Complication-related Removal of Central Venous Ports: Comparison of Subclavian and Upper Arm Approaches Authors: Tomokazu Takeuchi, Soma Kumasaka, Hiroyuki Tokue, Yuko Seki, Hayato Aoki, Ayako Taketomi-Takahashi, Yusuke Sato, Kei Shibuya, Masashi Ebara, Terutaka Yoshihara, Hiroyuki Yasui, Takayuki Yokota, Takayuki Suto, Yoshito Tsushima Published: 2026 ABSTRACT AbstractPurpose: To compare overall and cause-specific complication-related central venous port removal, particularly infection-related removal and catheter fracture, and time-to-event outcomes between the subclavian and upper arm approaches.Material and Methods: This retrospective study included 1,195 patients who underwent ultrasound-guided central venous port implantation via the subclavian (n = 706) or upper arm (n = 489) approach between 2012 and 2021. Comparative analyses were restricted to a maximum of 5 years after implantation. A total of 621,902 catheter-days were analyzed. The evaluated outcomes were overall complication-related central venous port removal and cause-specific removals, including infection-related removal, catheter fracture, malposition, occlusion, leakage, and venous thrombosis. Crude incidence rate ratios were estimated using Poisson regression, adjusted incidence rate ratios using a negative binomial model, and removal-free survival using Kaplan–Meier analysis.Results: Within 5 years, complication-related removals occurred in 50 patients (7.1%) in the subclavian group and 41 (8.4%) in the upper arm group. Crude incidence rates were 0.131 and 0.171 per 1,000 catheter-days, respectively (crude incidence rate ratio, 0.77; 95% confidence interval, 0.51–1.16; p = 0.211). Adjusted analysis showed no significant difference between approaches (adjusted incidence rate ratio, 1.09; 95% confidence interval, 0.59–2.00; p = 0.786). Infection-related removal was more frequent in the upper arm group (crude incidence rate ratio, 0.55; 95% confidence interval, 0.34–0.91; p = 0.020). Catheter fractures occurred only in the subclavian group (12/706, 1.7%; 0.031 per 1,000 catheter-days). Overall, removal-free survival did not differ significantly (p = 0.234), whereas infection-free and fracture-free survival differed significantly.Conclusions: Overall complication-related removal rates were similar between the approaches, but complication profiles differed.Keywords: central venous port, upper arm approach, catheter fracture, long-term outcomes Introduction With the global aging of the population, there has been an increasing demand for intravenous therapies, including chemotherapy for patients with cancer and long-term parenteral nutrition [1-3]. Consequently, the use of central venous (CV) port systems is increasing, and the optimal implantation approach for sustained clinical use requires careful evaluation [4]. Several methods for CV port implantation have been proposed, such as the subclavian vein approach and the upper arm vein approach. Although the subclavian vein approach has traditionally been widely used because of its technical familiarity and established clinical utility, perioperative adverse events have been reported [5, 6]. More recently, the upper arm vein approach has been increasingly adopted because of its procedural feasibility and favorable perioperative safety profile [7-12]. However, in some cases, complications may arise following implantation, necessitating port removal. Several studies have compared complications between the two approaches; however, available comparative data remain limited, and follow-up duration or time-at-risk have not always been fully characterized. One previous study reported 168,022 catheter-days with a mean follow-up of 398 days, another reported a median follow-up of 94 days, and another comparative series did not explicitly report the follow-up duration [13-15]. Given the increasing need for prolonged device use, further comparative data based on catheter-days and time-to-event analyses over clinically relevant observation periods are needed. The aim of this study was to compare complication-related CV port removal between the subclavian and upper arm approaches, with particular attention to differences in complication profiles and time-to-event patterns. Material and Methods Patients We conducted a retrospective review of all patients who underwent CV port implantation in our unit between January 2012 and December 2021. Patients who underwent CV port implantation via approaches other than the subclavian or upper arm approach were excluded from the study. We obtained age at implantation, sex, indication for CV port implantation (chemotherapy or nutrition), and history of diabetes mellitus from electronic clinical records. The primary endpoint of this study was complication-related CV port removal. Planned removals after completion of therapy were not considered complication-related events. The institutional review board approved all aspects of this study. Due to the retrospective nature of the investigation, the requirement for written informed consent was waived. This study was conducted using the opt-out method available on our hospital’s website. Placement of the CV port All CV port implantations were performed by interventional radiologists using one of several single-lumen CV port systems (BardPort, Bard Access Systems, Inc., Salt Lake City, UT, USA; PowerPort Implantable Port, Bard Access Systems, Inc.; PowerPort Slim Implantable Port, Bard Access Systems, Inc.; DewX Eterna, Terumo Corporation, Shibuya-ku, Tokyo, Japan; P-U Celsite Port, Toray Medical Co., Ltd., Chuou-ku, Tokyo, Japan). Based on the available product information, BardPort used a silicone catheter, whereas PowerPort Implantable Port, PowerPort Slim Implantable Port, DewX Eterna, and P-U Celsite Port used polyurethane catheters. Although the referring physician could specify the type of CV port, this was rarely the case in practice. In most instances, the interventional radiologist selected the port type according to the patient’s body habitus, planned implantation site, device availability, and operator discretion. No formal device-allocation protocol was used during the study period. The implantation approach was selected by the referring physician based on clinical judgment. All procedures were performed exclusively by interventional radiologists using a standardized technique. The upper arm approach has been used in our unit since 2016. The subclavian approach involved catheter implantation through the subclavian vein, whereas the upper arm approach involved catheter insertion via the basilic or cephalic vein. Both procedures were performed using real-time ultrasound guidance to ensure accurate venipuncture. For vein selection in the upper arm approach, vessels were selected so that catheter occupancy of the target vein was less than 45% to reduce the risk of catheter-related thrombosis [16-20]. Cause of CV port removal In this study, the reasons for removal were documented by extracting information from electronic medical records and radiology reports related to port removal or complication assessment. Infectious complications were defined as port removal due to positive catheter culture findings, clinically diagnosed sepsis, or clinically diagnosed local port-site infection. Local port-site infection included port-pocket infection or local inflammatory findings, such as erythema, swelling, tenderness, or purulent discharge around the port site, when these findings were clinically judged to be the reason for port removal. Cases with suspected infection alone were not classified as infectious complications unless these criteria were met. Catheter malposition, occlusion, fracture, and leakage were confirmed with radiography and contrast examination through the port system. Catheter fracture was defined as structural disruption of the catheter shaft, whereas leakage was defined as contrast extravasation from the port chamber without radiographic evidence of catheter shaft fracture. Venous thrombosis was confirmed with ultrasound. Statistical analysis For analysis, patients were divided into two groups according to the implantation approach: the subclavian group and the upper arm group. Baseline characteristics were compared using the Mann–Whitney U test for continuous variables and Fisher’s exact test for categorical variables. To improve comparability between groups and reduce bias related to unequal maximum follow-up duration, follow-up for comparative analyses was restricted to a maximum of 5 years after implantation. Thus, follow-up duration was defined as the date of implantation to the earliest of complication-related removal, 5 years after implantation, or the last known date the patient was alive on or before June 30, 2023. Crude incidence rates were calculated as events per 1,000 catheter-days. Crude incidence rate ratios (IRRs) were estimated using Poisson regression with implantation approach as the sole predictor and log-transformed follow-up duration as an offset. For the primary endpoint of total complication-related removals, adjusted IRRs were estimated using a negative binomial model including implantation approach, age, sex, diabetes mellitus, indication for CV port implantation, and implantation year, with log-transformed follow-up duration as an offset. Kaplan–Meier analysis with the log-rank test was used to compare complication-related removal-free survival, infection-free survival, and fracture-free survival between the two groups. Planned removals after completion of therapy were excluded from complication-related analyses. All tests were two-sided, with p-values < 0.05 considered statistically significant. All statistical analyses were performed using IBM SPSS Statistics for Mac, version 29.0 (IBM Corp., Armonk, NY, USA). Results Patient characteristics This study included 1,244 patients. Of these, 49 patients were excluded from further analysis for the following reasons: 1) the CV port was implanted in the femoral vein (n = 45) and 2) the CV port was implanted in the internal jugular vein (n = 4). Thus, a total of 1,195 patients met the study criteria and were divided into two groups: the subclavian group (n = 706) and the upper arm group (n = 489) (Figure 1). For comparative analyses, follow-up was restricted to a maximum of 5 years after implantation. The total follow-up time in the administratively censored cohort was 621,902 catheter-days, including 381,516 catheter-days in the subclavian group and 240,386 catheter-days in the upper arm group. In this cohort, the mean and median follow-up periods were 520.4 and 329 days, respectively (range, 2-1,827 days). In the subclavian and upper arm groups, the mean follow-up periods were 540.4 and 491.6 days, respectively, and the median follow-up periods were 311 days (range, 2-1,827 days) and 354 days (range, 3-1,826 days), respectively. Table 1 shows the baseline characteristics of the patients in the subclavian and upper arm groups. Flowchart of participants. Table 1.Baseline Characteristics of the Study Population.Subclavian group(n = 706)Upper arm group(n = 489) p-valueAge (years)65 (4-88)65 (13-86)0.588Sex, n (%)Male398 (56.4)279 (57.1)0.859Female308 (43.6)210 (42.9)Reason for CV port implantation, n (%)Chemotherapy477 (67.6)405 (82.8)<0.001Nutrition229 (32.4)84 (17.2)Diabetes mellitus, n (%)131 (18.6)74 (15.1)0.138CV port system, n (%)<0.001BardPort378 (53.5)0 (0)PowerPort Implantable Port220 (31.2)0 (0)PowerPort Slim Implantable Port0 (0)208 (42.5)DewX Eterna39 (5.5)155 (31.7)P-U Celsite Port69 (9.8)126 (25.8)Follow-up duration after implantation, days311 (2-1,827)354 (3-1,826)0.387Continuous variables are presented as median (range) and compared using the Mann–Whitney U test. Categorical variables are presented as n (%) and compared using Fisher’s exact test or chi-square test, as appropriate.CV: central venous Baseline Characteristics of the Study Population. Continuous variables are presented as median (range) and compared using the Mann–Whitney U test. Categorical variables are presented as n (%) and compared using Fisher’s exact test or chi-square test, as appropriate. CV: central venous Cause of CV port removal Within 5 years after implantation, a total of 91 patients underwent CV port removal because of complications, including 50 patients in the subclavian group and 41 patients in the upper arm group. Patient-based comparison showed no significant difference in the overall frequency of complication-related removals between the two groups (7.1% vs 8.4%, p = 0.438). The corresponding crude incidence rates were 0.131 and 0.171 per 1,000 catheter-days, respectively, with no significant difference in crude incidence rate ratio (IRR, 0.77; 95% CI, 0.51-1.16; p = 0.211). In adjusted negative binomial regression, the overall complication-related removal rate also did not differ significantly between the two approaches (adjusted IRR for subclavian vs upper arm, 1.09; 95% CI, 0.59-2.00; p = 0.786). The most frequent cause of complication-related removal was infection. Infection-related removal occurred in 29 patients (4.1%) in the subclavian group and 33 patients (6.7%) in the upper arm group (patient-based p = 0.047). The corresponding crude incidence rates were 0.076 and 0.137 per 1,000 catheter-days, respectively, and the crude IRR for the subclavian versus upper arm approach was 0.55 (95% CI, 0.34-0.91; p = 0.020). Catheter fracture occurred in 12 patients (1.7%) in the subclavian group, whereas no fractures occurred in the upper arm group (patient-based p = 0.002). The corresponding incidence rates were 0.031 and 0.000 per 1,000 catheter-days, respectively. Among the 12 catheter fracture cases, the median time from implantation to catheter fracture was 916.5 days (range, 453-1,798 days). Eleven cases involved PowerPort Implantable Port systems with polyurethane catheters, and one involved BardPort with a silicone Groshong catheter. Fractures occurred along the catheter shaft in seven cases and near the catheter–port junction in five cases. A history of power injection through the port system was documented in two cases. Detailed characteristics of the catheter fracture cases are shown in Supplementary Table 1. On chest radiographs, 10 fractures were located lateral to the junction of the clavicle and the first rib (Figure 2). The remaining two fractures were located near the clavicle–first rib crossing point. No significant differences were found in any other complications (Table 2). Chest radiographs showing catheter fractures in the subclavian group. (A) Chest radiograph shows a right subclavian CV port with a catheter fracture located lateral to the crossing point of the right clavicle and first rib (arrow). The distal catheter fragment is visible in the mediastinum (arrowhead). (B) Chest radiograph shows a left subclavian CV port with a catheter fracture located lateral to the crossing point of the left clavicle and first rib (arrow). The distal catheter fragment is visible in the left brachiocephalic vein (arrowhead). CV: central venous Table 2.Five-Year Complication-Related Removals after Central Venous Port Placement via the Subclavian and Upper Arm Approaches.Cause of removalSubclavian Group (n = 706)Upper arm Group (n = 489)Patient- basedp-value*Crude IRR(subclavian vs upper arm)95% CICrudep-valueTotal catheter-days381,516240,386n (%)Incidence rate per 1,000 catheter-daysn (%)Incidence rate per 1,000 catheter-daysInfection29 (4.1%)0.07633 (6.7%)0.1370.0470.550.34-0.910.020Venous thrombosis1 (0.1%)0.0031 (0.2%)0.0041.0000.630.04-10.070.744Catheter occlusion2 (0.3%)0.0054 (0.8%)0.0170.2340.320.06-1.720.182Catheter malposition2 (0.3%)0.0051 (0.2%)0.0041.0001.260.11-13.900.850Skin dehiscence4 (0.6%)0.0081 (0.2%)0.0040.6542.520.28-22.560.408Leakage0 (0%)0.0001 (0.2%)0.0040.409NENENECatheter fracture12 (1.7%)0.0310 (0%)0.0000.002NENENETotal complication- related removals50 (7.1%)0.13141 (8.4%)0.1710.4380.770.51-1.160.211Crude IRRs were calculated from catheter-day–based incidence rates, and 95% CIs and p-values were estimated using the Wald method.IRR: incidence rate ratio; NE: not estimable because of zero events in one or both groups*Patient-based p-values were calculated using a two-sided Fisher’s exact test. Five-Year Complication-Related Removals after Central Venous Port Placement via the Subclavian and Upper Arm Approaches. Crude IRRs were calculated from catheter-day–based incidence rates, and 95% CIs and p-values were estimated using the Wald method. IRR: incidence rate ratio; NE: not estimable because of zero events in one or both groups *Patient-based p-values were calculated using a two-sided Fisher’s exact test. Exploratory subgroup analyses stratified by indication for CV port implantation are shown in Tables 3 and 4. In the chemotherapy subgroup, catheter-day-based incidence rates of total complication-related removal and infection-related removal did not differ significantly between the approaches. In the nutritional-support subgroup, infection-related removal was also higher in the upper arm group in both patient-based and catheter-day–based analyses, and total complication-related removal was also higher in the upper arm group in catheter-day–based analysis. Because the number of events was limited in several strata, these subgroup analyses should be interpreted cautiously. Table 3.Subgroup Analysis of Causes of Removal in Patients with Catheter Placement for Chemotherapy.Cause of removalSubclavian Group(n = 477)Upper arm Group(n = 405)Patient-basedp-value*Crude IRR(subclavian vs upper arm)95% CICrude p-valueTotal catheter-days289,519211,389n (%)Incidence rate per 1,000 catheter-daysn (%)Incidence rate per 1,000 catheter-daysInfection25 (5.2%)0.08626 (6.4%)0.1230.4270.700.41-1.220.207Venous thrombosis1 (0.2%)0.0031 (0.2%)0.0051.0000.730.05-11.670.824Catheter occlusion0 (0%)0.0004 (1.0%)0.0190.044NENENECatheter malposition2 (0.4%)0.0071 (0.2%)0.0051.0001.460.13-16.100.757Skin dehiscence3 (0.6%)0.0101 (0.2%)0.0050.6292.190.23-21.060.497Leakage0 (0%)0.0001 (0.2%)0.0050.459NENENECatheter fracture11 (2.3%)0.0380 (0%)0.0000.001NENENETotal complication- related removals42 (8.8%)0.14534 (8.4%)0.1610.9040.900.57-1.420.655Crude IRRs were calculated from catheter-day–based incidence rates, and 95% CIs and p-values were estimated using the Wald method.IRR: incidence rate ratio; NE: not estimable because of zero events in one or both groups*Patient-based p-values were calculated using a two-sided Fisher’s exact test. Subgroup Analysis of Causes of Removal in Patients with Catheter Placement for Chemotherapy. Crude IRRs were calculated from catheter-day–based incidence rates, and 95% CIs and p-values were estimated using the Wald method. IRR: incidence rate ratio; NE: not estimable because of zero events in one or both groups *Patient-based p-values were calculated using a two-sided Fisher’s exact test. Table 4.Subgroup Analysis of Causes of Removal in Patients with Catheter Placement for Nutritional Support.Cause of removalSubclavian Group(n = 229)Upper arm Group(n = 84)Patient-basedp-value*Crude IRR (subclavian vs upper arm)95% CICrude p-valueTotal catheter-days91,99728,997n (%)Incidence rate per 1,000 catheter-daysn (%)Incidence rate per 1,000 catheter-daysInfection4 (1.7%)0.0437 (8.3%)0.2410.0100.180.05-0.620.006Venous thrombosis0 (0%)0.0000 (0%)0.0001.000NENENECatheter occlusion2 (0.9%)0.0220 (0%)0.0001.000NENENECatheter malposition0 (0%)0.0000 (0%)0.0001.000NENENESkin dehiscence1 (0.4%)0.0110 (0%)0.0001.000NENENELeakage0 (0%)0.0000 (0%)0.0001.000NENENECatheter fracture1 (0.4%)0.0110 (0%)0.0001.000NENENETotal complication- related removals8 (3.5%)0.0877 (8.3%)0.2410.1300.360.13-0.990.049Crude IRRs were calculated from catheter-day–based incidence rates, and 95% CIs and p-values were estimated using the Wald method.IRR: incidence rate ratio; NE: not estimable because of zero events in one or both groups*Patient-based p-values were calculated using a two-sided Fisher’s exact test. Subgroup Analysis of Causes of Removal in Patients with Catheter Placement for Nutritional Support. Crude IRRs were calculated from catheter-day–based incidence rates, and 95% CIs and p-values were estimated using the Wald method. IRR: incidence rate ratio; NE: not estimable because of zero events in one or both groups *Patient-based p-values were calculated using a two-sided Fisher’s exact test. Kaplan–Meier analysis showed no significant difference in complication-related removal-free survival between the two groups (log-rank p = 0.234; Figure 3A). In contrast, infection-free survival was significantly lower in the upper arm group (log-rank p = 0.048; Figure 3B), and fracture-free survival was significantly lower in the subclavian group (log-rank p = 0.020; Figure 3C). Kaplan–Meier curves after administrative censoring at 5 years. (A) Complication-related removal-free survival. (B) Infection-free survival. (C) Fracture-free survival. Discussion In this retrospective cohort study, with comparative analyses restricted to a maximum of 5 years after implantation, we compared complication-related CV port removals between the subclavian and upper arm approaches. Our results showed that the overall rate of complication-related port removal did not differ significantly between the two approaches after adjustment for clinical covariates and follow-up duration. However, the complication profile differed between the groups: infection-related removal was more frequent in the upper arm group, whereas catheter fracture occurred only in the subclavian group. Only a limited number of previous studies have compared complication-related outcomes between chest or subclavian ports and arm ports. Kuriakose et al. [13] reported a higher incidence of deep venous thrombosis with peripheral arm ports than with central chest ports. Akahane et al. [14] found that postprocedural adverse events were more frequent in the arm port group, although infectious adverse events and port removal rates were similar between subclavian and arm ports. Conversely, Shiono et al. [15] reported a lower overall postprocedural complication rate in the upper arm group, whereas infection and venous thrombosis rates were similar between groups, and catheter fracture was observed only in the subclavian group. In the present study, overall complication-related removal rates were similar between the two groups in both crude and adjusted analyses. In addition, Kaplan–Meier analysis showed no significant difference in complication-related removal-free survival. The discrepancy between the crude and adjusted IRRs may reflect differences in baseline characteristics between the groups, including implantation indication, implantation year, and patient background, although residual confounding cannot be excluded. These findings suggest that, within 5 years after implantation, neither approach was clearly superior with respect to overall complication-related port removal. Infection-related removal was more frequent in the upper arm group in both patient-based and catheter-day–based crude analyses, and infection-free survival was significantly lower in the upper arm group. This result differs from a previous study in which the upper arm approach showed a lower infection rate than the subclavian approach, although the difference was not statistically significant [15]. A recent narrative review of peripherally inserted central catheter (PICC)-associated complications emphasized that central line-associated bloodstream infection remains an important complication of upper-extremity CV access, with reported rates of approximately 1.4%-1.9%, and that infection risk is influenced by multiple factors, including malignancy, diabetes mellitus, obesity, chronic renal failure, immunosuppression, catheter dwell time, number of lumens, and device management practices [21]. Although PICCs and upper arm CV ports differ in structure and use, this evidence supports the concept that infection risk in upper-extremity venous access is multifactorial and cannot be attributed solely to the insertion route. The baseline differences between the cohorts should also be considered when interpreting infection-related outcomes. The subclavian group included a higher proportion of ports implanted for nutritional support, whereas the upper arm group included a higher proportion of ports implanted for chemotherapy. To address this imbalance in implantation indication, we performed exploratory subgroup analyses stratified by chemotherapy and nutritional support. In the chemotherapy subgroup, catheter-day–based incidence rates of infection-related removal did not differ significantly between the subclavian and upper arm groups. In the nutritional-support subgroup, infection-related removal was more frequent in the upper arm group; however, the number of events was limited, and this finding should be interpreted cautiously. These subgroup results suggest that the higher infection-related removal rate in the upper arm group may not be explained by indication alone. Several unmeasured mechanisms may have contributed to the higher infection-related removal rate in the upper arm group. Port location may affect infection risk because upper arm ports may be influenced by upper limb movement, sweating, friction from clothing, and difficulty maintaining a stable dressing. Technical and usage-related factors, including puncture difficulty, repeated puncture at the same site, access frequency, handling procedures, dressing technique, port-site management, and patient self-care, may also differ between approaches. In addition, device-related factors, such as port size, septum structure, puncture surface area, device profile, and leakage after repeated puncture, may influence local inflammation or infection risk. Because these factors were not systematically recorded in this retrospective study, the higher infection-related removal rate in the upper arm group should be interpreted cautiously and should not be attributed solely to the implantation approach. Catheter fracture was observed only in the subclavian group, and fracture-free survival was significantly lower in that group. Catheter fracture associated with the subclavian approach has classically been attributed to pinch-off syndrome, in which the catheter is compressed between the clavicle and the first rib [15, 22-24]. However, in our cohort, 10 of the 12 fractures occurred lateral to the clavicle–first rib crossing point. This distribution suggests that mechanisms other than classic pinch-off alone may have contributed to catheter fracture. Because our unit uses ultrasound-guided puncture rather than the landmark technique, the puncture site may be positioned more laterally, potentially allowing the catheter to traverse the pectoral musculature from a more lateral course. Repeated upper limb motion may therefore impose focal mechanical stress on the catheter and contribute to structural failure over time. This interpretation remains speculative, but the observed fracture distribution supports the possibility of a mechanism distinct from classic medial pinch-off alone. In our review of the catheter fracture cases, the median time from implantation to catheter fracture was 916.5 days (range, 453–1,798 days), suggesting that most fractures were delayed events rather than immediate procedural failures. Fractures occurred along the catheter shaft in seven cases and near the catheter–port junction in five cases. A history of power injection through the port system was documented in two cases. These findings suggest that fracture may reflect a combination of repeated bending or compression along the catheter course, stress concentration at the catheter–port connection, internal pressure stress in selected cases, and material fatigue, rather than classic pinch-off syndrome alone. Catheter material and device design may also have influenced the risk of catheter fracture. Based on available product information, BardPort used a silicone catheter, whereas PowerPort Implantable Port, PowerPort Slim Implantable Port, DewX Eterna, and P-U Celsite Port used polyurethane catheters. Among the 12 catheter fracture cases, 11 involved PowerPort Implantable Port systems with polyurethane catheters, and one involved BardPort with a silicone Groshong catheter. Silicone and polyurethane catheters differ in flexibility, tensile strength, and resistance to repeated bending or compression, and these material properties may affect long-term mechanical durability. However, because device selection was strongly related to implantation approach, these findings are descriptive and do not establish that polyurethane catheters or PowerPort systems were independently associated with fracture. The difference in implantation indication between the two groups should also be considered. The proportion of ports implanted for nutritional support was higher in the subclavian group, whereas the proportion implanted for chemotherapy was higher in the upper arm group. This imbalance was adjusted for in the primary regression analysis, and exploratory subgroup analyses stratified by implantation indication were also performed. However, because the implantation approach was selected according to clinical judgment rather than random allocation, residual selection bias and confounding related to clinical decision-making, indication for port placement, and clinical use may remain. The present study has several limitations. First, this was a retrospective single-center study, and the implantation approach was selected according to clinical judgment rather than random assignment. Although multivariable adjustment and exploratory subgroup analyses stratified by implantation indication were performed, residual selection bias and confounding related to clinical decision-making, indication for port placement, expected duration of use, access frequency, and port management may remain. Propensity score–based methods, such as matching or weighting, might have further improved comparability between groups but were not performed in this exploratory retrospective study. Second, although the distribution of CV port systems was summarized in Table 1 and catheter materials were identified based on available product information, device-related factors, including catheter material, device design, and catheter–port connection structure, were not included in the adjusted model. Because device selection was strongly related to implantation approach and device availability, residual device-related confounding cannot be excluded, particularly for catheter fracture. Third, complication-related removal was assessed based on events documented in our clinical setting, and complications managed entirely elsewhere may not have been captured. Fourth, comparative analyses were restricted to a maximum of 5 years after implantation to improve comparability between groups. Therefore, the present findings should be interpreted as reflecting medium-term rather than very late complication-related outcomes. In conclusion, within 5 years after CV port implantation, the overall rate of complication-related port removal was similar between the subclavian and upper arm approaches. However, the pattern of complications differed: infection-related removal was more frequent in the upper arm group, whereas catheter fracture was more frequent in the subclavian group. These findings suggest that the choice of implantation approach may influence the pattern of cause-specific complications rather than the overall complication-related removal rate. From a clinical perspective, this may support individualized selection of the implantation approach and tailored post-implantation surveillance strategies: infection-prevention measures should be emphasized for upper arm ports, whereas attention to catheter course, mechanical stress, and long-term monitoring for catheter fracture may be particularly important for subclavian ports. Author Contributions: Guarantor of the integrity of the entire study: SK. Study concepts and design: TT, SK, YT. Literature research: TT, SK. Clinical studies: SK, HT, KS, ME, TYos, HY, TYok, ATT. Experimental studies/data analysis: TT, YSe, HA, YSa, TS. Statistical analysis: TT. Manuscript preparation: TT. Manuscript editing: All authors. Conflicts of Interest: There are no conflicts of interest. Supplementary Material