94 Clinical response to incident af in a tertiary hospital is delayed and consequential. (6th June 2022)

• Record Type: Journal Article
• Authors: Prinz, Nick; Cotton, Caroline; Moorman, Randall; Monfredi, Oliver
• Is Part Of: Heart. Volume 108(2022)Supplement 1
• Page Start: A70
• Page End: A71
• Publication Date: 2022-06-06
• Subjects: Atrial Fibrillation -- Arryhthmia -- Continuous monitoring; Heart -- Diseases -- Treatment -- Periodicals; Cardiology -- Periodicals; 616.12
• DOI: 10.1136/heartjnl-2022-BCS.94
• Languages: English
• ISSNs: 1355-6037

Abstract

Abstract : Atrial fibrillation (AF) is common in hospitalized patients, occurring in c.10% of unselected inpatients. In a vulnerable hospital patient, the development and persistence of AF can precipitate acute hemodynamic decompensation and result in complications including stroke and heart failure. Early management of AF to anticoagulate, rate control or restore sinus rhythm, and manage underlying precipitants mitigates these risks, and prevents longer term anatomical and electrophysiological remodeling that otherwise favor AF recurrence/persistence. Yet we are not good at recognizing the development of AF in monitored hospitalized patients - prior work at our institution showed a median 1.5 day delay from automated algorithmic detection of AF to clinical response. The purpose of this work was to investigate delays in recognition and treatment of AF in floor patients, and their consequences. We undertook a retrospective review of 225 consecutively admitted patients from 12/25/20–3/29/21 who had automated algorithmic detection of AF episodes lasting at least 5 minutes. The algorithm employs both time and non-linear domain characteristics of the time series data to ensure highly sensitive and specific detection of AF. Time from alert to clinical identification, prespecified interventions and complications (Table 1), and prior history of AF were identified in each patient's chart. Patient age, race, sex and Area Deprivation Index were also recorded. The median duration of all AF episodes was 780 seconds. 94/225 patients had at least one AF episode that was completely clinically undetected. 78/225 patients had AF that was clinically detected prior to the first automated algorithmic alert (these patients were in units where cardiorespiratory monitoring was not subject to the AF detection algorithm, e.g. the emergency dept). 53/225 patients had initial clinical identification of AF after first automated detection, and we focused on these. Median delay from AF onset to clinical identification and intervention was 89.5 minutes. Median AF episode duration in this patient group was longer, 3960s, and most were males (70%), mean age 70.5 yrs (range 40–92). Critically, patients who experienced a delay from automated AF detection to medical intervention of >60 minutes were substantially more likely to suffer an AF-related complication (8 AF-related complications vs only 1 AF-related complication in patients who experienced a sub-60 minute delay to intervention, p=0.06). Complications included transition to an increased level of care, heart failure, and myocardial infarction. In summary, we present evidence that even in patients undergoing continuous cardiorespiratory monitoring, there are considerable and medically significant delays in the detection and management of incident AF. We are in the process of implementing an alerting system to immediately notify treating teams and clinicians of automatically detected significant AF episodes to investigate whether earlier clinical recognition leads to earlier intervention, and better outcomes. Conflict of Interest: None