Changing Treatment Patterns in Patients With Venous Thromboembolism in Taiwan

Abstract

Background: In Asia, little information is available about contemporary real-world treatment patterns for venous thromboembolism (VTE).

Methods and Results: Consecutive patients (n=11,414) from the Taiwan National Health Insurance Research Database with initial VTE and taking oral anticoagulants between May 1, 2014 and June 30, 2016 were included. The temporal trends of using oral anticoagulants and pharmacomechanical therapy during the study period were evaluated. The efficacy and safety of nonvitamin K antagonist oral anticoagulants (NOACs) vs. warfarin were compared. Propensity score analysis (NOACs n=3,647 vs. warfarin n=3,647) was used to balance covariates between groups, and Cox proportional hazards models with adjustment were used to estimate the risks of clinical outcomes. The use of NOACs increased from 0.3% to 60.2% for VTE treatment during the study period. Pharmacomechanical therapy was used in 9.60%, 8.22%, and 5.63% from 2014 through 2016. NOACs were associated with a 16% risk reduction (adjusted hazard ratio [aHR] 0.84, 95% confidence interval [CI] 0.77–0.93) in all-cause mortality and a 21% risk reduction (aHR 0.79, 95% CI 0.65–0.96) in recurrent VTE vs. warfarin. Overall, NOACs were associated with a lower risk of major bleeding compared with warfarin (aHR 0.804, 95% CI 0.648–0.998).

Conclusions: In real-world practice, NOACs have become the major anticoagulant used for Asians with VTE. Although NOACs had a lower risk of recurrent VTE and major bleeding compared with warfarin in Taiwan, we still need a large-scale randomized controlled trial to confirm the findings.

Venous thromboembolism (VTE), including deep vein thrombosis (DVT) and pulmonary embolism (PE), are serious and sometimes life-threatening conditions that require immediate medical attention.¹ Heparin or low-molecular-weight heparin and warfarin are the traditional standard treatments for VTE. Development of nonvitamin K dependent oral anticoagulants (NOACs), including dabigatran, rivaroxaban, apixaban and edoxaban, has revolutionized the treatment of VTE. Randomized clinical trials have demonstrated that NOACs are as effective as traditional standard anticoagulation therapy at reducing recurrent VTE with similar or reduced rates of major bleeding.^2–7 Although evidence of long-term efficacy is lacking, pharmacomechanical therapy with catheter-directed thrombolysis, thrombus aspiration or maceration, or percutaneous transluminal balloon venoplasty with or without stent placement, or a combination of procedures is used to treat selected cases of severe VTE for rapid symptom relief.⁸ The incidence of VTE in Asians is lower than that reported from Western populations.^9,10 However, our previous study in Taiwan showed that the cumulative rates of VTE recurrence at 6, 12, and 24 months after the index event were 6.7%, 9.4%, and 12.4%, which were comparable to Western populations.¹¹ Anticoagulant therapy reduces the risk of VTE recurrence but must be balanced against the associated increased risk of bleeding. Asian subgroup analysis from randomized clinical trials of NOACs for atrial fibrillation showed that NOACs carried a lower risk of major bleeding than warfarin.^12–15 Unfortunately, randomized clinical trials of NOACs for VTE in Asian-specific populations are lacking. There was 1 meta-analysis of randomized controlled trials to evaluate the efficacy and safety of NOACs in Asian (n=3,542) and non-Asian patients (n=23,481) with acute VTE, which showed that the efficacy of NOACs was comparable with warfarin irrespective of ethnicity, and NOACs could be safer alternatives in Asian patients.¹⁶ The clinical trials have selective inclusion criteria and the reproducibility of their findings needs to be reassessed in the broader patient populations that are seen in routine clinical practice. Most post-marketing prospective or retrospective cohort studies assessing the efficacy and safety of NOACs for treatment of acute VTE also come from the Western countries.^17–22 In fact, little information is available about the treatment of VTE in Asia in recent years. There are also no published data that directly compare efficacy and safety in Asians with VTE who are treated with NOACs vs. warfarin.

The purposes of this study were to investigate the pharmacoepidemiology of VTE patients who developed an initial DVT or PE from 2012 through 2016 by using the National Health Insurance Research Database (NHIRD) in Taiwan. The predisposing risk factors and associated comorbid diseases of these patients were evaluated. We also analyzed the trends of oral anticoagulant treatment and pharmacomechanical therapy during the study period. The risks of recurrent VTE, bleeding events, and all-cause death were compared between NOACs and warfarin in this real-world population of Asians with VTE.

Methods

Study Design

A nationwide population-based cohort study was performed to observe the trends of treatment of VTE and compare the efficacy and safety of NOACs and warfarin in adult patients with an initial VTE and at least 6-month follow-up in Taiwan. The study (B-EX-107-044) was approved by the Institutional Review Board of National Cheng Kung University Hospital, Tainan, Taiwan. This study used the NHIRD released by the Taiwan National Health Research Institute. The National Health Insurance in Taiwan is a mandatory universal health insurance program that offers comprehensive medical care coverage to almost all Taiwanese residents (>99% of Taiwan’s population). Our previous verification studies demonstrated a high validity for the diagnosis of common medical diseases and death in the NHIRD.^23,24 The accuracy for identifying procedures and medications was also high, with a positive predictive value >0.90.^23,24 In this cohort dataset, the patients’ original identification numbers were encrypted to protect privacy. By using a consistent encrypting procedure, it was feasible to link and continuously follow all of the claims belonging to the same patient within the NHIRD.

Study Population

The Taiwan NHIRD from January 1, 2008 to December 31, 2016 was used for this study. Because the first NOAC (rivaroxaban) was approved for reimbursement by the Taiwan National Health Insurance from May 1, 2014, adult patients (≥18 years old) with their first attack of VTE (DVT, PE, or both) treated with anticoagulant therapy between May 1, 2014 and June 30, 2016 (inclusive) were enrolled in the study. Inclusion criteria were hospitalization with the major diagnosis of VTE or major diagnosis of VTE 2 times consecutively within 1 month in outpatient clinics. The exclusion criteria were: (1) treatment with anticoagulants in previous 3 years before the index event; (2) VTE diagnosed only in outpatient clinic but not fulfilling the inclusion criteria; (3) death during admission for VTE or with incomplete follow-up information (Figure 1). The International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) diagnosis codes and ICD-9-CM treatment codes were used (Supplementary Table 1). For the first VTE event recorded in the database, and for each subsequent VTE event, the following information was captured: patient’s characteristics, medical history, provoking risk factors (e.g., major surgery/major trauma) within the previous 3 months, and type of VTE (DVT, PE or both) according to the ICD codes. We respectively ascertained the in-hospital and outpatient clinic use of warfarin, heparin/LMWH, antiplatelet agents, and NOACs by using the anatomic therapeutic chemical classification system. Recurrent VTE was defined as a new VTE hospitalization. The primary efficacy outcome was recurrent VTE, including DVT, PE or both. The breakdown incidences of recurrent DVT, recurrent PE, all-cause death, and myocardial infarction were also recorded. The primary safety outcome was major bleeding associated with hospitalization, including intracranial hemorrhage, major gastrointestinal (GI) bleeding, and other critical site bleeding (intraspinal, intraocular, pericardial, intra-articular, intramuscular with compartment syndrome, and retroperitoneal). The breakdown incidence of these bleeding events was also reported. Note that the same patient could have >1 study endpoint and have the same study endpoint many times during the study duration; only the study endpoint that appeared first was counted. The follow-up period was calculated from the index date until the occurrence of any study endpoint or to the end of the study period (December 31, 2016), whichever came first.

Figure 1.

Flow chart of patient disposition. NOAC, nonvitamin K antagonist oral anticoagulant; VTE, venous thromboembolism.

Statistical Analysis

Demographic data are expressed as mean (± standard deviation) or percentage. In general, differences in proportions were tested with the chi-square test or Fisher’s exact test, and differences in continuous variables were tested with an unpaired Student’s t-test. Because of the nonrandomized nature of the study, propensity score analysis was performed to minimize any selection bias caused by differences in the clinical characteristics between groups. The propensity score is defined to be the probability of exposure to the treatment conditional on a study subject’s baseline characteristics.^25–27 In this study, the propensity score for the likelihood of receiving NOACs or warfarin was computed using multivariate logistic regression analysis, conditional on covariates including the index year, age, sex, medications during index hospitalization, medications 90 days before the index hospitalization, VTE provoking factors, and comorbidities, including coronary artery disease, heart failure, peripheral artery disease, hypertension, diabetes mellitus, hyperlipidemia, chronic lung disease, peptic ulcer, liver cirrhosis, cancer, atrial fibrillation, chronic kidney disease, and endstage renal disease. We used a greedy-matching algorithm to generate matches with a caliper of 0.25 of the standard deviation of the logit of the propensity score. Time-to-event Kaplan-Meier curves for each cohort were constructed for death from any cause and VTE, compared by means of log-rank test. The estimates of relative risk, with 95% confidence interval (CI), were derived from Cox proportional hazards models, adjusted for potential confounders including comorbidities. Fine and Gray competing risk regression models were used to examine the risk of acute myocardial infarction, recurrent VTE and bleeding events accounting for the competing risk of death.²⁸ The SAS statistical software (version 9.4, SAS Institute, Cary, NC, USA) was used for data collection and all subsequent statistical analyses. Sensitivity analysis was used to examine the effect of other potential unmeasured confounders on the treatment groups. A sensitivity analysis was performed using the R package “obsSens” to estimate the range of hazard ratios (HRs) between the treatment groups for 3 outcomes, including recurrent VTE, all-cause death, and the primary safety endpoint. We added a hypothetical unmeasured confounder with a favorable protective effect to observe the range of HRs confounded by this add-on factor with different prevalence in the warfarin and NOACs groups. The duration of anticoagulation therapy could be widely different according to the risk of recurrence in the real world and these differences should be taken into account. Therefore, we performed a time-dependent Cox regression model for the NOACs effect on primary efficacy and safety outcomes after propensity score matching and fitting duration of anticoagulation therapy as a time-dependent covariate.

Results

Overall, identified in the database from May 1, 2014 to June 30, 2016 were 20,628 patients with a diagnosis of VTE. After exclusions, 11,414 patients (mean age 67.4±16.4 years, 44.4% male) were enrolled and formed the basis of this study (Figure 1). Figure 2 shows the temporal trends of oral anticoagulants and pharmacomechanical therapy for VTE from 2013 to 2016 in Taiwan. In anticoagulant-treated patients, warfarin use decreased progressively from 100% in 2013 Q1 to 39.8% in 2016 Q4, whereas NOACs use increased from 0% to 60.2% (Figure 2A). Among VTE patients in 2016 Q4, 56.3% of patients were treated with rivaroxaban, 3.1% with apixaban, and 0.8% with edoxaban. Unlike the gradually increasing use of NOACs, the proportion of patients who underwent pharmacomechanical therapy did not change significantly during the study period (Figure 2B).

Figure 2.

Distribution of antithrombotic drugs (A) and percentage of pharmacomechanical therapy (B) in patients with first venous thromboembolism from 2013 to 2016 in Taiwan.

Among the 11,414 patients enrolled in this study, 7,743 (68%) were given warfarin and 3,671 (32%) were given NOACs. The age, sex and other baseline clinical characteristics varied between the different treatment groups (Supplementary Table 2). After 1:1 propensity score matching, 7,294 patients were selected, and the baseline characteristics between patients treated with warfarin (n=3,647) and NOACs (n=3,647) became more balanced (Table 1). In the 7,294 oral anticoagulant-naive patients, 4,957 (68%) of them had DVT, 1,688 (23%) had PE, and 649 (9%) had concurrent DVT and PE. The mean duration of warfarin use was 122 days after the VTE event and the mean duration of NOACs was 142 days. The most common identifiable risk factors were major surgery (30.9%), cancer (22.3%), and major trauma (9.9%). In Table 2, we list the different cancer types in the 2 groups. Table 3 shows the number of events for the primary efficacy endpoint. The duration of follow-up was similar between the warfarin (488.59±229.39 days) and NOACs groups (474.06±223.74 days). Overall, 415 patients developed recurrent VTE during follow-up, with 233 (6.39%) in the warfarin group and 182 (4.99%) in the NOACs group. NOACs were associated with lower rates of recurrent VTE compared with warfarin (adjusted HR [aHR] 0.788, 95% CI 0.649–0.956). A competing risk analysis showed a similar result (Table 3). The effects of NOACs were consistent among subgroups that were defined according to age, sex, and comorbidities (Supplementary Figure 1). Restricting the analyses to recurrent DVT or PE yielded an aHR of 0.943 (95% CI 0.746−1.192) for recurrent DVT and an aHR of 0.535 (95% CI 0.366−0.781) for recurrent PE. For all-cause death, NOACs was associated with a lower risk compared with warfarin (aHR 0.844, 95% CI 0.765–0.930) (Table 3). Subgroup analysis found patients without cancer had more benefit in mortality risk reduction with NOACs (Supplementary Figure 2). Figure 3 displays the cumulative event-free curves and shows that recurrent VTE and all-cause death were significantly decreased in patients treated with NOACs. Table 4 shows the number of events for the primary safety endpoint. Overall, NOACs were associated with a lower risk of major bleeding compared with warfarin (aHR 0.804, 95% CI 0.648–0.998). If mortality was taken into consideration, a competing risk analysis also showed that NOACs were associated with a significantly lower risk of bleeding compared with warfarin (aHR 0.805, 95% CI 0.648–0.999). Subgroup analysis showed that the benefit of NOACs for the primary safety endpoint was more significant in male patients and in those who took antiplatelet agents concomitantly (Supplementary Figures 3–5).

Table 1. Baseline Clinical Characteristics of VTE Patients After Propensity Score Matching

Variables	All (n=7,294)	Warfarin (n=3,647)	NOACs (n=3,647)	P value
Index year				1.000
2014	1,410 (19.33)	705 (19.33)	705 (19.33)
2015	3,300 (45.24)	1,650 (45.24)	1,650 (45.24)
2016	2,584 (35.43)	1,292 (35.43)	1,292 (35.43)
Age, years (SD)	68.48 (16.06)	68.19 (16.10)	68.78 (16.03)	0.114
Male (%)	3,209 (44.00)	1,633 (44.78)	1,576 (43.21)	0.187
VTE type (%)				0.515
DVT	4,957 (67.96)	2,494 (68.38)	2,463 (67.53)
PE	1,688 (23.14)	842 (23.09)	846 (23.20)
DVT and PE	649 (8.90)	311 (8.53)	338 (9.27)
Unfractioned heparin/LMWH	3,724 (51.06)	1,904 (52.21)	1,820 (49.90)	0.052
Unfractioned heparin	1,317 (18.06)	752 (20.62)	565 (15.49)
LMWH	2,407 (33.00)	1,152 (31.59)	1,255 (34.41)
Risk factors of VTE (%)
Major trauma	726 (9.95)	351 (9.62)	375 (10.28)	0.368
Major surgery	2,256 (30.93)	1,128 (30.93)	1,128 (30.93)	1.000
Hormone therapy	143 (1.96)	70 (1.92)	73 (2.00)	0.866
Hypercoagulopathy	78 (1.07)	37 (1.01)	41 (1.12)	0.733
Varicose veins	275 (3.77)	142 (3.89)	133 (3.65)	0.623
Cancer	1,624 (22.26)	814 (22.32)	810 (22.21)	0.933
Pregnancy	42 (0.58)	18 (0.49)	24 (0.66)	0.439
Comorbidity (%)
Coronary artery disease	1,370 (18.78)	680 (18.65)	690 (18.92)	0.787
Atrial fibrillation	689 (9.45)	345 (9.46)	344 (9.43)	1.000
Peripheral artery disease	480 (6.58)	236 (6.47)	244 (6.69)	0.741
Ischemic stroke	582 (7.98)	281 (7.70)	301 (8.25)	0.412
Heart failure	1,135 (15.56)	577 (15.82)	558 (15.30)	0.561
Hypertension	4,113 (56.39)	2,057 (56.40)	2,056 (56.38)	1.000
Systemic lupus erythematosus	38 (0.52)	20 (0.55)	18 (0.49)	0.871
Diabetes mellitus	1,905 (26.12)	954 (26.16)	951 (26.08)	0.958
Dyslipidemia	1,763 (24.17)	864 (23.69)	899 (24.65)	0.352
Peptic ulcers	1,254 (17.19)	621 (17.03)	633 (17.36)	0.733
Chronic kidney disease	968 (13.27)	492 (13.49)	476 (13.05)	0.605
Chronic lung disease	222 (3.04)	104 (2.85)	118 (3.24)	0.376
Degenerative and paralytic neurologic disease	1,196 (16.40)	580 (15.90)	616 (16.89)	0.268
Chronic liver disease	499 (6.84)	257 (7.05)	242 (6.64)	0.516
Concomitant antiplatelet therapy (%)	2,829 (38.79)	1,410 (38.66)	1,419 (38.91)	0.848
Length of hospitalization, days	12.4 (11.04)	12.92 (11.35)	11.89 (10.72)	0.174

Data are presented as number (%) or mean value (± SD). DVT, deep vein thrombosis; LMWH low molecular weight heparin; PE, pulmonary embolism; VTE, venous thromboembolism.

Table 2. Different Cancer Type Distribution Between NOACs and Warfarin Users After Propensity Score Matching

	All cancer patients (n=1,624)	Warfarin (n=814)	NOACs (n=810)
Malignancy of oral cavity and pharynx	65 (4.00)	38 (4.67)	27 (3.33)
Malignancy of digestive organs	494 (30.42)	243 (29.85)	251 (30.99)
Malignancy of respiratory system	339 (20.87)	161 (19.78)	178 (21.98)
Malignancy of bone, soft tissue	173 (10.65)	92 (11.30)	81 (10.00)
Malignancy of genito-urinary system	398 (24.51)	204 (25.06)	194 (23.95)
Malignancy of nervous system	21 (1.29)	10 (1.23)	11 (1.36)
Malignancy of endocrine glands	20 (1.23)	8 (0.98)	12 (1.48)
Hematology malignancy	83 (5.11)	39 (4.79)	44 (5.43)
Othera	31 (1.91)	19 (2.33)	12 (1.48)

^aUnknown origin of malignancy.

Table 3. NOACs Effect on Primary Efficacy Outcomes for a Cohort With VTE When Compared With Warfarin After Propensity Score Matching

	Warfarin (n=3,647)	NOACs (n=3,647)
Primary efficacy endpoint
Recurrent VTE, n (%)	233 (6.39)	182 (4.99)
Crude HR (95% CI)	1	0.787 (0.648–0.955)
Adjusted HR (95% CI)	1	0.788 (0.649–0.956)
CRR (95% CI)	1	0.790 (0.651–0.959)
Recurrent DVT, n (%)	145 (3.98)	135 (3.70)
Crude HR (95% CI)	1	0.942 (0.745–1.190)
Adjusted HR (95% CI)	1	0.943 (0.746–1.192)
CRR (95% CI)	1	0.943 (0.746–1.193)
Recurrent PE, n (%)	76 (2.08)	42 (1.15)
Crude HR (95% CI)	1	0.559 (0.384–0.815)
Adjusted HR (95% CI)	1	0.535 (0.366–0.781)
CRR (95% CI)	1	0.555 (0.380–0.812)
All-cause death, n (%)	880 (24.13)	757 (20.76)
Crude HR (95% CI)	1	0.856 (0.777–0.944)
Adjusted HR (95% CI)	1	0.844 (0.765–0.930)
Myocardial infarction, n (%)	34 (0.93)	27 (0.74)
Crude HR (95% CI)	1	0.809 (0.488–1.341)
Adjusted HR (95% CI)	1	0.804 (0.485–1.334)
CRR (95% CI)	1	0.805 (0.486–1.333)

Patients treated with warfarin were recognized as reference group. Adjusted variables included sex, age, comorbidities, and concomitant medication. CI, confidence interval; CRR, competing risk regression; HR, hazard ratio; NOAC, nonvitamin K antagonist oral anticoagulant.

Figure 3.

Kaplan-Meier curves for event-free survival during follow-up for (A) recurrent VTE, (B) all-cause death, and (C) primary safety outcomes, including intracerebral hemorrhage, major gastrointestinal bleeding, and other critical site bleeding. NOAC, nonvitamin K antagonist oral anticoagulant; VTE, venous thromboembolism.

Table 4. NOACs Effect on Primary Safety Outcomes for a Cohort With VTE When Comparing With Warfarin After Propensity Score Matching

	Warfarin (n=3,647)	NOACs (n=3,647)
Primary safety endpoint
Major bleeding, n (%)	185 (5.26)	150 (4.25)
Crude HR (95% CI))	1	0.818 (0.659–1.014)
Adjusted HR (95% CI)	1	0.804 (0.648–0.998)
CRR (95% CI)	1	0.805 (0.648–0.999)
Intracranial hemorrhage, n (%)	62 (1.70)	45 (1.23)
Crude HR (95% CI))	1	0.734 (0.500–1.077)
Adjusted HR (95% CI)	1	0.712 (0.485–1.047)
CRR (95% CI)	1	0.712 (0.484–1.048)
Major GI bleeding, n (%)	78 (2.14)	67 (1.84)
Crude HR (95% CI))	1	0.867 (0.626–1.202)
Adjusted HR (95% CI)	1	0.861 (0.621–1.193)
CRR (95% CI)	1	0.859 (0.619–1.192)
Other critical site bleeding, n (%)	59 (1.62)	50 (1.37)
Crude HR (95% CI))	1	0.862 (0.592–1.257)
Adjusted HR (95% CI)	1	0.851 (0.584–1.241)
CRR (95% CI)	1	0.852 (0.585–1.240)

Patients treated with warfarin were recognized as reference group. Adjusted variables included sex, age, comorbidities, and concomitant medication. Abbreviations as in Table 3.

Sensitivity analysis was performed with an add-on unmeasured confounder for the 3 outcomes: recurrent VTE, all-cause death, and primary safety endpoint. Figure 4 displays the trend of estimates for the hazard using a multivariable-adjusted Cox proportional hazards model. For example, if all the patients in the warfarin group have the add-on unmeasured confounder (the prevalence of the confounder in the warfarin group is 1.0) and none of patients in NOACs group has this unmeasured confounder (the prevalence of the confounder in the NOACs group is 0.0), then the effect of NOACs group would be more beneficial for the 3 outcomes (recurrent VTE: HR=0.551, all-cause death: HR=0.59, and safety endpoint: HR=0.563, lefthand bottom line in Figure 4). The time-dependent Cox regression model for the NOACs effect on primary efficacy outcomes for a cohort with VTE when compared with warfarin after propensity score matching also showed that NOACs were associated with lower rates of recurrent VTE compared with warfarin (aHR 0.676, 95% CI 0.556–0.821) (Supplementary Table 3). NOACs were also associated with lower rates of major bleeding compared with warfarin (aHR 0.786, 95% CI 0.633–0.975) (Supplementary Table 4).

Figure 4.

Sensitivity analysis with the add-on of an unmeasured residual confounding risk factor showing trend estimates of the hazard ratio for recurrent VTE (A), all-cause death (B), and primary safety outcome (C), using a multivariable Cox regression model. NOAC, nonvitamin K antagonist oral anticoagulant; VTE, venous thromboembolism.

Discussion

This is the first population-based study to observe the changing pattern of VTE treatment and compare the effects of NOACs with warfarin on VTE in Asian patients. Our study showed that NOACs have become the major anticoagulants for VTE treatment in recent years in this Asian cohort. A relatively fixed ratio of patients still needs pharmacomechanical intervention to treat VTE. Compared with warfarin, NOACs seemed to be associated with lower risk for recurrent VTE, all-cause death, and major bleeding.

Overall, 94% of the NOACs-treated patients were given rivaroxaban, which was the first NOAC reimbursed by the Taiwan National Health Insurance from 2014 for VTE treatment. The EINSTEIN program, including EINSTEIN-DVT and EINSTEIN-PE studies, were randomized clinical trials that compared rivaroxaban with traditional therapy (enoxaparin/VKA) for treatment of DVT and PE, respectively.^4,5 Pooled analysis of the EINSTEIN-DVT and EINSTEIN-PE studies showed that the risk of recurrent VTE was similar between rivaroxaban and traditional therapy (HR 0.89, 95% CI 0.66–1.19) and major bleeding was significantly lower in the rivaroxaban group (HR 0.54, 95% CI 0.37–0.79).²⁹ In a meta-analysis of all clinical trials including dabigatran, rivaroxaban, apixaban, and edoxaban for VTE treatment, NOACs were shown to be as effective as traditional therapy to decrease recurrence of VTE, and the risk of major bleeding was lower in patients treated with NOACs.³⁰ More cancer patients were noted in our cohort (22%) compared with the EINSTEIN program (5.6%). A recent clinical trial recruiting only cancer patients with VTE demonstrated that rivaroxaban had a better effect in reducing recurrent VTE than dalteparin (HR 0.43, 95% CI, 0.19–0.99), but rivaroxaban showed a non-significant increase in major bleeding (HR 1.83, 95% CI 0.68–4.96).³¹ Another randomized clinical trial with edoxaban also showed similar results: the rate of recurrent VTE was lower but the rate of major bleeding was higher with edoxaban than with dalteparin in cancer patients with VTE.³² The higher proportion of cancer patients in our study could possibly explain the difference between our results and the data from randomized clinical trials. In addition, the quality of warfarin therapy is often represented by the time patients spend within the therapeutic range (percent time in therapeutic range, TTR). The correlation between TTR and the occurrence of complications during warfarin therapy has been established. According to previous trials (RELY, ROCKET-AF, ARISTOLE, ENGAGE AF-TIMI), warfarin TTR is lower in Asians than non-Asians (56.5% vs. 68.9%; 47.1% vs. 55.7%; 60% vs. 67%; 61.7% vs. 68.6%, respectively).^33–36 In the ROCKET-AF subgroup analysis, TTR in Taiwan, China, Singapore, Hong Kong was 38% vs. 47% vs. 64% vs. 66%, respectively.³⁷ Even in the randomized controlled trial, TTR in Taiwan was still far below the standard recommendation. We think that this could be an important reason why NOACs provide different results in Taiwan.

Despite advancements in vascular interventional therapy, only a minority of this study’s patients received pharmacomechanical therapy and that percentage did not increase in recent years. Our observation was similar to a recently published global registry study, The Global Anticoagulant Registry in the FIELD–Venous Thromboembolism (GARFIELD-VTE), showing that only 5.1% of patients received pharmacomechanical therapy.³⁸ Lack of robust evidence to prevent recurrent VTE and risk of bleeding are the major reasons preventing the wide-spread use of pharmacomechanical therapy. In the current study, the absolute risk of recurrent VTE was 6.39% in the warfarin group and 4.99% in the NOAC group, which were in line with our previous report of an approximately 5.1% recurrent rate of VTE per person-year in Taiwan.¹¹ The absolute risk of recurrent VTE was higher in our study compared with that in the EINSTEIN pooled analysis (2.3% in VKA and 2.1% in rivaroxaban group). This is probably because the mean age of our patients was older (57 years in the EINSTEIN pooled analysis vs. 68 years in our study) and more of our patients had cancer. These groups of patients carry a higher risk of VTE recurrence.

Most real-world studies of the use of NOACs in VTE treatment are from Western countries. In the Xa inhibition with rivaroxaban for Long-term and Initial Anticoagulation (XALIA) registry of DVT, rivaroxaban was associated with comparable effects in reducing recurrent VTE (HR 0.91, 95% CI 0.54–1.54) and major bleeding (HR 0.77, 95% CI 0.40–1.50) compared with warafrin.²¹ In the Danish health registries including only unprovoked VTE, the rate of recurrent VTE at 6 months was lower in the rivaroxaban-treated patients (HR 0.74, 95% CI 0.56–0.96), and major bleeding was similar between users of rivaroxaban and warfarin (HR 1.19, 95% CI 0.66–2.13).²⁰ Although there is increasing use of NOACs for VTE treatment in Asia, their efficacy and safety have only been investigated in very small case numbers. Large-scale real-world data for the efficacy and safety of NOACs in Asian VTE patients are also lacking. In the EINSTEIN program, 439 Chinese patients were included. The risks of recurrent VTE and clinically relevant bleeding were similar between the rivaroxaban and enoxaparin/warfarin groups.³⁹ The J-EINSTEIN study was a clinical trial that randomized only 81 Japanese patients with VTE to either rivaroxaban or intravenous unfractionated heparin (UHF) followed by warfarin. The composite endpoint of symptomatic recurrent VTE or asymptomatic deterioration occurred in 1.4% of rivaroxaban group and in 5.3% the UFH/warfarin group. No major bleeding occurred. Clinically relevant non-major bleeding occurred in 7.8% of the rivaroxaban group and 5.3% of the UFH/warfarin group.⁴⁰ Our study is the largest to compare NOACs with warfarin in an Asian population and reconfirmed the efficacy and safety of NOACs in Asian VTE patients. The main advantage of our study was the use of a nationwide database with a large case number.

Study Limitations

There are several limitations. First, the accuracy of VTE diagnosis has not been verified in the NHIRD, so potential disease misclassification bias could exist. Second, some detailed clinical information, such as the international normalized ratio, was not available from the NHIRD and, thus the TTR of warfarin could not be analyzed in our study. The favorable effects of NOACs with regard to lower risk of VTE recurrence could be partly caused by low TTR of warfarin, which is a common practice behavior of physicians in Asia. Third, although an extensive number of variables were selected in the propensity score-matching analysis, some residual unmeasured confounding factors cannot be completely excluded. Fourth, the NHIRD could not provide clear cause of death and we only recorded the incidence of all-cause death in this study. Fifth, we could not well explain why recurrent PE was lower in NOACs users than in those treated with warfarin, but the risk of recurrent DVT was comparable between these groups.

Conclusions

NOACs have become the major anticoagulants used for Asian patients with VTE in recent years. Only a minority of VTE patients were treated with pharmacomechanical therapy. Although NOACs had a lower risk of recurrent VTE and major bleeding compared with warfarin, we still need a large-scale randomized controlled trial in Taiwan to confirm the findings.

Addendum

C.-H.L. and C.-C.F. initiated and designed the study, obtained approvals, prepared the data, did the analysis and interpretation, and wrote the first draft of the paper. L.-M.T. and P.-S.C. contributed to the development of the protocol, to the design, analysis, and interpretation, and to drafting the paper. S.-H.L. contributed to the protocol, interpretation, and drafting the article. H.-W.L. undertook the data processing and statistical analysis. Y.-H.L. coordinated the conduct of study, established the strategy of data processing of the claims data, and participated in the manuscript preparation. All authors approved the final draft.

Role of Sponsors

This study was funded by National Cheng Kung University-Show Chwan Memorial Health Care System R&D Project (NCKUSCMH10811). The funding organizations had no role in the design, performance or analysis of this study or decision to submit the manuscript for publication.

Disclosures

The authors declare no conflicts of interest.

Supplementary Files

Please find supplementary file(s);

http://dx.doi.org/10.1253/circj.CJ-19-0741

References

• 1.   Silverstein MD, Heit JA, Mohr DN, Petterson TM, O’Fallon WM, Melton LJ 3rd. Trends in the incidence of deep vein thrombosis and pulmonary embolism: A 25-year population-based study. Arch Intern Med 1998; 158: 585–593.
• 2.   Schulman S, Kearon C, Kakkar AK, Mismetti P, Schellong S, Eriksson H, et al. Dabigatran versus warfarin in the treatment of acute venous thromboembolism. N Engl J Med 2009; 361: 2342–2352.
• 3.   Schulman S, Kakkar AK, Goldhaber SZ, Schellong S, Eriksson H, Mismetti P, et al. Treatment of acute venous thromboembolism with dabigatran or warfarin and pooled analysis. Circulation 2014; 129: 764–772.
• 4.   EINSTEIN Investigators, Bauersachs R, Berkowitz SD, Brenner B, Buller HR, Decousus H, Gallus AS, et al. Oral rivaroxaban for symptomatic venous thromboembolism. N Engl J Med 2010; 363: 2499–2510.
• 5.   EINSTEIN-PE Investigators, Büller HR, Prins MH, Lensin AW, Decousus H, Jacobson BF, Minar E, et al. Oral rivaroxaban for the treatment of symptomatic pulmonary embolism. N Engl J Med 2012; 366: 1287–1297.
• 6.   Agnelli G, Buller HR, Cohen A, Curto M, Gallus AS, Johnson M, et al. Oral apixaban for the treatment of acute venous thromboembolism. N Engl J Med 2013; 369: 799–808.
• 7.   Hokusai-VTE Investigators, Büller HR, Décousus H, Grosso MA, Mercuri M, Middeldorp S, Prins MH, et al. Edoxaban versus warfarin for the treatment of symptomatic venous thromboembolism. N Engl J Med 2013; 369: 1406–1415.
• 8.   Jarrett H, Bashir R. Interventional management of venous thromboembolism: State of the art. Am J Roentgenol 2017; 208: 891–906.
• 9.   Lee CH, Cheng CL, Lin LJ, Tsai LM, Yang YH. Epidemiology and predictors of short-term mortality in symptomatic venous thromboembolism. Circ J 2011; 75: 1998–2004.
• 10.   Hong J, Lee JH, Yhim HY, Choi WI, Bang SM, Lee H, et al. Incidence of venous thromboembolism in Korea from 2009 to 2013. PLoS One 2018; 13: e0191897.
• 11.   Lee CH, Lin LJ, Cheng CL, Chen JY, Tsai LM, Kao YH. Incidence and cumulative recurrent rates of venous thromboembolism in the Taiwanese population. J Thromb Haemost 2010; 8: 1515–1523.
• 12.   Hori M, Connolly SJ, Zhu J, Liu LS, Lau CP, Pais P, et al. Dabigatran versus warfarin: Effects on ischemic and hemorrhagic strokes and bleeding in Asians and non-Asians with atrial fibrillation. Stroke 2013; 44: 1891–1896.
• 13.   Wong KS, Hu DY, Oomman A, Tan RS, Patel MR, Singer DE, et al. Rivaroxaban for stroke prevention in East Asian patients from the ROCKET AF trial. Stroke 2014; 45: 1739–1747.
• 14.   Goto S, Zhu J, Liu L, Oh BH, Wojdyla DM, Aylward P, et al. Efficacy and safety of apixaban compared with warfarin for stroke prevention in patients with atrial fibrillation from East Asia: A subanalysis of the Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation (ARISTOTLE) Trial. Am Heart J 2014; 168: 303–309.
• 15.   Yamashita T, Koretsune Y, Yang Y, Chen SA, Chung N, Shimada YJ, et al. Edoxaban vs. warfarin in East Asian patients with atrial fibrillation: An ENGAGE AF-TIMI 48 subanalysis. Circ J 2016; 80: 860–969.
• 16.   Yamashita Y, Morimoto T, Toyota T, Shiomi H, Makiyama T, Ono K, et al. Asian patients versus non-Asian patients in the efficacy and safety of direct oral anticoagulants relative to vitamin K antagonist for venous thromboembolism: A systemic review and meta-analysis. Thromb Res 2018; 166: 37–42.
• 17.   Gaertner S, Cordeanu EM, Nouri S, Faller AM, Frantz AS, Mirea C, et al. Rivaroxaban versus standard anticoagulation for symptomatic venous thromboembolism (REMOTEV observational study): Analysis of 6-month outcomes. Int J Cardiol 2017; 226: 103–109.
• 18.   Keller L, Marten S, Hecker J, Sahin K, Tittl L, Beyer-Westendorf J. Venous thromboembolism therapy with rivaroxaban in daily-care patients: Results from the Dresden NOAC registry. Int J Cardiol 2018; 257: 276–282.
• 19.   Kucher N, Aujesky D, Beer JH, Mazzolai L, Baldi T, Banyai M, et al. Rivaroxaban for the treatment of venous thromboembolism: The SWIss Venous ThromboEmbolism Registry (SWIVTER). Thromb Haemost 2016; 116: 472–479.
• 20.   Larsen TB, Skjøth F, Kjældgaard JN, Lip GYH, Nielsen PB, Søgaard M. Effectiveness and safety of rivaroxaban and warfarin in patients with unprovoked venous thromboembolism: A propensity-matched nationwide cohort study. Lancet Haematol 2017; 4: e237–e244.
• 21.   Ageno W, Mantovani LG, Haas S, Kreutz R, Monje D, Schneider J, et al. Safety and effectiveness of oral rivaroxaban versus standard anticoagulation for the treatment of symptomatic deep-vein thrombosis (XALIA): An international, prospective, non-interventional study. Lancet Haematol 2016; 3: e12–e21.
• 22.   Wang L, Baser O, Wells P, Peacock WF, Coleman CI, Fermann GJ, et al. Overall effectiveness of rivaroxaban in patients with pulmonary embolism. Clin Ther 2017; 39: 1426–1436.
• 23.   Cheng CL, Lee CH, Chen PS, Li YH, Lin SJ, Yang YH. Validation of acute myocardial infarction cases in the national health insurance research database in Taiwan. J Epidemiol 2014; 24: 500–507.
• 24.   Cheng CL, Chien HC, Lee CH, Lin SJ, Yang YH. Validity of in-hospital mortality data among patients with acute myocardial infarction or stroke in National Health Insurance Research Database in Taiwan. Int J Cardiol 2015; 201: 96–101.
• 25.   Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika 1983; 70: 41–55.
• 26.   Rosenbaum PR, Rubin DB. Reducing bias in observational studies using subclassification on the propensity score. J Am Stat Assoc 1984; 79: 516–524.
• 27.   Reeve BB, Smith AW, Arora NK, Hays RD. Reducing bias in cancer research: Application of propensity score matching. Health Care Financ Rev 2008; 29: 69–80.
• 28.   Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc 1999; 94: 496–509.
• 29.   Prins MH, Lensing AW, Bauersachs R, van Bellen B, Bounameaux H, Brighton TA, et al. Oral rivaroxaban versus standard therapy for the treatment of symptomatic venous thromboembolism: A pooled analysis of the EINSTEIN-DVT and PE randomized studies. Thromb J 2013; 11: 21.
• 30.   van Es N, Coppens M, Schulman S, Middeldorp S, Büller HR. Direct oral anticoagulants compared with vitamin K antagonists for acute venous thromboembolism: Evidence from phase 3 trials. Blood 2014; 124: 1968–1975.
• 31.   Young AM, Marshall A, Thirlwall J, Chapman O, Lokare A, Hill C, et al. Comparison of an oral factor Xa inhibitor with low molecular weight heparin in patients with cancer with venous thromboembolism: Results of a randomized trial (SELECT-D). J Clin Oncol 2018; 36: 2017–2023.
• 32.   Raskob GE, van Es N, Verhamme P, Carrier M, Di Nisio M, Garcia D, et al. Edoxaban for the treatment of cancer-associated venous thromboembolism. N Engl J Med 2018; 378: 615–624.
• 33.   Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med 2009; 361: 1139–1151.
• 34.   Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med 2011; 365: 883–891.
• 35.   Granger CB, Alexander JH, McMurray JJ, Lopes RD, Hylek EM, Hanna M, et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med 2011; 365: 981–992.
• 36.   Giugliano RP, Ruff CT, Braunwald E, Murphy SA, Wiviott SD, Halperin JL, et al. Edoxaban versus warfarin in patients with atrial fibrillation. N Engl J Med 2013; 369: 2093–2104.
• 37.   Singer DE, Hellkamp AS, Piccini JP, Mahaffey KW, Lokhnygina Y, Pan G, et al. Impact of global geographic region on time in therapeutic range on warfarin anticoagulant therapy: Data from the ROCKET AF clinical trial. J Am Heart Assoc 2013; 2: e000067.
• 38.   Ageno W, Haas S, Weitz JI, Goldhaber SZ, Turpie AGG, Goto S, et al. Characteristics and management of patients with venous thromboembolism: The GARFIELD-VTE registry. Thromb Haemost 2019; 119: 319–327.
• 39.   Wang Y, Wang C, Chen Z, Zhang J, Liu Z, Jin B, et al. Rivaroxaban for the treatment of symptomatic deep-vein thrombosis and pulmonary embolism in Chinese patients: A subgroup analysis of the EINSTEIN DVT and PE studies. Thromb J 2013; 11: 25.
• 40.   Yamada N, Hirayama A, Maeda H, Sakagami S, Shikata H, Prins MH, et al. Oral rivaroxaban for Japanese patients with symptomatic venous thromboembolism: The J-EINSTEIN DVT and PE program. Thromb J 2015; 13: 2.