Delayed Cerebral Ischemia After Subarachnoid Hemorrhage

Abstract

Background and Purpose— Established predictors of delayed cerebral ischemia after aneurysmal subarachnoid hemorrhage are large amounts of extravasated blood and poor clinical condition on admission. The predictive value of other factors is uncertain. Methods— We searched MEDLINE (1960–2012) for clinical, laboratory, and radiological predictors routinely available within 72 hours after subarachnoid hemorrhage. The studies were categorized according to methodological quality. Crude data and effect estimates (odds ratio [OR], hazard ratios, and risk ratio) with 95% CI were extracted, (re-)calculated and pooled if possible. For every potential predictor we assessed all effect estimates on consistency (point estimates in equal direction) and clinical relevance (size and 95% CI). Results— Fifty-two studies on 33 potential predictors were included. There was strong evidence (≥3 high-quality studies) for a higher risk of delayed cerebral ischemia in smokers (pooled OR, 1.2; 95% CI, 1.1–1.4), and moderate evidence (2 high-quality studies) for an increased risk in patients with hyperglycemia (OR, 3.2; 1.8–5.8 and hazard ratios, 1.7; 1.1–2.5), hydrocephalus (OR, 1.3; 1.1–1.5 and OR, 2.6; 1.2–5.5), history of diabetes mellitus (pooled OR, 6.7; 1.7–26), and early systemic inflammatory response syndrome (pooled OR, 2.1; 1.4–3.3). Evidence was limited for increased risk in women (pooled OR, 1.3; 1.1–1.6) and in patients with history of hypertension (pooled OR, 1.5; 1.3–1.7). The evidence on initial loss of consciousness, history of migraine, previous use of selective serotonin reuptake inhibitors, hypomagnesemia, low hemoglobin, or high blood flow on early transcranial Doppler was also limited. Conclusions— There is strong evidence that smoking is a predictor of delayed cerebral ischemia. For several other potential predictions the evidence is moderate, limited, or inconsistent.

Early Prediction of Delayed Cerebral Ischemia After Subarachnoid Hemorrhage

“…In Model II, candidate predictors were added of strong or moderate evidence based on a previous systematic review (current smoking, admission glucose level, hydrocephalus on initial CT). 5 In Model III, candidate predictors with limited evidence in our systematic review were added (history of hypertension, initial loss of consciousness, and hemoglobin). 5 Restricted cubic spline functions and graphs were used to determine whether continuous variables (age, glucose, and hemoglobin) could be analyzed as linear terms or required transformation.…”

“…5 In Model III, candidate predictors with limited evidence in our systematic review were added (history of hypertension, initial loss of consciousness, and hemoglobin). 5 Restricted cubic spline functions and graphs were used to determine whether continuous variables (age, glucose, and hemoglobin) could be analyzed as linear terms or required transformation. 15 Missing values of patient characteristics were imputed by means of regression imputation.…”

“…We used predictors that are well known from the literature. 5 Finally, we developed a simple practical risk chart that can easily be used on admission to predict absolute risks of DCI in individual patients with SAH. Before application of the risk chart in clinical practice, it should be externally validated in another hospital setting, preferably in another healthcare system.…”

“…1 Established predictors of DCI are large amounts of subarachnoid blood and poor clinical condition on admission; [2][3][4] other potential predictors are current smoking, diabetes mellitus, presence of hyperglycemia on admission, hydrocephalus on admission, or early systemic inflammatory response syndrome. 5 Previous prediction models for DCI after SAH are impractical in daily practice or not confined to data present at admission, and most are not validated. [6][7][8] Furthermore, a risk prediction chart does not exist.…”

Inflammation, negative nitrogen balance, and outcome after aneurysmal subarachnoid hemorrhage

“…We hypothesized that the metabolic response and nutritional status after SAH would be dependent on the inflammatory status, and our decision to allocate measurements into phases by PBD was based on previous reports of inflammatory response after SAH. 17,38 While this may have introduced measurement bias into our results, we believe our overall method of repeating Receiver operating characteristic curve for C-reactive protein and transthyretin predicting poor outcome at 3 months…”