PracticeUpdate: Neurology - Winter 2018

ANZAN 2018 15

Better Screening for Atrial Fibrillation Needed in PatientsWho PresentWith TIA/Stroke Anticoagulant use among patients who present with transient ischemic attack (TIA) or stroke remains inadequate. Improved screening for patients presenting with TIA or stroke is needed, results of a retrospective chart review show. Z ainab Khan, MD, of Western Syd- ney University in Penrith, NSW, explained that atrial fibrillation raises has been questioned. We therefore mod- eled the trajectory of patients after such a stroke and counted both costs and how much quality adjusted life years changed when such a device was used.”

Known atrial fibrillation occurred in 34 (82.9%) stroke admissions and 8 (100%) admissions for TIA. Therefore, atrial fibril- lation was diagnosed post stroke in 17.1% of participants. Prior to stroke and TIA, 61.9% of patients were receiving anticoagulants. Treatment prior to stroke or TIA consisted of 11 (26.2%) on direct oral anticoagulants, 15 (35.7%) on warfarin, 5 (11.9%) on antiplate- lets, and 11 (26.2%) with no antithrombotics. Subtherapeutic international normalized ratios (<2.0) on admission were found in 60% of participants receiving warfarin. Among the 11 untreated patients, one was noncompliant, one had experienced recent subarachnoid hemorrhage, and no explanation was found for the remain- ing 9 participants. Anticoagulation rates improved on discharge with 40 (81.6%) patients prescribed either direct oral anti- coagulants or warfarin. Dr. Khan concluded that anticoagulation rates have improved since 2009 (61.9% vs 41%) in stroke/TIA admissions associated with atrial fibrillation. Use of anticoagulants in this population remains inadequate, how- ever. Of patients on warfarin, 60% exhibited subtherapeutic international normalized ratios on admission. A need remains for improved screening for atrial fibrillation. Only a slight decrease in rates of new diagnosis of atrial fibrilla- tion was observed vs the 2009 audit (from 22% to 17.1%). Detection of atrial fibrillation in patients with cryptogenic stroke using an inserta- ble cardiac monitor was evaluated in terms of cost-effectiveness. Vincent Thijs, MD, of the Florey Institute of Neuroscience and Mental Health, University of Melbourne, Victoria, explained that atrial fibrillation needs to be detected prior to ini- tiating oral anticoagulation after cryptogenic stroke. Paroxysmal atrial fibrillation can be dif- ficult to diagnose, however, with short-term cardiac monitoring. Dr. Thijs and colleagues set out to determine whether long-term continuous monitoring with an insertable cardiac monitor is cost-ef- fective for preventing recurrent stroke in patients with cryptogenic stroke. Dr. Thijs told Elsevier’s PracticeUpdate , “Whether loop recorders provide good value for money when used to detect atrial fibrillation after unexplained ischemic stroke

risk of potentially fatal or disabling ischemic stroke. Under 70% of eligible patients with atrial fibrillation have been noted to take oral anticoagulants. Dr. Khan and colleagues reviewed medical records of patients admitted with stroke/ TIA associated with a known or new diag- nosis of atrial fibrillation from 2016 to 2017. The data were compared with a similar audit conducted in 2009. Data collected included age, sex, history of atrial fibrillation, adequacy of anticoag- ulation (before and after admission), and vascular risk factors. A total of 49 patients (29 males, average age 76 years) were included. The incidence of atrial fibrillation was found to be 27.9% in stroke and 8.4% among TIA admissions. the positive likelihood ratio needed to be ≥5 and negative likelihood ratio, ≤0.1, to demonstrate clinical utility. A total of 36 studies described 6888 patients. The analysis included 23 studies not incorporated into previous meta-analyses. Hematoma grew at a frequency of 21%. Positive and negative likelihood ratios for the CTA spot sign indicating hematoma growth were 4.25 (95% CI 3.52–5.13) and 0.48 (95% CI 0.41–0.56), respectively. Sensitivity analyses showed a plateau in positive and negative likelihood ratios of 4.3 and 0.48, respectively. Heterogeneity chi square was 214.36–250.24 (P < .001) and inconsistency I2, 83.7–86%. Dr. Phan concluded that the positive like- lihood ratio for the CTA spot sign did not reach the threshold of 5 to demonstrate clinical utility. Heterogeneity regarding this estimate added to uncertainty. The high negative likelihood ratio sug- gested that absence of the CTA spot sign does not guarantee a lack of growth of intracerebral hemorrhage. CTA spot sign predicted neither mortal- ity from nor expansion of spontaneous

A lifetime Markov model was developed to simulate patient follow-up. Long-term continuous monitoring with an insertable cardiac monitor was compared with con- ventional monitoring. A linked evidence approach was used to estimate rates of recurrent stroke when atrial fibrillation detection led to initiation of oral anticoagulation, as detected using the insert- able cardiacmonitor during the lifetime of the device vs detection during usual care. Diagnostic and patient management costs were modeled. Other model inputs were determined by literature review. Probabilistic sensitivity analysis was under- taken to explore the effect of parameter uncertainty according to the Congestive heart failure; Hypertension; Age ≥75 years; Diabetes mellitus; prior Stroke, TIA, or thromboembolism (CHADS 2 ) score; and oral anticoagulation treatment effect. In the base-case analysis, the model pre- dicted an incremental cost-effectiveness ratio of $29,570 AUD per quality-adjusted life year. Among CHADS 2 subgroup analy- ses, the incremental cost-effectiveness ratio ranged from $26,342 AUD per quality-ad- justed life year (CHADS 2 = 6) to $42,967 AUD per quality-adjusted life year (CHADS 2 = 2). Probabilistic sensitivity analysis suggested that the probabilities of a strategy employing an insertable cardiac monitor being cost-ef- fective were 53.4% and 78.7% at thresholds of $30,000 AUD (highly cost-effective) and $50,000 AUD per quality-adjusted life year (cost-effective), respectively. Dr. Thijs concluded that long-term contin- uous monitoring with an insertable cardiac monitor is cost-effective in preventing recurrent stroke in patients following cryp- togenic stroke in the Australian context. “The results demonstrated,” he said, “that loop recorders provide good value for the money in the Australian context. They iden- tify patients who need anticoagulation to prevent recurrent stroke. We hope clinicians adopt this strategy to help reduce the bur-

den of stroke in Australia.” www.practiceupdate.com/c/68172

intracerebral hemorrhage. www.practiceupdate.com/c/68177

VOL. 3 • NO. 3 • 2018