Is radiologic placement of an arm port mandatory in oncology patients?

Analysis of a large bi-institutional experience




The objective of the current study was 2-fold: to evaluate a radiologically placed percutaneous arm port device (PRAPD) in a large series of 1000 consecutive cancer patients undergoing chemotherapy (in terms of safety, efficacy, complications, and quality of life [QoL]) and to propose future recommendations.


From 1998 to August 2002, all patients who had cancer required chemotherapy underwent insertion of a PRAPD and were prospectively included. All patients were followed for technical feasibility, overall device-related complications, and QoL.


Technical failures (6.3%) were caused by the inability to perform an arm venogram in 22 patients or to catheterize the brachial vein in 41 patients. Septic complications (3.2%) included septicemia (n = 7 patients), catheter sepsis (n = 9 patients), and febrile neutropenia (n = 16 patients). Mechanical complications (4%) included a twisted port (n = 2 patients), extravasation (n = 7 patients), catheter leaks (n = 7 patients), port obstruction (n = 7 patients), skin dehiscence of the port (n = 11 patients), catheter rupture and occlusion (n = 5 patients), and median nerve compression (n = 1 patient). Central venous thrombosis occurred in 12 patients (1.2%), and arm phlebitis occurred in 7 patients (0.7%). Procedure-related death occurred in 0.4%. Early port removal was performed in 5.3% of patients. Good QoL was reported at port removal.


The PRAPD was found to be safe, effective, and well tolerated in oncology patients. PRAPD could be recommended in selected patients instead of a surgical port device. Cancer 2007. © 2007 American Cancer Society.

Venous port device placement techniques have been established as safe and effective procedures for chemotherapy infusion in cancer patients. Initiated in the 1980s by surgeons,1 this technique advantageously replaced the external tunnelized catheter2 by inserting a subclavian port through a subclavicular or jugular approach.3 The disadvantage of the cephalic vein cutdown approach is a significantly high rate of failure of approximately 20%. Nevertheless, recent series have reported an excellent rate of success for the cephalic vein cutdown approach.4–6 This improved success rate is because of the recent use by surgeons of guidewires used in interventional radiology to catheterize small veins and/or the use of an external jugular approach in case of failure of the cephalic vein cutdown approach. Unfortunately, the risk of pneumothorax still is approximately 1% with this technique.7–9 In the 1990s, imaging eliminated complications such as pneumothorax related to the type of approach.10, 11 It also was possible to develop an imaging-guided antebrachial and brachial approach12–16 and to manage mechanical complications using a “minimally invasive strategy”.17–19 In the current study, port device placement and tunneling were performed after venography-guided percutaneous puncture of the basilic vein to gain access to the brachial vein with minimal dissection of the cellulous subcutaneous tissue.20 The objective of the current bi-institutional study was 2-fold: 1) to evaluate radiologic placement of a percutaneous arm port device in a large series of 1000 consecutive cancer patients undergoing intravenous chemotherapy in terms of safety, efficacy, complications, and quality of life (QoL); and 2) in the light of the results, to propose future recommendations.


Study Design and Parameter Analysis

The study design was reviewed formally and was approved by the local ethical committees. All patients provided their informed consent before entering the trial. Complete medical records were available for review from a prospective databank.

Technically successful port placement, procedure duration, acute and late complications, and QoL were recorded. All relevant discharge summaries, outpatient notes, laboratory reports, and chest radiograms were reviewed. Technical success was accomplished when the target vein (basilic or cephalic vein) was punctured successfully and catheterized with no more than 3 attempted needle passes, and when the port device was placed effectively in a subcutaneous pocket and was connected to the catheter extending up to the atrium-superior vena cava junction. Procedure duration was defined as the time elapsed from the skin incision at the vein access site to the pocket suture, as recorded in the nurse's or radiographer's notes.

Acute (procedural) and late complications were recorded. Major complications were defined as procedural or port device complications that prolonged the patient's hospital stay, required hospitalization of the patient, or resulted in the patient's death. Minor complications were defined as self-limiting events. Complications included symptomatic hematoma, device-related sepsis (port infection, febrile neutropenia, septicemia), skin dehiscence, catheter occlusion, deep venous thrombosis, and rupture of the catheter.

Evaluation of QoL was performed at port removal. The Visual Analog Scale 1 (VAS 1) was used to assess QoL related to activities of daily living with the device in the arm, and QoL was rated from 0 of 10, with 0 considered the worst QoL and 10 considered the perfect QoL. The VAS 2 also was used to assess QoL related to port device use. QoL scores using the VAS were recorded by a neutral individual.21 Finally, the 2 senior radiologists retrospectively analyzed the indications for the use of a radiologically placed arm port device, trying to define criteria for the absolute or relative use of an extrathoracic port device.


From April 1998 to August 2002, an arm port device was placed under imaging guidance in 1000 consecutive cancer patients (577 women and 423 men) with a mean age of 62.5 years (range, 19–90 years) in 2 French anticancer institutes. The indications were: an imperative need for venous access, a surgical chest port was contraindicated or port placement had failed, and unavailabilty of surgeons. In patients with breast cancer who required axillary lymph node dissection, in patients with pace-makers, and in patients who had a history of a peripheral or thoracic venous catheter, the arm port device was implanted in the contralateral arm; otherwise, the patient's nondominant arm was used whenever possible. Exclusion criteria were symptomatic venous thrombosis or contralateral lymphedema. Port devices were used to administer chemotherapy in all patients. Table 1 depicts the baseline characteristics of patients. Greater than 33% of the participants were women with breast cancer. During this study, patients did not receive prophylactic antibiotherapy for the procedure nor prophylactic anticoagulatory treatment during follow-up. Twelve patients had a history of deep vein thrombosis, and 6 patients had received oral anticoagulatory drugs before port placement that were continued after port placement.

Table 1. Baseline Characteristics of Patients Undergoing Arm Port Placement
Primary cancer locationNo. of patients
  • *

    Twenty-nine port devices were implanted on the previously treated side (bilateral cancer).

  • Five patients had 2 synchronous carcinomas, and 1 patient had 3 synchronous carcinomas.

Breast cancer*371
Head and neck cancer121
Genitourinary cancer107
Lung cancer107
Brain tumor82
Gynecologic cancer81
Colon cancer75
Miscellaneous tumor63

Port Placement

Patients who were coagulopathic or thrombocytopenic received blood products to correct deficiencies before port placement. No routine prophylactic antibiotherapy was administered. All patients received cetirizine dichlorhydrate at a dose of 10 mg (at mid-day and in the evening on the day before the procedure and in the morning on the day of the procedure). In the interventional radiology suite, the patient was positioned with the arm abducted and externally rotated. Under venography guidance, port device placement and tunneling were performed after percutaneously puncturing the basilic vein to gain access to the brachial vein with a small dissection of the cellulous subcutaneous tissue. All details of the procedure have been described previously.20 The port was a plastic, 25 × 10-mm high, magnetic resonance imaging, low-profile device (weight, 3.2 g; internal volume, 0.3 mL; septal dimension, 10.8 mm; Bard Inc., Salt Lake City, Utah) and an attachable, 7-French Groshong catheter (inner dimension, 1.3 mm).

Statistical Analysis

All data were entered into a computerized database and were analyzed using SPSS 11 statistical software for Windows. The chi-square statistic test or Fisher exact test was used to establish differences in the distribution of discontinuous variables, and the Student t test or the Mann-Whitney U test was used to compare continuous variables. Tests of significance were 2-tailed and considered significant with an α level of .05.


One thousand seven devices were placed in 1000 consecutive patients at 2 anticancer institutes. In terms of results, there was no statistical impact noted with regard to the involved center. Port placement was repeated twice in 9 patients and 3 times in 2 patients during follow-up, contralateral to the previous placement site in 4 patients, and ipsilateral in 7 patients. The mean follow-up of catheter use was 255.9 days (range, 4–699 days). Two hundred one patients died under therapy with functional ports. In 295 patients, the port was removed after completion of therapy. All of the remaining patients had functioning ports between 31 days and 543 days after placement. Port devices were removed because of complications in 5.3% of patients. Chemotherapy was initiated on the same day as port placement in 22% of patients. The technical success rate (access) was 93.7% for arm port device placement. The basilic vein was accessed in 92.6% of patients, the brachial vein was accessed in 7.2% of patients, and the cephalic vein was accessed in 0.2% of patients. Venography could not be performed in 22 patients (2.2%), and the target vessel in the arm could not be punctured and/or catheterized in 41 patients (4.1%). Regarding these technical failures, no statistically significant difference was observed between treatment-naive patients and treatment-experienced patients.

Symptomatic hematoma that required local care and analgesics occurred in 9 patients, mostly at the beginning of our experience (0.9%). Thirty-two patients (3.2%) developed clinically suspected, device-related sepsis. Local infection of the port occurred in 9 patients (0.9%). All patients received broad-spectrum intravenous antibiotics and local care. Six devices had to be removed. Seven patients (0.7%) developed septicemia, all of whom had received vancomycin over 8 to 15 days through the port. Sixteen patients experienced febrile neutropenia 6 to 12 days after intravenous chemotherapy, all of whom had received intravenous, broad-spectrum antibiotherapy. Three devices were explanted, but bacteriologic cultures were negative. Procedure-related deaths are summarized in Table 2.

Table 2. Procedure-related Death Rate, 0.4%
Life-threatening complicationsLethal outcome
Skin necrosis/drug extravasationCardiovascular collapse (n = 2 at D 8 and 10)
Febrile neutropenia/Pseudomonas septicemiaSeptic shock (n = 1 at D 8)
Central venous thrombosisPulmonary embolism (n = 1 at D 24)

Central venous thrombosis occurred at a mean of 42 days after port placement (range, 24–761 days) in 12 patients (1.2%). Skin dehiscence occurred at the port site in 11 patients and required removal of the device in 9 patients. Torquing of the catheter occurred in 2 patients (0.2%).

Leakage occurred in 7 patients (0.7%) at the site where the catheter and the port chamber were connected (n = 4 patients) or a few centimeters away from there, at the point where the catheter was inserted into the vessel (n = 3 patients). The mean time to the occurrence of leakage was 127 days (range, 17–400 days). In 5 patients, a second device was placed in the contralateral arm. Therapy was administered peripherally in the remaining 2 patients (during the last course of chemotherapy in 1 patient and for an intravenous infusion for resuscitation in the other patient). Catheter occlusion occurred in 7 patients (0.7%). Five patients had asymptomatic catheter migration after port device placement (0.5%). No arrythmias and no clinical incidents of pulmonary embolism because of catheter migration were documented. The catheter was removed percutaneously with the loop-snare technique without any complication by using right femoral venous access.22 Seven patients had extravasation (0.7%). One patient had median nerve compression by the port (0.1%) observed at Day 10. All symptoms disappeared 3 weeks after explantation.

QoL assessment was performed in 295 patients. The mean VAS 1 and VAS 2 scores were 9.2 of 10 (range, from 5 to 10 of 10) and 8.7 of 10 (range, from 0.2 to 10 of 10), respectively.


To begin this discussion, it is important to note and specify the unique nature of the complications that we encountered with these ports. In the current study, the technical failure rate with the exclusive use of arm venography and arm placement was 6.3% in oncology patients, which is lower than the rate reported by Starkhammar et al. in their first experience with the technique (16%).23 Compared with recent series on radiologically placed brachial devices, these rates are roughly comparable, because earlier series either combined venographic and ultrasonographic guidance24 or included ipsilateral forearm port placement,10 ipsilateral axillary vein placement,25 or contralateral arm port placement.24

The 2 major drawbacks of the brachial approach are the risk of humeral artery puncture (the brachial vein is located behind this artery in 6% of patients) and the risk of median nerve injury (0.1% compared with 0.6%).24 Because of these disadvantages, the use of the basilic vein and the placement of a low-profile, internal venous access port seem preferable.

The complication rate in our study was in accordance with the Journal of Vascular Interventional Radiology guidelines described in Table 3.26 The complication rate in our series was 10.4% or 0.4 of 1000 patient-days, which is higher than the rate reported by Bow et al. (0.23 of 1000 patient-days)27 but slightly lower than reported in larger series.3, 28

Table 3. Specific Major Complications for Image-guided Central Venous Access Through an Arm Port Device: Comparison With the Journal of Vascular Interventional Radiology Guidelines
Major complicationsCurrent study rate, %Reported rate, %Suggested rate,%
Wound dehiscence1.112
Arterial injury0.40.51

The sepsis-related (3.2%) and thrombus-related (1.9%) complication rates in the current study were similar to or better than the rates reported in the literature in smaller series.10, 12, 13, 15, 16, 29–31 The 5.3% premature port removal rate compared favorably with the mean score of 7.1% (range, 3.3%–18%) reported in the literature.

Systematic, prophylactic antibiotherapy remains controversial. The very low early local sepsis rate (0.4%) appears to indicate that prophylactic antibiotherapy is unnecessary.

In the current study, no patient received prophylactic anticoagulatory therapy. Nevertheless, the rate of symptomatic deep vein thrombosis (1.2%) was far lower than that observed in most reported series, although the population was composed exclusively of patients with malignancies. Indeed, Kuriakose et al. observed a significantly greater difference in the rate of venous thrombosis between the brachial approach versus the subclavicular approach (11.8% vs 5.2%) in similar patients.32 Only 7 incidents of ipsilateral brachial phlebitis were observed clinically during the follow-up period in our study, which confirmed the low thrombogenic quality of the polyurethane brachial catheter, particularly compared with the cephalic approach,33, 34 and which fully justified the preferential choice of access through the basilic vein and venographic monitoring. The presence and the length of the endovenous catheter, the ratio between the implanted catheter dimension and the target vein dimension, venous stasis, aggression of the venous wall (by the catheter or by chemotherapy), and the presence of thrombogenic neoplastic factors34 are many elements that may cause venous thrombosis. The texture of the catheter (polyvinyl chloride, polyethylene) also may be responsible. Silicone and polyurethane are considered less thrombogenic.35 Under venographic guidance, the clinician can select the adequate or optimal vein to puncture (which is a decisive advantage over the surgical technique) and also can catheterize occlusive venous thrombosis—in particular, pre-existent, latent, asymptomatic subclavian vein thrombosis (3 patients in the current study)—by using hydrophilic guidewires. Thus, ipsilateral port placement is possible without subsequent worsening of thrombosis.

Deep vein thrombosis certainly was underestimated in the current study. Indeed, Luciani et al. reported asymptomatic venous thrombosis in 76% of patients during the systematic, prospective ultrasound follow-up of subclavicular catheters in a head and neck cancer population.36 However, in their study, only 1 patient died of pulmonary embolism, and that patient had symptomatic, extensive venous thrombosis. Otten et al. reported a pulmonary embolism rate of 8.7% for thrombosis of the superior vena cava or of the innominate venous trunk.37 In the study by Monreal and Davant, pulmonary embolism occurred in up to 23% of patients, and lethal embolism occurred in 2.3% of patients despite adequate heparinotherapy.35

The surgical and radiologic techniques for arm port device placement are roughly similar (Table 4); however, very few studies have indicated an interest in surgical port placement using the brachial approach.23, 38 The main difference in the procedures is the use of arm venography, which may depict the most appropriate access route, a venous spasm (2%), asymptomatic central thrombosis (0.3%), central venous kinking, or congenital abnormalities (an episode of duplication was observed in our study). The use of venography may explain the significantly lower rate of central venous thrombosis after radiologically inserted devices. In fact, surgeons currently prefer the chest port,39–41 because it tends to be used more frequently and is more convenient during surgical procedures; however, nurses have pointed out that implanted ports are more stable.42 When surgical subclavicular placement is carried out by cannulating the cephalic arch in the deltopectoral groove, the average failure rate of venous catheterization is 19.2% (range, 6%–28%).6, 43 Directly puncturing the subclavian vein can cause a pneumothorax in 3.4% of patients40; in addition, the risk of pinch-off syndrome (POS) is estimated at 0.8%.44, 45 The use of endothelial denudation of the external jugular vein is less common.4, 6 The subclavicular port device does not have a propensity to give rise to leakage and occlusion of the catheter, which are attributed to POS in slightly greater than 50% of such incidents.45, 46 The percentage of patients with POS in our study was 0.5%, which was lower than the rate observed when a chest port device was used (range, 0.8%–1.1%).45, 47, 48 The average duration was 189 days for the subclavian device versus 200 days for the chest port. The 2 episodes of early embolism most likely were related to a faulty connection between the port chamber and the catheter; whereas the late complications were caused by the premature aging of the equipment as a result of repeated microtrauma, especially during certain physical activities, and exaggerated bending of the catheter at the vein entry point.49 This is supported by the finding that the cracks in the catheters (n = 5 patients) were observed only at the connection point between the catheter and the port chamber at the vein entry point. The 5 patients who had cracked catheters were not particularly overractive during their professional or sports activities. They received adequate treatment with percutaneous catheter withdrawal using the loop-snare technique under fluoroscopic guidance, and the interventional radiologist removed the brachial chamber. Percutaneous management of the mechanical complications by the interventional radiologist unquestionably is an advantage of radiologic port placement over surgical placement.22

Table 4. Surgical and Radiologic Results With Subcutaneous Peripheral Arm Ports
StudySpecialtyNo. of patientsSuccess rate, %Premature port removal, %Mean catheter use, DCentral venous thrombosis, %Postinsertion phlebitis, %Catheter sepsis, %
  1. S indicates surgery; R, radiology; F/US, fluoroscopic/ultrasound; ND, not determined.

Starkhammar et al., 199216S74844.915958.24.2
Foley, 200210R1501003.316127.33.3
Beheshti et al., 199829F/US521004372222
Lyon et al., 199915F/US2041009.71694.4ND3.4
Hills et al., 199713R981006.2246017
Bodner et al., 200012US109100182669.8ND9.9
Lersch et al., 199930R1009651278ND10
Hata et al., 199831R10493.75.787.705.70
Current studyR100793.75.32561.20.70.9

The key to success is achieving minimal dissection of the subcutaneous tissue. The size of this dissection has to correspond to that of the port chamber. The chamber lies on a stable, level musculoaponevrotic plane without in-depth fixation. Suturing has to be on 2 planes (subcutaneous with an absorbable wire and cutaneous with a nonabsorbable wire) and parallel with areas of tension in the arm. If the subcutaneous pocket is too large, then twisting of the port device may occur. Conversely, if the subcutaneous pocket is not large enough, then there is a risk of skin dehiscence. This occurred in 11 of our patients, mainly at the beginning of our experience (9 of the first 110 patients), because the scar was under tension. In 9 of those patients, the port device had to be removed. It led us to suture on 2 planes (subdermal and subcuticular) that were remote from the chamber on a lax area of skin (not under tension).

The device had to be withdrawn in 3 patients, because the port chamber was not rinsed after chemoinfusion. The episodes of transitory occlusion were managed adequately by the nurse and were not taken into account in the analysis. The use of a 1-eyed catheter with a side valve (Groshong catheter), adequate rinsing of the catheter at the end of each procedure, aspiration with the Huber needle under positive pressure, and x-ray inspection to verify proper positioning of the catheter,36, 50 explain this low incidence of catheter occlusion (0.3%).

The analysis of QoL scores produced excellent results and a better appreciation of the practical aspect of chemoinfusion. The patients particularly appreciated the ease of accessibility of the arm port device (simply rolling up the sleeve was sufficient), no uncosmetic scar at the incision site for the women, and especially not having to expose their chest during chemotherapy. The discomfort that patients felt when dressing or walking explains the lower QoL scores for activities of daily living. In fact, the indications rarely were preferences of the patient, who followed the advice and the opinion of their practitioners (who, in turn, were dependent on local practices in each hospital and on the availability of the different specialists in those hospitals). We analyzed the indications for arm port device placement retrospectively. According to our experience, and in the opinion of the patients (VAS 1 and VAS 2), the indications for arm port device placement may include the following: 1) women with breast cancer who wear garments with a low neckline and are concerned by the cosmetic aspect (physical and psychic trauma); 2) patients who previously received irradiation to the neck or the chest (skin brittleness, delayed wound healing, a higher risk of skin dehiscence); 3) patients who present with cervical venous thrombosis or who have pacemakers (up to 39% may present with total or partial subclavicular thromboses51); 4) patients with head and neck cancer (tracheostomy, cervical radiodermatitis, cervical dissection); 5) patients with high-risk respiratory conditions or with major kyphosis (may benefit from a radiologic brachial implantation in an upright posture); and 6) the presence of pulmonary carcinomatous lymphangitis and/or pericardial or pleural effusions (the risk of iatrogenic pneumothorax is estimated between 1% and 3% for chest port placement according previously published series39–41, 52). All of these types of patients represent > 50% of the patients who were included in our study. Conversely, patients who suffer from known venous brittleness without an identifiable surface venous network and those who have venous spasm, who are agitatedwith neuropsychiatric signs and symptoms (peripheral neuropathy, multiple sclerosis), or who require prolonged chemotherapy (particularly during the night) should benefit from a subclavicular port device.11, 52

In conclusion, the current study indicated that arm port device placement is a safe and effective technique in oncology patients and is well tolerated. In our opinion, some selected patients may benefit from this technique.


We thank Lorna Saint Ange for editing.