This study was approved by the institutional review board (No.19-22), and informed consent was obtained from all eligible patients. It was registered in the UMIN Clinical Trial Registry (UMIN 000036040). This study was conducted from January to August 2019 at the Sakurabashi-Watanabe Hospital, Osaka, Japan. We prospectively examined the PAC location at the level of the tricuspid annulus in 135 consecutive adult patients undergoing cardiovascular surgery. We excluded cases whose PAC was placed before surgery, whose 3D TEE view was not clear, whose PAC had to be placed through the left internal jugular vein, whose data had a missing value, and those who had a history of tricuspid annuloplasty.

The electrocardiogram, invasive arterial blood pressure, oxygen saturation, and endo-tidal carbon dioxide were monitored on all patients. After induction of general anesthesia with midazolam of 0.1mg/kg, fentanyl of 200μg, and vecuronium of 0.1mg/kg, mechanical ventilation was started following tracheal intubation. Anesthesia was maintained with propofol or sevoflurane combined with remifentanil and fentanyl. The PAC (continuous cardiac output/SvO2 Catheter 744HF75, Edwards Lifescience, Irvine, CA, USA) was inserted through the right internal jugular vein and placed by one of the three staff anesthesiologists in our hospital. First, the introducer sheath was placed through the right internal jugular vein and the PAC was started floating through the sheath until the pressure waveform changed to the central venous pressure pattern. Then, the balloon was inflated with 1.5 ml of air. With inflated balloon, the catheter was floated into the pulmonary artery. Once the waveform of the pulmonary artery was first observed, we inserted the catheter approximately 2–3cm forward and deflated the balloon.

The time required for the placement of PAC was measured as follows. The right atrium time (RAT) and the right ventricle time (RVT) were defined as the duration of time required for the catheter to float from the CVP position through the tricuspid valve to the right ventricle and that from the right ventricle through the pulmonary valve to the pulmonary artery, respectively.

The 3D TEE studies were performed using the Phillips iE90 ultrasound system (Philips Healthcare, Eindhovent, Netherlands). The image of the tricuspid valve was created using the mid-esophageal 4-chamber view in full volume 3D mode wherein the pyramidal scan volume could capture the whole tricuspid valve, mitral valve, and aortic valve (Fig. 1). This image makes it easier to determine the orientation of the tricuspid valve. After the gain was optimized, the image was cropped to visualize the tricuspid valve in the surgeon’s view from right atrium perspective. Originally, we planned to let the tricuspid valve be oriented with the septum in 6 o’clock position in accordance with the American Society of Echocardiography guideline [8], but we found that the images were easier to understand using a view that depicted the three valves all together as we are accustomed to use.

The 3-dimensional echocardiographic image of the tricuspid valve enface view from right atrium. The tricuspid valve is seen to the left, the mitral valve on the right, and the aortic valve below the two valves

The tricuspid valve has three leaflets, anterior, posterior, and septal. The anterior leaflet is the largest and covers the infundibulum anteriorly to the inferolateral wall posteriorly. The posterior leaflet has many variations because of the multiple scallops. The septal leaflet is the smallest and arises directly from the annulus above the interventricular septum [9]. We divided the PAC position at the tricuspid annulus into four categories (Fig. 2). If it was in the commissure, the position was defined according to the commissural location between the respective two tricuspid leaflets, antero-posterior leaflets (AP), antero-septal leaflets (AS), and septal-posterior leaflets (SP), respectively. Everything else was defined as the center. After the PAC was placed, TEE probe was inserted. We acquired the 3D image of tricuspid valve and determined the PAC position before the surgery was started.

Schematic view of the tricuspid valve with leaflets A and pulmonary artery catheter location on 3D transesophageal echocardiography B. Example of four categories of the catheter locations at the tricuspid valve annulus. AP anterior-posterior leaflets, SP septal-posterior leaflets, AS anterior-septal leaflets, center center of the three leaflets

Based on our previous study [1], we planned to collect 100 cases and then calculate the sample size. The sample size was calculated with 80 % power (a=0.05 and b=0.20) to detect the mean difference of RVT between the AP position and the SP position, and it was calculated that 128 patients would be necessary to demonstrate a significant difference.

Data were expressed as mean ± SD or as a median and interquartile range as appropriate. The preoperative patient data were analyzed by the analysis of variance or Kruskal-Wallis test as appropriate. The placement time was analyzed by Kruskal-Wallis test followed by Mann-Whitney test with Holm multiple comparisons to specify differences between groups. The statistical analysis was performed by SPSS (IBM Corporation, USA) version 20.0. P < 0.05 was considered statistically significant.