Background: Arterial pressure monitoring using the a continuous noninvasive arterial pressure (CNAP) device during general anaesthesia is known to be interchangeable with continuous invasive arterial pressure (CIAP) monitoring. Agreement with invasive measurements in cardiovascular postsurgical intensive care patients has not been assessed.
Objective: The objective of this study is to assess the agreement and interchangeability of CNAP with CIAP in cardiovascular postsurgical patients and to determine the effects of cardiac arrhythmia, catecholamine dosage, respiratory weaning and calibration intervals on agreement.
Design: A prospective observational study.
Setting: German university hospital cardiovascular ICU. Data were collected from April 2010 to December 2011.
Patients: From 110 enrolled patients, 104 were included. Inclusion criteria were American Society of Anaesthesiologists (ASA) physical status III or IV patients undergoing controlled ventilation. Exclusion criteria included emergencies, complete heart block and marked arterial pressure differences greater than 10 mmHg in the two arms.
Main outcome measures: Bland-Altman plots, bias, precision, 95% limits of agreement, percentage error and agreement : tolerability indexes (ATIs) were estimated to determine clinical agreement.
Results: From 11 222 arterial pressure readings, biases (SD) for CIAP-CNAP for systolic arterial pressure (SAP), diastolic arterial pressure (DAP) and mean arterial pressure (MAP) for all patients were 4.3 (11.6), -9.4 (8) and -6 (7.6) mmHg, respectively. Cardiac arrhythmia (4.1 (13.1), -14.4 (8.3), -9.5 (8.9) mmHg) and long interval to last calibration [4.5 (15), -9.8 (9.5), -6.4 (9.1) mmHg] impaired the accuracy of CNAP with failed interchangeability criteria defined by the percentage error. In contrast, use of catecholamines (epinephrine or norepinephrine infusions >0.1 μg kg min), short calibration intervals and weaning conditions did not affect accuracy, interchangeability and agreement, especially of MAP. Agreement was defined as acceptable for MAP for all data and subgroups (ATI 0.8 to 1.0) and at worst, marginal for SAP and DAP (ATI 0.9 to 1.6).
Conclusion: CNAP showed acceptable agreement defined by the ATI with invasive measurements for MAP and partially for DAP, but there was considerable variability for SAP. MAP should be preferred for clinical decision making. Cardiac arrhythmia, in contrast to catecholamine dosage or weaning procedures, impaired the accuracy, agreement and interchangeability of CNAP.
Trial registration: Clinical trials.gov identifier NCT01003665.