Annals of Internal Medicine
Death and Cardiac Arrest in U.S. Triathlon Participants, 1985 to 2016
A Case Series
Kevin M. Harris, MD; Lawrence L. Creswell, MD; Tammy S. Haas, RN; Taylor Thomas, BS; Monica Tung, BA; Erin Isaacson, BS;
Ross F. Garberich, MS; and Barry J. Maron, MD
Background: Reports of race-related triathlon fatalities have
raised questions regarding athlete safety.
Objective: To describe death and cardiac arrest among triath-
lon participants.
Design: Case series.
Setting: United States.
Participants: Participants in U.S. triathlon races from 1985 to
2016.
Measurements: Data on deaths and cardiac arrests were as-
sembled from such sources as the U.S. National Registry of Sud-
den Death in Athletes (which uses news media, Internet
searches, LexisNexis archival databases, and news clipping ser-
vices) and USA Triathlon (USAT) records. Incidence of death or
cardiac arrest in USAT-sanctioned races from 2006 to 2016 was
calculated.
Results: A total of 135 sudden deaths, resuscitated cardiac ar-
rests, and trauma-related deaths were compiled; mean age of
victims was 46.7 12.4 years, and 85% were male. Most sudden
deaths and cardiac arrests occurred in the swim segment (n =
90); the others occurred during bicycling (n = 7), running (n =
15), and postrace recovery (n = 8). Fifteen trauma-related deaths
T riathlon is a public competitive sport combining
swimming, bicycling, and running (1). Since its ori-
gin in the United States in the 1970s, this sport has
become an increasingly popular endurance activity
worldwide, particularly among adult recreational ath-
letes, as well as in the Olympic Games. For example,
according to data from USA Triathlon (USAT), the U.S.
governing body for the sport, more than 3200 sanc-
tioned adult events involving more than 460 000 partic-
ipants took place in 2015 (1). However, over the years,
the number of race-related fatalities has generated
concern regarding athlete safety during this competi-
tive sport.
Previous preliminary reports suggested that triath-
lon may not be entirely safe (2, 3). Because participation
in triathlon events has grown steadily, we considered
the possibility that the frequency and demographics of
the risks associated with this sport may have changed.
Therefore, we believe our expanded prole, involving
more than 9 million participants over 30 years of com-
petition, is a timely effort to dene the public health
implications of triathlon participation more completely.
METHODS
We prospectively and retrospectively identied
race-related deaths and cardiac arrests occurring dur-
Annals.org
Downloaded From: https://annals.org/ by a Columbia University User on 09/19/2017
ORIGINAL RESEARCH
occurred during the bike segment. Incidence of death or cardiac
arrest among USAT participants (n = 4 776 443) was 1.74 per
100 000 (2.40 in men and 0.74 in women per 100 000; P <
0.001). In men, risk increased substantially with age and was
much greater for those aged 60 years and older (18.6 per
100 000 participants). Death or cardiac arrest risk was similar for
short, intermediate, and long races (1.61 vs. 1.41 vs. 1.92 per
100 000 participants). At autopsy, 27 of 61 decedents (44%) had
clinically relevant cardiovascular abnormalities, most frequently
atherosclerotic coronary disease or cardiomyopathy.
Limitations: Case identication may be incomplete and may
underestimate events, particularly in the early study period. In
addition, prerace medical history is unknown in most cases.
Conclusion: Deaths and cardiac arrests during the triathlon are
not rare; most have occurred in middle-aged and older men.
Most sudden deaths in triathletes happened during the swim
segment, and clinically silent cardiovascular disease was present
in an unexpected proportion of decedents.
Primary Funding Source: Minneapolis Heart Institute
Foundation.
Ann Intern Med. doi:10.7326/M17-0847 Annals.org
For author afliations, see end of text.
This article was published at Annals.org on 19 September 2017.
ing triathlons in the United States between 1985 and
2016. The study group was assembled from USAT re-
cords (3); the U.S. National Registry of Sudden Death in
Athletes (4, 5), which uses news media reports, Internet
searches, LexisNexis archival databases, and news clip-
ping services; and personal reports from interested
parties. Specic keywords and phrases, including triath-
lon death; triathlon fatality; sudden cardiac death, triath-
lon; and cardiac arrest, triathlon, were entered into 2
independent public search engines: LexisNexis and
Google. Deaths occurring during triathlon training,
USAT training camps, events other than traditional
swim bikerun triathlons, and races occurring outside
the United States were excluded. Triathlon events were
classied on the basis of swim segment length: short
(750 m), intermediate (751 m to 1 mile), and long (>1
mile). Deaths were divided into sudden nontrauma-
and trauma-related deaths.
See also:
Editorial comment . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Summary for Patients . . . . . . . . . . . . . . . . . . . . . . . . . 2
Web-Only
Supplement
Annals of Internal Medicine 1
 ORIGINAL RESEARCH Death and Cardiac Arrest in U.S. Triathlon Participants, 1985 to 2016

Table 1. Summary of Deaths and Cardiac Arrests Occurring During Triathlons*

Variable All Deaths/Cardiac Deaths (n 122) Cardiac Arrest Men (n 115) Women (n 20)
Arrests (n 135) Survivors (n 13)
Men 115 (85) 105 (86) 9 (75) 115 (100) 0 (0)
Women 20 (15) 17 (14) 3 (25) 0 (0) 20 (100)
Mean age (SE), y 46.7 12.4 47.2 12.6 41.8 8.0 46.9 12.3 44.2 13.3
Race segment
Swim 90 (67) 85 (70) 5 (38) 78 (68) 12 (60)
Bike 22 (16) 20 (16) 2 (15) 17 (15) 5 (25)
Run 15 (11) 12 (10) 3 (23) 13 (11) 2 (10)
Post race 8 (6) 5 (4) 3 (23) 7 (6) 1 (5)
Race length
Sprint 66 (54) 60 (54) 6 (60) 56 (54) 10 (53)
Intermediate 33 (27) 29 (26) 4 (40) 29 (28) 4 (21)
Long 23 (19) 23 (21) 0 (0) 18 (18) 5 (26)
First triathlon 26 (39) 26 (39) 0 (0) 21 (37) 5 (50)
Body of water
Lake/reservoir 43 (48) 41 (48) 2 (40) 39 (50) 4 (33)
Ocean/gulf/bay/harbor 31 (34) 30 (35) 1 (20) 26 (33) 5 (42)
River/dam 13 (14) 12 (14) 1 (20) 11 (14) 2 (17)
Pool 3 (3) 2 (2) 1 (20) 2 (3) 1 (8)
* Values are numbers (percentages) unless otherwise indicated. Percentages may not sum to 100 due to rounding.
Includes 15 deaths due to blunt trauma.
Data available for 122 participants.
Data available for 67 participants.
Of the 90 cardiac arrests/deaths that took place in the water. Six starts were time trials, 5 were mass, 4 were deep water, 1 was rolling, and 1 was
staggered in pool.

Additional information regarding the medical this study and contacted next of kin by mail for consent
background of victims and race conditions was ob- to participate in the study. The funding organizations
tained through contact with next of kin or race had no role in the analysis, manuscript preparation, or
directors. For triathletes who died during USAT events decision to submit the manuscript for publication.
(2006 to 2015), USAT contacted decedents' next of kin
through a letter explaining this research effort and
seeking permission to release contact information to RESULTS
the Minneapolis Heart Institute Foundation. If the next Description of Observed Sudden Deaths and
of kin granted permission, he or she was contacted by Cardiac Arrests
the research team and asked to complete a question-
We identied 135 race-related sudden deaths (n =
naire regarding the medical background of the dece-
107), resuscitated cardiac arrests (n = 13), and trauma
dent. The research team contacted race directors di-
deaths (n = 15) in triathlons from 1985 through 2016
rectly for details regarding race conditions. Next of kin
(Table 1). Average age of the victims was 46.7 12.4
provided the research team with additional information
years (range, 15 to 80 years); 115 (85%) were male, and
through a questionnaire or by phone in 15 cases, and
20 (15%) were female. Of the 135 deaths and cardiac
race directors provided additional data regarding 12
arrests, 90 (67%) happened during the swim, 22 (16%)
races. Autopsy results were requested from medical
during the bike, and 15 (11%) during the run segments
examiner ofces, where this information was publicly
and 8 (6%) occurred during postrace recovery. Fifteen
available.
of the 22 deaths during the bike segment were the
This study was approved by the Schulman Institu-
result of blunt trauma incurred in collisions (10 involv-
tional Review Board.
ing motor vehicles and 5 in falls to the ground or colli-
The number of nishers in USAT-sanctioned triath-
sions with stationary objects, such as guard rails). Ath-
lons was calculated by using USAT records from 2006
letes whose deaths were trauma related (12 male and 3
to 2016 (2, 3). The number of race nishers was used as
female) were younger than those who died of other
a surrogate for the number of participants, with the un-
causes (40.5 13.5 vs. 47.3 1.1 years). None of the
derstanding that doing so slightly underestimated par-
deaths occurred among those considered to be elite or
ticipation because of the relatively small number of ath-
professional athletes.
letes who did not nish the race. Age and sex data for
Of the 135 deaths and cardiac arrests, 66 (49%)
each participant were available from 2010 to 2016, and
occurred in short-, 33 (24%) in intermediate-, and 23
sudden deaths during that period were used to gener-
(17%) in long-distance triathlons; for the remaining 13,
ate incidence rates by age and sex.
the race distance was unknown. Of the 68 participants
Role of the Funding Source whose previous race experience was known, 26 (38%)
The Minneapolis Heart Institute Foundation pro- were competing in their rst triathlon. Of these rst-
vided Dr. Harris with data collection and statistical anal- time triathletes, most (69%) were competing in sprint
ysis. USA Triathlon independently collected data for rather than intermediate- (15%) or long-distance (15%)
2 Annals of Internal Medicine Annals.org

Downloaded From: https://annals.org/ by a Columbia University User on 09/19/2017

 Death and Cardiac Arrest in U.S. Triathlon Participants, 1985 to 2016
races. Eleven victims were participating in a 2- or
3-member relay team, with each competing in 1 or 2
segments of the race. The number of participants
in races associated with a death or arrest was
1225 1144 (range, 20 to >3000), and the average air
temperature was 21.9 C 13.1 C (71.5 F 8.4 F)
(range, 6.7 C to 28.9 C [44 F to 84 F]).
Average water temperature, available for 41 of the
swimming-related fatalities, was 22.1 C 13.2 C
(71.8 F 8.2 F) (range, 10.6 C to 31.7 C [51 F to
89 F]). Twenty-three victims were known to be wearing
wetsuits. Of the victims in the swim segment, 9 were
rst observed to be struggling by other competitors or
rescue personnel, 9 signaled for help, 30 were found
unconscious, and 4 collapsed immediately after exiting
the water.
Incidence
From 2006 to 2016, 4 776 443 persons partici-
pated in USAT-sanctioned events. The incidence of all
fatal and nonfatal events and of fatal events only was
1.74 and 1.47, respectively, per 100 000 participants.
Incidence per 100 000 participants was greater during
the swim segment (1.17) than during the bike segment
(0.27), during the run segment (0.19), or after the race
(0.10) (P < 0.001) but did not differ signicantly among
short, intermediate, and long races (1.61, 1.41, and
1.92, respectively).
The event rate for men greatly exceeded that for
women (2.40 vs. 0.74 per 100 000) and varied during
the study period, ranging from a low of 0.39 per
100 000 participants in 2010 to 3.43 per 100 000 in
2012. Mean age of victims was 50 11 years, com-
pared with 38 11 years for all participants.
Sex
The vast majority of events (115 [85%]) occurred in
men, whereas only 20 (15%) occurred in women.
Women and men who died or survived cardiac arrest
did not seem to differ with respect to age, percentage
of deaths during the swim segment, race length, trau-
matic death, participation in rst triathlon, or survival of
cardiac arrest.
The rate of deaths and cardiac arrests per 100 000
participants was 3.3-fold higher for men aged 40 years
and older than for those younger than 40 years (8.25
vs. 2.49) (Table 2). Men's risk increased incrementally
Table 2. Triathlon Death/Cardiac Arrest Incidence, by Age and Sex*
Age Overall
Groups
Deaths, Participants, Death Rate Deaths, Participants,
n n per 100 000 n n
Participants
<30 y 5 351 593 1.42 4 180 139
3039 y 11 550 854 2.00 8 302 603
4049 y 20 505 114 3.96 18 295 951
5059 y 16 239 866 6.67 14 145 677
60 y 8 62 130 12.9 8 42 988
* Data limited to USA Triathlon events between 2010 and 2016.
Sixty of the overall 135 sudden deaths, resuscitated cardiac arrests, and traumatic deaths occurred between 2010 and 2016, the period during
which death incidence could be calculated by age and sex.
Annals.org
Downloaded From: https://annals.org/ by a Columbia University User on 09/19/2017
ORIGINAL RESEARCH
with each decade: 40 years and older to 50 years and
older to 60 years and older (6.1 to 9.6 to 18.6 per
100 000).
Causes of Death
Autopsy reports were available in 61 cases. Of the
remaining cases, autopsy was not performed in 21; an
autopsy was performed in 23, but reports could not be
obtained; and it is unknown whether an autopsy was
performed in the other 17. According to the available
autopsy reports, cardiovascular abnormalities were
identied in 27 decedents (44%), including 18 with sig-
nicant atherosclerotic coronary artery disease (dened
as coronary arterial narrowing greater than 50% in the
left main coronary artery or greater than 70% in the left
anterior descending, circumex, or right coronary
artery).
Three other decedents had evidence of hypertro-
phic cardiomyopathy (Table 3). In addition, 2 athletes
had mitral valve prolapse (1 of whom had a history of
WolffParkinsonWhite syndrome), and 1 each had as-
cending aortic dissection with rupture, spontaneous re-
nal artery dissection, arrhythmogenic right ventricular
cardiomyopathy, and congenital coronary anomaly.
Other causes of death included heat stroke (n = 2),
rhabdomyolysis (n = 1), and traumatic injuries incurred
in a bicycle crash (including 1 in a participant with hy-
pertrophic cardiomyopathy) (n = 2). Of the 6 partici-
pants who died during the run segment, 4 had athero-
sclerotic coronary artery disease. All 18 persons with
atherosclerotic disease were men. Average heart
weight for the 61 decedents who underwent autopsy
was 470 83 g (range, 290 to 675 g).
Cardiac Arrest Survivors
Of the 135 events, 107 were sudden deaths and 13
were race-related cardiac arrests in persons who sur-
vived because of timely cardiopulmonary resuscitation
and debrillation. Of the 13 cardiac arrests, 5 occurred
during the swim segment (including 1 in a swimming
pool), 2 during the bike segment, 2 during the run seg-
ment, and 4 at the nish line or immediately after the
race. Mean age of the cardiac arrest survivors was
42 8 years, compared with 38 11 years for all par-
ticipants, but the survivors were younger than the tri-
athletes who died (50 11 years). Four of the cardiac
arrest survivors subsequently had percutaneous or sur-
Men Women
Death Rate Deaths, Participants, Death Rate
per 100 000 n n per 100 000
Participants Participants
2.22 1 171 454 0.58
2.64 3 248 251 1.21
6.08 2 209 163 0.96
9.61 2 94 189 2.12
18.61 0 19 142 0
Annals of Internal Medicine 3
 ORIGINAL RESEARCH Death and Cardiac Arrest in U.S. Triathlon Participants, 1985 to 2016

Table 3. Causes of Death in 61 Triathletes at Autopsy
Cause of Death Sudden Deaths,
n*
Cardiovascular disease 27
Atherosclerotic coronary artery disease 18
Probable hypertrophic cardiomyopathy 3
Congenital coronary anomaly 1 middle-aged athletes, the risk for sports-related sud-
Myxomatous mitral valve 2 den cardiac death may be severalfold greater during
Arrhythmogenic right ventricular cardiomyopathy 1 competitive versus recreational activities (12).
Aortic/vascular rupture 2**
Noncardiovascular causes 5
Heat stroke 2 for sudden death for a middle-aged person in the gen-
Rhabdomyolysis 1 eral population. The risk associated with triathlon par-
Trauma 2 ticipation also exceeds previous estimates for long-
* No denitive cause of death was evident at autopsy in remaining
cases.
Includes 2 participants with left ventricular hypertrophy.
Dened as heart weight >600 g and left ventricular thickness >20
mm.
One participant had a fatal bike crash.
Acutely angled left coronary artery with slit-like orice.
One participant had coexistent WolffParkinsonWhite syndrome.
** One participant had aortic dissection, and 1 had retroperitoneal
hematoma due to idiopathic spontaneous renal artery dissection.
Multiple trauma, including head injury.
gical coronary revascularization, and 1 had corrective
surgery for congenital aortic valve disease.
DISCUSSION
Triathlon is an increasingly popular and particularly
vigorous public athletic competition, with about
500 000 participants annually in more than 3200 events
in the United States alone (1). For this reason, the pub-
lic health issue of participant safety has become a par-
amount consideration. In our survey of more than 9
million triathletes, sudden death, cardiac arrest, and
trauma-related death were not rare; these events af-
fected 135 athletes, with an incidence of 1.74 per
100 000 higher than earlier estimates and exceeding
the incidence reported for marathon racing (1 per
100 000) (2, 6). On average, decedents were about 12
years older than survivors, and men aged 40 years and
older were at greatest risk.
We also found the incidence of cardiovascular
events to be strikingly lower in female triathletes3.5-
fold less than in menand women made up only 15% of
the present study group. The infrequency of adverse
events in female triathletes is consistent with the sub-
stantial data on sudden death reported in young com-
petitive athletes participating in a wide variety of sports
(4, 5, 7), including marathon racing (6, 8, 9), as well as
those reported in a French national study of partici-
pants in recreational and competitive sports (in which
deaths in females were up to 30-fold less common than
those in males) (10). This sex difference may relate to
the predominance of coronary disease noted in the
male triathletes. Also plausible is that certain, as yet
undocumented, protective metabolic mechanisms are
operative in females that might conceivably suppress
arrhythmogenic and other risks during intense physical
exertion.
The data regarding triathlon risk may be inter-
preted in the context of previous estimates of risk for
4 Annals of Internal Medicine
sports-related sudden cardiac death in young and
middle-aged athletes. Several studies and a recent
meta-analysis reported this risk to be approximately 0.7
per 100 000 per year in the general population (11)
(Table 4), higher in older athletes and those engaged
in less habitual exercise (12, 13). Moreover, among
Thus, the risk associated with participation in a sin-
gle triathlon seems to exceed the expected annual risk
distance running races, including marathon (6, 8, 9). It
is unfortunate that no data are available comparing risk
for sudden death specically during pool or open-
water swimming or during cycling, the other triathlon
components. Finally, although previous reports esti-
mated the risk per hour of participation in such activi-
ties as jogging (14), translating such comparisons to
triathlon is difcult, because most deaths occur during
the early minutes of the race during the swim segment,
and the precise duration of participation before death
is unknown.
As noted previously (2), the vast majority of sudden
cardiac events occurred during the swim, the segment
that initiates the triathlon, and several causative hypoth-
eses have emerged to explain this observation. First,
catecholamine release likely peaks in the early phase of
competition and may play a role in triggering arrhyth-
mias, particularly in athletes with underlying (but unsus-
pected) cardiovascular disease (15). Second, some par-
ticipants may be unfamiliar with and untrained for
open-water swimming and therefore may have dif-
culty dealing with adverse environmental conditions
(such as large waves and cool water temperatures).
Moreover, collisions among swimmers are routine,
given their proximity to one another. The latter risks are
notable, because a substantial proportion of the victims
were rst-time triathlon competitors. Third, water res-
cue is logistically complex, given an athlete's difculty
in resting or signaling for assistance if an emergency
situation arises. On-water rescuers also face challenges
in recognizing and reaching swimmers in distress, as
well as transporting them to a setting suitable for ad-
vanced resuscitation.
Nevertheless, 5 triathletes in this study had cardiac
arrest while swimming and were fortunate to survive
thanks to immediate recognition and resuscitation due
to their proximity to rescuers (in a pool in 1 case). Fur-
thermore, several survivors (8 of 13) had cardiac arrest
while engaged in nonswimming events, underscoring
the benet of more rapid accessibility to resuscitation
and debrillation. For the triathlon, the ratio of cardiac
arrest survival to sudden death (11%) is much less fa-
vorable than that reported for the marathon (29%) (6)
and likely relates to the proximity of rescuers during the
marathon (9) as well as the greater difculties inherent
in water rescue (16, 17). Of note, the incidence of death
and cardiac arrest during the triathlon was not directly
Annals.org

Downloaded From: https://annals.org/ by a Columbia University User on 09/19/2017

 Death and Cardiac Arrest in U.S. Triathlon Participants, 1985 to 2016
related to the length of competition, with most fatalities
paradoxically occurring during the more common,
shorter races.
We cannot be denitive regarding the mechanisms
of death during swimming in the triathlon. Most of
these events were associated with pulmonary edema
and may have resulted from drowning (1518). How-
ever, whether other variables, such as arrhythmias due
to underlying cardiovascular disease or blunt trauma,
served as precipitating factors is unknown. Other theo-
ries offered to explain deaths during the swim segment
include seizures; hypothermia or hyperthermia (related
to wetsuit use); pulmonary edema (swimming-induced
pulmonary edema); and arrhythmias potentially related
to autonomic conict, in which high sympathetic stim-
ulation (such as exertion, anxiety, or physicality) is met
with a sudden parasympathetic surge (such as breath
holding or pharyngeal wetting) (15, 19, 20). Neverthe-
less, in this triathlon cohort, apart from 1 athlete with a
history of WolfParkinsonWhite syndrome, no one was
known to have a preexisting proarrhythmic substrate.
A surprising and important observation of this in-
vestigation, on the basis of autopsy reports, was the
high frequency of clinically silent cardiac abnormalities
(present in about 50% of the cases with an autopsy re-
port available) that may have caused or contributed to
sudden cardiac death. In particular, atherosclerotic cor-
onary artery narrowing was identied in 30% of the de-
cedents (including 2 with thromboses, 1 in the left an-
terior descending artery and 1 in the right coronary
artery), in addition to hypertrophic cardiomyopathy, ar-
rhythmogenic right ventricular dysplasia, congenital
coronary anomaly, and aortic dissection and rupture (4,
5, 21). These pathologic ndings raise the question of
whether detection on preparticipation screening might
have prevented these deaths. Unlike high school and
college athletes, who are required to undergo system-
ORIGINAL RESEARCH
atic screening examinations by their educational insti-
tutions to participate in sanctioned sports, adult triath-
letes act as their own agents in this regard, without
having to meet screening requirements by any supervi-
sory body. How many of the triathletes in our study had
a family history, risk factors, or symptoms of cardiac
disease is unknown, and it should be underscored that
triathletes may seem t while harboring underlying, po-
tentially lethal cardiac disease, which is often difcult to
detect prospectively. Nevertheless, on the basis of our
data, men older than 40 years should consider the po-
tential risks of triathlon competition and the value of
screening for cardiovascular disease by their physicians
(2224).
Practical strategies to reduce the risks of triathlon
competition largely focus on improving event organiza-
tion. In this regard, USAT has promoted a series of spe-
cic recommendations to improve race safety (1, 3).
This framework includes minimum standards for medi-
cal presence at races, water quality, on-water rescuers,
and prospective notication of local emergency medi-
cal services and hospitals; rules regarding wetsuit use
that are designed to prevent hypo- or hyperthermia
during the swim segment; and ensuring safe water
temperatures for races (3). Elimination of mass starts at
Ironman events (with >2000 competitors) would allow
easier recognition of troubled or lifeless participants
during the swim segment.
Future prudent efforts to reduce the number of
swim-related fatalities should focus on targeted, robust,
coordinated, and practiced safety responses to identify
distressed participants promptly, leading to timely car-
diopulmonary resuscitation, debrillation, airway man-
agement, and advanced medical care. In addition, ath-
letes should be educated and trained to prepare
adequately for their event (especially the swim) in sim-
ulated race conditions and should be encouraged to

Table 4. Estimated Mortality Risks for Young and Middle-Aged Participants in Various Sporting Activities, Including Triathlon
and Marathon

Study, Year (Reference) Country Study Years Population Age Range, y Risk for Death
Marathon/triathlon
Kim et al, 2012 (6) United States 20002010 Marathon 2265 1.0*
Redelmeier and Greenwald, 2007 (8) United States 19752004 Marathon Mean: 41 0.8*
Harris et al, 2017 (present study) United States 19852016 Triathlon 1580 1.74*

Young athletes
Maron et al, 2016 (5) United States 19802006 High school and college athletes 39 0.6
Maron et al, 2013 (25) United States 19862011 High school athletes (Minnesota) 1218 0.7
Van Camp et al, 1995 (26) United States 19831993 High school/college athletes 1322 Men: 0.75
Women: 0.1
Maron et al, 2009 (27) United States 19932004 High school/college athletes (Minnesota) 0.9
Holst et al, 2010 (28) Denmark 20002006 Danish athletes 1235 1.2
Harmon et al, 2011 (7) United States 20032013 NCAA athletes 1724 1.86
Corrado et al, 2006 (29) Italy 19792004 Young athletes (Veneto region) 1235 1.9

General athlete population

Bohm et al, 2016 (30) Germany 20122014 Sports participants (Germany) 1075 0.120.15
Marijon et al, 2011 (10) France 20052010 Sports participants (France) 1075 0.46
Marijon et al, 2015 (31) United States 20022013 Sports participants (Oregon) 3565 2.17
NCAA = National Collegiate Athletic Association.
* Per 100 000 participants.
Per 100 000 person-years.

Annals.org Annals of Internal Medicine 5

Downloaded From: https://annals.org/ by a Columbia University User on 09/19/2017

 ORIGINAL RESEARCH
do prerace warm-up swims and wear brightly colored
swim caps to make them easier to recognize. Further-
more, the trauma-related deaths during the bike seg-
ment that are cited here should raise awareness of the
potentially adverse consequences of allowing motor
vehicle trafc to continue on roads during races.
Our use of race nishers in the present analysis
may slightly underestimate the actual number of race
participants. However, using nishers as a surrogate for
participants is unavoidable and similar to the method-
ology used in studies of sudden death and cardiac ar-
rest in long-distance running races (6). Although we
used comprehensive search techniques, we may have
underestimated the total number of deaths, particularly
in the early years of the study, or those that occurred
after the races were completed. In addition, prerace
medical history is unknown for most of the victims.
Taken together, these observations raise the possi-
bility that many deaths occurring during the triathlon,
often in rst-time participants, theoretically may be pre-
vented by better race organization and rescue during
the swim segment as well as by participants' awareness
of potentially lethal cardiovascular disease (2531).
From Minneapolis Heart Institute Foundation at Abbott North-
western Hospital, Minneapolis, Minnesota; University of Mis-
sissippi Medical Center, Jackson, Mississippi; and Tufts Med-
ical Center, Boston, Massachusetts.
Acknowledgment: The authors acknowledge the substantial
help of USAT in providing data for this investigation.
Financial Support: By the Minneapolis Heart Institute
Foundation.
Disclosures: Dr. Creswell reports personal fees from USA Tri-
athlon outside the submitted work and has been an annual
member, a medical review panel member, and a committee
member of USA Triathlon. Authors not named here have dis-
closed no conicts of interest. Disclosures can also be viewed
at www.acponline.org/authors/icmje/ConictOfInterestForms
.do?msNum=M17-0847.
Reproducible Research Statement: Study protocol: See the
Supplement (available at Annals.org). Statistical code: Not ap-
plicable. Data set: Not available.
Requests for Single Reprints: Kevin M. Harris, MD, Minneapo-
lis Heart Institute at Abbott Northwestern Hospital, 920 East
28th Street, Suite 300, Minneapolis, MN 55407; e-mail,
kevin.harris@allina.com.
Current author addresses and author contributions are avail-
able at Annals.org.
References
1. USA triathlon 2015 membership report. 2017. Accessed at www
.teamusa.org/usa-triathlon/about/multisport/demographics on 21
June 2017.
6 Annals of Internal Medicine
Downloaded From: https://annals.org/ by a Columbia University User on 09/19/2017
Death and Cardiac Arrest in U.S. Triathlon Participants, 1985 to 2016
2. Harris KM, Henry JT, Rohman E, Haas TS, Maron BJ. Sudden
death during the triathlon. JAMA. 2010;303:1255-7. [PMID:
20371783] doi:10.1001/jama.2010.368
3. USA triathlon fatality incidents study. Accessed at www.teamusa
.org/USA-Triathlon/News/Articles-and-Releases/2012/October/25
/102512-Medical-Panel-Report on 13 August 2017.
4. Maron BJ, Doerer JJ, Haas TS, Tierney DM, Mueller FO. Sudden
deaths in young competitive athletes: analysis of 1866 deaths in the
United States, 1980-2006. Circulation. 2009;119:1085-92. [PMID:
19221222] doi:10.1161/CIRCULATIONAHA.108.804617
5. Maron BJ, Haas TS, Ahluwalia A, Murphy CJ, Garberich RF. De-
mographics and epidemiology of sudden deaths in young compet-
itive athletes: from the United States National Registry. Am J Med.
2016;129:1170-7. [PMID: 27039955] doi:10.1016/j.amjmed.2016.02
.031
6. Kim JH, Malhotra R, Chiampas G, dHemecourt P, Troyanos C,
Cianca J, et al; Race Associated Cardiac Arrest Event Registry
(RACER) Study Group. Cardiac arrest during long-distance running
races. N Engl J Med. 2012;366:130-40. [PMID: 22236223] doi:10
.1056/NEJMoa1106468
7. Harmon KG, Asif IM, Klossner D, Drezner JA. Incidence of sudden
cardiac death in National Collegiate Athletic Association athletes.
Circulation. 2011;123:1594-600. [PMID: 21464047] doi:10.1161
/CIRCULATIONAHA.110.004622
8. Redelmeier DA, Greenwald JA. Competing risks of mortality with
marathons: retrospective analysis. BMJ. 2007;335:1275-7. [PMID:
18156224]
9. Roberts WO, Maron BJ. Evidence for decreasing occurrence of
sudden cardiac death associated with the marathon [Letter]. J Am
Coll Cardiol. 2005;46:1373-4. [PMID: 16198859]
10. Marijon E, Tafet M, Celermajer DS, Dumas F, Perier MC, Mus-
tac H, et al. Sports-related sudden death in the general population.
Circulation. 2011;124:672-81. [PMID: 21788587] doi:10.1161
/CIRCULATIONAHA.110.008979
11. Mohananey D, Masri A, Desai RM, Dalal S, Phelan D, Kanj M,
et al. Global incidence of sports-related sudden cardiac death [Let-
ter]. J Am Coll Cardiol. 2017;69:2672-3. [PMID: 28545643] doi:10
.1016/j.jacc.2017.03.564
12. Risgaard B, Winkel BG, Jabbari R, Glinge C, Ingemann-Hansen
O, Thomsen JL, et al. Sports-related sudden cardiac death in a com-
petitive and a noncompetitive athlete population aged 12 to 49
years: data from an unselected nationwide study in Denmark. Heart
Rhythm. 2014;11:1673-81. [PMID: 24861446] doi:10.1016/j.hrthm
.2014.05.026
13. Siscovick DS, Weiss NS, Fletcher RH, Lasky T. The incidence of
primary cardiac arrest during vigorous exercise. N Engl J Med. 1984;
311:874-7. [PMID: 6472399]
14. Thompson PD, Funk EJ, Carleton RA, Sturner WQ. Incidence of
death during jogging in Rhode Island from 1975 through 1980.
JAMA. 1982;247:2535-8. [PMID: 6978411]
15. Asplund CA, Creswell LL. Hypothesised mechanisms of
swimming-related death: a systematic review. Br J Sports Med. 2016;
50:1360-6. [PMID: 26941276] doi:10.1136/bjsports-2015-094722
16. Szpilman D, Bierens JJ, Handley AJ, Orlowski JP. Drowning.
N Engl J Med. 2012;366:2102-10. [PMID: 22646632] doi:10.1056
/NEJMra1013317
17. Jeong J, Hong KJ, Shin SD, Ro YS, Song KJ, Lee EJ, et al. Rela-
tionship between drowning location and outcome after drowning-
associated out-of-hospital cardiac arrest: nationwide study. Am J
Emerg Med. 2016;34:1799-803. [PMID: 27342967] doi:10.1016
/j.ajem.2016.06.008
18. Idris AH, Berg RA, Bierens J, Bossaert L, Branche CM, Gabrielli A,
et al; American Heart Association. Recommended guidelines for uni-
form reporting of data from drowning: the Utstein style. Circula-
tion. 2003;108:2565-74. [PMID: 14623794]
19. Shattock MJ, Tipton MJ. Autonomic conict: a different way to
die during cold water immersion? J Physiol. 2012;590:3219-30.
[PMID: 22547634] doi:10.1113/jphysiol.2012.229864
Annals.org
 Death and Cardiac Arrest in U.S. Triathlon Participants, 1985 to 2016
20. Moon RE, Martina SD, Peacher DF, Kraus WE. Deaths in triath-
letes: immersion pulmonary oedema as a possible cause. BMJ Open
Sport Exerc Med. 2016;2:e000146. [PMID: 27900191]
21. Harris KM, Tung M, Haas TS, Maron BJ. Under-recognition of
aortic and aortic valve disease and the risk for sudden death in com-
petitive athletes [Letter]. J Am Coll Cardiol. 2015;65:860-2. [PMID:
25720631] doi:10.1016/j.jacc.2014.09.094
22. Maron BJ, Friedman RA, Kligeld P, Levine BD, Viskin S, Chait-
man BR, et al; American Heart Association Council on Clinical Car-
diology, Advocacy Coordinating Committee, Council on Cardiovas-
cular Disease in the Young, Council on Cardiovascular Surgery and
Anesthesia, Council on Epidemiology and Prevention, Council on
Functional Genomics and Translational Biology, Council on Quality
of Care and Outcomes Research, and American College of Cardiol-
ogy. Assessment of the 12-lead ECG as a screening test for detection
of cardiovascular disease in healthy general populations of young
people (12-25 years of age): a scientic statement from the American
Heart Association and the American College of Cardiology. Circula-
tion. 2014;130:1303-34. [PMID: 25223981] doi:10.1161/CIR.0000
000000000025
23. Fletcher GF, Ades PA, Kligeld P, Arena R, Balady GJ, Bittner VA,
et al; American Heart Association Exercise, Cardiac Rehabilitation,
and Prevention Committee of the Council on Clinical Cardiology,
Council on Nutrition, Physical Activity and Metabolism, Council on
Cardiovascular and Stroke Nursing, and Council on Epidemiology
and Prevention. Exercise standards for testing and training: a scien-
tic statement from the American Heart Association. Circulation.
2013;128:873-934. [PMID: 23877260] doi:10.1161/CIR.0b013e31
829b5b44
24. Maron BJ, Araujo CG, Thompson PD, Fletcher GF, de Luna AB,
Fleg JL, et al; World Heart Federation. Recommendations for pre-
participation screening and the assessment of cardiovascular dis-
ease in masters athletes: an advisory for healthcare professionals
Annals.org
Downloaded From: https://annals.org/ by a Columbia University User on 09/19/2017
ORIGINAL RESEARCH
from the working groups of the World Heart Federation, the Interna-
tional Federation of Sports Medicine, and the American Heart Asso-
ciation Committee on Exercise, Cardiac Rehabilitation, and Preven-
tion. Circulation. 2001;103:327-34. [PMID: 11208698]
25. Maron BJ, Haas TS, Ahluwalia A, Rutten-Ramos SC. Incidence of
cardiovascular sudden deaths in Minnesota high school athletes.
Heart Rhythm. 2013;10:374-7. [PMID: 23207138] doi:10.1016/j.hrthm
.2012.11.024
26. Van Camp SP, Bloor CM, Mueller FO, Cantu RC, Olson HG. Non-
traumatic sports death in high school and college athletes. Med Sci
Sports Exerc. 1995;27:641-7. [PMID: 7674867]
27. Maron BJ, Haas TS, Doerer JJ, Thompson PD, Hodges JS. Com-
parison of U.S. and Italian experiences with sudden cardiac deaths in
young competitive athletes and implications for preparticipation
screening strategies. Am J Cardiol. 2009;104:276-80. [PMID:
19576360] doi:10.1016/j.amjcard.2009.03.037
28. Holst AG, Winkel BG, Theilade J, Kristensen IB, Thomsen JL,
Ottesen GL, et al. Incidence and etiology of sports-related sudden
cardiac death in Denmarkimplications for preparticipation screen-
ing. Heart Rhythm. 2010;7:1365-71. [PMID: 20580680] doi:10.1016
/j.hrthm.2010.05.021
29. Corrado D, Basso C, Pavei A, Michieli P, Schiavon M, Thiene G.
Trends in sudden cardiovascular death in young competitive ath-
letes after implementation of a preparticipation screening program.
JAMA. 2006;296:1593-601. [PMID: 17018804]
30. Bohm P, Scharhag J, Meyer T. Data from a nationwide registry on
sports-related sudden cardiac deaths in Germany. Eur J Prev Car-
diol. 2016;23:649-56. [PMID: 26130495] doi:10.1177/2047487
315594087
31. Marijon E, Uy-Evanado A, Reinier K, Teodorescu C, Narayanan K,
Jouven X, et al. Sudden cardiac arrest during sports activity in mid-
dle age. Circulation. 2015;131:1384-91. [PMID: 25847988] doi:10
.1161/CIRCULATIONAHA.114.011988
Annals of Internal Medicine 7
 Current Author Addresses: Dr. Harris, Ms. Haas, and Mr. Gar-
berich: Minneapolis Heart Institute at Abbott Northwestern
Hospital, 920 East 28th Street, Suite 300, Minneapolis, MN
55407.
Dr. Creswell: Department of Cardiothoracic Surgery, Univer-
sity of Mississippi, 2500 North State Street, Jackson, MS
39216.
Ms. Thomas: Mayo School of Medicine, 200 1st Street South-
west, Rochester, MN 55905.
Ms. Tung: Johns Hopkins University School of Medicine, 733
North Broadway, Baltimore, MD 21205.
Ms. Isaacson: Medical College of Wisconsin, 8701 West Wa-
tertown Plank Road, Milwaukee, WI 53226.
Dr. Maron: Hypertrophic Cardiomyopathy Institute, Tufts
Medical Center, 800 Washington Street, Boston, MA 02111.
Annals.org
Downloaded From: https://annals.org/ by a Columbia University User on 09/19/2017
Author Contributions: Conception and design: K.M. Harris,
L.L. Creswell, E. Isaacson, B.J. Maron.
Analysis and interpretation of the data: K.M. Harris, L.L.
Creswell, E. Isaacson, B.J. Maron.
Drafting of the article: K.M. Harris, L.L. Creswell, T. Thomas,
B.J. Maron.
Critical revision for important intellectual content: K.M. Harris,
L.L. Creswell, B.J. Maron.
Final approval of the article: K.M. Harris, L.L. Creswell, T.S.
Haas, T. Thomas, M. Tung, E. Isaacson, R.F. Garberich, B.J.
Maron.
Provision of study materials or patients: K.M. Harris, T.S. Haas.
Statistical expertise: R.F. Garberich.
Obtaining of funding: K.M. Harris.
Administrative, technical, or logistic support: L.L. Creswell,
T.S. Haas, M. Tung.
Collection and assembly of data: K.M. Harris, L.L. Creswell,
T.S. Haas, T. Thomas, M. Tung, E. Isaacson, B.J. Maron.
Annals of Internal Medicine