Sage Journals: Discover world-class research

Abstract

Purpose:

Direct oral anticoagulants (DOACs) have been shown to be as effective or superior to warfarin, but warfarin use remains constant. Knowledge regarding the patient population who have switched from a DOAC to warfarin is limited. The objective of this study was to identify clinical predictors which may influence a patient’s likelihood of switching from a DOAC to warfarin for atrial fibrillation (AF) or venous thromboembolism (VTE).

Methods:

In this single-center, case-control study, patients who switched from a DOAC to warfarin were compared with patients who remained on a DOAC. Baseline demographics were compared between the switch and control groups. Independent factors that increased the likelihood of switching from a DOAC to warfarin were analyzed using logistic regression.

Results:

A total of 150 patients were included in the control (n = 100) and switch (n = 50) groups. Patients switched from a DOAC to warfarin had more medications at baseline (9 [7, 13] vs 11 [8, 18], P = 0.009). The presence of heart failure (HF) increased the likelihood of switching (odds ratio [OR] = 3.95, confidence interval [CI] = 1.70-9.21, P = 0.002), and for every 10 mL/min increase in creatinine clearance (CrCl), the likelihood of switching decreased (R = 0.89 [0.80-0.99], P = 0.026). Patients with pulmonary embolism (PE) were less likely to switch from a DOAC to warfarin (OR = 0.20, CI = 0.05-0.86, P = 0.031). Explicitly listed reasons for switching included left ventricular assist device (LVAD) implantation (20%) and valve replacement procedures (20%).

Conclusion:

Congestive HF was a clinical predictor associated with an increased likelihood of switching from a DOAC to warfarin. Anticoagulation therapy for PE and higher CrCl was associated with a decreased likelihood in switching from DOAC to warfarin.

Keywords

Direct oral anticoagulant warfarin switch atrial fibrillation venous thromboembolism

Introduction

Atrial fibrillation (AF) and venous thromboembolism (VTE) are among the leading causes of morbidity and mortality in the United States.^1-3 For many years, warfarin, a vitamin K antagonist (VKA), has been the mainstay of anticoagulation therapy for AF and VTE.⁴ While efficacious, anticoagulation with warfarin is associated with several challenges. Warfarin therapy requires routine monitoring of international normalized ratio (INR) to maintain a patient-specific therapeutic target. Warfarin metabolism is also greatly sensitive to variations in diet and numerous drug interactions. Therapeutic INR values may be difficult to sustain as evidenced by an average time in therapeutic range of 55% for US patients.⁴

Anticoagulation management for AF and VTE has expanded given the advent of the direct oral anticoagulants (DOACs).⁵ Unlike warfarin, DOACs do not require INR monitoring to ensure therapeutic levels. Clinical trials for the DOACs have shown a significant decrease in risk of bleeding, particularly intracranial hemorrhage, when compared with warfarin.^6-9 For this reason, many patients and physicians desire therapy with DOACs over warfarin when choosing anticoagulation therapy for the management of AF or VTE, and often switch to DOAC therapy even when established on a VKA.¹⁰

Factors associated with the utilization of a DOAC over warfarin at the start of anticoagulation therapy include younger age, male sex, prior ischemic stroke, lower risk scores of stroke or bleeding, fewer comorbidies or comedications, superior renal function, prescription insurance, and prescription by a cardiologist.¹¹ The safety profile and less burdensome monitoring parameters associated with DOACs compared with warfarin may be desirable for patients; however, there are several clinical scenarios in which a patient may not be a suitable candidate for anticoagulation therapy with a DOAC. This may include patients with renal dysfunction, mechanical heart valves, moderate-to-severe mitral stenosis, dual antiplatelet therapy for coronary stents in addition to anticoagulation therapy, left ventricular device (LVAD) implantation, or those who cannot afford the cost of DOACs.^4,12,13

Current knowledge about choice of anticoagulation for the management of AF or VTE is mainly derived from studies done in populations initiating anticoagulation therapy or switching from warfarin to DOAC.^13,14 Little is known regarding the patient population who have followed the reverse path of switching from a DOAC to warfarin.^15,16 In a recent review of 40 patients, bleeding, rethrombosis, and worsening renal function were the most common reasons for switching. Knowledge surrounding this topic is vital when selecting optimal anticoagulation therapy for AF or VTE. Awareness of these clinical predictors may help facilitate decisions when choosing anticoagulation therapy as well as reduce the need for therapy modification.

The aim of this study was to describe and identify clinical predictors which may influence a patient’s likelihood of switching from a DOAC to warfarin for AF or VTE.

Methods

This was a single-center, case-control study at a large academic medical center, the University of Michigan. Patients from the institution prescribed a DOAC for AF or VTE from January 1, 2014 to December 31, 2017 were included in the screening portion of the analysis. Data were collected from the electronic medical record. Patients included in the study were required to be 18 years or older and have a diagnosis of AF or VTE confirmed through ICD-9 or ICD-10 codes. Patients were excluded if they were on DOAC therapy less than 14 days, transitioned from a DOAC to a non-VKA or back to a DOAC, or had insufficient data. Those switched from a DOAC to warfarin were placed in the “switch” group. A group of control patients were collected at a 2:1 ratio consisting of patients who remained on a DOAC (Figure 1). A total of 176 patients met study criteria and data collection concluded when a total of 150 patients were collected chronologically between the two groups. Anticoagulation initiation and monitoring at the institution was primarily led by providers and pharmacists. Dosing of DOACs and warfarin was driven by national guidelines for AF and VTE.^17-19 Provider documentation, lab values, and medication administration records were used to determine study endpoints. The Cockcroft-Gault equation was used to calculate creatinine clearance (CrCl) in milliliters per minute (mL/min).

Figure 1.

Data collection timeline.

Data collected from patient records consisted of age, sex, height, body weight, serum creatinine, comorbidities, indication for anticoagulation, number of medications at baseline, concurrent antithrombotic usage, surgical history within 30 days of switch to warfarin if applicable, insurance coverage, and specialty of prescribing provider. Explicitly stated reasons for switching from DOAC to warfarin were documented if listed within the chart. Defined agents for anticoagulation therapy included dabigatran, rivaroxaban, apixaban, edoxaban, and warfarin. Baseline values for patients initiated on a DOAC at an outside facility were collected at the time of first contact to the study health system.

Demographic and clinical characteristics were compared between the control group and the switch group. Student t-test or Wilcoxon rank sum test, as appropriate, was used to compare continuous variables. Chi-square and Fisher exact test were used for categorical variables. Multivariable logistic regression was used to identify clinical predictors of switching from a DOAC to warfarin using variables with a P-value less than 0.1 in the univariate analysis. All statistics were analyzed using SAS statistical software (version 9.4).

Results

A total of 176 patients were included in the initial screening portion of the study. Of these, 150 patients were included in the switch (n = 50) and control (n = 100) groups (Figure 2). Baseline characteristics are shown in Table 1. The majority of patients were men with median age of 66 years. The groups were well matched in weight, concurrent antithrombotic use, insurance coverage, and DOAC prescriber specialty. Most patients were treated for AF, followed by deep vein thrombosis (DVT), and pulmonary embolism (PE).

Figure 2.

Patient enrollment.

Table 1.

Baseline demographics.^a

	Switch group(n = 50)	DOAC control(n = 100)	P-value
Men, n (%)	32 (64)	65 (65)	0.904
Age, years	67 [53, 71]	66 [56, 74]	0.617
Weight, kg	93 [75, 112]	92 [78, 110]	0.922
Number of medications	11 [8, 18]	9 [7, 13]	0.009
Baseline DOAC, n (%)			0.336
Apixaban	26 (52)	42 (42)
Rivaroxaban	21 (42)	54 (54)
Dabigatran	3 (6)	4 (4)
Edoxaban	0 (0)	0 (0)
Concurrent antithrombotics, n (%)			0.764
Zero	29 (58)	64 (64)
One	20 (40)	34 (34)
Two	1 (2)	2 (2)
Patient insurance, n (%)
Medicare	31 (62)	59 (59)	0.724
Medicaid	10 (20)	9 (9)	0.056
Private insurance	10 (20)	32 (32)	0.123
DOAC prescriber, n (%)
Cardiology	27 (54)	47 (47)	0.419
Internal medicine	8 (16)	23 (23)	0.318
Primary care	0 (0)	4 (4)	0.302
Other	15 (30)	26 (26)	0.698
Renal function
Serum creatinine, mg/dL	1.20 [0.90, 1.50]	0.88 [0.76, 1.10]	<0.001
eGFR > 60 mL/min, n (%)	24 (48)	75 (75)	0.001
CrCl, mL/min	57 [42, 74]	74 [50, 100]	0.001
Indication for anticoagulation,^b n (%)
Atrial fibrillation	45 (90)	89 (89)	0.851
Deep vein thrombosis	8 (16)	20 (20)	0.553
Pulmonary embolism	3 (6)	17 (17)	0.062
Patient comorbidities, n (%)
BAVR	0 (0)	2 (2)	0.553
Diabetes	18 (36)	22 (22)	0.068
Liver disease	2 (4)	1 (1)	0.258
CHF	31 (62)	24 (24)	<0.001
Solid tumor	2 (4)	3 (3)	1
Myocardial infarction	6 (12)	8 (8)	0.552
COPD	4 (8)	8 (8)	1
PVD	0 (0)	3 (3)	0.551
CVA/TIA	6 (12)	14 (14)	0.734
Peptic ulcer disease	1 (2)	0 (0)	0.333
Hypertension	32 (64)	61 (61)	0.721

Abbreviations: DOAC, direct oral anticoagulant; eGFR, estimated glomerular filtration rate; CrCl, creatinine clearance; BAVR, bioprosthetic valve replacement; CHF, congestive heart failure; COPD, chronic obstructive pulmonary disease; PVD, peripheral vascular disease; CVA, cerebrovascular accident; TIA, transient ischemic attack; IQR, interquartile range.

Data expressed as median (IQR) or n (%).

The percentages do not add up to 100% because patients can have more than one indication for anticoagulation.

Of the DOACs, a majority of patients in the control and switch groups were prescribed rivaroxaban (42% vs 54%) and apixaban (52% vs 42%) followed by dabigatran (6% vs 4%) (P = 0.336). No patients in the study were prescribed edoxaban.

Patients in the switch group were prescribed more medications at baseline (11 [8, 18] vs 9 [7, 13], P = 0.009). Renal function was better in the control group as assessed by baseline serum creatinine (1.20 [0.90, 1.50] vs 0.88 [0.76, 1.11], P ⩽ 0.001), estimated glomerular filtration rate (eGFR) greater than 60 mL/min (48% vs 75%, P = 0.001), and CrCl (mL/min; 57 [42, 74] vs 74 [50, 100], P = 0.001). The switch group was found to have a larger population of patients with congestive heart failure (CHF) compared with that of the control group (62% vs 24%, P ⩽ 0.001) (Table 1).

Multivariable logistic regression revealed several significant clinical predictors of switching from a DOAC to warfarin. Diagnosis of HF increased the odds of switching from a DOAC to warfarin by almost four times (odds ratio [OR] = 3.95, confidence interval [CI] = 1.70-9.21, P = 0.002). The analysis identified a 11.2% decrease in the likelihood of switching from a DOAC to warfarin with every 10 mL/min point increase in CrCl. Compared with other indications, treatment with anticoagulation for PE reduced the odds of switching from a DOAC to warfarin by 80% (OR = 0.20, CI = 0.05-0.86, P = 0.031). Diagnosis of diabetes (OR = 1.89, CI = 0.76-4.70, P = 0.172), medicaid insurance (OR = 2.95, CI = 0.91-9.61, P = 0.072), and number of medications at baseline (OR = 1.00, CI = 0.92-1.08, P = 0.936) were found to be insignificant in the logistic regression model (Table 2).

Table 2.

Multivariable analysis.

Covariate	Odds ratio (95% CI)	P-value
Creatinine clearance, per 10 mL/min	0.89 (0.80-0.99)	0.026
Diabetes	1.89 (0.76-4.70)	0.172
CHF	3.95 (1.70-9.21)	0.002
Pulmonary embolism	0.20 (0.05-0.86)	0.031
Medicaid insurance	2.95 (0.91-9.61)	0.072
Medications at baseline	1.00 (0.92-1.08)	0.936

Abbreviations: CI, confidence interval; CHF, congestive heart failure.

In addition to our quantitative analysis, explicitly stated reasons for switching from DOAC to warfarin listed within each individual chart were assessed. The majority of patients were documented to have switched secondary to left ventricular assist device (LVAD) implantation (20%), valve replacement procedures (20%), acute kidney injury (8%), cost (8%), and DOAC failure in the setting of thrombosis (8%) (Table 3).

Table 3.

Explicit reasons for switching from direct oral anticoagulant.^a

Reason for switching	n = 50
LVAD implantation	10 (20)
Valve replacement	10 (20)
Acute kidney injury	4 (8)
Cost	4 (8)
DOAC failure (thrombosis)	4 (8)
Renal dysfunction	2 (4)
Intolerance to DOAC	2 (4)
Bleeding event	2 (4)
Lack of reversal agent	2 (4)
Recent surgery	3 (6)
Other^b	3 (6)
Unknown	4 (8)

Abbreviations: LVAD, left ventricular assist device; DOAC, direct oral anticoagulant.

Data expressed as n (%).

Compliance, transplant patient, and liver dysfunction.

Discussion

Our study identified clinical predictors of switching from DOACs to warfarin in patients treated for AF or VTE. Identifying specific clinical indicators which may increase the likelihood of switching from a DOAC to warfarin may help guide initial anticoagulation management and prevent the need for therapy modification. Patients who switched from a DOAC to warfarin had more medications at baseline. Brais et al similarly identified that DOACs were more commonly prescribed to patients with fewer concomitant medications (OR = 0.30, 95% CI = 0.10-0.89). Though not assessed within our study, additional medications may increase the incidence of drug-drug interactions with DOACs and result in a switch to warfarin therapy where the impact of these interactions can be monitored and adjusted for via INR monitoring. More baseline medications may also suggest patients with more acute illness which could result in complex medication regimens, end organ dysfunction, and increased risk for bleeding where physicians feel more comfortable regularly monitoring anticoagulation therapy through INR testing.

Baseline renal function plays a significant role in the risk of switching from a DOAC to warfarin. Barrett et al¹⁶ found that 9 out of 40 patients (17.5%) were switched from a DOAC to warfarin due to renal function. Baseline renal function was better in the switch group than that of the control group across all measures within our study. Metabolism of the DOACs relies heavily on renal function which may explain these trends. At the time of initiation, renal function parameters did not preclude the use of DOACs, suggesting that patients with compromised renal function at the start of DOAC therapy may be more likely to switch to warfarin as time progresses and renal function declines.

A significant number of patients switching from a DOAC to warfarin had a diagnosis of CHF. Patients with heart failure often require use of diuretics which can result in fluctuating fluid balance and renal function. As CHF progresses, patients may develop cardiorenal syndrome, subsequently increasing the risk for LVAD implantation. Fluctuation and worsening of renal failure in the setting of CHF may explain the increased likelihood of switching from a DOAC to warfarin found within our study. Regarding the use of DOACs in LVAD devices, a study by Terrovitis et al is one of the few to have explored the safety and efficacy of a DOAC (dabigatran) as second-line anticoagulation with HeartMate II (HMII) LVAD implantation to date. Of the seven patients assessed within the study, two experienced gastrointestinal bleeding events, while another two patients experienced device thrombosis and ischemic stroke. The authors concluded that larger, prospective, randomized control trials were needed to provide definite answers regarding the efficacy and safety of dabigatran use in patients with HMII devices.²⁰ Due to the paucity data surrounding DOAC use in LVAD populations, current guidelines do not recommend their use in this patient population.¹²

Patients treated for PE were found to have a decreased risk of switching from a DOAC to warfarin, which may be explained by more defined treatment durations in PE compared with that of AF. There was however no significant difference in the risk of switching in patients treated for DVT, thus indicating that further research is needed to better understand the role of anticoagulation indication and risk of switching from DOAC to warfarin.

Additional data were collected to provide information about why providers may switch patients from DOACs to warfarin. Valve replacement procedures (20%) and LVAD implantation (20%) were two of the prominent reasons explicitly stated for switching therapy from DOACs to warfarin within the patient charts. Diagnosis of CHF can be directly linked to LVAD implantation and can explain the relationship between these results. Overall, these results align with the recommendations against the use of DOACs in valvular heart disease and LVAD devices.^4,12,21

Unlike warfarin, the DOACs remain under drug patent, which can often result in a higher cost to the patient. The minimum of 14 days prior to switch from DOAC to warfarin was chosen to reduce the number of patients switched to warfarin solely due to cost. Instead, we sought to compare insurance coverage between the control and the switch populations in the hope of identifying whether specific insurance coverage resulted in an increased likelihood of switching from DOAC to warfarin. Since there was no statistically significant difference in the primary insurance types between the two groups, it remains inconclusive whether insurance coverage alone increases the likelihood of switching from a DOAC to warfarin.

Overall, data identifying clinical predictors of switching from DOACs to warfarin are limited. Manzoor et al retrospectively evaluated patterns of prescription switching in non-valvular AF patients initiated on a DOAC and previously naïve to anticoagulation therapy through the use of insurance claims data. Of the 6613 patients who switched from their index DOAC to an alternative anticoagulant, 44.5% of the patients were switched to warfarin. Overall, the study found that older age and female sex were significantly associated with a higher likelihood to switch index DOAC therapy.¹⁵ Results from our study did not find a significant difference between baseline age and sex in relation to risk for switching which could potentially be explained by our small study population, single-center design, and the large number of patients on apixaban or rivaroxaban. Results from our study are unique in the fact that they are derived from real-world evidence and manual chart review as opposed to insurance claims data. Furthermore, results from this study focus on switches directly from DOACs to warfarin.

This study was limited by its single-center, retrospective design. Our study included a large number of patients switched from DOAC to warfarin secondary to LVAD implantation. Many facilities do not perform LVAD implantation in patients, and these results may not be generalizable to all locations or institutions. In addition, demographics were assessed at baseline but not at the time of switch from DOAC to warfarin. Unknown changes in baseline characteristics, such as renal function, from the time of DOAC initiation until the switch to warfarin may have contributed to reasons for the change in therapy. The writers acknowledge the potential for selection bias in this study as patients switched from a DOAC to warfarin may have had more acute problems at the time of anticoagulation switch compared with those who remained on a DOAC. Furthermore, it is difficult to extrapolate these results to dabigatran and edoxaban because the primary DOACs in our sample were apixaban and rivaroxaban. Finally, endpoints of this research topic were not well published or documented prior to the undertaking of this study making it difficult to calculate statistical power.

Conclusions

This study assessed clinical predictors of switching from a DOAC to warfarin in a real-world patient population. Clinical predictors associated with an increased risk of switching from a DOAC to warfarin in AF or VTE included more medications at baseline, diagnosis of CHF, and lower creatinine clearance at baseline. Anticoagulation therapy for PE was associated with a decreased risk in switching from DOAC to warfarin. Awareness of these clinical predictors may aid in the facilitation of anticoagulation choice at the start of AF or VTE therapy in addition to reducing the need for therapy modification. Replication of this clinical study on a larger scale should be done to corroborate the results of this research.

Footnotes

Funding:

The author(s) received no financial support for the research,authorship,and/or publication of this article.

Declaration of conflicting interests:

The author(s) declared the following potential conflicts of interest with respect to the research,authorship,and/or publication of this article: Michael Dorsch has received research funding from Johnson and Johnson/Jansen.

Author Contributions

AR and MD researched literature and conceived the study. AR was involved in protocol development,gaining ethical approval,patient recruitment,and data analysis. AR wrote the first draft of the manuscript. All authors reviewed and edited the manuscript and approved the final version of the manuscript.

Ethical Approval

Ethical approval for this study was obtained from the University of Michigan Institutional Review Board (HUM00136309).

ORCID iD

Kristen Pogue

References

Patel

Deshmukh

Pant

et al . Contemporary trends of hospitalization for atrial fibrillation in the united states, 2000 through 2010 implications for healthcare planning. Circulation. 2014;129:2371–2379.

Mahan

Borrego

Woersching

et al . Venous thromboembolism: annualised United States models for total, hospital-acquired and preventable costs utilising long-term attack rates. Thromb Haemost. 2012;108:291–302.

Fitch

Broulette

Kwong

WJ.

The economic burden of ischemic stroke and major hemorrhage in medicare beneficiaries with nonvalvular atrial fibrillation: a retrospective claims analysis. Am Heal Drug Benefits. 2014;7:200–209.

Lopes

Guimaraes

Kaatz

Practical and clinical considerations in assessing patients with atrial fibrillation for switching to non-vitamin K antagonist oral anticoagulants in primary care. Int J Gen Med. 2015;8:283.

Barnes

Lucas

Alexander

Goldberger

ZD.

National trends in ambulatory oral anticoagulant use. Am J Med. 2015;128:1300–1305.

Granger

Alexander

McMurray

et al . Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2011;365:981–992.

Connolly

Ezekowitz

Yusuf

et al . Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med. 2009;361:1139–1151.

Patel

Mahaffey

Garg

et al . Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med. 2011;365:883–891.

Agnelli

Buller

Cohen

et al . Oral apixaban for the treatment of acute venous thromboembolism. N Engl J Med. 2013;369:799–808.

10.

Alalwan

Voils

Hartzema

Trends in the utilization of warfarin and the new oral anticoagulants in atrial fibrillation patients. Pharmacoepidemiol Drug Saf. 2015;24:133–134.

11.

Brais

Larochelle

Turgeon

et al . Predictors of direct oral anticoagulants utilization for thromboembolism prevention in atrial fibrillation. J Pharm Pharm Sci. 2017;20:8–14.

12.

Feldman

Pamboukian

Teuteberg

et al . The 2013 International Society for Heart and Lung Transplantation Guidelines for mechanical circulatory support: executive summary. J Hear Lung Transplant. 2013;32:157–187.

13.

AbuDagga

Stephenson

A-C

Kwong

Tan

Weintraub

WS.

Characteristics affecting oral anticoagulant therapy choice among patients with non-valvular atrial fibrillation: a retrospective claims analysis. BMC Health Serv Res. 2014;14:310.

14.

Lauffenburger

Farley

Gehi

Rhoney

Brookhart

Fangg

HHS public access. Am J Cardiol. 2015;115:1095–1101.

15.

Manzoor

Walton

Sharp

Galanter

Lee

Nutescu

EA.

High number of newly initiated direct oral anticoagulant users switch to alternate anticoagulant therapy. J Thromb Thrombolysis. 2017;44:435–441.

16.

Barrett

Moore

Ferrins

Thornton

Murphy

Quinn

From a direct oral anticoagulant to warfarin: reasons why patients switch. Ir J Med Sci. 2018;187:719–721.

17.

Kearon

Akl

Ornelas

et al . Antithrombotic therapy for VTE disease: CHEST guideline and expert panel report. Chest. 2016;149:315–352.

18.

Guyatt

Akl

Crowther

Gutterman

Schunemann

HJ.

Executive summary: antithrombotic therapy and prevention of thrombosis, 9th ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest. 2012;141:7S–47S.

19.

January

Wann

Alpert

et al . 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the Heart Rhythm Society. Circulation. 2014;130:e199–e267.

20.

Terrovitis

Ntalianis

Kapelios

et al . Dabigatran etexilate as second-line therapy in patients with a left ventricular assist device. Hell J Cardiol. 2015;56:20–25.

21.

Eikelboom