Assessing agreement between simultaneous BP measurements obtained from cuff non-invasive blood pressure (NIBP) and radial arterial invasive blood pressure (AIBP) devices to determine whether systolic blood pressure (SBP) was reliable in acute spontaneous intracerebral (sICH). A total of 303 (39.5%) of 766 prospectively screened sICH subjects had NIBP and AIBP measurements. The first 24 hours saw the collection of 2,157 simultaneous paired measurement readings. Paired NIBP/AIBP measurements were used in a Bland-Altman technique with 95% agreement limits and regression coefficients derived from a bootstrap procedure. SBP variation was 66.1 mmHg, greater than the 44.3 mg Hg for diastolic blood pressure (DBP) or the 46.1 mmHg for mean arterial pressure (MAP). When mean biases were compared pairwise, there was a significant difference between SBP and DBP (P<0.0001) or MAP (P<0.0001). The mean bias was not different between DBP and MAP (P=0.68). Over the range of mean SBP, regression-based Bland Altman analysis revealed significant bias (slope −0.16, 95% CI −0.23, 0.09, P<0.05). Bias over the mean DBP or MAP range was not significant. In patients with sICH, researchers concluded that SBP was an unreliable blood pressure measurement.