Endurance State of the Evidence v0.1

Endurance Events — State of the Evidence

Bottom line for practitioners

Endurance racing is a low-incidence, high-consequence environment: the events that kill are rare, cluster late in the course and at the finish, and are survivable when bystander CPR, AEDs, and on-site providers are already in position. The two hard numbers the corpus can defend — cardiac-arrest incidence and outcomes from the RACER registry ^[1] and exercise-associated hyponatremia (EAH) from the 2002 Boston cohort ^[2] — both point the same way: plan for the finish line, treat overdrinking as a modifiable hazard, and build a heat plan around acclimatization and cooling ^[3]. Everything downstream of those three anchors — staffing ratios, aid-station density, ultra and cold-weather profiles — remains largely unextracted here and should be read as a gap, not as reassurance.

What we know

Cardiac arrest (RACER). Across US marathons and half-marathons from 2000–2010, the RACER registry recorded 59 arrests among 10.9 million participants — an incidence of 1 per 184,000, or 0.54 per 100,000 (95% CI 0.41–0.70) ^[1]. Risk concentrates sharply: marathons ran 1.01/100,000 versus 0.27 for half-marathons, and male marathoners were the highest-risk group at 1.41/100,000 ^[1]. Case fatality was 71% (42/59), yet that compares favorably with out-of-hospital arrest generally (~92%), a gap attributed to spectator density, on-site services, and early CPR/AED use ^[1]. Cause splits by outcome: among deaths with complete data, definite or possible hypertrophic cardiomyopathy dominated (15/23), while ischemic disease led among survivors (5/8) — with no acute plaque rupture found, suggesting demand ischemia rather than thrombosis ^[1]. The strongest survival predictors were bystander CPR initiation and a diagnosis other than HCM; a shockable initial rhythm (VF/VT) and more prior races completed were also protective ^[1]. Absolute counts rose over the decade only because participation doubled; incidence itself was stable ^[1].

Exercise-associated hyponatremia (Boston cohort). Of 488 sampled 2002 Boston finishers, 13% (62) were hyponatremic (Na ≤135 mmol/L) — 22% of women, 8% of men — and 0.6% (3 runners) were critically hyponatremic at ≤120 mmol/L ^[2]. On multivariate analysis the dominant, and modifiable, driver was overdrinking: weight gain during the race carried an OR of 4.2 (per-kg OR 2.0), and racing longer than 4 hours an OR of 7.4 versus sub-3:30 ^[2]. Critically, female sex, fluid composition (water versus sports drink), and NSAID use were not significant after adjustment — the message to the finish line is that volume consumed, not electrolyte brand or sex, is what to counsel against ^[2].

Heat doctrine (Racinais consensus). The multi-society consensus names heat acclimatization — repeated exercise-heat exposure over 1–2 weeks — as the single most important strain-reducing intervention ^[3]. Athletes should start euhydrated and minimize dehydration; pre- and per-cooling (e.g., cooling vests) can extend heat-storage capacity; and organizers are advised to provide shade, cooling and rehydration facilities, and to schedule events to minimize risk — explicitly at mass-participation events and on the first hot days of the season ^[3].

Finish-line demand concentration. The corpus supports treating the finish as the center of gravity. RACER arrests cluster in the latter portions of the course and the finish area, which the authors use to justify concentrating finish-line medical and AED coverage ^[1]. Exercise-associated collapse — the most common presentation in the finish tent — occurs after completion, now attributed principally to transient postural hypotension from lower-extremity venous pooling once the runner stops, not to dehydration or hyperthermia; providers must nonetheless screen every collapse for arrest, heat stroke, and EAH first ^[4].

The endurance demand profile

What the corpus does not yet supply here is a defensible staffing model for races. The marathon staff fraction, aid-station spacing, and presentation-rate norms that planners want are not present in extracted content: the Twin Cities 12-year profile [roberts-2000-twin-cities-marathon unextracted] and the surge-planning literature [chiampas-2009-preparing unextracted] are cited here for existence only. General mass-gathering forecasting methods exist and are complementary — retrospective review outperforming a-priori models on a recurring event in one head-to-head ^[5] — but that comparison was run on a multi-day agricultural show, not an endurance race, and its PPR/transport figures do not transfer. Treat any race PPR or staffing ratio as pending extraction, not as an evidence-based number.

What's contested or fragile

Two fault lines matter. First, guideline scope conflation: the team-physician consensus statements govern participants and explicitly disclaim spectator care ^[6], yet event plans routinely blend participant and spectator provision. The distinction is load-bearing — the denominators, risk profiles, and legal duties differ — and the 2022 update inherits a lineage [herring-2004-participation unextracted] that has always drawn that line. Second, single-cohort dependence: the EAH incidence and odds ratios everyone quotes rest on one race, one year — 2002 Boston ^[2]. The 13% figure and the weight-gain OR are internally strong but externally unreplicated in this corpus; a different course, climate, or finisher mix could move them.

What we don't know

The unextracted backlog is substantial and named: Twin Cities profiling ^[7], surge preparation ^[8], the Boston care model ^[9], ultramarathon services ^[10], the Pittsburgh experience ^[11], and the World Athletics medical standards ^[12,13] are all present as citations without extracted content. Whole sub-domains are dark: ultra-endurance (distances and durations that break the RACER 4-hour marathon assumption), and cold-weather endurance, have no grounded numbers here at all. A contemporary RACER-successor registry exists [kim-2025-cardiac unextracted] and should be extracted before any incidence claim is treated as current.

How MGMI operationalizes

MGMI encodes what the evidence supports and flags what it cannot. The heat plan follows the consensus: acclimatization guidance, cooling/rehydration infrastructure, and schedule-risk framing surface directly from ^[3], and WBGT activity tiers are staged for display pending physician sign-off on threshold values — the corpus grounds the doctrine but not a validated numeric cut table. The composer gates provision by participant-versus-spectator scope, mirroring the consensus disclaimer so that a race plan does not silently import spectator assumptions ^[6]. The demand calculator carries endurance-specific limits, but the intended reference for those limits (Nable) is not in the citable whitelist; until it is extracted, the calculator's endurance output must be labeled provisional rather than evidence-cited. Finish-line and AED placement defaults follow the RACER clustering finding ^[1].

Reading pathway

Start with ^[1] for the arrest incidence and survival logic, then ^[2] for the overdrinking mechanism, then ^[3] for the heat plan and ^[4] for the finish-tent differential. Read ^[6] to fix the participant/spectator scope before applying any of it.

Citations

^[1] — Cardiac Arrest during Long-Distance Running Races, NEJM 2012 (RACER). Extraction/full-text.
^[2] — Hyponatremia among Runners in the Boston Marathon, NEJM 2005. Extraction/full-text.
^[3] — Consensus Recommendations on Training and Competing in the Heat, Sports Med 2015. Domain summary + key findings.
^[4] — Exercise-associated collapse: an evidence-based review, BJSM 2011. Summary (qualitative).
^[5] — Forecasting Medical Work at Mass-Gathering Events, PDM 2005. Extraction; non-endurance setting.
^[6] — Mass Participation and Tournament Event Management (2022 Update), MSSE 2023. Summary (scope).
^[14] — Mass Participation Event Management Consensus, MSSE 2004. unextracted.
^[7] — 12-yr Twin Cities Marathon profile, MSSE 2000. unextracted.
^[8] — Preparing for the Surge, CSMR 2009. unextracted.
^[9] — Boston Marathon medical care model, 2012. unextracted.
^[10] — Medical services at ultramarathons (Western States), 2015. unextracted.
^[11] — Marathon runners and medical care: Pittsburgh, 1986. unextracted.
^[12] — Competition Medical Guidelines for World Athletics Series Events, 2nd ed., 2020. unextracted.
^[13] — Label Road Races Regulations C1.6, 2025. unextracted.
^[15] — Cardiac Arrest During Long-Distance Running Races, JAMA 2025. unextracted.