The aim of this study was to evaluate the accuracy and precision of non-invasive continuous blood pressure measurement by applanation tonometry (AT) in awake or anaesthetised cardiological intensive care patients. Patients suffering from highly impaired left ventricular function atrial fibrillation or severe aortic valve stenosis were included into the study. Arterial blood pressure was recorded by applanation tonometry (T-Line 400, Tensys Medical®, USA) and an arterial line in awake or anaesthetised patients. Discrepancies in mean (MAP), systolic (SAP), and diastolic (DAP) arterial pressure between the two methods were assessed as bias, limits of agreement and percentage error respectively. In 31 patients a total of 27,900 measurements were analyzed. The concordance correlation coefficient was 0.23, 0.45 and 0.06 for MAP, SAP and DAP, respectively. For all patients bias for MAPAT compared to MAPAL was 14.96 mmHg (SAPAT 4.51 mmHg; DAPAT 19.12 mmHg) with limits of agreement for MAPAT of 46.25 and - 16.33 mm Hg (SAPAT 48.00 and - 38.98 mmHg; DAPAT 50.12 and - 11.89 mmHg). Percentage error for MAPAT was 56.8% (42.7% for SAPAT; 75.2% for DAPAT). We conclude that the AT method is not reliable in ICU patients with severe cardiac comorbidities.