FormalPara Key Summary Points

Why carry out this study?

Patients with venous thromboembolism (VTE) at increased risk of bleeding or recurrences present challenges for clinical management.

There is currently insufficient evidence about non-vitamin K antagonist oral anticoagulants (NOACs) in patients with VTE at increased risk of bleeding or recurrences as these patients have been excluded or underrepresented in NOAC VTE trials.

What was learned from the study?

Patients with apixaban (vs warfarin) had a lower risk of recurrent VTE, major bleeding, and clinically relevant non-major bleeding events.

The treatment effects were generally consistent across subgroups of patients with VTE at increased risk of bleeding or recurrences.

Introduction

Venous thromboembolism (VTE), which includes deep vein thrombosis (DVT) and pulmonary embolism (PE), is a major cause of morbidity and mortality [1]. Current clinical guidelines recommend initial anticoagulation for 3–6 months after an initial VTE event [2]. Vitamin K antagonists (VKAs), such as warfarin, used to be the treatment of choice for anticoagulation among patients with VTE [3]. However, warfarin therapy must be monitored and adjusted within a narrow therapeutic index of International Normalized Ratio (INR) results. The need for regular monitoring, the potential risk of hemorrhage, and poor control of INR levels may lead to medication discontinuation or non-adherence. Over the last 15 years, non-VKA oral anticoagulants (NOACs)—including dabigatran, apixaban, rivaroxaban, and edoxaban—were approved in the USA for the treatment of VTE [4,5,6,7]. Clinical trials demonstrated that NOACs were as effective as warfarin and had similar or less major bleeding (MB) in patients with VTE [8].

Patients with VTE at increased risk of bleeding present great challenges for OAC management. Clinical factors such as history of bleeding and thrombocytopenia can further exacerbate the risk of bleeding associated with anticoagulant treatment in patients with VTE [9]. Evidence about the effectiveness and safety of OACs in these high risk patients is needed to help inform clinical decisions.

Patients with a current VTE who are also at increased risk of recurrent VTE present challenges for OAC management. Several clinical conditions are associated with risk of recurrences such as thrombophilia, chronic liver disease, and immune-mediated disorders (e.g., rheumatoid arthritis and inflammatory bowel disease) [10,11,12,13]. Patients with VTE and these conditions are at increased risk for recurrent VTE events, and effective OAC treatments are needed to minimize this risk. There is currently insufficient evidence about NOACs in patients with VTE at increased risk of bleeding or recurrences as these patients have been excluded or underrepresented in NOAC VTE trials. For example, the AMPLIFY trial excluded patients with high risk of bleeding [12]. None of the pivotal NOAC VTE trials required testing for primary hypercoagulable state and the proportions of patients with thrombophilia in these trials were low [14]. We have previously published data on other risk factors such as obesity, age, and chronic kidney diseases; however, there is still a lack of literature focusing on risk factors like thrombocytopenia and immune-mediated disorders [15,16,17]. The purpose of this study was to evaluate the effectiveness and safety of apixaban vs warfarin among subgroups of patients with VTE who have risk factors for bleeding (e.g., history of bleeding or thrombocytopenia) or for recurrences (e.g., thrombophilia, chronic liver disease, or immune-mediated disorders).

Methods

Study Population

Patients with at least one medical claim for VTE in any position (index VTE event) in the inpatient or outpatient setting were identified from Centers for Medicare & Medicaid Services fee-for-service Medicare (study period March 1, 2014–December 31, 2017), IBM® MarketScan® Commercial Claims and Encounter and Medicare Supplemental and Coordination of Benefits (MarketScan; study period March 1, 2014–September 30, 2018), IQVIA PharMetrics Plus™ (PharMetrics; study period March 1, 2014–March 31, 2019), Optum Clinformatics™ Data Mart (Optum; study period March 1, 2014–December 31, 2018), and Humana® Research Database data (Humana; study period March 1, 2014–December 31, 2018). Patients with at least one pharmacy claim for warfarin or apixaban during the 30-day period following the index VTE event were identified, and the first warfarin or apixaban prescription date was designated as the index date. Patients with an outpatient prescription for another oral or parenteral anticoagulant on the index date or during the period between the index VTE event and the index date or baseline period were excluded, see Fig. 1 for additional selection criteria that were applied. The reason for the exclusion of patients with outpatient prescription for another oral or parental anticoagulant was to avoid the potential impacts of other anticoagulants on outcomes. The baseline period was defined as 6 months prior to and including the index date. The follow-up period was defined as the day after the index date through the earliest of the end of the subsequent 6-month period, index therapy discontinuation, switch to another OAC or parenteral anticoagulant treatment, health plan disenrollment, death, or study end.

Fig. 1
figure 1

Patient selection criteria. *Identification periods were MarketScan, September 1, 2014–September 30, 2018; Optum and Humana, September 1, 2014–December 31, 2018; PharMetrics, September 1, 2014–March 31, 2019; Medicare, September 1, 2014–December 31, 2017. VTE venous thromboembolism. ◊Defined as at least two cancer diagnoses or at least one cancer diagnosis and a cancer-related treatment (e.g., chemotherapy, radiation, cancer-related surgery)

Full size image
Fig. 2
figure 2

Risk of recurrent VTE for apixaban vs warfarin among patients with VTE stratified by factors that increased risk of bleeding or recurrences. CI confidence interval, VTE venous thromboembolism

Full size image
Fig. 3
figure 3

Risk of MB for apixaban vs warfarin among patients with VTE stratified by factors that increased the risk of bleeding or recurrences. CI confidence interval, MB major bleeding, VTE venous thromboembolism

Full size image
Fig. 4
figure 4

Risk of CRNM bleeding for apixaban vs warfarin among patients with VTE stratified by factors that increased the risk of bleeding and/or recurrences. CI confidence interval, CRNM clinically relevant non-major bleeding, VTE venous thromboembolis

Full size image

Two subgroups of patients with VTE with a risk factor for bleeding were evaluated in this study: history of bleedingFootnote 1 (vs no history of bleeding) and thrombocytopenia (vs no thrombocytopenia). Three subgroups of patients with VTE with a risk factor for recurrences were assessed: thrombophiliaFootnote 2 (vs no thrombophilia), chronic liver disease (vs no chronic liver disease), and immune-mediated disordersFootnote 3 (vs no immune-mediated disorders). Risk factors were determined using baseline claims with International Classification of Diseases, 9th/10th Revision (ICD-9/10) diagnosis and procedure codes, detailed in Supplemental Table 1.

Study Measures

Clinical outcomes including recurrent VTE, MB, and clinically relevant non-major (CRNM) bleeding were evaluated during the follow-up period, and all the outcomes were measured independently [18,19,20,21]. Recurrent VTE was defined using ICD-9/10 codes for VTE in the primary (Medicare) or first-listed (commercial database) diagnosis in an inpatient setting, and admissions that occurred within 7 days after the index VTE encounter were excluded. Similarly, MB events were identified using ICD-9/10 codes for MB in the primary/first-listed diagnosis in the inpatient setting and stratified by gastrointestinal (GI) bleeding, intracranial hemorrhage (ICH), and bleeding at other sites (genitourinary bleeding, respiratory tract bleeding, ocular bleeding, joint bleeding/hemarthrosis, transfusion of blood and blood components, other bleeding, or no bleeding site specified). CRNM bleeding was defined as either an inpatient admission with a secondary diagnosis code for non-critical sites of bleeding (excluded if MB occurred before the CRNM bleed or during the same hospitalization) or a diagnosis code for gastrointestinal bleeding or another selected non-critical site of bleeding in the outpatient setting. Demographic (age, gender, etc.) and VTE-related variables were measured on the index VTE date and clinical characteristics were evaluated during the baseline period.

Statistical Methods

Stabilized inverse probability treatment weighting (IPTW) was conducted within each database to balance patient characteristics between the treatment cohorts. Covariates included in the IPTW model were demographics, clinical characteristics, and VTE-related variables. Each patient was weighted by the inverse of the probability of their treatment option (weight = 1/propensity score), and the weights were stabilized by multiplying the original weights with a constant—which is equal to the expected value of being in the treatment or comparison cohorts, respectively [22,23,24]. Balance of characteristics was assessed by standardized differences with values less than 10 suggesting balance between cohorts. After IPTW patients were pooled for further analysis.

The risk of recurrent VTE, MB, and CRNM bleeding in patients with VTE who initiated apixaban vs warfarin was evaluated using Cox proportional hazard models. To evaluate the impacts of bleeding risk factors on outcomes associated with apixaban compared to warfarin, interaction analysis was conducted to evaluate whether treatment effects differed by history of bleed (yes/no) and by thrombocytopenia (yes/no). To assess the impacts of recurrences risk factors on outcomes associated with apixaban vs warfarin, interaction analyses was conducted to evaluate treatment effects by thrombophilia (yes/no), by chronic liver disease (yes/no), and by immune-mediated disorders (yes/no) [10, 11, 25]. For each analysis, a statistical significance (p < 0.10) of the interaction between the treatment and subgroup strata on effectiveness (recurrent VTE) and safety (MB and CRNM bleeding) was evaluated.

All components of this study are STROBE compliant.

Compliance with Ethics Guidelines

This study was performed in accordance with the Helsinki Declaration of 1964, and its later amendments. Since this study did not involve the collection, use, or transmittal of individually identifiable data, it was deemed exempt from institutional review board review by Solutions IRB. Both the data sets and the security of the offices where analysis was completed (and where the data sets are kept) meet the requirements of the Health Insurance Portability and Accountability Act of 1996. Solutions IRB determined this study to be EXEMPT from the Office for Human Research Protections (OHRP)’s Regulations for the Protection of Human Subjects (45 CFR 46) under Exemption 4: Research involving the collection or study of existing data, documents, records, pathological specimens, or diagnostic specimens, if these sources are publicly available or if the information is recorded by the investigator in such a manner that subjects cannot be identified, directly or through identifiers linked to the subjects. The HIPAA Authorization Waiver was granted in accordance with the specifications of 45 CFR 164.512(i). This project was conducted in full accordance with all applicable laws and regulations, and adhered to the project plan that was reviewed by Solutions Institutional Review Board.

Results

After the inclusion and exclusion criteria were applied, a total of 155,119 patients with VTE, including 60,786 (39.2%) who initiated apixaban and 94,333 (60.8%) who initiated warfarin, were identified in the pooled databases (Fig. 1). The baseline characteristics of patients with VTE receiving apixaban or warfarin before and after IPTW are shown in Table 1.

Table 1 Baseline characteristics of patients with VTE initiating apixaban or warfarin: pre/post IPTW
Full size table

Post-IPTW baseline characteristics were well balanced (Tables 2, 3). For the two treatment cohorts, 20.7% of patients had a history of bleed and 5.4% had thrombocytopenia. Additionally, nearly 4.0% of patients had thrombophilia, 6.7% (warfarin) and 7.8% (apixaban) had chronic liver disease, and 5.3% (apixaban) and 7.0% (warfarin) had immune-mediated disorders during the baseline period. During follow-up, apixaban was associated with a significantly lower risk of recurrent VTE (hazard ratio [HR] 0.72; 95% confidence interval [CI] 0.67–0.78), MB (HR 0.70; 95% CI 0.64–0.76), and CRNM bleeding (HR 0.83; 95% CI 0.80–0.86) compared to warfarin.

Table 2 Post-IPTW baseline characteristics by presence of risk factors for bleeding
Full size table
Table 3 Post-IPTW baseline characteristics by presence of risk factors for recurrences
Full size table

Table 2 shows patient characteristics stratified by history of bleeding and thrombocytopenia. Patients with a history of bleeding were more likely to have an inpatient VTE diagnosis, provoked VTE, and higher mean Deyo-Charlson comorbidity index (Deyo-CCI) score compared to patients without a history of bleeding. Similar differences in patient characteristics were observed for patients with thrombocytopenia vs those without thrombocytopenia.

Patient characteristics by thrombophilia, chronic liver disease, and immune-mediated disorders are presented in Table 3. Patients with these conditions were more likely to have inpatient VTE diagnosis than those without the condition. The mean Deyo-CCI score was also higher for patients with chronic liver disease and patients with immune-mediated disorders.

Figures 2, 3, and 4 show incidence rates and hazard ratios of recurrent VTE, MB, and CRNM bleeding for apixaban vs warfarin stratified by history of bleed, thrombocytopenia, thrombophilia, chronic liver disease, and immune-mediated disorders. No significant interaction was observed between treatment and history of bleed for recurrent VTE (interaction p = 0.309); however, significant interactions were observed between treatment and history of bleed for MB (interaction p = 0.001) and CRNM bleeding (interaction p < 0.001). Apixaban patients had a lower risk of MB and CRNM bleeding among patients regardless of history of bleed status, but the magnitude of the treatment effects was larger for patients without a history of bleeding (HR 0.62 and 0.79, respectively) than those with a history of bleeding (HR 0.84 and 0.91, respectively). When stratified by diagnosis of thrombocytopenia, no significant interaction was observed between treatment and thrombocytopenia for recurrent VTE (interaction p = 0.349) or CRNM bleeding (interaction p = 0.628). A significant interaction was observed between treatment and thrombocytopenia for MB (interaction p = 0.006): apixaban had a lower risk among patients without thrombocytopenia and similar risk among patients with thrombocytopenia compared to warfarin.

When stratified by the presence and absence of thrombophilia, no significant interaction was observed between treatment and thrombophilia for recurrent VTE (interaction p = 0.598) or MB (interaction p = 0.500). A significant interaction was observed between treatment and thrombophilia for CRNM bleeding (interaction p = 0.078). Apixaban (vs warfarin) was associated with a lower risk of CRNM bleeding across patients regardless of thrombophilia status, but treatment effects were larger for patients with thrombophilia (HR 0.71) than without thrombophilia (HR 0.84). No significant interaction was observed between treatment and chronic liver disease for recurrent VTE (interaction p = 0.853), MB (interaction p = 0.389), or CRNM bleeding (interaction p = 0.652). Similarly, no significant interaction was observed between treatment and immune-mediated disorders for recurrent VTE (interaction p = 0.692), MB (interaction p = 0.211), or CRNM bleeding (interaction p = 0.906).

Discussion

There has been a lack of evidence about the effectiveness and safety of NOACs among subgroups of patients with VTE at increased risk of bleeding and/or recurrences. To help fill the gap, this real-world observational study compared the risk of recurrent VTE, MB, and CRNM bleeding among patients with VTE who initiated apixaban vs warfarin stratified by factors that increased the risk of bleeding or factors that increased the risk of recurrences. In general, beneficial effects of apixaban vs warfarin on recurrent VTE, MB, and CRNM bleeding were consistently observed in patients with VTE with or without history of bleeding, thrombocytopenia, thrombophilia, chronic liver disease, or immune-mediated disorders.

Patients with VTE at increased risk of bleeding and recurrences constitute a vulnerable population and must be treated with anticoagulant therapy with utmost care considering the benefit-to-risk ratio for recurrent VTE as well as their bleeding risk. However, these patients have not been well represented in randomized clinical trials. The AMPLIFY trial excluded patients with VTE at a high risk of bleeding [26]. In a systematic review and meta-analysis of NOACs in patients with VTE with thrombophilia, no data were identified about the use of apixaban in patients with VTE and thrombophilia [27]. The findings from this real-world observational study help address some of the data gaps in existing evidence.

Furthermore, these patients are difficult to manage because those with risk factors for bleeding are often at increased risk of recurrences and vice versa as seen in this study. Patients with a history of bleeding and thrombocytopenia had a higher frequency of recurrences than those without those risks; and bleeding was more common in those with immune-mediated disorders.

Risk of bleeding associated with anticoagulant treatment is a major concern for patients with VTE with a risk factor for bleeding such as those with a history of bleeding or thrombocytopenia. In this study, we found that apixaban was generally associated with a lower risk of MB and CRNM bleeding vs warfarin among patients with VTE with and without history of bleeding. However, the reduction in the risk of MB and CRNM bleeding for apixaban vs warfarin was smaller for patients with history of bleeding vs patients without history of bleeding. Apixaban was also associated with a consistent treatment effect on CRNM bleeding compared to warfarin among patients with and without thrombocytopenia. However, a significant interaction was observed for MB: apixaban patients with thrombocytopenia had a similar risk of MB whereas apixaban patients without thrombocytopenia had a lower risk of MB vs warfarin. As prior research showed that bleeding risk increases among those with thrombocytopenia in patients with VTE, further research is needed to understand the different effects of apixaban among VTE patients with and without thrombocytopenia [28].

Minimizing the risk of recurrent VTE is an important consideration when selecting anticoagulant treatment for patients at increased risk of recurrences. Our comparison of apixaban and warfarin stratified by three different risk factors for recurrences including thrombophilia, chronic liver disease, and immune-mediated disorders did not find significant interactions between the treatment and any of the risk factors on recurrent VTE. Apixaban was associated with lower risk of recurrent VTE regardless of whether patients had any of these risk factors.

Limitations

This study has certain limitations, most of which are inherent to claims databases. First, associations can be inferred from this retrospective analysis, but causal relationships cannot be established. Second, there were no lab values collected; therefore, no data were available regarding international normalized ratio monitoring for warfarin nor for laboratory assessment of renal impairment. Third, although IPTW has been used to balance patient characteristics between treatment cohorts, residual confounding may persist. Fourth, the definitions for recurrent VTE and MB events were identified on the basis of primary/first-listed ICD-9/10 diagnosis codes, and the number of events may be different compared with those identified in clinical trials which were based on adjudication. It is also possible that symptomatic events, for which diagnoses were not attempted, were not captured in the study. The definition for CRNM bleeding has not been validated in the literature, although attempts were made to follow the definition as suggested by the International Society on Thrombosis and Haemostasis. Fifth, thrombocytopenia and thrombophilia were identified on the basis of diagnosis codes. A previous study found that the sensitivity for using diagnosis codes to identify thrombophilia was low [29]. Thrombophilia and thrombocytopenia may be underestimated in this study. Sixth, other factors associated with bleeding and recurrences were not assessed such as concomitant medications and the presence of cancer. Finally, the study results are not generalizable to uninsured populations or those with Medicaid.

Conclusion

In this study of patients with VTE at increased risk of bleeding or recurrences, differences in patient characteristics were observed in patients with risk factors for bleeding (history of bleeding and thrombocytopenia) and recurrences (thrombophilia, chronic liver disease, and immune-mediated disorders) relative to those without. Patients with prescription fills for apixaban had a lower risk of recurrent VTE, MB, and CRNM bleeding compared to those with warfarin. The treatment effects of apixaban vs warfarin were generally consistent across subgroups of patients at increased risk of bleeding or recurrences. More studies are needed to determine optimal VTE management strategies for patients at elevated risk for bleeding or recurrences.