Detection of paroxysmal atrial fibrillation in patients with newly diagnosed ischaemic stroke

Key findings

There is insufficient evidence to draw robust conclusions about the comparative clinical and cost-effectiveness of different types of ambulatory electrocardiogram to detect PAF. Our review compared Holter devices, event recorder technologies, and patch technologies.

NICE medical technologies guidance published in 2020 recommends the patch technology, Zio XT, as an option for people with suspected cardiac arrhythmias who would benefit from ambulatory ECG monitoring for longer than 24 hours. The guidance states that data should be collected on factors such as long-term clinical outcomes and resource use.

A health technology assessment published in 2016 by the Canadian Agency for Drugs and Technologies in Health (CADTH) recommends 7 days of continuous outpatient cardiac monitoring with either an ambulatory Holter monitor or external loop recorder for patients who have been discharged from hospital after a stroke or TIA and who did not undergo continuous cardiac monitoring while in hospital.

Two systematic reviews reported that the detection rate for any new atrial fibrillation (AF) was greater in selected (defined as ‘prescreened’) patients following stroke or TIA, compared with unselected patients, and greater again in patients with cryptogenic strokes. The definition of ‘selected’ patients varied and included older age, more extensive testing for arrhythmias before enrolment, or presumed cardioembolic/cryptogenic cause. One review reported lower AF detection rates in TIA cohorts compared with cohorts including both stroke and TIA patients.

There is insufficient evidence to define the optimal duration of long-term monitoring. The length of AF that warrants treatment with anticoagulants is not clear from
the published evidence.

In patients with cryptogenic stroke, extended ECG monitoring for AF detection may be economically worthwhile when traditional willingness-to-pay thresholds are adopted. However, there was substantial variation in the reported incremental cost effectiveness ratios (ICERs). The feasibility of direct comparison of cost-effectiveness across technologies is also limited by heterogeneity in modelling assumptions.

What were we asked to look at?

We were asked to review the published evidence on the clinical and cost-effectiveness of various strategies to detect paroxysmal atrial fibrillation (PAF) in patients with newly diagnosed ischaemic stroke who are selected to have prolonged electrocardiogram (ECG) monitoring

Why is this important?

There is variation across NHSScotland in provision of prolonged ECG monitoring for patients with ischaemic stroke. Some NHS boards refer all patients with ischaemic stroke/transient ischaemic attack (TIA) for prolonged monitoring, whilst other boards refer very few patients. The decision to refer patients is partly dependent on access to monitoring, and there is variation across boards in the criteria for selecting patients. There is also variation across boards in the technologies used to monitor patients, and how the technologies are used (for example, how long patients are monitored for). In addition, access to analyst time – with shortages in some NHS boards – can also lead to variation in monitoring practice.

Referred by

National Planning atrial fibrillation subgroup