The Undiagnosed Epidemic: Why AFib Detection Matters
Atrial fibrillation (AFib or AF) is the most common cardiac arrhythmia worldwide, affecting an estimated 33.5 million people globally and 6–7 million Americans. Yet an estimated 3–5 million Americans remain undiagnosed — carrying a significantly elevated stroke risk that a simple diagnosis would allow physicians to treat. The consequence of missed diagnosis is stark: AFib is responsible for approximately 20–25% of all ischaemic strokes, and AFib-related strokes are among the most devastating — larger, more disabling, and more often fatal than strokes from other causes.
The diagnostic challenge is the paroxysmal nature of the arrhythmia. Paroxysmal AFib — episodes that begin and self-terminate, sometimes lasting minutes, sometimes hours — may be completely asymptomatic or produce only vague symptoms (palpitations, mild fatigue, slight breathlessness) easily dismissed. A patient in sinus rhythm at the time of an annual physical will have a normal 12-lead ECG and leave with no diagnosis. Consumer wearable devices have created, for the first time, a plausible pathway to passively screen the population during everyday activities — and the Apple Heart Study demonstrated this is not merely theoretical.
The Physiology of Atrial Fibrillation
In normal sinus rhythm, the sinoatrial (SA) node in the right atrium fires at a regular rate (60–100 bpm), generating a P wave visible on ECG, followed by orderly conduction through the AV node and bundle of His to the ventricles. In atrial fibrillation, the atria fire chaotically at 400–600 electrical impulses per minute from multiple re-entrant circuits and ectopic foci. This disorganised electrical activity means no coordinated atrial contraction — the atria quiver rather than pump. The AV node is bombarded with impulses and conducts them to the ventricles in an irregularly irregular pattern — the hallmark of AFib on ECG and rhythm monitoring.
The consequences are twofold. Haemodynamically: loss of the 'atrial kick' (the 15–20% of ventricular filling provided by organised atrial contraction) reduces cardiac output, causing dyspnoea and fatigue, especially at higher heart rates. Thrombogenically: the stagnant blood in the non-contracting left atrial appendage (a small pouch in the left atrium) is prone to clot formation. These clots can embolise to the cerebral circulation, causing the large-territory ischaemic strokes that characterise AFib-related cerebrovascular events.
How Optical PPG Detects Irregular Rhythm
Photoplethysmography (PPG) — the optical sensor technology in Apple Watch, Fitbit, Garmin, Samsung Galaxy Watch, and Withings devices — uses LED light (typically green for heart rate measurement; some devices additionally use infrared and red for SpO2) shone through the wrist skin, with the reflected and transmitted light measured by a photodetector. Blood pulsing through the microvasculature with each heartbeat changes the light absorption, allowing measurement of pulse timing (the R-R interval — the time between successive heartbeats).
In normal sinus rhythm, R-R intervals vary smoothly with respiration (respiratory sinus arrhythmia — heart rate rises slightly during inspiration and falls during expiration) and are otherwise regular. In AFib, R-R intervals are irregularly irregular — there is no discernible pattern, because ventricular activation timing depends on unpredictable AV node filtering of the chaotic atrial input. Algorithms trained on large labelled datasets can distinguish the signature of AFib irregular rhythm from the regular irregularity of normal breathing variation or the more structured irregularity of PACs/PVCs.
The Apple Heart Study: Population-Scale Validation
The Apple Heart Study (Perez et al., New England Journal of Medicine, November 2019) was the largest cardiovascular screening study ever conducted. Conducted by Apple and Stanford Medicine, it enrolled 419,297 Apple Watch users through a smartphone app and monitored them passively for an average of 117 days. An irregular pulse notification algorithm ran periodically when the wearer was sedentary, analysing 30-minute windows of PPG data.
Key results: 2,161 participants (0.52%) received an irregular pulse notification. Of those who were eligible and willing to wear a follow-up ECG patch (AliveCor KardiaMobile, worn for 1 week), among those who wore the patch simultaneously with a notification event, the positive predictive value (PPV) for AFib was 84% — meaning 84% of notifications coinciding with patch-worn periods reflected genuine AFib. Among those aged ≥65 receiving notifications, the PPV was 84.6%. Of participants who showed AFib on the ambulatory patch and subsequently received a telehealth physician consultation, 57% were recommended for initiation of anticoagulation — treatments they would not have received without the watch notification.
The study also demonstrated the challenge of passive screening: 34% of participants who received notifications showed AFib on the patch, but a large proportion were already in sinus rhythm when they received their notification — illustrating the paroxysmal nature of the arrhythmia and the importance of extended monitoring to catch it.
KardiaMobile: The Medical-Grade Alternative
AliveCor's KardiaMobile device produces a single-lead ECG recording by placing fingers on metal electrode pads (most commonly positioned on either side of a smartphone case or as a standalone device). The signal quality is typically superior to watch PPG because skin impedance is lower with deliberate finger contact and the electrical signal path is more direct. Sensitivity 98%, specificity 97% for AFib versus 12-lead ECG reference standard in the largest validation study (n=2,269). The KardiaMobile 6L adds a 6-lead capability. Unlike the Apple Watch's passive background monitoring, KardiaMobile requires the user to initiate a recording.
Clinical use: KardiaMobile is the preferred option for capturing episodes during symptoms — if a patient experiences palpitations, they can immediately record a KardiaMobile ECG that a physician can review. It is also widely used for monitoring patients on antiarrhythmic medications, patients with known AFib, and post-ablation surveillance. Many cardiologists recommend the combination: Apple Watch (or similar) for passive continuous screening plus KardiaMobile for symptom-triggered documentation.
Stroke Risk Stratification: The CHA2DS2-VASc Score
Not all AFib carries the same stroke risk — individual risk is stratified by the CHA2DS2-VASc score, which assigns points for congestive heart failure (1), hypertension (1), age ≥75 (2), diabetes (1), prior stroke or TIA (2), vascular disease (1), age 65–74 (1), and female sex (1). The ACC/AHA guidelines recommend anticoagulation for males with score ≥2 and females with score ≥3 (excluding sex-only score). At these thresholds, oral anticoagulation reduces stroke risk by 60–68% — one of the most impactful preventive interventions in cardiovascular medicine.
Direct oral anticoagulants (DOACs — apixaban, rivaroxaban, dabigatran, edoxaban) have replaced warfarin as the standard of care for non-valvular AFib. They require no INR monitoring, have more predictable pharmacology, and have lower intracranial haemorrhage rates than warfarin. The net clinical benefit of anticoagulation in high-risk AFib is substantial and well-established — which is why identifying previously undiagnosed AFib in a person with risk factors (elderly, hypertensive, diabetic) through wearable screening has genuine mortality-saving clinical value.
Sleep Apnoea: The Most Treatable AFib Risk Factor
Obstructive sleep apnoea (OSA) is present in approximately 40–50% of patients with AFib — one of the highest rates of any AFib comorbidity. OSA causes repetitive hypoxia and hypercapnia during sleep, activating the sympathetic nervous system and producing oxidative stress and inflammation in atrial tissue. This remodels atrial tissue in ways that promote re-entrant circuit formation — the electrophysiological substrate for AFib. Critically, treating OSA with CPAP significantly reduces AFib burden and recurrence rates after cardioversion or ablation. All AFib patients should be screened for OSA (Berlin Questionnaire, followed by overnight polysomnography or home sleep test if positive). Treating it is not only beneficial for sleep quality but is a direct antiarrhythmic intervention.
The practical implication is that wearable ECG monitoring alone is insufficient for comprehensive AFib prevention — it must be paired with sleep assessment. A wearable that detects intermittent AFib in a patient with undiagnosed severe OSA is capturing the downstream effect while missing the upstream cause. The SAVE trial (McEvoy et al. 2016, NEJM) and subsequent OSA-AFib analyses have reinforced that CPAP adherence is associated with a meaningful reduction in AFib recurrence rates post-ablation, with benefits roughly proportional to the severity of the untreated OSA and the degree of nocturnal desaturation. For individuals using wearable ECG monitoring who also report significant snoring, witnessed apnoeas, or excessive daytime sleepiness, the highest-yield next step is often a sleep study rather than further cardiac testing.