Methods
This prospective single-centre study was performed in the Department of Anaesthesiology and Intensive Care Medicine between March 2006 and April 2008. The local ethics committee approved the study protocol.
To increase the clinical relevance, the study included patients receiving ventilatory support [long-term ventilation at the intensive care unit (ICU; 53.1%), CVC insertion as part of the preoperative management after induction of anaesthesia (32.7%), tracheal intubation in the emergency admission unit with subsequent CVC placement (9.2%)] and patients with spontaneous respiration (5%).
In total, 260 patients were enrolled in the study, of which 10 had two CVC insertions performed. One hundred and eleven patients (42.7%) were female and 149 (57.3%) were male. Baseline patient characteristics of the study population are summarized in Table 1.
In general, we aimed to achieve a balanced use of the different venous access points.
Even though the decision about the CVC insertion site was always taken on clinical grounds, if all large central veins were available, the aim to have a similar number of access procedures for each site would have influenced the final choice in a non-randomized fashion. Most of the time, however, clinical necessity (change of site when changing catheter, SC vein preferred access point in the ICU, unsuccessful venous puncture attempts, and predefined venous access points for certain clinical or surgical settings) or anatomical circumstances (thrombosed vein, pneumothorax, damaged vessels, anatomical anomalies, and body habitus) dictated which side and vein were chosen. Nevertheless, the distribution of access points in this study was balanced between the central veins on both sides, albeit there was a slight preference for the SC vein (Fig. 1).
(Enlarge Image)
Figure 1.
Distribution of access points (LIJV, left internal jugular vein; LSV, left subclavian vein; LINV, left innominate vein; RIJV, right internal jugular vein; RSV, right subclavian vein; RINV, right innominate vein).
Interoperator variability was eliminated as all catheters were inserted by the same anaesthetist following a standardized sequence of steps using anatomical landmarks and a sterile Seldinger technique. Two-dimensional imaging ultrasound guidance was only used in difficult cases and not real time. A detailed account of the insertion technique can be found in the Supplementary Appendix.
After insertion of each catheter, the correct position and depth of insertion were estimated by two different ECG-based methods using the Certodyn® universal adaptor (Fa. B. Braun Melsungen AG, Germany). The first measurement used the Seldinger wire as intra-atrial ECG lead. For the second, the guide wire was removed and replaced by a 10% saline lock as a conducting medium using the Alphacard-System® (B. Braun Melsungen AG). The line tip was placed at the point of maximum P-wave amplitude for each measurement and the depth measured. At this position, the line tip lies in close proximity to the crista terminalis, at the junction between the SVC and right atrium. The saline-based reading was considered to be optimal and the line secured.
In 23 CVC insertions (8.5%), TOE was performed for diagnostic indications. In these cases, it was possible to assess the catheter tip position sonographically. Providing the catheter tip was within 0.5 cm above or below the base of the crista terminalis, the position was considered to be accurate.
If the patient had a diagnostic chest X-ray (anterior–posterior, patient recumbent) within 24 h after line insertion, it was also used to assess the CVC position.
Forty-three X-rays were unavailable for analysis as the image had not been digitalized or could not be evaluated for other reasons. The line tip was considered to be in the correct position if placed within 55 mm below the carina. In addition, the angle between the CVC tip and vessel wall was measured and regarded as correct if <30° (Fig. 2).
(Enlarge Image)
Figure 2.
Chest X-ray with correctly positioned CVC. Catheter tip marked with an arrow.
Statistics
All data were analysed using SAS® software version 9.1 (SAS Institute Inc., Cary, NC, USA, 2003). The differences in the results between the two methods were compared using the χ test. The absolute values were compared applying the SAS MIXED procedure (F-test). Patient characteristic data were grouped according to venous access sites and gender and comparisons between groups made using the Wilcoxon rank-sum test. A value of P<0.05 was considered statistically significant.