BACKGROUND: Blood pressure (BP) varies considerably during general anesthesia. Accurate BP measurement is critical for appropriate treatment, especially during hypotension and hypertension. Here we evaluated whether the noninvasive oscillometric BP measurement technique accurately reflects BP measured by the direct intraarterial technique. METHODS: A total of 256 samples were extracted from 10 patients operated on under general anesthesia. Systolic, diastolic and mean BP were analyzed according to the level of BP; hypotension, normotension, and hypertension. Repeatability of the noninvasive BP measurement were analyzed with repeatability coefficients and percent errors. Agreement between the two BP measurements were analyzed with a Bland-Altman and Modified Bland-Altman analysis. RESULTS: The repeatability coefficient for mean BP of the noninvasive oscillometric BP measurement was 6.34. Percent errors of mean BP were smaller than those of systolic and diastolic BP. All biases were less than 5 mmHg at all BP levels. Most BP agreements were larger than 8 mmHg except all-range mean BP (7.7), hypotensive diastolic BP (6.7), and mean BP (6.2). During hypotension, only mean BP was within the 95% of confidence interval (CI) of bias and limit of agreement. During hypertension, only systolic BP could meet the criteria. CONCLUSIONS: Only mean BP during hypotension and systolic BP during hypertension accurately reflect the direct intraarterial BP. But, diastolic BP does not reflect the direct intraarterial BP over all BPs.