Addressing the burden of hospital‐related venous thromboembolism: the role of extended anticoagulant prophylaxis

Abstract

Hospitalization increases the risk of venous thromboembolism (VTE) by at least 8‐fold. The magnitude and duration of the VTE risk are modulated by patient‐ and disease‐specific factors in the medically ill, and by perioperative factors in surgical patients. The interaction among these risk factors is greatest when patients are in the hospital, which explains why the risk of VTE is highest in acutely ill medical patients, and in the immediate postoperative period in surgical patients. However, the risk of VTE persists for several weeks after hospital discharge in patients with comorbidities, such as heart failure, limb paresis or paralysis, or active cancer. Thus, in the USA, it is estimated that more than half of the 500 000 VTE events that occur each year are related to recent hospitalization or institutionalization 1.Heit J.A. Crusan D.J. Ashrani A.A. Petterson T.M. Bailey K.R. Effect of a near‐universal hospitalization‐based prophylaxis regimen on annual number of venous thromboembolism events in the US.Blood. 2017; 130: 109-14Crossref PubMed Scopus (74) Google Scholar. Current guidelines recommend routine in‐hospital thromboprophylaxis for most surgical and for at‐risk medical patients. Understanding why this approach fails to reduce out‐of‐hospital VTE is critical for developing better strategies to address this ongoing problem. There is ample evidence that in‐hospital anticoagulant prophylaxis reduces the risk of symptomatic VTE and fatal pulmonary embolism (PE). Although extended thromboprophylaxis is recommended after certain types of surgery, it is not currently recommended for medical patients because of concerns about bleeding. Is this the right approach? A recent study by Heit et al. suggests that it is not 1.Heit J.A. Crusan D.J. Ashrani A.A. Petterson T.M. Bailey K.R. Effect of a near‐universal hospitalization‐based prophylaxis regimen on annual number of venous thromboembolism events in the US.Blood. 2017; 130: 109-14Crossref PubMed Scopus (74) Google Scholar. By identifying Olmsted County residents with incident or recurrent VTE that occurred between 2005 and 2010 and determining how many of them were hospitalized during that time‐period, Heit and colleagues examined the effect of increased uptake of in‐hospital thromboprophylaxis on the risk of VTE. Although the use of in‐hospital thromboprophylaxis increased from about 20% to 90% over the 10‐year study period, there was no change in the incidence of in‐hospital VTE or VTE related to hospitalization, which was defined as VTE occurring from the inpatient period to 92 days after discharge 1.Heit J.A. Crusan D.J. Ashrani A.A. Petterson T.M. Bailey K.R. Effect of a near‐universal hospitalization‐based prophylaxis regimen on annual number of venous thromboembolism events in the US.Blood. 2017; 130: 109-14Crossref PubMed Scopus (74) Google Scholar. Furthermore, the study confirmed previous reports that even though the risk of VTE is highest when patients are in hospital, about 75% of VTEs related to hospitalization occurred after discharge and about half of these events manifested as PE. This is an important finding because the case‐fatality rate with PE is about twice that with deep vein thrombosis. It is unlikely that over‐diagnosis of PE because of the widespread use of multi‐detector computed tomography scanners was responsible for the high rate of PE after discharge because the community VTE rate was unchanged during this period. Therefore, near universal provision of in‐hospital thromboprophylaxis is less effective than expected, and we need to understand why. There are several potential reasons why near universal in‐hospital thromboprophylaxis does not eliminate VTE. First, the median duration of hospitalization was only 70 h. Therefore, many patients received an abbreviated course of thromboprophylaxis. It is not surprising that such a short course of anticoagulation therapy is less effective than the minimum 10‐day course used in randomized clinical trials examining the efficacy of prophylaxis in medically ill patients. Second, with the aging of the population, hospitalized patients are increasingly older and sicker. Thus, the comorbidity burden of hospitalized Olmsted County patients increased from 2005 to 2010, as evidenced by a significant increase in the Charlson and Elixhauser comorbidity indices. Sicker patients are at higher risk of VTE, and a short course of anticoagulant thromboprophylaxis may be insufficient to prevent VTE in those with ongoing risk factors. What can we learn from this study? First, it highlights the significant burden of hospital‐related VTE (Figure 1). The data suggest that hospitalization contributes to half of the 500 000 VTE events that occur each year in the USA, and of these 250 000 VTE events, about 200 000 occur after hospital discharge. Unfortunately, the distribution of patients hospitalized for medical illness or for surgery was not provided, but it is likely that the majority were medical patients. The failure of in‐hospital prophylaxis to reduce the burden of VTE after discharge probably reflects an inadequate duration of thromboprophylaxis in patients with ongoing risk factors. This leaves us with the question as to whether or not to extend anticoagulant thromboprophylaxis and, if so, how and in which patients. Most hospitalized patients have their anticoagulant prophylaxis stopped when they are discharged. Continuing thromboprophylaxis for 10 to 14 days is reasonable, but routine extended anticoagulant prophylaxis for up to 6 weeks has an even greater potential to reduce the burden of hospital‐related VTE if it can be easily implemented, is cost‐effective, and if there is evidence of net clinical benefit. The introduction of direct oral anticoagulants (DOACs) such as dabigatran, apixaban, rivaroxaban, edoxaban and betrixaban has the potential to simplify out‐of‐hospital anticoagulant prophylaxis because these agents can be given orally in fixed doses without routine coagulation monitoring. Nonetheless, robust evidence of a clear net benefit of extended thromboprophylaxis is limited and these agents are costly, albeit less expensive than low‐molecular‐weight heparin. There is little doubt that high‐risk surgical patients, such as those undergoing elective hip or knee arthroplasty, benefit from extending thromboprophylaxis for up to 45 days. Thus, a recent Cochrane meta‐analysis revealed that extended anticoagulant prophylaxis with DOACs in such patients was associated with a significant reduction in symptomatic VTE (from 1.2% to 0.3%) and a small increase in clinically relevant bleeding (from 2.7% to 3.3%) 2.Forster R. Stewart M. Anticoagulants (extended duration) for prevention of venous thromboembolism following total hip or knee replacement or hip fracture repair.The Cochrane Database of Systematic Reviews. 2016; 3: CD004179PubMed Google Scholar. With advances in thromboprophylaxis and improvements in perioperative care, mortality rates from PE or other causes 90 days after elective hip or knee arthroplasty are now less than 0.5%. Consequently, the benefit of extended anticoagulant thromboprophylaxis in these patients may be diminished. In support of this concept, a recent trial in patients undergoing elective hip or knee arthroplasty showed that after initial thromboprophylaxis with rivaroxaban, extended out‐of‐hospital prophylaxis with aspirin was non‐inferior to extended prophylaxis with rivaroxaban for the prevention of symptomatic VTE. Therefore, clinicians may now have a choice of using aspirin or rivaroxaban for extended out‐of‐hospital thromboprophylaxis in such patients. Robust evidence of a net clinical benefit of extended thromboprophylaxis with other types of surgery is lacking. What about medical patients? Until recently, there was no evidence of a net clinical benefit with extended thromboprophylaxis in medically ill patients. Thus, the results of trials comparing a 30‐day course of thromboprophylaxis with rivaroxaban or apixaban with a 10‐day course of enoxaparin were disappointing because extended anticoagulation was associated with a 2‐fold increase in the rate of major bleeding, which negated the benefit of the reduction in symptomatic VTE (Table 1). So where do we go from here?Table 1Trials evaluating extended thromboprophylaxis in acutely ill medical patientsStudy (no. of patients)InterventionControlPrimary efficacy outcome (VTE rates) Intervention vs. ControlSymptomatic VTE rates Intervention vs. ControlMajor bleeding rates Intervention vs. ControlCommentsEXCLAIM(n = 5963)Extended duration enoxaparin(up to 28 days)Enoxaparin(10 ± 4 days)2.5% vs 4.0%*Significant P‐value; ns = non‐significant P‐value;0.2% vs. 1.2%#P‐value not reported. IMPROVE VTE RAM, International Medical Prevention Registry on Venous Thromboembolism Risk Assessment Model; VTE, venous thromboembolism.0.8% vs 0.3%*Significant P‐value; ns = non‐significant P‐value;No net benefitADOPT(n = 6528)Extended duration apixaban(30 days)Enoxaparin(6–14 days)2.71% vs. 3.06%ns0.37% vs.0.73%#P‐value not reported. IMPROVE VTE RAM, International Medical Prevention Registry on Venous Thromboembolism Risk Assessment Model; VTE, venous thromboembolism.0.47% vs. 0.19%*Significant P‐value; ns = non‐significant P‐value;No net benefitMAGELLAN(n = 8101)Extended duration rivaroxaban(35 ± 4 days)Enoxaparin(10 ± 4 days)4.4% vs. 5.7%*Significant P‐value; ns = non‐significant P‐value;1.3% vs. 2.0%#P‐value not reported. IMPROVE VTE RAM, International Medical Prevention Registry on Venous Thromboembolism Risk Assessment Model; VTE, venous thromboembolism.1.1% vs. 0.4%*Significant P‐value; ns = non‐significant P‐value;No net benefitAPEX(n = 7513)Extended duration betrixaban(35 to 42 days)Enoxaparin(10 ± 4 days)Cohort 1:6.9% vs. 8.5%nsCohort 2:5.6% vs. 7.1%*Significant P‐value; ns = non‐significant P‐value;Overall:5.3% vs. 7.0%*Significant P‐value; ns = non‐significant P‐value;Cohort 1:1.3% vs. 1.9%nsCohort 2:1.0% vs. 1.4%nsOverall:0.9% vs. 1.5%*Significant P‐value; ns = non‐significant P‐value;Cohort 1:0.6% vs. 0.7%nsCohort 2:0.7% vs. 0.6%nsOverall:0.7% vs. 0.6%nsEnrichment design based on clinical risk factors and D‐dimer.The consistency of results in the 3 analysis cohorts and no excess in major bleeding suggest a net benefitMARINER(n= up to 12 000)Ongoing trialEvent drivenExtended duration rivaroxaban(45 days)PlaceboResults AwaitedEnrichment design based on use of IMROVE VTE RAM and D‐dimer at time of discharge* Significant P‐value; ns = non‐significant P‐value;# P‐value not reported. IMPROVE VTE RAM, International Medical Prevention Registry on Venous Thromboembolism Risk Assessment Model; VTE, venous thromboembolism. Open table in a new tab Extending thromboprophylaxis in all medical patients is not an option because the excess of bleeding is likely to negate any benefit. However, if medical patients at high risk of VTE can be identified, extending anticoagulant prophylaxis may reduce the VTE burden while incurring a more acceptable bleeding cost. Emerging data provide support for such a strategy. The Acute Medically Ill Venous Thromboembolism Prevention with Extended Duration Betrixaban (APEX) trial prospectively evaluated the efficacy and safety of betrixaban thromboprophylaxis for up to 42 days with a 10‐day course of enoxaparin prophylaxis in medical patients who were deemed at risk of VTE based on advanced age, history of previous VTE, certain comorbidities and a 2‐fold or greater increase in D‐dimer level above the upper limit of normal 3.Cohen A.T. Harrington R.A. Goldhaber S.Z. Hull R.D. Wiens B.L. Gold A. Hernandez A.F. Gibson C.M. Extended thromboprophylaxis with betrixaban in acutely Ill medical patients.N Engl J Med. 2016; 375: 534-44Crossref PubMed Scopus (334) Google Scholar. Extended prophylaxis with betrixaban resulted in a numerically lower rate of the composite outcome of asymptomatic proximal DVT and symptomatic VTE than that with enoxaparin in the restricted cohort of patients with elevated D‐dimer levels (6.9% and 8.5%, respectively; P = 0.054), but this reduction was statistically significant in the overall population. Whereas the primary efficacy outcome was dominated by asymptomatic events, additional exploratory analyses focusing on patient‐important outcomes not only demonstrated a significant 36% reduction in symptomatic VTE with betrixaban compared with enoxaparin (0.9% and 1.5%, respectively; P = 0.04), but extended thromboprophylaxis with betrixaban also reduced re‐hospitalization for VTE at 77 days by 56% compared with enoxaparin (0.45% and 1.04%, respectively; P = 0.006), and ischemic stroke by 47% (0.48% and 0.91%, respectively; P = 0.026) in the overall population 4.Gibson C.M. Chi G. Halaby R. Korjian S. Daaboul Y. Jain P. Arbetter D. Goldhaber S.Z. Hull R. Hernandez A.F. Gold A. Bandman O. Harrington R.A. Cohen A.T. Extended‐duration betrixaban reduces the risk of stroke versus standard‐dose enoxaparin among hospitalized medically ill patients: an APEX trial substudy (acute medically ill venous thromboembolism prevention with extended duration betrixaban).Circulation. 2016; 135: 648-55Crossref PubMed Scopus (55) Google Scholar. Importantly, the reduction in VTE was not associated with an excess of major bleeding (0.7% and 0.6% in the betrixaban and enoxaparin groups, respectively), although there was an increase in clinically relevant bleeding with betrixaban compared with enoxaparin (3.1% and 1.6%, respectively; P < 0.001). Therefore, the results of the APEX trial provide proof of concept that targeting medical patients whose risk of VTE is sufficiently high to justify the increased risk of bleeding associated with extended anticoagulant prophylaxis is a promising approach to optimizing the benefit–risk profile. The data also suggest that such an approach has the potential to reduce the burden of VTE related to hospitalization because it is feasible, is likely to be cost‐effective because of its benefits beyond the reduction in symptomatic VTE, and because it appears to confer a net benefit, particularly if the reductions in re‐hospitalization and ischemic stroke also are considered. Two potential options to enhance the benefit–risk profile of extended thromboprophylaxis are currently under evaluation. The first involves the use of more refined risk assessment models, whereas the second focuses on the development of new anticoagulants that have the potential to attenuate thrombosis with little or no disruption of hemostasis. Progress in the development of risk assessment models includes validation of standardized VTE risk assessment tools in medical patients, and the addition of elevated D‐dimer levels to improve their utility for identifying high‐risk patients. The ease of implementing these risk assessment tools in clinical practice requires further evaluation but their implementation is facilitated by the fact that health agencies mandate their use in many countries. Using the International Medical Prevention Registry on Venous Thromboembolism (IMPROVE VTE) risk score in combination with determination of D‐dimer levels in hospitalized medical patients to identify those at high risk, the ongoing Medically Ill Patient Assessment of Rivaroxaban versus Placebo in Reducing Post‐Discharge Venous Thromboembolism Risk (MARINER) trial is comparing the efficacy and safety of a 45‐day course of rivaroxaban (at a dose of 10 mg once daily reduced to 7.5 once daily in patients with a creatinine clearance between 30 and 50 mL/min) started at hospital discharge with placebo 5.Raskob G.E. Spyropoulos A.C. Zrubek J. Ageno W. Albers G. Elliott C.G. Halperin J. Haskell L. Hiatt W.R. Maynard G.A. Peters G. Spiro T. Steg P.G. Suh E.Y. Weitz J.I. The MARINER trial of rivaroxaban after hospital discharge for medical patients at high risk of VTE. Design, rationale, and clinical implications.Thromb Haemost. 2016; 115: 1240-8Crossref PubMed Scopus (56) Google Scholar. To develop potentially safer anticoagulants, attention is now focusing on factor (F) XI and FXII, clotting factors upstream to FXa and thrombin. Knockout or knockdown of these factors in animal models attenuates thrombosis without increasing bleeding. Furthermore, in a proof of concept study in patients undergoing elective knee arthroplasty, knockdown of FXI with an antisense oligonucleotide reduced postoperative VTE to a greater extent than enoxaparin without increasing the rate of bleeding. The introduction of antibodies, aptamers and small molecule inhibitors of FXIa provides an array of compounds that may be at least as effective but safer than currently available anticoagulants. With long‐acting FXI‐directed antibodies, a single injection has the potential to provide prophylaxis for a month or longer. The efficacy and safety of such an approach in medical or surgical patients require investigation. Despite near universal in‐hospital thromboprophylaxis, the burden of hospital‐related VTE persists. Factors contributing to this burden include shortening hospital stays leading to inadequate durations of thromboprophylaxis, and the increased comorbidity burden of hospitalized patients. Although the relative contribution of medical and surgical patients to the burden of hospitalization‐related VTE is unclear, these changes are likely to have the greatest impact on medical patients because surgical patients have benefited from advances in anesthesia, less invasive operative techniques and earlier postoperative mobilization. Safely reducing VTE in the acutely ill medical population depends on identifying high‐risk patients and providing them with extended thromboprophylaxis with anticoagulants with favorable benefit–risk profiles. Such a strategy worked with betrixaban, and the results of the MARINER study will tell us whether it also works with rivaroxaban. Future research will determine whether upstream inhibitors of FXIIa or FXIa provide advantages over FXa inhibitors for this indication.