FormalPara Key Summary Points

Why carry out this study?

Since the initiation and continuous use of anticoagulant therapy is key to adequate disease control among patients with venous thromboembolism (VTE), poor compliance with therapy may lead to increased risk of complications, mortality, and hospitalization, representing a major public health concern.

This study sought to describe and compare the risk of VTE recurrence among patients who abandoned their first direct oral anticoagulant (DOAC) fill relative to patients who had the first DOAC prescription filled.

What was learned from the study?

In this study of patients with VTE in real-world clinical practice, abandonment of the first DOAC prescription fill was associated with a significantly higher risk of VTE recurrence in the short term, which was also sustained for 2 years after abandonment.

Relative to non-abandoners, the risk of recurrence was 72% higher among abandoners at 3 months and 53% higher at 12 months.

These findings suggest that ensuring proper access and encouraging early and continuous use of DOACs may help prevent severe and fatal complications among patients with VTE.

Introduction

Venous thromboembolism (VTE) is a chronic condition that comprises deep vein thrombosis (DVT) and pulmonary embolism (PE) [1]. VTE occurs in 1–2 per 1000 individuals each year, with 1.2 million people reported to have VTE in the USA in 2016 [2, 3]. VTE can recur frequently, with an average 10-year recurrence rate of 30%, which can vary according to existing risk factors [2]. Individuals at high risk of incident and/or recurrent VTE include those with older age, active cancer, antiphospholipid syndrome, prolonged immobility, major trauma, or surgery [1, 4].

Given that 20% of patients with PE die within 1 year of presentation, it is of utmost importance to diagnose and treat VTE in a timely fashion [1, 2]. Without treatment, VTE can lead to cardiopulmonary collapse, risk of long-term complications, and death [1]. Additionally, if left untreated, up to 50% of patients with symptomatic proximal DVT or PE experience a recurrence within 3 months [5].

Treatment guidelines recommend the use of direct oral anticoagulants (DOACs) instead of vitamin K antagonists for the management of VTE because DOACs are equally effective at decreasing the risk of recurrence and death but are also associated with a significantly reduced risk of major bleeding [3]. Four DOACs are currently approved by the US Food and Drug Administration (FDA) for the treatment of VTE, namely rivaroxaban, apixaban, dabigatran, and edoxaban [6,7,8,9]. Since the initiation and continuous use of anticoagulant therapy is key to adequate disease control among patients with VTE, poor compliance with therapy may lead to increased risk of complications, mortality, and hospitalizations, representing a major public health concern [10, 11].

While DOAC nonadherence has been well characterized in the management of atrial fibrillation [12], there is limited literature evaluating this topic among patients with VTE. Additionally, most research has focused on secondary nonadherence (i.e., patient stops filling prescriptions after having previously filled their first prescription) rather than primary nonadherence or abandonment (i.e., patient receives a prescription for a new medication but does not fill, or abandons, the first prescription or does not obtain an appropriate alternative within a reasonable time period) [13]. Notably, prescription abandonment has been found to be more likely to occur among new users of medications than prevalent users [14], highlighting the need for further characterization of primary medication abandonment. One claims-based study found that DOACs were associated with a significantly higher rate of primary abandonment compared to warfarin (26.7% vs 14.0%), but this study was conducted among any patients treated with oral anticoagulants, regardless of diagnosis [15].

As such, there is a need for further research on the primary abandonment of DOACs and its effects on clinical outcomes, specifically among the VTE population. This study was conducted to address this gap in knowledge by comparing the risk of VTE recurrence among patients who abandoned their first DOAC fill relative to patients who did not.

Methods

Data Source

Data from Symphony Health, an ICON plc Company, PatientSource®, April 1, 2017 to October 31, 2020, were used. The database links health care data for the US population from three basic sources: pharmacy point-of-service, switch/network transactions, and additional direct prescription, medical, and hospital claims data. The data source includes patient demographics, medical and procedure claims, and prescription drug claims, including status of prescription drug claims (i.e., approved, rejected, abandoned). The database captures more than 75% of all US retail prescription claims, representing over three-quarters of the US population of patients annually across multiple payer channels (i.e., commercial, Medicare, Medicaid).

Study Design

A retrospective longitudinal design was used. The index date was the date of the first DOAC fill (i.e., apixaban, dabigatran, or rivaroxaban); only patients with the status of the first fill being “approved” or “abandoned” were retained. Approved fills were claims submitted by a pharmacy and approved for payment by health plans after claims adjudication. Abandoned fills were adjudicated claims for prescriptions, approved by health plans that patients decided to abandon, usually due to cost or non-compliance. The baseline period was defined as the 6-month period of clinical activity before the index date. Clinical activity was based on the first and last patient-level activity flags, with activity defined as either a pharmacy or medical claim. The follow-up period started on the index date and spanned until the earliest of the end of clinical activity or data availability.

Sample Selection

Patients were included in the study if they had (1) ≥ 1 diagnosis of VTE (International Classification of Diseases, 10th Revision, Clinical Modification [ICD-10-CM]: I26, I80.1, I80.2, I80.3, I82.4, I82.6, I82.A1, I82.B1, I82.C1, I82.90) in any care setting, with the first diagnosis defined as the index VTE; (2) the first claim for apixaban, dabigatran, or rivaroxaban within 28 days after the index VTE diagnosis; (3) no claims for other oral anticoagulants (i.e., betrixaban, edoxaban, warfarin) before the index date; (4) ≥ 6 months of clinical activity before the index date (i.e., washout and baseline period); and (5) ≥ 18 years old on the index date. Patients were excluded from the study if they had (1) ≥ 1 claim with a diagnosis of atrial fibrillation before or on the index date; (2) organ or tissue replaced by transplant at baseline; (3) pregnancy at baseline or during the follow-up period; (4) ≥ 1 claim with a diagnosis of VTE in an inpatient setting on the index date; (5) > 1 DOAC (i.e., apixaban, betrixaban, dabigatran, edoxaban, rivaroxaban) claimed on the index date; or (6) > 1 final claim status (e.g., approved and abandoned) for the index DOAC on the index date.

On the basis of the status of the index claim, patients were classified into the following mutually exclusive cohorts: (1) patients with primary non-adherence or those who abandoned their first DOAC fill (abandoner cohort), and (2) patients with an approved first DOAC fill (non-abandoner cohort).

Outcome Measures

VTE recurrence was measured in the abandoner and non-abandoner cohorts separately and was defined as the first DVT or PE event in an inpatient setting during the follow-up period. DVT and PE events were reported as a composite outcome (i.e., DVT/PE) and separately.

Statistical Analysis

All analyses were conducted using SAS Enterprise Guide software Version 7.1. To balance baseline characteristics between cohorts, inverse probability of treatment weighting (IPTW) was used. The propensity score was computed from a logistic regression model adjusting for demographic characteristics (i.e., age, sex, region of residence, insurance plan type, index year), comorbidities (Quan Charlson Comorbidity Index [Quan-CCI] [16], Registro Informatizado de Enfermedad TromboEmbolica Venosa [RIETE] score for risk of major bleeding [17]), site of care for the index VTE diagnosis, number of unique prescription drugs used, index DOAC medication, diagnosis of congestive heart failure, number of inpatient days, and patient out-of-pocket paid amount for the index DOAC medication. The balance of baseline characteristics was assessed using the standardized difference, where less than 10% indicated balance [18]. Continuous variables were described using means, standard deviations (SDs), and medians, while categorical variables were described using frequencies and proportions.

The probability of the first VTE recurrence event during the follow-up period was described in each cohort using weighted Kaplan–Meier survival analysis. Risk of recurrence was compared between the cohorts using weighted Cox proportional hazards models and reported as the hazard ratio (HR) with its 95% confidence interval (CI) and p value. The time to the first recurrent event was measured from the index date; patients for whom the event was not observed during the follow-up period were censored at the end of the follow-up period.

Sensitivity Analysis

A sensitivity analysis using a censored follow-up period was conducted to assess VTE recurrence while the abandoner cohort remained untreated, and the non-abandoner cohort remained on DOAC treatment. The censored follow-up period began on the index date and spanned until the earliest of (1) the first non-oral anticoagulant or warfarin claim; (2) the first approved DOAC fill (abandoner cohort only); (3) the date of DOAC discontinuation, defined as the last day with therapy supply before a 15-day gap without DOAC medication (non-abandoner cohort only); (4) the end of clinical activity; or (5) the end of data availability.

Compliance with Ethics Guidelines

Data were de-identified and complied with the patient requirements of the Health Insurance Portability and Accountability Act (HIPAA) of 1996; therefore, no review by an institutional review board was required per Title 45 of CFR, Part 46.101(b)(4) [19].

Results

Among a total of 306,374 identified patients with VTE who met all criteria for the study, 25,780 (8.4%) abandoned their index DOAC fill (abandoner cohort) and 280,594 (91.6%) had an approved index DOAC fill (non-abandoner cohort; Fig. 1).

Fig. 1
figure 1

Identification of DOAC abandoners and non-abandoners among patients with VTEa. AF atrial fibrillation, DOAC direct oral anticoagulant, VTE venous thromboembolism. aPatients were excluded from the study if they had (1) ≥ 1 claim with a diagnosis of atrial fibrillation before or on the index date; (2) organ or tissue replaced by transplant at baseline; (3) pregnancy at baseline or during the follow-up period; (4) ≥ 1 claim with a diagnosis of VTE in an inpatient setting on the index date; (5) > 1 DOAC (i.e., apixaban, betrixaban, dabigatran, edoxaban, rivaroxaban) claimed on the index date; or (6) > 1 final claim status (e.g., approved and abandoned) for the index DOAC on the index date

Full size image

Study Population and Weighted Baseline Characteristics

After the abandoner and non-abandoner cohorts were weighted to balance baseline characteristics, 152,443 patients comprised the abandoner cohort and 153,931 patients comprised the non-abandoner cohort. On the basis of the standardized differences, patient characteristics between the weighted cohorts were balanced (Table 1).

Table 1 Selected baseline characteristics in the weighted abandoner and non-abandoner cohortsa,b
Full size table

Mean age was 60.1 years in the abandoner cohort and 60.3 years in the non-abandoner cohort, and 52.8% and 52.9% were female, respectively (Table 1). Close to half of patients were covered by commercial insurance (47.7% in both cohorts), just over one-third were covered by Medicare (37.5% and 37.6%), and the remaining patients were covered by Medicaid (13.0% and 13.2%) or another type of insurance (1.8% and 1.6%). Over half of the patients (52.4% among abandoners and 54.3% among non-abandoners) had DVT as their index VTE diagnosis, followed by PE (36.3% and 34.6%, respectively), and the remaining patients had both DVT and PE (11.3% and 11.0%). The mean time from the index VTE diagnosis to index date was 5 days in both cohorts.

VTE Recurrence, Composite Outcome

Mean [standard deviation] duration of follow-up was 14.8 [10.6] months in the abandoner cohort and 15.4 [10.4] months in the non-abandoner cohort. At 3 months of follow-up, the probability of VTE recurrence as a composite outcome of DVT or PE was 7.74% in the abandoner cohort compared to 4.65% in the non-abandoner cohort (log-rank p value < 0.0001; Fig. 2). The higher probability of recurrence in the abandoner compared to the non-abandoner cohort was sustained at later points of follow-up; for example, at 12 months, it was 9.91% in the abandoner and 6.89% in the non-abandoner cohort (log-rank p value < 0.0001).

Fig. 2
figure 2

VTE recurrence (DVT or PE) probability after first DOAC fill in weighted abandoner versus non-abandoner cohortsa. CI confidence interval, DOAC direct oral anticoagulant, DVT deep vein thrombosis, PE pulmonary embolism, VTE venous thromboembolism. aVTE recurrence was identified in an inpatient setting. The time to the first recurrent event was measured from the index date; patients for whom the event was not observed during the follow-up period were censored at the end of the follow-up period

Full size image

The risk of recurrence was 72% higher in the abandoner cohort than the non-abandoner cohort at 3 months, 62% higher at 6 months, 53% higher at 12 months, and 49% higher at 24 months of follow-up (all p values < 0.0001; Fig. 3).

Fig. 3
figure 3

Risk of VTE recurrence in weighted abandoner versus non-abandoner cohorts.a *p value < 0.05; CI confidence interval, DVT deep vein thrombosis, HR hazard ratio, PE pulmonary embolism, VTE venous thromboembolism. aVTE recurrence was identified in an inpatient setting. HRs were generated using univariate weighted Cox proportional hazard models

Full size image

VTE Recurrence, DVT and PE Separately

The probabilities of DVT and PE recurrence were similar within each cohort during the follow-up period. At 3 months of follow-up, the probability of DVT alone was 4.60% in the abandoner cohort compared to 2.98% in the non-abandoner cohort; at 12 months, these probabilities were 5.95% and 4.48%, respectively (both log-rank p values < 0.0001; data not shown). At 3 months of follow-up, the probability of PE alone was 4.56% in the abandoner cohort compared to 2.45% in the non-abandoner cohort; at 12 months, these probabilities were 5.77% and 3.63%, respectively (both log-rank p values < 0.0001; data not shown). The similar probabilities of DVT and PE recurrence within each cohort may be explained by the identification of recurrence only in the inpatient setting. Since DVT is more often treated in outpatient settings than PE [20], DVT recurrence may have been underestimated in this study.

The risk of DVT alone was 38–58% higher in the abandoner cohort than in the non-abandoner cohort at 3, 6, 12, and 24 months of follow-up (all p values < 0.0001). The risk of PE alone was 62–91% higher in the abandoner cohort than in the non-abandoner cohort at 3, 6, 12, and 24 months of follow-up (all p values < 0.0001; Fig. 3).

Sensitivity Analysis Using a Censored Follow-up Period

When assessing VTE recurrence while the abandoner cohort remained untreated and the non-abandoner cohort remained on DOAC treatment, the probabilities of recurrence in the two cohorts were similar to the main analysis (Fig. 4). The risk of VTE recurrence as a composite outcome of DVT or PE was 104–111% higher in the abandoner cohort than in the non-abandoner cohort at 3, 6, 12, and 24 months of follow-up (all p values < 0.0001; Fig. 5).

Fig. 4
figure 4

VTE recurrence (DVT or PE) probability after first DOAC fill in weighted abandoner versus non-abandoner cohorts (censored follow-up period)a,b. CI confidence interval, DOAC direct oral anticoagulant, DVT deep vein thrombosis, PE pulmonary embolism, VTE venous thromboembolism. aVTE recurrence was identified in an inpatient setting. The time to the first recurrent event was measured from the index date; patients for whom the event was not observed during the follow-up period were censored at the end of the follow-up period. bThe censored follow-up period began on the index date and spanned until the earliest of (1) the first non-oral anticoagulant or warfarin claim; (2) the first approved DOAC claim (abandoner cohort only); (3) the date of DOAC discontinuation, defined as the last day with therapy supply before a 15-day gap without DOAC medication (non-abandoner cohort only); (4) the end of clinical activity; or (5) the end of data availability

Full size image
Fig. 5
figure 5

Risk of VTE recurrence in abandoner versus non-abandoner cohorts (censored follow-up period)a,b. *p value < 0.05; CI confidence interval, DVT deep vein thrombosis, HR hazard ratio, PE pulmonary embolism, VTE venous thromboembolism. aVTE recurrence was identified in an inpatient setting. HRs were generated using univariate weighted Cox proportional hazard models. bThe censored follow-up period began on the index date and spanned until the earliest of (1) the first non-oral anticoagulant or warfarin claim; (2) the first approved DOAC claim (abandoner cohort only); (3) the date of DOAC discontinuation, defined as the last day with therapy supply before a 15-day gap without DOAC medication (non-abandoner cohort only); (4) the end of clinical activity; or (5) the end of data availability

Full size image

Discussion

In this real-world study, first DOAC fill abandonment was associated with statistically significant short- and long-term elevated risk of VTE recurrence for DVT and PE, both combined and separately. The results remained robust when assessing VTE recurrence while the abandoner cohort remained untreated, and the non-abandoner cohort remained on DOAC treatment. These findings highlight the need to improve access and primary adherence to DOACs among patients with VTE as abandonment of the first fill could translate into missing one-third of the recommended 3–6 months treatment regimen [3].

There is scarce literature describing the primary abandonment of DOACs among patients with VTE in the USA. In one claims-based retrospective cohort study of adults with at least one prescription for an oral anticoagulant, a significantly higher proportion of DOAC prescriptions were primarily abandoned compared to warfarin (26.7% vs 14.0%) [15]. Of note, several sociodemographic factors predicted abandonment in this study, including higher copays and tier 4 drug status. This association was corroborated by a more recent US study that also found that larger copayments were associated with lower adherence and higher discontinuation of DOACs [21]. However, neither of these two studies was conducted exclusively among patients with VTE, and clinical outcomes like recurrence were not evaluated. Since patient out-of-pocket paid amount for the first DOAC prescription was used as a balancing variable in the propensity score in the current study, further research is warranted to confirm the association between drug costs and primary abandonment and to identify other predictors of DOAC abandonment specifically among patients with VTE.

As observed in the present study and in the literature, lack of anticoagulant treatment for VTE can result in serious clinical implications, including recurrent events and death [5, 22]. While there are no recent estimates of recurrence rates in untreated patients, older studies have reported recurrence rates of 26% within 14 days of initial PE [22] and up to 50% within 3 months of initial symptomatic proximal DVT or PE [5]. These estimates are higher than those of the current study among the abandoner cohort (7.74% for DVT or PE and 4.56% for PE alone at 3 months), possibly because of differing levels of severity of the initial VTE event in the patient populations (e.g., the current study captured index VTE in any care setting, while 32% of untreated patients were deemed to have “severe” PE in a prior study [22]). Patients who initiate anticoagulation and discontinue are also at risk of experiencing a VTE recurrence. In a systematic literature review and meta-analysis of the long-term risk of recurrent VTE, patients who completed at least 3 months of anticoagulant therapy had a 10% risk of recurrence in the first year after treatment discontinuation and a 16% risk at 2 years [23]. Importantly, shorter duration of anticoagulation (i.e., up to 6 months) was associated with a statistically significant increased risk of recurrent VTE (adjusted HR [95% CI] = 1.39 [1.08, 1.80]) in a separate analysis [24]. Taken together, the present study findings and the prior literature emphasize the importance of initiating and adhering to DOAC therapy to minimize the risk of downstream VTE recurrence.

The clinical and economic implications of VTE recurrence are well documented. A meta-analysis of 17 studies estimated that the pooled case fatality rate of recurrent VTE after discontinuation of anticoagulation was 3.8% [23]. Additionally, a separate registry-based study found that both recurrent PE and DVT were significantly associated with subsequent mortality risk [25]. VTE recurrence was also reported to be associated with considerable healthcare resource utilization (HRU) and costs [26, 27].

Given the large economic burden associated with VTE recurrence, it is particularly important to promote primary adherence to DOACs to reduce the incidence of costly VTE recurrences. As such, future research is needed to gain a better understanding of the factors associated with primary abandonment and how they affect recurrence rates and subsequent clinical and economic outcomes in patients with VTE. This insight would be instrumental for the development of policies aimed at optimizing VTE management and for informing treatment decision-making in clinical practice.

Limitations

Some limitations of the study should be noted. The database does not capture services received from providers outside of the network, so the use of pharmacological treatments and medical services could have been underestimated. Specifically, some patients may have appeared as less compliant to DOAC therapy if they changed to another claims transaction network for their prescriptions. As with all claims or transaction database studies, prescription fills did not account for whether the medication dispensed was taken as prescribed, which may have led to an overestimation of treatment continuation. Given the possibility of treatment discontinuation after the first fill among non-abandoners or the initiation of DOAC treatment after the abandoned first fill among abandoners, these results are likely a conservative estimate. Additionally, drug samples and over-the-counter medications were not captured in the data. Indeed, abandoners may have received free samples instead of filling their first DOAC prescription, which may have biased results towards the null and led to a more conservative effect observed. Although the database captures more than 75% of all US retail prescription claims, its representativeness of the total US population was not assessed in this study and thus cannot be concluded. Moreover, the reporting of VTE recurrence using ICD-10-CM codes in the database was not validated. Lastly, the analysis may have been subject to residual confounding due to unmeasured confounders (i.e., information not available in the transactions data).

Conclusions

In this study of patients with VTE in real-world clinical practice, apparent abandonment of the first DOAC prescription fill was associated with a significantly higher risk of VTE recurrence in the short term, which was also sustained for 2 years after abandonment. Therefore, ensuring proper access and encouraging early and continuous use of DOACs may help prevent severe and fatal complications among patients with VTE.