RESUMO
BACKGROUND: Hyperglycemia is a marker of poor outcome in acute ischemic stroke (IS) patients. We aimed at evaluating the effect of combined HbA1c and first glucose measurement values on 3-month mortality prediction. METHODS: In a prospective analysis, 1,317 first-ever IS patients with HbA1c values were classified by first glycemia value (<155, 155-199, ≥200 mg/dl). Three-month mortality was analyzed by glycemia category in nondiabetics, diabetics with good previous glucose control (PGC) (HbA1c <7%), and diabetics with poor PGC (HbA1c ≥7.0%). RESULTS: Mortality at 3 months was 13.1%, with no differences (p = 0.339) between non-diabetes mellitus (DM) (12.3%), good PGC-DM (12.4%), and poor PGC-DM (15.6%) patients. The unadjusted relative risk of 3-month mortality for patients with glucose ≥200 mg/dl was 3.76 (95% CI 1.48-9.56) in non-DM, 6.10 (95% CI 1.76-21.09) in good PGC-DM, and 1.44 (95% CI 0.77-2.69) in poor PGC-DM. Glycemia cutoffs most highly correlated with mortality increased as PGC declined: 107 mg/dl in non-DM, 152 mg/dl in good PGC-DM, and 229 mg/dl in poor PGC-DM patients. Glycemia correlated with stroke severity in nondiabetics and diabetic patients with good PGC, but not in those with poor PGC. CONCLUSIONS: HbA1c determination combined with first measured glucose value is useful to stratify mortality risk in IS patients: hyperglycemia is a poor prognostic marker in non-DM and DM patients with good PGC; results are inconsistent in poor PGC-DM patients. Our data suggest the relationship between hyperglycemia and poor outcome reflects stress response rather than a deleterious effect of glucose.