Arm-inserted catheters may outperform neck catheters for leukemia chemotherapy support

When a patient with acute myeloid leukemia begins remission induction chemotherapy, one of the first practical decisions is not about the drugs at all. It is about the tube. Every one of these patients needs a central venous catheter, a device that will deliver cytotoxic chemotherapy, enable daily blood sampling, carry transfusions of red cells and platelets, support weeks of intravenous antibiotics, and, when necessary, provide parenteral nutrition. The catheter is the physical infrastructure of supportive care, and its failure can interrupt treatment, force additional invasive procedures, prolong hospitalization, and delay the leukemia therapy itself. A new retrospective study from LMU University Hospital in Munich, published in Supportive Care in Cancer, now offers some of the clearest real-world evidence yet on which type of catheter serves AML patients best during this vulnerable phase.

The Munich team compared the two dominant catheter options. Centrally inserted central catheters, known as CICCs, are the traditional choice, placed through the internal jugular or subclavian vein in the neck or upper chest. Peripherally inserted central catheters, or PICCs, are the newer alternative, threaded into the basilic or brachial vein of the upper arm and advanced until the tip rests at the cavoatrial junction or in the lower third of the superior vena cava. Because PICCs avoid puncturing the neck or thorax, patients can be discharged with the device in place and continue treatment as outpatients, which is one reason they are often preferred by patients themselves. Yet the evidence comparing the two in AML has been limited and inconsistent, with some studies reporting lower infection and thrombosis rates for PICCs and others finding a significantly increased risk of catheter-related thrombosis.

The researchers analyzed 116 adults with newly diagnosed AML who received intensive remission induction chemotherapy between June 2019 and March 2024. Fifty-nine patients received a PICC and 57 received a CICC. Patients with acute promyelocytic leukemia, additional catheters such as ports, or other central devices were excluded, and only the first catheter was considered. The observation period ran from the day of placement to the day of removal. Catheter type was chosen based on patient preference, feasibility, and physician judgment rather than randomization, which the authors acknowledge as a key limitation. All insertions followed standardized institutional protocols with maximal sterile barrier precautions, including surgical hand disinfection, sterile gloves and long-sleeved gowns, caps and face masks, alcohol- or iodine-based skin antisepsis, chlorhexidine-impregnated dressings, and real-time ultrasound guidance using the Seldinger technique.

The technical details of insertion differed in ways that matter. PICCs were placed exclusively in the radiology department, while CICCs were inserted across various clinical settings, including intensive and intermediate care units, anesthesiology, and the hematology ward. The internal jugular vein was used in all but one CICC patient, and most catheters in both groups went in on the right side. For fixation, CICCs received sutures while PICCs used a sutureless securement device. Catheter tip position was confirmed by electrocardiography and chest X-ray. Notably, patients in the CICC group had higher C-reactive protein levels on the day of insertion, and significantly more of them received at least one platelet transfusion before the procedure, 36.8 percent versus 16.9 percent in the PICC group, despite comparable platelet counts between the groups.

The headline finding concerns dwell time. PICCs remained in place for a median of 16 days, significantly longer than the 11 days for CICCs, a difference with a p-value of 0.004. Because catheters that stay in place longer without complications reduce the need for replacement procedures, the authors argue this likely represents a clinically relevant advantage for PICCs. The interval from insertion to the start of chemotherapy was also shorter for CICC patients, one day versus three days, reflecting the faster availability of bedside insertion when treatment needed to begin urgently.

On the critical safety question of catheter-related bloodstream infections, the two catheter types performed essentially identically. CRBSI rates were 2.8 per 1000 catheter days for PICCs and 2.6 per 1000 catheter days for CICCs, with an incidence rate ratio of 1.08 and a p-value of 0.93. Only five CRBSI events occurred in the entire cohort, all caused by Gram-positive bacteria. In the PICC group, Staphylococcus haemolyticus was identified in two patients and Staphylococcus aureus in one, while both CICC infections were caused by Staphylococcus epidermidis. The diagnosis followed the 2012 criteria of the AGIHO working party of the German Society of Hematology and Medical Oncology, counting both definite cases, where identical pathogens were cultured from catheter tip and peripheral blood or where the differential time to positivity exceeded two hours, and probable cases involving typical catheter-associated pathogens without an alternative source.

The most striking secondary result was the rate of exit site infections, defined as clinical inflammation at the insertion site with erythema, warmth, swelling, pain, or purulent discharge. These were significantly less common with PICCs, at 7.4 per 1000 catheter days versus 23.1 for CICCs, a p-value of 0.01. The authors point to earlier studies showing higher skin and catheter colonization rates in the cervical and thoracic area, which may explain why neck-inserted catheters develop more local site complications. Given the profound neutropenia that AML induction therapy produces, with absolute neutrophil counts below 500 per microliter for extended periods, reducing exit site infection risk carries real clinical weight. Interestingly, the higher exit site infection rate in the CICC group did not translate into more bloodstream infections, probably because those catheters were promptly removed whenever local signs appeared.

Bleeding events were more frequent in the PICC group, 13.9 versus 6.4 per 1000 catheter days, but the difference was not statistically significant, and all but one event were minor, mostly documented directly after insertion. The authors interpret the heavier platelet transfusion use before CICC insertion as a more cautious approach by clinicians who perceive higher bleeding risk in the cervical region. While transfusions increase procedural safety, they also consume platelet concentrates, so catheter choice has implications for transfusion stewardship and healthcare resource utilization, not just patient safety. Mechanical complications were rare in both groups, with no pneumothorax reported, one case of nerve injury in the PICC group, and two cases of symptomatic catheter-related venous thrombosis, also in the PICC group.

The thrombosis finding deserves careful interpretation. Because the study was retrospective and thrombosis screening was performed only in symptomatic patients, asymptomatic clots may have gone undetected, and previous literature, including a systematic review and meta-analysis in the Lancet, has described a higher incidence of venous thromboembolism with PICCs. All patients consistently received prophylactic anticoagulation with low-molecular-weight heparin when platelet counts exceeded 30 times 10 to the ninth per liter, which may have reduced thrombotic risk. The authors also note a practical constraint: PICCs, which typically have one or two lumens, may be inadequate when clinically stable patients deteriorate and need rapid multi-infusion access, a situation relevant because up to 20 to 30 percent of AML induction patients require at least one intensive care stay.

Exploratory univariable Poisson regression found no baseline characteristic significantly associated with CRBSI, though trends emerged toward higher infection rates in patients with adverse-risk AML by the European LeukemiaNet 2017 classification, fever before insertion, and age above 65. With only five events, multivariable analysis was not feasible. The authors conclude that PICCs are a good alternative for clinically stable patients undergoing planned AML remission induction chemotherapy, offering longer dwell times and fewer exit site infections without increased bloodstream infection risk. Ultimately, they argue, the choice of catheter should be individualized and guided by the patient’s clinical condition and anticipated supportive care requirements. As hypothesis-generating evidence from a single center, the findings now await confirmation in larger prospective studies, but they suggest that for many AML patients, the arm may be the better entry point.

Subject of Research: Catheter choice in AML induction chemotherapy

Article Title: Catheter-related complications in AML remission induction chemotherapy: a retrospective comparison of PICCs and CICCs

Article References: Ehrlich, S., Aytan, F., Eufinger, J., Ruzicka, M., Doedens, D., Seidensticker, M., Herold, T., Subklewe, M., Bücklein, V., von Bergwelt-Baildon, M., & Spiekermann, K. (2026). Catheter-related complications in AML remission induction chemotherapy: a retrospective comparison of PICCs and CICCs. Supportive Care in Cancer, 34(9), Article 906. https://doi.org/10.1007/s00520-026-11146-3

Image Credits: AI Generated

DOI: 10.1007/s00520-026-11146-3

Keywords: AML, PICC, CICC, central venous catheter, catheter-related bloodstream infection, supportive care in cancer, Catheter-related, complications, remission, induction, chemotherapy, retrospective