Effectiveness and safety of oral anticoagulants in patients with sickle cell disease and venous thromboembolism: a retrospective cohort study

Abstract

Patients with sickle cell disease (SCD) experience initial and recurrent venous thromboembolism (VTE) more commonly and at a younger age than the general population, and it confers a higher mortality for patients with SCD. However, limited evidence is available to guide anticoagulant use for VTE treatment in this population. The primary objective of this study is to characterize the effectiveness and safety of direct oral anticoagulants (DOAC) and warfarin for VTE treatment among patients with SCD. This single-center retrospective study includes adult patients with SCD who were diagnosed with VTE. Data was obtained from review of electronic health records for the 6 months after VTE diagnosis. Among the 22 patients treated initially with a DOAC, 6 (27%) developed recurrent VTE, none experienced major bleeding, and 3 (14%) experienced clinically relevant non-major bleeding (CRNMB). Similarly, of 15 patients initially treated with warfarin, 3 (20%) developed a recurrent VTE, 1 (7%) experienced major bleeding, and 2 (13%) experienced CRNMB. Twelve patients received more than one oral anticoagulant during the study period, most commonly due to a recurrent VTE, concern for non-adherence, or subtherapeutic INR. Overall, the incidence of VTE recurrence and bleeding events were similar between groups, but occurred at a higher rate than those found in major clinical trials of anticoagulant agents. Prescribers should continue to individualize therapeutic decision-making regarding oral anticoagulant therapy for VTE treatment for individuals with SCD based on patient-specific factors and anticipated ability to adhere to the drug regimen or required monitoring.

Introduction

Among patients with sickle cell disease (SCD), the frequency of venous thromboembolism (VTE) is greater and the age of onset is younger than the general population [1–4]. These patients also experience a higher rate of VTE recurrence (25–29%) [3, 5] and increased mortality compared to patients without SCD [6, 7].

The high VTE incidence and recurrence among patients with SCD is attributed to a pro-thrombotic state induced by iron overload and alterations in nearly every component of hemostasis [4]. In addition, the common use of central venous catheters (CVC) and frequent hospitalizations and surgeries contribute to the increased incidence of VTE [4].

Despite the heightened incidence, recurrence, and mortality of VTE in SCD, there is limited evidence to guide treatment. Current oral anticoagulant options within VTE treatment guidelines include warfarin or direct oral anticoagulants (DOACs): apixaban, dabigatran, edoxaban, or rivaroxaban [8]. Warfarin use is constrained by its narrow therapeutic index, unpredictable pharmacokinetics and pharmacodynamics, and food and drug interactions [9]. A single institution study of warfarin in patients with SCD and upper extremity deep vein thrombosis (UEDVT) found a low time within the therapeutic INR range (TTR) of 25% [5]. The inflammatory manifestations of SCD [10] also pose concern for alterations in warfarin pharmacokinetics based on in vitro data [11]. Thus, the potential use of DOACs for patients with SCD is appealing, theoretically avoiding the need for monitoring due to predictable pharmacokinetics [9]. Results of a murine model of SCD suggest that DOACs could be used in SCD to reduce thrombotic complications [12]. A single study has evaluated the use of DOACs among patients with SCD and VTE, demonstrating clinical failure in 2 of 7 patients [13]. The goal of this study is to evaluate the effectiveness and safety of warfarin and DOACs for VTE treatment among patients with SCD.

Materials and methods

Study population and design

This retrospective study was conducted using patient records from an academic healthcare system from July 1, 2012 to June 30, 2016. Patients were identified based ICD codes for SCD and VTE and medication orders for oral anticoagulant agents or anticoagulant reversal agents. Inclusion criteria were confirmed for patients 18 years of age or older with SCD diagnosis established by hemoglobin electrophoresis, incidence of a new, acute VTE verified by diagnostic imaging, and administration of a DOAC or warfarin in the inpatient setting or prescription from affiliated outpatient provider clinics. Individuals who were pregnant, treated for an isolated distal VTE, or receiving anticoagulant agents for indications other than VTE were excluded. Institutional Review Board approval was obtained.

Outcomes and definitions

All effectiveness and safety outcomes were evaluated for the 6 months following the index VTE event and identified by review of the electronic health record. VTE was defined as a pulmonary embolism (PE), proximal lower extremity (LEDVT) or upper extremity (UEDVT). The primary effectiveness outcome was the incidence of recurrent VTE. The secondary effectiveness outcome was hospitalization for VTE complications.

Safety outcomes coincide with clinical trials of DOACs [14–16]. The primary safety outcome was incidence of major bleeding and secondary safety outcomes included clinically relevant non-major bleeding (CRNMB) and any bleeding. Drop in hemoglobin or incidence of transfusions as indicators of bleeding were not incorporated for safety analysis, as these are not necessarily indicative of bleeding among patients with SCD [17].

Additional data points were evaluated to further characterize the population. Patients receiving more than one oral anticoagulant during the study period, had outcomes evaluated based solely on the first oral anticoagulant agent received.

Statistical analysis

Descriptive statistics were used to describe baseline characteristics. For differences between the 2 groups, Chi square tests, Fisher’s exact tests, and t tests were used, as appropriate. A Kaplan–Meier curve and log-rank test were used to describe the time to recurrence between the two groups. p values < 0.05 were considered statistically significant. Analyses were performed with Microsoft Excel 2010 and SPSS for Mac®, version 24 (SPSS Inc., Cary, NC).

Results

Thirty-seven patients were included, of whom 22 received a DOAC and 15 received warfarin (Online Resource 1). Within the DOAC group, 15 patients received rivaroxaban, 6 received apixaban, and 1 received dabigatran. One patient reported Pacific Islander, while the remaining were African American. Additional baseline characteristics were similar between groups and without statistically significant differences (Online Resource 2).

All warfarin-treated patients received bridge therapy with enoxaparin (13/15, 87%) or unfractionated heparin (2/15, 13%). The median duration was 24 days (1st–3rd quartile range of 10–46) among the 11 patients for whom bridge therapy was clearly documented. The TTR determined based on standard methods [18] beginning on day 15 of therapy ranged from 0 to 39%. DOAC regimens were concordant with FDA-approved dosing in majority (16/19, 84%) of patients; 3 patients received rivaroxaban at dosages lower than FDA-approved without documented rationale for off-label dosing. None of these 3 patients experienced recurrent VTE. Warfarin and DOAC patients reported non-adherence rates of 64% (9/14) and 50% (8/16), respectively.

The primary effectiveness outcome of recurrent VTE within 6 months occurred in 24% of the study population (Online Resource 3). The rate of VTE recurrence was similar among groups and the overall mean time to recurrence was 89 days (Fig. 1). VTE recurrences included development of a new PE (56%) and a new DVT (44%). One patient with a recurrent VTE on apixaban was transitioned to enoxaparin as recommended by VTE treatment guidelines [8]. Other patients with recurrent VTE were continued on the current anticoagulant agent (3/9, 33%) or switched to another oral anticoagulant (5/9, 56%). One patient receiving warfarin for bilateral PE treatment died after 124 days; however, cause of the patient’s death was not elucidated, due to fatality occurrence outside of the study institution and limited documentation. Although not statistically significant (p = 0.3, Fisher’s exact test), warfarin treated patients were hospitalized more frequently for VTE complications compared with DOAC treated patients (Online Resource 3 and 4). In addition, concomitant medications with SCD modifying properties or drug–drug interactions were evaluated. Among the 9 patients with recurrent VTE 4 patients were received hydroxyurea and 1 patient was received aspirin. No patient with recurrent VTE received concomitant hydroxyurea and aspirin (Online Resource 4).

Fig. 1.

Time to venous thromboembolism recurrence

Findings of safety outcomes indicate 1 patient in the warfarin group experienced major bleeding, a vitreal hemorrhage. At the time of the hemorrhage, the patient had an INR of 1.36 and aPTT of 105.3 s (goal 60–100 s) while receiving unfractionated heparin at 14 units/kg/h as bridge therapy. In each group, 13% of patients experienced CRNMB, while a greater proportion of the warfarin group experienced any bleeding (53.3 vs. 27.3%, p = 0.169, Fisher’s exact test). Bleeding events included epistaxis, abnormal vaginal bleeding, lower extremity ulcer site bleeding, and external ear bleeding (Online Resource 3 and 4).

Discussion

DOAC use in patients with SCD reflects current national trends in oral anticoagulant use, but evaluation of use in this population is limited [13]. This study provides additional clinical data regarding use of DOACs and warfarin among patients with SCD [5, 19]. VTE recurrence was similar between DOAC and warfarin groups, but the incidence was 10-fold higher than observed in landmark studies in patients without SCD [14–16].

The cause of the high recurrence rate cannot be determined from the current data, but it is possible that in addition to the pro-thrombotic conditions associated with SCD [4] and reported non-adherence [20], pharmacokinetics or pharmacodynamics may be altered in this generally younger patient population. Benefits of dose intensification and monitoring drug levels or coagulation assays warrants further investigation, particularly given the use of DOACs in obese patients [21]. The frequency of dosing discordance is consistent with published data [22, 23]. Variation from approved doses may confer altered risk of treatment failure or adverse events.

The low TTR for patients prescribed warfarin in this study is consistent with prior studies in patients with SCD [5]. In this study, the overall rate of patient reported missed doses was 56.6%. This contrasts with the over 60% achievement of INR within the therapeutic window and 80% or greater adherence rates of over 94% of patients in clinical trials [14–16]. Patients with SCD may not adhere well to prescribed treatments; as evidenced by studies describing hydroxyurea adherence [24, 25]. Attributing low TTR to adherence versus other factors that may affect patients with SCD disproportionately such as inflammation, changes in general activity level, wellness, diet, or drug interactions [12, 26, 27] presents a challenge to treating clinicians. Theoretically, intermittently non-adherent patients may benefit from warfarin, while those with labile INRs for other reasons may benefit from a DOAC.

Safety outcomes were similar between DOAC and warfarin groups and the incidence of major bleeding was comparable to clinical trials [14–16]. A higher incidence of CRNMB in this study population (13.5 vs. 4 to 10%) was not readily explicable.

This is the largest study assessing effectiveness and safety of oral anticoagulant use among patients with SCD, but it includes small population from a single healthcare system. Attribution of outcomes would be best assessed prospectively. In addition, nearly 25% of the study population was lost to follow-up, possibly resulting in the underreporting of outcomes. Receipt of VTE prophylaxis at time of the index VTE event was not assessed within this study, yet is reported as frequently utilized and safe [28]. Further, even if prescribing bias occurred, the high incidence of recurrence engenders a sense of urgency to better treat VTE in SCD.

In conclusion, the rate of VTE recurrence in patients with SCD was high and similar among patients treated with warfarin or a DOAC. Future assessment of treatment, including drug choice, anticoagulation intensity, and adherence are necessary to guide optimal therapy. Until more data becomes available, warfarin and DOACs appear to be equally viable options to treat VTE in patients with SCD, but dosing and adherence should be carefully assessed.