dc.description.abstract | Background
Warfarin is an oral anticoagulant widely prescribed for several thromboembolic indications
including stroke associated with atrial fibrillation (AF). Warfarin control is influenced by a
number of genetic and environmental factors which, combined with a narrow therapeutic
index, results in a need for close monitoring using the International Normalised Ratio (INR).
Maintaining INR values within the therapeutic range can reduce adverse effects such as
bleeding and stroke, with time in therapeutic range (TTR) commonly used as a measure of
the quality of anticoagulation with warfarin. Increased TTR directly correlates to the
efficacy and safety of warfarin and improved patient outcomes.
Many patients with AF are either ineffectively managed with warfarin or not candidates for
warfarin therapy. This led to the development of alternate non-vitamin K oral
anticoagulants (NOACs) which were compared to warfarin in patients with AF through large
randomised controlled trials. Meta-analyses of these trials demonstrated the NOACs to be
either non-inferior or slightly superior to warfarin in terms of stroke and associated with a
lower risk of intracranial haemorrhage. However, sub-analyses of these trials highlighted
concerns regarding the mean warfarin TTR in these trials and the differences in TTR
according to geographical location and management systems. Variation in warfarin TTR
impacts the clinical outcomes of warfarin therapy including the comparative outcomes with
NOACs. The choice between anticoagulants requires consideration of likely warfarin control
which may be influenced by patient demographics and clinical characteristics together with
health care management systems. In Australia, this prompted the Australian Government
to commission a review of anticoagulant therapies in AF. This review identified a need for national guidelines for AF and therapeutic management algorithms for individual
anticoagulants including warfarin. However, this review also identified a barrier to
producing these management algorithms was the lack of information regarding the quality
of warfarin management in Australia and factors which may influence TTR including patient
variables and management systems.
Aims
The overall aim of this study was to determine predictors of warfarin control in patients
with AF in South-East Queensland and Singapore. Singapore was chosen as a comparator
site in the Asia-Pacific region to assist with applicability of results across the multi-cultural
population of Australia given potential factors influencing warfarin TTR include patient
demographics, ethnicity and management systems. Specific objectives of the study were to:
(1) establish the prescribing of warfarin in AF since the introduction of the NOACs (2)
determine the quality of warfarin control in Queensland achieved by different warfarin
management systems; (3) identify factors influencing warfarin TTR in patients with AF in
Queensland as compared to Singapore; and (4) establish the efficacy of risk models
recommended in AF guidelines to predict warfarin control.
Methods
Ethics approval was granted (PHM/09/14HREC,2015/863 and 2015/2435). A retrospective
observational study was conducted of patients receiving warfarin for AF at two study sites,
Sullivan Nicolaides Pathology (SNP) in Queensland and the National Heart Centre Singapore
(NHCS). PBS data was utilised to establish prescribing rates of oral anticoagulants since the introduction of the NOACs, Data was collected as of 30 September 2014 at SNP and between 1 January and 30 June
2014 at the NHCS. Data collected was INR test results, together with patient demographics
such as age, gender, co-morbidities and concurrent medications. Warfarin control for
individuals was determined by calculation of percentage of INR tests within target range and
TTR via a linear interpolation method. Patients were excluded if TTR could not be
calculated, that is less than two INR tests, or if less than thirty days of treatment. Patient
data categorised patients according to specific factors, including gender, age, concurrent comorbidities
and medication, bleed an stroke risk scores, and a predictor score for warfarin
control. Mean patient data was used for analysis and comparison between groups with
analysis of specific factors on TTR conducted using chi-squared, Tukey-Kramer, or a one-way
analysis of variance tests via nonparametric methods including Kruskal-Wallis test and
Dunn’s multiple comparisons test with GraphPad Instat Version 3. Significance was defined
at p<0.05(*), p<0.01(**), p<0.001(***), and p<0.0001(****).
Results
Warfarin prescribing decreased since introduction of the NOACs but it remains widely
prescribed with almost 1.5 million units prescribed in Australia (2017/8 financial year), of
which 250,478 units were prescribed in Queensland. In Queensland, the number of patients
with AF managed by SNP decreased from 10,806 patients in July 2012 to 5524 patients in
July 2017. During this time, 3036 patients exited the warfarin program to commence NOAC
therapy but almost 5% (n=141) reverted to warfarin. In these patients, no significant
difference in mean TTR was found before or after NOAC treatment but significantly more frequent testing and lower doses were required to obtain this level of control.
In Queensland, warfarin control in patients with AF was above recommended minimum
targets of 65% with a mean TTR of 68.5±16.2% by general practitioner (GP) and 81.5±9.1%
by SNP with significant differences (p < 0.0001) in frequency of testing between
management systems. Warfarin control in Singapore was significantly lower than
Queensland (mean TTR 82.3±15.6% vs 57.6±34.2%, p < 0.0001). There were significant
differences (p < 0.0001) in mean frequency of testing between Queensland and Singapore
(16.9 vs 29.3 days). Patients with chronic kidney disease in both Queensland and Singapore
had significantly lower TTR (77.2±16.8%, p<0.01 and 50.9±32.9%, p<0.05 respectively). Age
less than 60 years (p<0.05) in Queensland and use of a platelet inhibitor in Singapore (p <
0.01) also significantly reduced TTR. After excluding interacting drugs with warfarin, a
significantly reduced TTR (p<0.05) was found in Queensland for patients taking concurrent
non-steroidal inflammatory drugs (NSAIDs) and aspirin In Singapore, a predictor score
could identify poor warfarin control as patients with a score > 2 had significantly lower TTR
than other patients (55.8±34.1% vs 63.2±34.1%, p<0.001).
Discussion and conclusion
Warfarin requires regular monitoring to maintain INR levels within range and minimum
targets of 65% TTR are recommended to optimise benefits from warfarin therapy. This
study demonstrated that warfarin is well controlled in Queensland with TTR levels above
this minimum by both GP and dedicated warfarin management systems but, in contrast,
Singapore has poor control. Frequency of INR testing strongly contributed to warfarin
control with increased warfarin monitoring leading to improved control in Queensland
compared to Singapore but also by the dedicated warfarin management system compared to GP management in Queensland.
In this study, patient factors which consistently influenced warfarin TTR in both Queensland
and Singapore were chronic kidney disease and the concurrent administration of NSAIDs
and aspirin, whilst age < 60 years also influenced TTR in Queensland. Based on these
factors, prescribing guidelines have been proposed which highlight these subsets of patients
as being at risk of poor warfarin control and who may benefit from the additional
intervention of a dedicated warfarin management system to optimise control. In addition, a
risk model combining these factors into the mnemonic WARFARIN (Warfarin management
program available, AF valvular, Renal Function normal, Age > 60 years, Race = Caucasian,
Intolerance to NOAC, NSAIDs or aspirin not concurrently used) has been proposed as a
method of identifying patients likely to obtain good warfarin control and best suited to
warfarin therapy.
In conclusion, despite the introduction of the NOACs, warfarin remains widely prescribed
and is currently the only anticoagulant option available for patients with valvular AF.
Dedicated warfarin management programs can improve warfarin control and use of
proposed algorithms or guidelines can identify subsets of patients at risk of reduced control
who may benefit from additional intervention to improve warfarin TTR. This addresses
recommendations from an Australian government report and may help optimise anticoagulant therapy for patients with AF. | |