Abstract 3051: Blood Pressure Thresholds to Predict the Cause of Intracerebral Hemorrhage
Introduction: Chronic hypertension is a common cause of spontaneous intracerebral hemorrhage (ICH), but not all patients who are hypertensive on hospital presentation have an ICH caused by hypertension (htnICH). We sought to determine blood pressure (BP) thresholds that correlated with a presumed htnICH in a prospective cohort.
Methods: The NIH-funded Diagnostic Utility of MRI in Spontaneous Intracerebral Hemorrhage (DASH) study prospectively enrolled consecutive ICH patients to determine the utility of routine MRI in the diagnosis and management of these patients. Contrast angiography was pursued in a predefined patient subset. At 3 months, ICH cause was determined by the treating stroke physician after review of all clinical information, including MRI in the acute and chronic phase, pathology, and clinic follow-up, as available. Statistical analyses were done using SPSS: χ2; 2 tailed t-tests; and Mann-Whitney U tests were used as appropriate. Receiver operator characteristic (ROC) curves were created and results expressed as area under curve (AUC).
Results: We included 136 patients in this report (age: 63±17yrs; ICH volume: 22±27cc; NIHSS: 9±8; GCS: 13±3). Of these, 70% had a history of hypertension, 40% had an admission SBP> 180mmHg, and 22% an admission SBP >200mmHg. Sixty patients (44%) had htnICH as their final diagnosis. A history of hypertension was associated with htnICH (χ2=11.8, p<0.001), but 48% (46/95) of patients with a history of hypertension did not have a htnICH. Patients with a htnICH had: higher SBP (189 vs 157mmHg, p<0.0001); higher MAP (131 vs 110mmHg, p<0.0001), higher NIHSS (12 vs 6.5, p<0.0001); smaller ICH volumes (16.6 vs 26.0cc, p=0.03); and non-lobar hematomas (χ2=62.3, p180mmHg was 74% specific and 58% sensitive for a subsequent diagnosis of htnICH, a SBP> 200mmHg was 90% specific and 37% sensitive, and a MAP >132mmHg was 90% specific and 45% sensitive. Using ROC analysis, MAP predicted hypertensive etiology with an AUC of 0.75 (p<0.0001, 95% CI: 0.67-0.83). Patients with htnICH and MAP 132mmHg based on age, NIHSS, gender, ICH volume, or ICH location (lobar vs non-lobar). Conversely, patients with non-htnICH and MAP >132mmHg on admission tended to have larger ICH volumes (44 vs 24cc, p=0.07) and higher NIHSS (10 vs 6, p=0.21), but did not differ by location from those with a MAP<132mmHg. Non-lobar location alone was 74% specific and 93% sensitive for htnICH, and a non-lobar ICH or a MAP>132mmHg was 64% specific and 95% sensitive for htnICH.
Conclusions: In our cohort, a history of hypertension predicted htnICH no better than chance. An admission SBP >200mmHg or a MAP >132mmHg predicted htnICH in 90% of these patients, but missed two-thirds of cases. ICH location alone was a sensitive predictor of htnICH, but blood pressure thresholds were more specific.
- © 2012 by American Heart Association, Inc.