|| Checking for direct PDF access through Ovid
Background: The role of hyperglycemia in acute intracerebral hemorrhage (ICH) is controversial: Some studies find an association with poor outcomes; others report neutral relationship. All interventions aiming at glucose lowering in acute ICH have yielded neutral or negative results. We attempt an explanation of the causal role of hyperglycemia in ιCH outcome using path analysis.Methods: 235 consecutive primary ICH patients (69.2±13.2 years, 44% female) admitted to an academic hospital in Boston. We recorded hematoma volume (HV) on presentation, mean 72-hour glucose and age for each patient as continuous variables. Intraventricular extension (IVE), infratentorial location, coagulopathy (INR≥1.4 or platelets <100,000), hematoma expansion (HE, defined as >33% relative increase in HV) and presence of diabetes were recorded as dichotomous variables. 90-day functional outcome was dichotomized as favorable vs unfavorable (modified Rankin Scale of ≤ 2 vs >2). We used generalized structural equation modelling for model estimation.Results: Mean 72-hour glucose was 143± 35 mg/dl, median HV was 23.3 cm3. 66 (28%) were diabetic, 103 (44%) had IVE, 56 (26%) had HE, 29 hemorrhages (12%) were infratentorial. 60 (26%) patients had favorable functional outcome. HV (p=0.001), diabetes (p<0.0001), IVE (p=0.02) but not HE were significant predictors of hyperglycemia. HV and IVE were associated with unfavorable outcome but the effect was not mediated by elevated glucose levels. On bivariate analysis, higher glucose levels were negatively correlated with favorable outcome (p=0.002, 95% CI 0.97-.099) but this effect lost its significance when introducing the rest of the variables in our model (figure).Conclusion: Elevated glucose in ICH appears to be strongly influenced by markers of ICH severity; this is a possible explanation for the lack of response to intensive glucose lowering interventions. Its direct effect on functional outcome in ICH merits further investigation.