Treatment of central venous catheter‐associated deep venous thrombosis in cancer patients with rivaroxaban

To the Editor: Central venous catheters (CVC) are an important tool in ongoing cancer therapy. However, central venous catheter-related, upper extremity deep venous thrombosis (CVC-UEDVT) is a common complication in patients with cancer 1. These thromboses often lead to loss of the CVC, which presents an obstacle to ongoing treatment 2. There is little prospective data and limited clinical trials on appropriate anticoagulation management of CVC thrombosis. However, low molecular weight heparin (LMWH) has been recommended and routine removal of the CVC has not been recommended 1-5. There are no data available yet about the use of rivaroxaban for CVC-UEDVT. Since January 1, 2014, all rivaroxaban use at Memorial Sloan Kettering Cancer Center (MSKCC) is being monitored under an Internal Review Board (IRB) approved Quality Assessment Initiative. For this analysis, we have identified all patients with active cancer and a CVC-UEDVT from January 1, 2014 through February 24, 2016, treated with rivaroxaban. A CVC-UEDVT was identified by imaging study (CT, MRI, and/or ultrasound) and related clinical notes. This was a retrospective analysis. The primary endpoint was preservation of line function through 90 days. Secondary endpoints included removal of central line for other medical reason, major bleeding (MB), clinically relevant non-major bleeding leading to discontinuation of rivaroxaban, death, and development of other venous thromboembolic event. Patients were censored if they reached an endpoint, or discontinued rivaroxaban prior to completion of 90 days. All clinical notes were reviewed by a combination of automated text search with predefined terms, followed by review by a study physician. During the study period, we identified a key cohort of 83 patients with active cancer and a CVC-UEDVT, in whom the central line was present and functional at initiation of rivaroxaban. Patients whose CVC had been removed prior to initiation of anticoagulation were not included in this cohort. Table 1 summarizes the baseline characteristics of the 83 patients. Forty one of the CVC-UEDVT events were identified incidentally on routine imaging studies and 42 were identified following symptoms (N = 40) or line dysfunction (N = 2). Most of the thromboses were found in an internal jugular vein (N = 37), the superior vena cava (N = 16), and subclavian vein (N = 10). An indwelling port was the predominant central access device. In 55 patients rivaroxaban was the sole anticoagulant and 5 additional patients received less than 7 days of LMWH prior to transition to rivaroxaban. The remaining patients transitioned to rivaroxaban after initial anticoagulation with LMWH. In three patients, the CVC-UEDVT developed in patients already on another anticoagulant for either atrial fibrillation or a previous thromboembolic event, and the patients were transitioned to rivaroxaban when the new CVC-UEDVT was identified. The majority of patients were in an advanced cancer stage; 60% stage IV and 13% stage III. Our analysis focused on the 90-day period after the thrombosis. Within the 90-day period, in only three patients was the CVC line removed due to development of line dysfunction. These three patients developed inability to aspirate from a Port type central line on day 15, 20, and 36 of rivaroxaban anticoagulation. Fifty-three patients (64%) completed a follow-up time of 90 days without the removal of their central line, or reaching another endpoint (Table 1). In addition, nine other patients had their CVC lines removed within the 90-day period, but not due to line failure. These were for end of cancer treatment (N = 6), infection (N = 1), thrombocytopenia (N = 1), and patient preference (N = 1). Other primary endpoints of note are listed in Table 1(B), including six deaths, three new VTE at other sites, two major bleeds, and one clinically relevant non-major bleeding leading to discontinuation of rivaroxaban. In this single institutional experience, rivaroxaban appears to be a good choice for treatment of a CVC-UEDVT. The failure rate at three months of treatment with rivaroxaban in this cohort is low, with only 3 patients out of 83 (3.6%) requiring CVC line removal due to development of line dysfunction. The overall rate of CVC line removal for any cause in our rivaroxaban cohort was 12 of 83 (14%). Our cohort study does not lend itself to direct comparisons with previous reports. With that limitation in mind, in the previously published Catheter Study of LMWH followed by warfarin, the overall rate of CVC line removal was 43% 3. The safety profile of rivaroxaban use for CVC-UEDVT was encouraging. Major bleeding events occurred in two patients treated with rivaroxaban, with an estimate of 2.4%. In The Catheter Study and the upper-Extremity DVT arm of the RIETE trial, major bleeding was reported in 10.9% and 2.1%, respectively 3, 6. Overall the safety and efficacy of rivaroxaban use in patients with active cancer for treatment of central venous catheters associated upper extremity deep venous thrombosis is very favorable in this single institutional cohort. Nevertheless, randomized controlled trials are needed to confirm these results. Eva S. Laube,1 Simon Mantha,1 Patrick Samedy,2 Jonathan Wills,3 Stephen Harnicar,4 and Gerald A. Soff1* 1Hematology Service; 2Division of Quality and Safety; 3Data Solution Group; 4Pharmacy Department, Memorial Sloan Kettering Cancer Center, New York, New York