Totally Implantable Venous Access Devices

Retrospective Analysis of Different Insertion Techniques and Predictors of Complications in 796 Devices Implanted in a Single Institution

Elisa Granziera; Marco Scarpa; Angelo Ciccarese; Bogdan Filip; Matteo Cagol; Valentina Manfredi; Rita Alfieri; Connie Celentano; Sandra Cappellato; Carlo Castoro; Muzio Meroni


BMC Surg. 2014;14(27) 

In This Article


Patient Characteristics

Overall characteristics of the study population and divided by the insertion technique are shown in Table 1 and Table 2. The majority of patients had solid organ cancer; only 29 (3.64%) had hematologic malignancies (lymphoma). The median follow up was 35.1 (23.0–55.4) weeks. During the period of our study three different implantation techniques were used. In the first part of our study, we utilised the surgical cutdown or blind percutaneus according to surgeon and anestesiologist preference. In the second part of our study, all the consecutive TIVADs were inserted using the ultrasound technique. The anatomical choice for venous access is shown in Table 3. In the surgical cutdown technique, the cephalic vein was used in almost all cases (92/102 patients), in the rest of the cases (10 patients) the external jugular vein was used. The access site for blind cutaneus technique was the subclavian vein in 41 cases or the jugular vein in 5 cases (2 cases missing). The majority of the access sites were, the distal internal jugular vein or brachiocephalic vein through the supraclavicular approach (515/646 patients), then the subclavian vein (29 cases) or the distal internal jugular vein (100 cases).

Clinical Outcome

In 102 patients placement of TIVAD via the cephalic vein cutdown approach was initially attempted, 87 patients underwent successful insertion while 15 required conversion to a percutaneous approach (success rate 85.3%), 11 using the blind technique, 3 using the ultrasound guided technique. In one patient, it was impossible to access any central vein due to the anatomical pitfalls and a peripheral inserted central venous access (PICC) was inserted under ultrasound guidance. Lack of success was related to small size of vein (n = 6) or inability to locate the vein (n = 8) or impossibility to transverse the angle of insertion of cephalic vein into the subclavian vein (n = 1). The percutaneous blind technique was initially used in 48 patients, and was successful in 45 patients (93.7%), 3 patients required conversion to a surgical approach. Lack of success was related to occurrence of pneumothorax attempting subcavian vein cannulation (n = 2) or inability to locate the vein (n = 1). The ultrasound guided approach was used as a primary technique in 646 patients with a success rate of 99.7% (644/646). In 13 patients, a second attempt changing puncture site was necessary, while 2 patients required conversion to a surgical approach because of fibrotic or collapsed vein. Mean operating time was 51 minutes for cephalic vein cutdown approach, 49 minutes for blind percutaneous technique and 33 minutes for ultrasound guided technique.

The cost, expressed as cost of time of use of the operative room of US-guided TIVAD placement, was significantly lower compared to cutdown approach and to percutaneous blind technique. In fact, in spite of the use of the same device, the duration of the procedure was significantly shorter in patients who had a US guided TVAD placement.

The overall early complication rate including changing the puncture site using then same technique was 7.2% (57 of 796 cases) as listed in Table 4. The only severe intraoperative complications requiring immediate treatment were three of the four pneumothoraxes and one of the two arrhythmias (a supraventricular paroxistic tachycardia requiring pharmacological treatment). All the other complications did not require any treatment except for a longer monitoring before discharge. The technique failure was significantly reduced using ultrasound compared with the surgical technique (p < 0,001). The occurrence of pneumothorax and arterial puncture were significantly higher in the landmark group as compared to the ultrasound group. Interestingly, our data shows a significantly increased number of sepsis in the surgical group compared with those documented in the ultrasound group, although the overall rate of late complications was significantly lower in this group.

Late complications occurred in 49 of 796 patients (6.1%), requiring removal of the TIVAD in 43 (5.2%) as listed in Table 5. The microorganisms isolated in TIVAD related sepsis were candida glabrata, candida parapsilosis, bacillus species, streptococcus mitis, pseudomanas aeruginosa, and staphylococcus epidermidis. The other cases of late complications that did not require removal of the TIVAD were: one case of sepsis treated with intravenous antibiotics, one thrombosis that required anticoagulants with remission, one malfunction, 2 cases of extravasation or severe pain. As shown in Table 6, we performed a logistic regression for prediction of late complications that lead to TIVAD removal. Only the open approach and low levels of white blood cells resulted to be independent predictor of urgent TIVAD removal. Risk rate analysis of TIVAD removal for sepsis showed a peak between 20 and 30 weeks and a second one after 60 weeks DVT, decubitus and dislocation had a risk rate peak after 60 weeks. Risk analysis for TIVAD urgent removal is shown in Figure 1. Late misplacement of TIVAD occurred but they did not require any further intervention unless it precluded its function. Some misplacement cases are shown in Figure 2.

Figure 1.

Risk analysis for TIVAD urgent removal.

Figure 2.

Late TIVAD misplacement that did not compromise their function.