MGMIMass Gathering Medicine

← State of the Evidence

Cardiac State of the Evidence v0.1

Cardiac Events at Mass Gatherings — State of the Evidence

Bottom line for practitioners

Cardiac arrest at mass gatherings is rare, survivable, and time-governed. On the participant side, the RACER registry puts race-related arrest incidence at 1 per 184,000 marathon/half-marathon participants with 71% case fatality — a fatality rate that nonetheless compares favorably with out-of-hospital arrest generally (median ~92%), attributed to spectator density, on-site medical services, and early CPR/AED [1]. On the spectator side, the European (EACPR/ESC) consensus sets the operative doctrine: BCLS plus AED is the minimum acceptable level of care at any mass gathering, with defibrillation within 5 minutes of collapse and AED counts scaled to arena size [2]. The oldest evidence in this domain already showed the model works: at Expo '86, trained lay security personnel defibrillated two VF arrests with AEDs and both patients survived [3]. The planning question is not whether to deploy AEDs but whether your geography, staffing, and drill tempo actually deliver a shock inside the window.

What we know

Participant side. RACER captured 59 cardiac arrests (40 marathon, 19 half-marathon) among 10.9 million US race participants, 2000–2010: overall incidence 0.54/100,000 (95% CI 0.41–0.70), marathon 1.01 vs half-marathon 0.27, men 0.90 vs women 0.16. Male marathoners were the highest-risk group (1.41/100,000), with incidence rising from 0.71 (2000–2004) to 2.03 (2005–2010) [1]. Case fatality was 71% (42/59); sudden-death incidence 1 per 259,000. Etiology splits sharply by outcome: among deaths with complete data, definite or possible hypertrophic cardiomyopathy dominated (15/23), while survivors were mostly ischemic heart disease (5/8) — with no angiographic acute plaque rupture in any runner with coronary atherosclerosis, suggesting demand ischemia rather than plaque events. Strongest survival predictors were bystander CPR initiation and a diagnosis other than HCM; initial shockable rhythm strongly favored survival (VF/VT OR for death 0.040). Arrests cluster in the latter race quartiles and finish area, justifying finish-line-weighted medical and AED coverage. Hyponatremia and hyperthermia caused arrests but were uncommon [1]. For context beyond the racecourse, sports-associated arrest in middle age (35–65) is about 5% of all community SCA (63/1247; incidence 21.7 per million per year), markedly male-skewed, and far more likely to be witnessed than non-sports arrest (87% vs 53%) [4] — witnessed events are precisely the ones a prepared venue can convert to survivors.

Spectator side. The Borjesson consensus is the only comprehensive standard for spectator cardiovascular safety at arenas ≥1,000 capacity. Its core numbers: spectator SCA incidence of roughly 1 per 500,000–600,000 spectators at major European soccer arenas (one study as high as 1 per 260,000); the critical arrhythmia-to-shock window is ~3–5 minutes; the standard is defibrillation ≤5 minutes from collapse [2]. Its Table 2 minimums by arena size: <10,000 spectators — 1–2 AEDs, 1 physician, 1 nurse, 2 medical technicians, 0–1 ambulances; 10,000–50,000 — 4 AEDs, 2 physicians, 1–5 nurses, 2–10 MTs, 1–2 ambulances; >50,000 — 8 AEDs, 2–4 physicians, >5 nurses, >10 MTs, >2 ambulances. AED counts are sized via the Motyka/Crocco method to a 3-minute retrieval-and-response time. It mandates a written 13-item Medical Action Plan (named medical director, arena map, event-specific planning, level of care, personnel, equipment, communication, treatment facilities, transport, documentation, EMS/hospital collaboration, quality improvement, external information), reviewed at least annually. Supporting outcome data within the document: the Fritz-Walter Stadium series (13 witnessed spectator SCAs over 80 months, all VF; defibrillation plus ALS within 4 minutes; 77% ROSC, 62% neurologically intact survival) and a 190-arena survey showing the implementation gap — only 64% had a written MAP, 72% a defibrillator, and 25% of arenas with >5-minute hospital transport had no AED at all [2].

The participant/spectator distinction

These are different populations under different doctrine, and MGMI keeps them separate. Race participants are self-selected exercisers whose arrests concentrate in young-to-middle-aged men, cluster near the finish, and skew toward structural disease (HCM) in fatal cases [1]; the medical plan is course-linear, finish-weighted, and built around mobile response. Arena spectators are a demographic cross-section whose arrests are essentially community out-of-hospital arrests occurring inside a controllable perimeter — witnessed, VF-predominant in the reported series, and won or lost on AED geography and staff drill [2]. Weaver's Expo data foreshadowed the spectator doctrine — lay responders with AEDs at a fixed venue converting VF arrests to survivors [3] — and the Melbourne Cricket Ground tiered-response experience is positioned by its title as a forerunner of public-access defibrillation at stadia, though its outcomes remain unextracted [5]. Conflating the two populations produces plans that put physicians where AEDs are needed and vice versa.

What's contested or fragile

  • Borjesson verification status. The DOI (10.1093/eurheartj/ehr178) is correct and the full text was re-verified verbatim against the publisher HTML on 2026-08-08, but the automated verification gate shows `doi_verified: 0` — a title word-order miss, not a content problem. The entry is a whitelist candidate pending gate correction; treat the extraction note as the grounding of record until then.
  • Single-registry dependence. Essentially all US participant-side incidence doctrine flows from one registry (RACER), which relies on media/keyword case-finding with denominator data from industry participation statistics [1]. Misses are plausible; confidence intervals are wide in subgroups.
  • The RACER-2 update is unextracted. The 2025 JAMA follow-up covers 2010–2023, a period with >29 million US marathon/half-marathon finishers (~3× the 2000–2009 volume), explicitly framed as answering whether contemporary incidence and outcomes have changed [6]. Its abstract in our file truncates before results; until extraction, MGMI's participant-side constants remain anchored to 2000–2010 data.
  • AED effectiveness at population scale is sobering. The AHA's 2024 symposium reports AEDs are used in only 4% of US out-of-hospital arrests despite three decades of public-access defibrillation, with survival "disappointingly low" [7]. Venue programs are the favorable special case, not the general condition.

What we don't know

  • Spectator arrest incidence at US venues. The 1-per-500,000–600,000 figure is European soccer data [2]; no extracted US arena incidence exists in this domain.
  • AED geography optimization. The Motyka/Crocco 3-minute sizing method is consensus-endorsed but we hold no extracted validation of AED placement algorithms at venues; the foundational PAD trial is in the corpus but unextracted [8].
  • The unextracted backlog, named honestly: Page's venue-AED review [9] (unextracted); the MCG tiered-response outcomes [5] (unextracted); the PAD trial [8] (unextracted); AHA lay-rescuer program legislative components [10] (unextracted); the Drezner inter-association school/college SCA preparedness consensus [11] (unextracted); Zeitz's mass-gathering resuscitation case series [12] (unextracted); and RACER-2 results [6] (abstract only). Seven of eleven entries cannot yet ground a number.

How MGMI operationalizes this

The MGMI coverage model's AED line item cites the AHA public-access-defibrillation entries [10] (unextracted; program-components authority) and [7] for program design and the honest base rate of real-world AED use, with counts per venue size taken from Borjesson Table 2. The response-target constant in the coverage model is anchored to defibrillation timing: shock ≤5 minutes from collapse, AED retrieval sized to 3 minutes [2]. Participant-event modules weight coverage toward the final race quartiles and finish area, and treat bystander-CPR density as a modifiable survival input [1].

Reading pathway

Start with [2] (doctrine and minimums), then [1] (participant epidemiology), then [7] (why deployment alone is insufficient). Add [4] for middle-aged sports-arrest context and [3] for the founding venue-AED proof of concept. Extraction priorities, in order: [6], [5], [8], [11], [12], [10], [9].

Citations

  • [1] — Kim JH, Malhotra R, Chiampas G, et al. (RACER Study Group). Cardiac Arrest during Long-Distance Running Races. NEJM 2012. Full-text extraction.
  • [2] — Borjesson M, et al. Consensus document regarding cardiovascular safety at sports arenas (EACPR/ESC). Eur Heart J 2011. Full-text extraction; DOI correct, gate verification pending (word-order miss).
  • [3] — Weaver WD, et al. Emergency medical care requirements for large public assemblies… Ann Emerg Med 1989. Summary + key findings.
  • [4] — Marijon E. Sudden Cardiac Arrest During Sports Activity in Middle Age. Circulation 2015. Abstract summary.
  • [6] — Kim JH, Rim AJ, Miller JT. Cardiac Arrest During Long-Distance Running Races. JAMA 2025. Abstract only; results unextracted.
  • [7] — AHA AED Symposium: Summary and Recommendations. JAHA 2025. Guideline; summary.
  • [10] — Community Lay Rescuer AED Programs: Key State Legislative Components. 2006. Position statement; unextracted.
  • [8] — Hallstrom AP, et al. Public-access defibrillation and survival after out-of-hospital cardiac arrest. NEJM 2004. Unextracted.
  • [5] — Wassertheil J, et al. Cardiac arrest outcomes at the Melbourne Cricket Ground… Resuscitation 2000. Unextracted.
  • [11] — Drezner JA. Inter-Association Task Force Recommendations on Emergency Preparedness and Management of SCA in High School and College Athletic Programs. Clin J Sport Med 2007. Unextracted.
  • [12] — Zeitz KM. Chain of Survival at Mass Gatherings: A Case Series of Resuscitation Events. Prehosp Disaster Med 2007. Unextracted.
  • [9] — Page RL. Use of automated external defibrillators at large public venues. 2000. Unextracted.