Early Removal of Central Venous Catheters Does Not Increase Risk of PE

Among patients with a hematologic malignancy and central venous catheter–associated upper extremity deep vein thrombosis (UEDVT), removal of catheters within 48 hours did not increase the risk of pulmonary embolism (PE), compared with delayed or no removal, according to a study presented at the 2019 ASH Annual Meeting by Damon E. Houghton, MD, from the Mayo Clinic in Rochester, Minnesota.

Catheter-related thrombosis is a known complication of these devices in patients with cancer, Dr. Houghton said during his presentation. While guidelines recommend anticoagulation for at least 3 months, immediate catheter removal may be required in certain circumstances, and “it is unknown if catheter removal causes PE, or whether a short course of anticoagulation prior to removal decreases that risk,” he explained.

In this multicenter, retrospective cohort study, researchers reviewed data from patients with hematologic malignancies who had experienced a catheter-related UEDVT between 2010 and 2016 and were treated at one of 10 institutions participating in VENUS (the Venous thromboEmbolism Network U.S.). Patients with UEDVT that was considered not catheter-provoked or was concurrent with lower-extremity DVT and/or PE were excluded from the analysis.

Dr. Houghton and researchers recorded anticoagulant start and finish dates, as well as the time of the catheter removal, total follow-up, and death. Study outcomes included PE and all-cause mortality within 7 days of catheter-related thrombosis (CRT) diagnosis.

Of 663 patients with CRT identified, 151 patients received no anticoagulation, including 119 who underwent catheter removal alone as treatment for DVT. The remaining 512 started anticoagulation at the time of their diagnosis, with either:

  • delayed (>48 hour) or no catheter removal (n=200; mean age = 52 years)
  • early (<48 hours) catheter removal (n=312; mean age = 53 years)

Age and sex did not differ between patients with early versus delayed or no catheter removal, however, the type of hematologic malignancy (p=0.015), type of central catheter (p<0.001), and DVT location (p=004) were significantly different between the groups. In the delayed or no removal cohort, the most common hematologic malignancy was lymphoma (48%; n=96); in the early removal cohort, the most common diagnosis was leukemia/myelodysplastic syndromes (MDS; 45%; n=140). Patients with peripherally inserted central catheter (PICC) lines also appeared to be more likely to have early catheter removal (71% vs. 49%).

In those receiving anticoagulation, the most commonly prescribed agents were low-molecular-weight heparins, and the authors noted that anticoagulation treatment did not differ between groups. Median platelet counts did not significantly differ among the anticoagulation cohorts but were lower in patients treated with catheter removal only and no anticoagulation.

In the primary analysis (delayed vs. early catheter removal), there was no significant difference in incidence of PE with either approach:

  • early catheter removal: 2 patients (0.64%)
  • delayed or no catheter removal: 1 patient (0.5%; p=1.0 for comparison)

The same was true for rates of PE and all-cause mortality (secondary endpoint): 3 patients in each group (1.5% of delayed cohort and 1.0.% of early cohort) died or experienced a PE within 7 days of UEDVT diagnosis (p=0.68). All 3 patients with PE within 7 days had PICC lines, leukemia/MDS, and the sites of most proximal DVT involvement were brachiocephalic veins (n=2) and subclavian vein (n=1), the authors noted.

Among patients who did not receive anticoagulation and underwent catheter removal only, there were no PEs within 7 days, although 3 patients (2.5%) in this cohort died.

Overall, the risk of symptomatic PE after central venous catheter removal is low in patients who received anticoagulation and in those who did not receive anticoagulation because of a high risk for bleeding or contraindication.

However, the implications of these findings are limited by the retrospective design and the lack of imaging to identify subclinical PE after catheter removal. “Furthermore,” Dr. Houghton noted, “our study only included patients with hematologic malignancies. Therefore, the generalizability of our results to other cancer patients or to patients without cancer might be limited.”

When asked by an audience member about the implications for catheter removal in the clinic, Dr. Houghton responded that these results, as well as results from ongoing studies by this research group, suggest that there is not necessarily a need to remove these catheters. “Patients could receive anticoagulation, and that decision to remove the catheter could very well be postponed,” he said. Conditions for removal include if the catheter is not working, if there is a concern for infection, or if the line is malpositioned.

The authors report no relevant conflicts of interest.


Houghton DE, Billett HHH, Gaddh M, et al. Optimal timing for removal of an upper extremity central catheter when associated with a deep vein thrombosis: a venous thromboembolism network US multicenter retrospective cohort study. Abstract #325. Presented at the 2019 American Society of Hematology Annual Meeting, December 7, 2019; Orlando, FL.