Accuracy of totally implanted ports, tunnelled, single- and multiple-lumencentral venous catheters for measurement of central venous pressure

Objective: To verify the accuracy of totally implanted ports, tunnelled cen tral venous catheters (CVC), widely used in cancer patients, and multi-lume n catheters, used in intensive care units (ICUs), in measuring central veno us pressure (CVP), using right atrial pressure (RAP) measured in a Swan-Gan z catheter as the reference standard. Design: A prospective study, over a 10-month period. Setting: A medical-sur gical ICU in a comprehensive cancer centre. Patients and participants: Patients who had both (1) a Swan-Ganz catheter a nd (2) either a tunnelled catheter, a single or a multi-lumen catheter, or a totally implanted port. Interventions: RAP and CVP were measured simultaneously in each patient. Measurements and results: Fifty-six pairs of RAP-CVP measurements were perf ormed in 35 patients: 6 tunnelled catheters, 6 non-tunnelled single-lumen c atheters, 26 multiple-lumen catheters and 18 totally implanted ports were s tudied. RAP measured in the Swan-Ganz catheter and CVP measured in the CVC were strongly correlated (r = 0.94, p < 0.01), whatever the type of cathete r studied. The mean difference between RAP and CVP was -0.39 +/- 1.73 (SD) mmHg. In 51 cases (91 %), the difference was within the limits of agreement (-3.78 to 3.00 mmHg. Bland and Altman method). For the five cases with a d ifference of 4 mmHg (three totally implanted ports, one double- and one tri ple-lumen catheter), CVP was greater than RAP. Conclusions: CVP can be accurately measured in totally implanted ports, tun nelled or non-tunnelled single-lumen and multiple-lumen catheters. When the difference exceeds the limit of agreement, the discrepancy between the two measurements has limited significance in most cases.