Use of non‐vitamin K antagonist oral anticoagulants in Colombia: A descriptive study using a national administrative healthcare database

Abstract Purpose We aimed to describe time‐trends in the use of NOACs among a group of ambulatory patients with nonvalvular atrial fibrillation (NVAF) in Colombia and to describe treatment patterns and user characteristics. Methods Using the Audifarma S.A administrative healthcare database in Colombia, we identified 10 528 patients with NVAF aged at least 18 years between July 2009 and June 2017 with a first prescription (index date) for apixaban, dabigatran or rivaroxaban (index NOAC) and followed them for at least year (max, 8.0 years, mean 2.2 years). We described patient characteristics, NOAC use over time, and the dose of the first NOAC prescription. Results A total of 2153 (20.5%) patients started on apixaban, 3089 (29.3%) on dabigatran and 5286 (50.2%) on rivaroxaban. The incidence of new users of apixaban and rivaroxaban increased over study years while for dabigatran it decreased. Mean age at the index date was: 78.5 years (apixaban), 76.5 years (dabigatran), 76.0 years (rivaroxaban). The percentage of patients started NOAC therapy on the standard dose was: apixaban 38.0%, dabigatran 30.9%, rivaroxaban 56.9%. The percentage still prescribed their index NOAC at 6 months was apixaban 44.6%, dabigatran 51.4%, rivaroxaban 52.7%. Hypertension was the most common comorbidity (>80% in each NOAC cohort). Conclusion During the last decade, the incidence of NOAC use in patients with NVAF affiliated with a private healthcare regime in Colombia has markedly increased. Future studies should evaluate whether the large number of patients with NVAF starting NOAC treatment on a reduced dose are done so appropriately.


| INTRODUCTION
Atrial fibrillation (AF) is a common cardiac arrhythmia with a prevalence that increases with age. 1 It is estimated that one in four middle-aged adults in Europe and the United States will develop AF in their lifetime. 2 The arrhythmia is associated with a 4-to 5-fold increase in the risk of ischaemic stroke 3 and a 1.5-to 2-fold increased risk of all-cause mortality. 2 Most epidemiological data on AF have come from Western populations; however, analysis of national healthcare databases show that AF also represents a substantial public health burden in Latin America, with estimated country-specific prevalences of around 13% among individuals aged 70 years or more. 4 Moreover, evidence suggests that the prevalence of AF, stroke, and associated mortality has increased dramatically in Latin America, likely due to the combined effect of the aging population and poor control of major risk factors such as hypertension. 5 However, little is known about the management of patients with AF in this area of the world since the introduction of non-vitamin K antagonist oral anticoagulants (NOACs) as an alternative option for stroke prophylaxis in this patient population in the last decade. Non-vitamin K antagonist oral anticoagulants have been shown to be noninferior to warfarin in reducing the risk of stroke and systemic embolism in patients with AF, and to have a superior safety profile. This has been shown in the overall pivotal clinical trial populations in which their approval was based, [6][7][8] as well as in subanalyzes of these trials restricted to participants in Latin America. 9 Unlike warfarin, the fixed dose regimens and predictable pharmacoki-   10 The Audifarma database has been validated in multiple studies that show how medications are used in the Colombian population. [11][12][13] The source population included patients in the contributory regime aged at least 18 years between July 2009 and June 2017 with at least 1 year of enrollment with their insurance provider and at least 1 year of available data following their first recorded outpatient health contact to guarantee a certain level of continuity with health services. No patient identifying information was used in this study. The study protocol was approved by the bioethics committee of the Universidad Tecnológica de Pereira, Colombia.

| NOAC study cohorts
From within the source population, three mutually exclusive cohorts of first-time users of NOACs (apixaban, dabigatran or rivaroxaban) were identified with the date of first NOAC prescription (index NOAC) set as the index date. If a patient had a prescription for another anticoagulant (eg, warfarin) in the year before their index date they were classified as non-naïve, while patients with no prescription for another anticoagulant in the year before their index date were classed as naïve. We excluded patients who were prescribed two different NOACs on the same day. Patients who qualified as a first-time user of more than one NOAC at different times during the study period (ie, switchers) were assigned to the cohort of the first prescribed NOAC. For each NOAC cohort we subsequently only retained patients with a record of AF (ICD-10 code I48) before the index date or in the 2 weeks after the index date. Patients with a record of mitral

KEY POINTS
• The marked increase in the use of rivaroxaban and apixaban to patients with NVAF affiliated with a private healthcare regime in Colombia over the last decade indicates growing confidence in the prescribing of these two NOACs among physicians in the country.
• As substantial numbers of patients with NVAF affiliated with a private healthcare regime in Colombia appear to be prescribed a reduced dose NOAC, studies are now warranted to evaluate the extent to which this is done appropriately-in accordance with the drug labels. stenosis (ICD-10 codes: I050, I05X, I052), valvular replacement (ICD-10 codes: Z952-Z954), and others stenosis (ICD-10 codes: I058, I059, I080, I081, I083, I088) during this time interval were excluded to identify only those patients with NVAF because there are no specific ICD-10 codes for NVAF. All patients were followed up for at least 1 year from the index date, until leaving the health plan, death or end of study data collection (December 2017).

| Characteristics of first-time NOAC users with NVAF
We extracted data on patient demographics (age and sex), comorbidities in the year before the index date including cardiovascular comorbidities (myocardial infarction, heart failure, ischaemic stroke, haemorrhagic stroke, venous thromboembolism (VTE) and hypertension) and other comorbidities (diabetes mellitus, chronic obstructive pulmonary disease, gastrointestinal bleeding, severe renal failure, and cancer). We also extracted data on the following medications prescribed in the year before the index date: other anticoagulants including warfarin and low-molecular-weight heparin (LMWH), antiplatelets (low-dose aspirin and clopidogrel), antiarrhythmic drugs, antihypertensive drugs, statins, antidiabetic drugs, nonsteroidal antiinflammatory drugs, acid-suppressive drugs, antidepressants and oral steroids (see Table S1). Polypharmacy was assessed as the number of different medications prescribed in the 2 months before the index date. We also identified patients with a prescription for another anticoagulant drug (including warfarin and low-molecular weight heparin) at any time before the index date, and classed these patients as anticoagulant non-naïve; all other patients were classed as anticoagulant naïve.

| Characteristics of the index NOAC prescription
For the index NOAC prescription and for subsequent NOAC prescriptions, we extracted information on the number of pills prescribed, the dose and the posology, and estimated the duration of each prescription using these instructions. If there was missing information on the daily dose, we made an assumption about the most likely dose based on assessment of the timing of subsequent prescriptions to that patient. We assessed the dose of the index NOAC prescription as well as the dose prescribed 3 months later. For all patients, we calculated the duration of the first episode of continuous NOAC treatment. Continuous treatment was when there was either no gap in treatment of >30 days between the end of the supply of a prescription and the start of the next prescription for the same NOAC, or no further prescription after the end of the previous one.

| Statistical analysis
For each NOAC cohort, the characteristics of patients and of the index NOAC prescription (at the start of follow-up and at 3 and 6 months) were described using frequency counts and percentages for categorical data, and means with SD for age. To evaluate trends in NOAC prescribing for stroke prevention in AF over time, the number of patients with NVAF newly prescribed a NOAC was described for each study year. In the calculation of incidence rates, we used only F I G U R E 1 Flowchart depicting the identification of the three NOAC study cohorts. *Patients dispensed two different NOACs on the same date. †Patients were excluded if they had a code for mitral stenosis/valvular replacement before the index date or in the 2 weeks after the index date, or if they had a NOAC dispensation with no associated diagnosis. NOAC, non-vitamin K antagonist oral anticoagulant; NVAF, nonvalvular atrial fibrillation [Colour figure can be viewed at wileyonlinelibrary.com] patients enrolled with two of the five healthcare providers contributing to the database (Salud Total and Compensar, corresponding to 3.6 of the 4.8 million patients in this study) for the numerator. This was because the denominator for this calculation-the total number of patients in the database (the exact number of individuals affiliated with the insurance regime in each year)-was only available for these two providers. Incidence rates of new users of NOACs with NVAF were calculated for each study year and were expressed per 10 000 individuals. We also calculated the percentage use for each OAC dispensed out of all OACs for each study year. In an analysis comparing the proportions of use of the two insurance companies vs the other three, in which the exact number of the total membership was not known, it was found that the proportions of use of NOACs was the same. All data analysis was conducted using SPSS Statistics Version 25 (IBM) for Windows.  In the year before the index date. b On the index date or in the year before the index date.

| Characteristics of first-time users of a NOAC with NVAF
Characteristics of the three study cohorts are shown in Tables 1 and   2. There were slightly more males than females in each cohort and the mean age was similar between cohorts, albeit slightly higher among apixaban users (apixaban 78. 5

| Characteristics of NOAC use
Details about the index NOAC prescription are shown in Table 3.
Rivaroxaban was mostly prescribed once daily (97.2%), which is the correct posology for stroke prevention in AF.  29,33 Heart failure was present in about one-third of patients with NVAF, which is both higher than some previous findings, 29,33-35 but lower than others, 23

| CONCLUSION
We conclude that over the last decade, rivaroxaban has been the most commonly prescribed NOAC, followed by dabigatran, among patients with NVAF affiliated to a private health insurer in Colombia.
Approximately half of patients continue to receive NOACs 6 months after the start of treatment, which suggests a certain level of adherence and tolerability, and a substantial percentage of patients, especially those starting therapy on apixaban, are prescribed a reduced dose. Studies are now needed focusing on the real-world effectiveness and safety of NOACs in Colombia, as well as an evaluation of the appropriateness of reduced dosing.