Peri-Procedural Anticoagulation Interruption in Acute Cancer Associated Thrombosis

Background

Cancer associated thrombosis (CAT) is a frequent complication in patients with malignancy and is associated with higher rates of breakthrough thrombosis and bleeding during anticoagulation compared to the general population. Performing procedures in patients with CAT reflects a high-risk period for bleeding and thrombotic complications. There is paucity in the literature to guide clinical decision-making regarding outcomes in patients with acute CAT that require urgent procedures or surgery. We aimed to examine the bleeding and recurrent thrombosis outcomes of peri-procedural anticoagulation interruption and whether certain exposures were predictors of these adverse outcomes in acute CAT.

Methods

A retrospective cohort study was performed at Beth Israel Deaconess Medical Center to examine outcomes of peri-procedural anticoagulation interruption in acute CAT. Patients with malignancies who developed a new venous thromboembolic (VTE) event treated with anticoagulation between 1/1/2017 and 12/19/2023 and subsequently underwent a procedure within 30 days of the index VTE were included. Patients who had superficial or catheter associated thrombosis were excluded, along with those who underwent procedures related to their index VTE or a bleeding complication.

Outcomes included the cumulative incidence of recurrent VTE or bleeding events within 30 days of the index procedure. Bleeds were classified, per ISTH criteria for surgical bleeding, as major bleeding, clinically relevant non major bleeding and a composite of both: clinically relevant bleeding (CRB). Data was extracted by manual chart review by two trained abstractors. Procedures were classified by as low, moderate or high risk per ISTH Cancer Procedure Guidance Statements (Wang et al, JTH 2022). We employed univariate cause-specific Cox proportional hazard regression to assess the associations between exposures of interest (sex, procedure risk, cancer site, anticoagulant agent used and the interval the anticoagulant was stopped prior to procedure) and CRB.

Results

Of 75 patients included the median age was 69 years (IQR 62, 77.5) and 59% were female. The malignancies included were 21% (n=16) genitourinary/gynecologic (GYN-GU), 21% (n=16) pancreatic/hepatobiliary, 19% (n=14) luminal gastrointestinal, 15% (n=11) lung, 12% (n=9) hematologic and 12% (n=9) other cancers. There were 41% (n=31) low risk, 37% (n=28) moderate risk and 21% (n=16) high risk procedures performed. Heparin infusion (75%, n=56) was the most frequent anticoagulation used prior to the index procedure followed by therapeutic enoxaparin (21% n=16). Anticoagulation was held anywhere from less than 12 hours (76%, n=57), one day (21%, n=16) and >1 day (3%, n=2) prior to procedure. One inferior vena cava (IVC) filter was placed for the acute VTE prior to the procedure.

There were 2 recurrent thromboses (a pulmonary embolism (PE) at 2 days after the index procedure and a case of deep vein thrombosis and PE at 29 days) and 14 clinically relevant bleeding events (these included 4 clinically relevant non-major bleeds such as genital or intramuscular bleed, as well as 10 major bleeds with the most common being upper GI (n=6) and retroperitoneal (n=2) bleed). The median time to CRB was 3 days (IQR 1, 6). The cumulative incidence of CRB was 19% (CI 11%-28%) and recurrent thrombosis was 1.3% (CI 0.11%-6.5%) at 30 days.

Patients with hematologic malignancy had not experienced any CRBs within 30 days unlike the solid cancers (P <0.001), the sub-groups of which had similar risk of CRB. Use of heparin infusion prior to index procedure was associated with a higher risk of bleed compared to other anticoagulants which were not associated with any CRB (p<0.001). There was no difference in bleed when anticoagulants were held for greater than 12 hours (HR 1.3, CI 0.42-4.05; p=0.7) compared when the last dose was given less than 12 hours prior to procedure.

Conclusion

Patients with acute CAT requiring urgent procedures within 30 days of VTE diagnosis had high rates of post-procedural CRB including major bleeding. A factor associated with increased CRB risk was the use of heparin infusion prior to the procedure and underlying hematologic malignancy was associated with lower risk of CRB. These results support the need to develop individualized procedural protocols to manage anticoagulation during this particularly high risk period in patients with CAT.