Catheter-related thrombosis in cancer patients with tunnelled non-cuffed centrally inserted central catheters: a retrospective cohort study

Abstract

Background

Peripherally inserted central catheters (PICCs) are widely used in oncology but are associated with higher rates of catheter-related thrombosis compared to other central venous access devices. Tunnelled non-cuffed centrally inserted central catheters (tnc-CICC) may offer a safer alternative in patient at high risk of thrombosis.

Methods

This retrospective cohort study evaluated 193 tnc-CICCs placed in 182 cancer patients at a tertiary oncology unit between January 2021 and December 2022. Primary outcomes is catheter-related thrombosis; secondary outcomes are procedural duration and complication rates. Data were analysed using non-parametric statistical methods.

Results

The overall thrombosis rate was 1% (0.18 per 1,000 catheter days), significantly lower than reported PICC-associated thrombosis rates. No major intraoperative complications occurred. The overall complication rate was 19.7%, with catheter malfunction (9.3%), migration (5.2%) and bloodstream infections (4.1%) being the most common. Median procedural time was 34 min, with no significant difference based on operator experience.

Conclusions

Off label use of PICCs as tnc-CICCs seems safe and effective. It provides an alternative to PICCs in oncology patients at high risk of thrombosis, even when placed by operators with low experience levels. These findings support the inclusion of tnc-CICCs in vascular access algorithms for cancer patient at high risk of thrombosis and suggest feasibility for nurse-led insertion programs.

Background

Central venous access is essential for cancer patients receiving chemotherapy, as it minimizes caustic damage to peripheral veins and facilitates frequent blood sampling. Unfortunately, central venous access is associated with thrombosis. Catheter-related venous thrombosis leads to treatment delays, a reduction in patient quality of life and an increase in healthcare costs [1–4]. Compared to totally implantable venous access devise and tunnelled cuffed centrally inserted central catheters, peripherally inserted central catheters (PICC) have a higher risk of thrombosis [5]. This is caused by the small vein size at the insertion site and the long and winding intravascular tractjectory [5]. Still, many oncologists prefer PICCs as central venous access device because of the ease of insertion as well as removal and the low costs [3, 6].

A possible strategy to reduce the risk of thrombosis while maintaining many of the advantages of a PICC is the “off-label” use of PICCs, by inserting them as tunnelled non-cuffed centrally inserted central catheters (tnc-CICCs), also known as ‘atypical’ PICC [7, 8]. This technique was first described in small infants to avoid the small peripheral veins [7, 9, 10]. More recently, 2 small cohort studies in adults showed that the technique is safe, well accepted and has low complication rates [11, 12]. Lastly, one study by Tran et al. showed a reduction in catheter-related thrombosis when using tnc-CICCs compared to PICCs in a haematology population [4]. Yet, the impact of tnc-CICCs on thrombosis in other types of malignancies has not been reported.

This retrospective study aims to evaluate tnc-CICC use in a specific oncology and haematology population at high risk of thrombosis at a tertiary oncology referral unit. We hypothesise that tnc-CICCs are less prone to venous thrombosis and mechanical problems, with little increase in technical complications. This cohort study is reported in line with the STROCSS guidelines [13].

Methods and materials

Population and setting

Data for this retrospective cohort study were collected from January 2021 until December 2022 from the University Hospitals Leuven electronic patient records. The University Hospitals Leuven is a large tertiary academic teaching facility, running a dedicated multidisciplinary and interprofessional vascular access team, consisting of surgical oncologists, vascular surgeons, interventional radiologists and specialist nurses. This team inserts approximately 1500 PICCs and 250 tunnelled cuffed and non-cuffed catheters yearly. Complications are managed by a vascular access nurse-specialist team.

tnc-CICCs are mainly inserted for 2 indications. The first indication is autologous stem cell transplantation, for which a large-bore (6 French) double-lumen central catheter is needed to ensure rapid infusion of stem cells [4, 5]. The second indication is ovarian cancer and germline tumours, which are both tumour types associated with a high risk of venous thrombosis [5, 14, 15]. Of note, a third indication, which is not tumour type-related, is the inaccessibility of peripheral veins for PICC insertion.

Adult cancer patients who underwent a tnc-CICC insertion were included into this trial. Patients under 18 years of age, catheters inserted in the inferior vena cava and patients who were not treated for a malignancy were excluded.

Data collection and definitions

Data on patient demographics (age, BMI, diagnosis), insertion (insertion side, operator experience), device (size, brand, number of lumina) and catheter dwell time were collected. Study data were collected and managed using REDCap electronic data capture tools hosted at Yale University [16, 17].

The primary outcome is catheter-related venous thrombosis, defined as any symptomatic thrombotic event in a vein containing a catheter. In case of a thrombosis, patient files were also screened for embolic events. Secondary outcomes are procedural duration and intra- and postoperative complications. Procedural time was defined as the time from door-to-door, i.e. including patient transfer and preparation. Patient files were screened for any complication related to the tnc-CICC such as but not limited to haematoma, pneumothorax, haemothorax, thrombosis, infection, catheter migration and catheter malfunction. Catheter function was evaluated using the ‘Catheter Injection and Aspiration” (CINAS) classification tool [18]. Diagnosis of central line-associated blood stream infection was based on the description of any infectious episode attributed to the presence of the catheter by the treating physicians. Differential time to positivity was not available from the patients’ files, so a confirmation of catheter-related blood stream infection could not be made. Catheter malposition was defined as an abnormal tip position. The surgical experience of the operator level was divided into 3 levels: surgical trainees have a maximum of 6 years of surgical experience, fellows 7–8 years and senior staff > 8 years. Most surgical trainees have limited to no experience in vascular access surgery, when joining the department of Surgical Oncology. Before they are allowed to insert catheters, they get a theoretical and dry lab training under supervision of an experienced operator.

Surgical technique

The applied technique is like the techniques described by Bernasconi or Lawson in children [7, 10]. Catheters are inserted under local anaesthesia, in a dedicated operating room. No antibiotics are administered. A single-lumen (Medcomp^®, Pro-PICC CT, 5 French) or double lumen (BBraun^®, Celsite PICC-cel, 6 French/ Medcomp^®, Pro-PICC CT, 5 French) polyurethane catheter was used. The internal jugular vein or brachiocephalic vein is approached using an in-plane short-axis real time ultra-sound guided vein puncture. The catheter is tunnelled from the infraclavicular region (exit site) to the venotomy site, using a BD Angiocath^® 10 Gauge needle with an antegrade technique (Fig. 1A, B). A modified Seldinger technique is used to insert the catheter into the vein using a micro-introducer set. Guided by intra-operative fluoroscopy, the catheter tip is positioned two vertebrae below the bronchial carina (Fig. 1C) [19]. The catheter is then secured with an appropriately sized subcutaneous anchoring device (Securacath^®)(Fig. 1D) and the exit site is covered with a gauze dressing. The neck incision is closed using steristrips (Fig. 1E). Catheters are locked with a 0.9% sodium chloride solution. No peri-operative anti-coagulants are administered. The postoperative care is the same as described in guidelines for PICC [20].

Fig. 1.

tnc-CICC insertion technique, (A, B) Tunnelling technique, (C) Fluoroscopic tip positioning, (D) Subcutaneous anchoring device, (E) Dressing

Statistical analysis

Given the retrospective nature of the study, no sample size calculation was performed. Statistical analysis was conducted using SPSS (IBM v29). Categorical variables are presented as frequencies and percentages, and continuous variables as medians with interquartile range (Q1-Q3). Group comparisons on continuous variables were performed using the Kruskal Wallis test for the comparison of more than 2 groups. The analysis was performed on the number of devices.

Ethical approval

Ethical approval was obtained on 02-06-2022 from the University hospitals Leuven Ethics Committee for research (S66620). The study was performed in accordance with relevant guidelines and regulations as described in the declaration of Helsinki. Due to the retrospective nature of the study, no informed consent was attained in concordance with the advice of our ethics committee.

Results and discussion

A total of 193 tnc-CICCs were inserted in 182 patients. Eleven out of 182 patients (6%) underwent a second procedure, 6 of which occurred during the first treatment episode due to catheter-related complications (n = 2), inadvertent dislodgement (n = 3), or erroneous catheter type placement (single-lumen instead of double-lumen) (n = 1). An additional five patients required a second treatment episode for disease recurrence or tandem autologous stem cell transplantation.

Patient, device and procedural characteristics (Table 1)

Table 1.

Patient, device and procedural characteristics

Patient and device characteristics
Variables	Stem cel transplantation n (%)	Chemotherapy, n (%)	Total, n (%)
Gender (N = 182)	85 (46,7)	97 (53,3)
Female	32 (37,7)	82 (84,5)	114 (62,6)
Male	53 (62,3)	15 (15,5)	68 (37,4)
Diagnosis (N = 182)
Hematologic malignancy			85 (46,7)
Multiple myeloma	49 (59,8)
Acute lymphoma	29 (35,4)	3 (3,1)
Monoclonal Gammopathy	2 (2,4)
Plasmacytoma	2 (2,4)
Solid tumour			97 (53,3)
Gynaecologic	2 (2,4)	73 (75,3)	75 (41,2)
Germ line (seminoma/non-seminoma)	1 (1,2)	14 (14,4)	15 (8,2)
Gastro-intestinal		1 (1)	1 (0.6)
Breast		6 (6,2)	6 (3.3)
Therapy (N = 182)
Autologous stem cell transplantation	85 (100)		85 (46,7)
Taxol-Carboplatin (q1w18)		40 (41,2)	40 (22,0)
Bleomycin-etoposide-cisplatin (q3w3)		15 (15,5)	15 (8,2)
Cisplatin (q1w6)		28 (28,9)	28 (15,4)
Other		14 (14,4)	14 (7,7)
Catheter type (N = 193)
Medcomp®, Pro-PICC CT, 5 Fr, Single lumen	1 (1,1)	98 (97)	99 (51.3)
BBraun®, Celsite PICC-cel, 6Fr, Double lumen	91 (98,9)	2 (2)	93 (48,2)
Medcomp®, Pro-PICC CT, 5 Fr, Double lumen		1 (1)	1 (0,5)
Side of insertion (N = 193)
Left	42 (45,7)	49 (48,5)	102 (52.8)
Right	50 (54,3)	52 (51,5)	91 (47.2)
Surgical experience (N = 193)
Staff	39 (42,4)	39 (38,6)	78 (40.4)
Fellow	33 (35,9)	27 (26,7)	60 (31.1)
Trainee	20 (21,7)	35 (34,7)	55 (28.5)

The median age was 30 years (44.8–66.1), with a median BMI of 25.6 kg/m² (23.1–29.2). Patients (N = 182) were grouped based on treatment indication: chemotherapy (n = 97) or autologous stem cell transplantation (n = 85). The chemotherapy group was predominantly female (84.5%), consistent with the high prevalence of gynaecological malignancies, while the transplantation group had a slight male majority (62.3%). Gynaecological or germ cell cancers accounted for 89.7% of chemotherapy cases, and 96.4% of transplantations were performed for haematological malignancies. Device selection aligned with clinical needs: single-lumen catheters were used in the chemotherapy group, while double-lumen catheters were placed for stem cell transplantation. The median catheter dwell time across the cohort was 39 days (Q1 21–Q3 83). As expected, dwell times differed significantly between the chemotherapy group (median 73 days, Q1 52 –Q3 128) and transplantation group (median 21 days, Q1 20 –Q3 24.5)(p < 0.001). A total of 11,318 catheter days were recorded.

Intraoperative outcomes and technical performance

Intraoperative outcomes were uniformly favourable. No major complications, such as arterial puncture, pneumothorax, or haemothorax occurred. All catheter tips were correctly positioned during insertion. These findings challenge long-standing perceptions that central catheter insertion is inherently more complex or risk-prone than PICC insertion [21]. We attribute this to several key factors: the use of real-time ultrasound guidance for vein puncture, standardized insertion protocols, rigorous operator training, and real-time tip position confirmation. These safeguards likely mitigated the risks typically associated with central venous access, even though many operators had limited experience.

A key advantage of PICCs has been their nurse-led placement, improving accessibility and reducing costs. However, the low complication rates in our cohort—even with inexperienced operators—suggest that, given proper training, nurses could safely perform CICC insertions as well. This is supported by the study of Alexandrou et al., who previously demonstrated that nurse-led CICC insertions are safe and even associated with lower infection rates than those performed by medical staff [22].

The median procedural duration was 34 min (Q1: 24, Q3: 41). The median operation duration was 31 min (23-38.3) for senior staff, 34.5 min (22-42.5) for fellows and 36 min (30–42) for trainees. There was no significant difference in procedural duration based on operator experience (p = 0.09).(Fig. 2) Procedure duration in our study was longer compared to those typically reported for PICC insertion, which range from 8 to 26 min [23–26]. This time difference may reflect the inherently greater complexity of tnc-CICC placement. However, several study-specific factors likely contributed to the extended times observed. Most reports measure skin-to-skin duration, whereas this study recorded door-to-door times, which include patient preparation and transfer. Additionally, the teaching hospital setting, operator inexperience, and cases involving concurrent procedures or priorly failed PICC attempts may have further prolonged procedural duration. Our Insertion protocol may be further optimised by using intracavitary ECG as tip positioning method better aligning with current guidelines. This may further shorten the insertion duration. We expect only slightly longer operation duration of tnc-CICC than PICC in comparable situations.

Fig. 2.

Comparison of procedural duration between operator experience

Postoperative complications and catheter safety (Table 2)

Table 2.

Complications

	Thrombosis n (%)	Migration n (%)	CLABSI n (%)	Malfunction n (%)	Total n (%)
Stem cell transplantation	0 (0)	8 (8.7)	2 (2.2)	3 (3,3)	13 (14.1)
Chemotherapy	2 (2)	2 (2.0)	6 (5.9)	15 (14.9)	25 (24.7)
Total	2 (1)	10 (5.2)	8 (4.1)	18 (9.3)	38 (19.7)

A total of 38 complications occurred in 37 catheter insertions, resulting in an overall complication rate of 3.4 per 1,000 catheter days. This complication rate of 19.7% is lower compared to the 30% overall complication rate reported by Chopra et al. with PICC in oncology patients [3]. Eleven catheters out of 193 catheters (5.7%) were removed before therapy completion due to dislodgement (n = 5), infection (n = 5), thrombosis (n = 1), mechanical failure (n = 1) and 1 case involving both thrombosis and migration. These numbers are lower than the reported rate of PICC failure in oncology patients, which is 12.5% [3, 27].

Catheter-related thrombosis occurred in 2 out of 193 catheters (1%), both in the chemotherapy group. This corresponds to an incidence of 0.18 per 1000 catheter days overall, and 0.22 per 1000 catheter days within the chemotherapy subgroup. The longer dwell time in the chemotherapy group may partially explain the higher observed incidence of venous thrombosis. One case involved a catheter tip thrombus complicated with endocarditis, requiring early catheter removal and anticoagulation. The second case presented with internal jugular vein thrombosis, manifesting as cervical swelling and pain, managed conservatively with anticoagulation therapy.

Globally, the study population represents a cohort with a particularly high baseline risk for catheter-related venous thrombosis, as all patients were actively receiving chemotherapy, which is known to increase thrombosis risk by up to 6.5-fold [28, 29]. Furthermore, the underlying malignancies were predominantly haematological, gynaecological, or germline in origin known to be associated with increased venous thrombotic risk. In haematological malignancies, thrombosis rates can reach up to 34% in the context of central venous access for stem cell transplantation [4, 30]. Gynaecological cancers, such as ovarian and endometrial carcinoma, are associated with venous thromboembolism rates of 11–25%, while germline tumours, especially those treated with cisplatin, have thrombosis rates of 10–15% [5, 15, 31]. Still, our thrombosis rate is lower than the reported PICC related thrombosis in adult oncology patients of up to 7.8% in haematologic patients and 2.2–5.9% in oncologic populations [3, 4, 27, 32]. Even when only considering the chemotherapy group, 0.2 thrombosis /1000 catheter days is a quarter of the reported thrombosis rates with PICC of 0.9/1000 catheter days a study by Bertoglio [27]. Although in recent years some technical improvements in catheter insertion techniques have been shown to reduce the risk of PICC related thrombosis to 3–6%, this remains higher than our tnc-CICC related thrombosis rate [21, 33, 34]. Similar to our results, Tran et al. reported a thrombosis rate of 0.4% with tnc-CICCs in a population of paediatric haematology patients [4]. These results support our hypothesis that tnc-CICCs lead to a lower risk of venous thrombosis compared to PICCs. The lower thrombosis risk is likely due to several factors. Insertion in large central veins with higher blood flow and a lower catheter-to-vein ratio reduces stasis and vessel wall irritation. Also, the shorter and less winding intravascular trajectory reduces vein wall damage. Lastly, the use of subcutaneous tunnelling provides mechanical stability and a barrier against infection—both known as contributors to thrombosis. Of note is that tunnelling PICCs also reduces the risk of thrombosis [35].

Bloodstream infections occurred at a rate of 0.7 per 1,000 catheter days in our study. This is lower than the rates reported for PICCs in cancer patients, typically ranging between 1.0 and 7.7 per 1,000 catheter days [3, 36]. In a similar population of haematology patients, Caris et al. reported an infection rate of 4.0 per 1000 catheter days [37]. Alexandrou et al. demonstrated that infections could be significantly reduced with structured insertion protocols and experienced operators, which very likely also contributed to the favourable outcomes in this study [22]. The tunnelled nature of tnc-CICCs probably provides a protective barrier against skin flora migration, further lowering infection risk.

In our cohort catheter malfunction was observed in 18 out of 193 catheters (9.3%), corresponding to 1.6 events per 1,000 catheter days. Malfunctions included 6 complete occlusions (IN3AS3) and 12 cases of withdrawal occlusion (11 IN1AS3, one IN1AS2). Four catheters were removed because therapy had been completed at the time of malfunction. All other malfunctions were successfully managed with thrombolytic therapy (urokinase). In comparison, literature reports PICCs to be associated with higher malfunction rates of up to 30% in oncology settings [1]. The use of larger bore catheters (5–6 Fr) compared to most studies on PICCs using 4 Fr catheters might in part explain this difference. No previous data on catheter malfunction in tnc-CICC has been published.

In this study, PICCs were used as tnc‑CICCs, constituting off‑label use, as the instructions for use specify peripheral insertion and do not address tunnelling. PICC lines are used since no dedicated non-cuffed catheter is available for longer-term central venous access. Although acute central venous catheters could be used, this would also be off-label as they are not validated for prolonged use, and standard CICC insertion kits do not include the micro-introducer systems preferred for the modified Seldinger technique. This deviation from the intended use may alter the device’s safety profile. The present study provides important data on the safety of this practice in a clinical setting. Under the European Medical Device Regulation (EU) 2017/745, such use places increased responsibility on healthcare professionals and institutions to ensure a well‑documented risk–benefit assessment, appropriate documentation, and adequate follow‑up.

This study has several methodological strengths. It represents a relatively large real-world cohort with detailed prospective data collection on both surgical outcomes and complications over more than 11,000 catheter days. The inclusion of high-risk oncology populations—such as patients with haematological, gynaecological, and germline malignancies—enhances the clinical relevance of our findings, particularly regarding thrombosis risk. Moreover, all procedures were performed using a standardized technique with ultrasound and real time tip positioning. However, the study also has limitations. It is a single-centre study, conducted in a teaching hospital setting, which may affect generalizability, especially regarding procedure duration and operator experience. The observational design limits causal inferences, and some complications very likely have been underreported, e.g. if they occurred after catheter removal or outside of the follow-up period. A last limitation is the time effect of the used techniques. During the study period fluoroscopy was used to determine the catheter tip position although nowadays intracavitary ECG is advised as primary tip positioning method. Intracavitary ECG is feasible with tnc-CICC insertion, and we would advise this technique over fluoroscopy in the future. While this part of the technique may be outdated, the tip position was accurately determined and did not influence the results of our study. Lastly the retrospective character of the study impairs direct comparison to PICC or t-PICC. Despite these limitations, our findings offer valuable insights into the safety and particularly the thrombosis profile of tnc-CICCs in high-risk oncology patients.

Many of the benefits historically ascribed to PICCs, such as outpatient feasibility, nurse-led insertion, and cost-effectiveness, are also achievable with tnc-CICCs, but with a significantly lower risk of catheter-related thrombosis. This suggests that tnc-CICCs could be the preferable venous access option in oncology populations at high risk of thrombosis. Therefore, further research should focus on a direct, prospective comparison between PICC/t PICC and tnc-CICC in cancer patients.

Conclusions

This study demonstrates that tnc-CICCs are safe and effective when placed using a standardised, ultrasound-guided technique with real time tip localisation, even when performed by less experienced operators. The overall complication and venous thrombosis rates were low, with excellent procedural success.

These findings suggest that tnc-CICCs could be considered as an alternative to PICCs. Therefore, we suggest adding tnc-CICCs to device selection algorithms for patient populations at elevated risk of thrombotic events. Moreover, the reproducibility of safe outcomes with structured training supports the viability of expanding tnc-CICC placement to nurse-led teams, mirroring current best practices with PICC. This paradigm shift has the potential to improve vascular access outcomes even further, while reducing healthcare costs and procedural variability.

Abbreviations

PICC

Peripherally inserted central catheter

tnc-CICC

Tunnelled non-cuffed centrally inserted central catheters

Authors’ contributions

All authors were responsible for conseptualisation and approved the final manuscript. TD was responsible for data collection and analysis. TD and LG were responsible for manuscript preparation.

Funding

No Funding was received for this study.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Ethical approval was obtained on 02-06-2022 from the UZL Ethics Committee for research (S66620). The study was performed in accordance with relevant guidelines and regulations. Due to the retrospective nature of the study, no informed consent was attained.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.