infection during a stroke admission increases the risk of 30dRAs, even after accounting for disease severity and other predictors of readmission, and these associations remain regardless of a 7, 30, or 60 days readmission. This paper supports previous reported evidence, in which a stroke favors an inflammatory reaction that triggers the appearance of infection. Such infection increases mentioned inflammatory reaction, causing a negative feedback on stroke outcomes. There is previous evidence, which has reported post-stroke infection as a marker of severity, rather than a consequence of it. The infection in general and the urinary one in particular, are related to hospital readmission within the first thirty days after stroke onset. This leads to higher morbidity and worsening prognosis, which is related to a health spending increment. Therefore, infection is a life-threatening complication during hospital admission, but its influence goes beyond discharge. Does this mean that we should change post-stroke infections assessment during admission? Should we control the inflammatory cascade after cerebral ischemic injury in order to avoid the unchained consequences of it? Probably, this manuscript, provides more indirect evidence about immune cascade importance after stroke in patients’ prognosis.