Observational study of potential risk factors of immersion pulmonary edema in healthy divers: exercise intensity is the main contributor
Boussuges et al. Sports Medicine - Open
Observational study of potential risk factors of immersion pulmonary edema in healthy divers: exercise intensity is the main contributor
A. Boussuges 0 3
K. Ayme 0 3
G. Chaumet 2
E. Albier 1
M. Borgnetta 1
O. Gavarry 4
0 UMR MD2, Dysoxie-Suractivité, Aix-Marseille Université et Institut de Recherche Biomédicale des Armées (IRBA), Faculté de Médecine Nord , Marseille , France
1 Institut National de Plongée Professionnelle , Port de la Pointe Rouge, Marseille , France
2 Altrabio , Lyon, SA , France
3 UMR MD2, Dysoxie-Suractivité, Aix-Marseille Université et Institut de Recherche Biomédicale des Armées (IRBA), Faculté de Médecine Nord , Marseille , France
4 Laboratoire HandiBio EA 4322, Université de Toulon , La Garde , France
Background: The risk factors of pulmonary edema induced by diving in healthy subjects are not well known. The aim of the present study was to assess the parameters contributing to the increase in extravascular lung water after diving. Methods: This study was carried out in a professional diving institute. All divers participating in the teaching program from June 2012 to June 2014 were included in the study. Extravascular lung water was assessed using the detection of ultrasound lung comets (ULC) by chest ultrasonography. Clinical parameters and dive profiles were recorded using a questionnaire and a dive computer. Results: One-hundred six divers were investigated after 263 dives. They used an open-circuit umbilical supplying compressed gas diving apparatus in 202 cases and a self-contained underwater breathing apparatus in 61 cases. A generalized linear mixed model analysis was performed which demonstrated that the dive induced a significant increase in ULC score (incidence rate ratio: 3.16). It also identified that the predictive variable of increased extravascular lung water after the dive was the exercise intensity at depth (z = 3.99, p < 0.0001). The other parameters studied such as the water temperature, dive profile, hyperoxic exposure, or anthropometric data were not associated with the increase in extravascular lung water after the dive. Conclusions: In this study, the exercise intensity was the main contributor to the increase in extravascular lung water in healthy divers. To improve the prevention of immersion pulmonary edema, the exercise intensity experienced during the dive should thus be adapted to the aerobic fitness level of the divers.
Diving medicine; Pulmonary edema; Ultrasound; Chest ultrasonography
This is the first prospective study investigating the
factors implicated in the increase in extravascular
lung water after diving.
An important increase in extravascular lung water
rarely occurs after a well-controlled dive.
The main contributor to the increase in extravascular lung water in a healthy diver is the exercise intensity.
Pulmonary edema (PE) induced by SCUBA diving has
been described in 1981 by Wilmshurst et al. [
this initial report, several cases of PE induced by various
aquatic activities such as swimming, aquagym,
breathhold diving, and SCUBA diving have been reported [
The injuries may be under-reported, as the intensity of the
clinical disorders may vary from minimal symptoms
(transitory cough) to acute respiratory distress. Consequently,
some divers may not apply for medical attention.
However, diagnosis remains important because when recurrent
injuries have been reported, the second episode of PE can
be fatal . In some cases, PE occurred in divers with
cardiac disease [
]. Most frequently, the injury affected
divers with an apparently normal heart [
]. A deeper
understanding of the pathophysiology of the SCUBA
divinginduced PE is important to better prevent the injury. Risk
factors probably included environmental stressors and
individual factors of susceptibility. Environmental
stressors experienced during SCUBA diving such as water
immersion, cold exposure, exercise intensity, hyperoxia,
and decompression have been implicated in the
]. Furthermore, the injury occurred more
frequently in one sole diver among a group of divers
performing the same dive profile, suggesting that some
individual risk factors existed. In a recent review, Peacher
et al. [
] reported that among recreational divers, risk
factors of cardio-respiratory disorders are more frequent
than initially estimated (from 44 to 72%). Other risk
factors such as female gender [
], advanced age ,
consumption of fish oil [
] or anti-platelet agent [
], and obesity [
] have also been
For a better understanding of the pathogenesis of this
injury, a prospective study is needed. Nevertheless,
immersion pulmonary edema (IPE) is a rare event. The
frequency of the disease has been assessed in an
epidemiologic study performed on 1250 divers and swimmers.
From this group, 1.1% presented symptoms related to IPE
]. To undertake a prospective study, a procedure that
can detect clinically silent PE would be advantageous.
Chest ultrasonography is recognized as an interesting tool
to visualize alveolo-interstitial syndrome using specific
images called ULC (Ultrasound lung comets, Fig. 1) [
It has been used to quantify extravascular lung water
Fig. 1 Two ultrasound lung comets (arrows) arising from the pleural line
(EVLW) in detecting minor PE in both climbing victims
of mountain sickness [
] and athletes performing
heavy exercise at sea level [
]. Chest ultrasonography
has also been performed after apnea competitions
demonstrating that breath-hold diving could increase
the ULC score (ULCs) particularly in divers with
clinical troubles [
The present study was therefore designed to assess the
environmental and individual factors promoting an
increase in EVLW in a large population of professional
divers with no medical history of cardio-respiratory
disease. We hypothesized that risk factors suggested in
previous studies such as cold temperature, heavy
exertion, and stressors induced by the dive profile are
implicated in the increase in EVLW after diving.
This prospective study consisted in an assessment of
EVLW using ultrasonographic chest examinations in
healthy volunteers who performed dives in the
Our study was an observational study performed during
the daily activity of the French National Institute of
Professional Diving. All divers participating in the teaching
program (8 weeks) were included in the study from June
2012 to June 2014. While the frequency of the injury is
rather low (around 1%), the increase in EVLW,
suggesting minor pulmonary edema, is more frequent. For
example, in the study of Castagna et al. [
], an EVLW
accumulation was reported in 73% of divers.
Consequently, the investigation of a population larger than 100
individuals was considered to be appropriate for the aim
of our study.
Each subject passed a screening examination, which
included a physical examination and a review of their
medical history. The baseline investigations, including
chest ultrasonography, were performed between the
third and the eighth day after entering the diving center,
in a period without diving activity. During the study, the
subjects were investigated directly in the boat, inside a
cabin with a controlled ambient temperature (25 °C),
before and after the dives. The number of dives per
subject varied according to the availability of the
investigators and the weather conditions. In the case of several
examinations for the same diver, the recordings were
performed before and after each dive. The post-dive
investigations, which included chest ultrasonography, were
recorded between 5 and 60 min after the end of the dive.
This delay was considered appropriate because Ljubkovic
et al. [
] have reported that ULC disappeared within 2 or
3 h after surfacing. The diving parameters were also
recorded (depth, duration at the bottom, duration of
immersion, breathing apparatus, decompression stop
duration, gas at depth and during decompression stops,
diving suit type, and water temperature). The dive profile
was recorded using a dive computer (Galileo Sol,
Scubapro-Uwatec, Antibes, France). Furthermore, the
occurrence of respiratory symptoms such as coughing,
shortness of breath, expectorations or hemoptysis, during
or after the dive was systematically researched. Lastly, the
exercise intensity (using the Borg Rating of Perceived
Exertion Scale—RPE) and cold perception (using a visual
analog scale—VAS cold) experienced during the dive were
quantified. The training program and the dive profiles
were decided by the diving instructors. Consequently, the
investigators were blinded to the dive profiles as well
as to the stressors experienced by the divers during
exercise at depth.
The ultrasonographic examinations were carried out by
experienced investigators using a commercially available
Doppler echocardiograph (Esaote Mylab 30CV, Genoa,
Italy) connected to a 2.5–3.5 MHz transducer array. The
researchers were blinded to the clinical status as well as
to the results of the questionnaire. Detection of ULC
was performed using a recognized method [
16, 19, 23
previously used for breath-hold divers [
]. A small
number of ULC can be found in healthy subjects. In our
work, according to past studies [
], an increase in
ULC number of more than four artifacts when
compared with the examination performed before the
corresponding dive and a score greater than 5 was considered
as a sign of increase in EVLW. The investigations were
recorded on computers and were subsequently read by
two independent observers.
Continuous variables were expressed as mean ± standard
deviation. All the statistical analyses were performed
with R statistical software [
]. Each subject served as
his own control. Two series of measurements were
obtained: the first as a control before each dive and the
second after the dive. We then searched for the factors
associated to the increase in ULC number. First, linear
relationship between our variables was described by
using Spearman’s correlation.
Ultrasound examinations of the chest were performed
between 2 and 12 times per subject, before and after
diving. Number of diving instances per subject varied
between one and six. Thus, we needed to take this source
of randomness into account and assess the implied
unbalance of our protocol. Because Ancova based on
linear model with imbalanced data produced numerous
], we used the most appropriate statistical
tool, i.e., a mixed model with the following fixed effects:
protocol factor, diving, individual factors such as age,
weight, etc. and characteristics of the dive profile (e.g.,
duration of the dive, water temperature, etc.) and
random effect: subjects.
Wald statistic was calculated and reported (z value) for
each fixed effect. The z value is the Wald statistic for
testing the hypothesis that the corresponding parameter
(regression coefficient) is zero. Under the null hypothesis, it
has an approximately N(0,1) distribution. P(> |z|) is the tail
area in a two-tail test, i.e., the test within a two-sided outer
hypothesis. Our dependent variable, a sign of increase in
EVLW, is considered a rare event in a healthy population.
Thus, we used a negative binomial model that takes into
account overdispersion in the distribution of events.
We compared two models by Akaike information
criterion (AIC) [
]: the first was composed by all the
predictive variables with no interaction and a second with all
the variables plus age crossed with the physical activity
intensity level (Borg scale). Model with the lowest AIC was
kept. The lme4 package [
] was used for these analyses.
The incidence rate ratio was calculated with the exponent
of the estimate when appropriate. To assess the stability of
our results, we produced power calculation through
simulation process: 100 new dependant variables were
generated and 100 new mixed model calculations were
computed thus proportion of success (p value equal or
inferior to 0.05) defines as power.
One-hundred six divers (104 men, 2 women) were
included in the study. Their mean age was 31 ± 7 years
(from 19 to 58) with most divers (75%) between 27 (first
quartile) and 35 years old (third quartile). The mean
weight was 80 ± 12 kg, and the mean height was
178 ± 6 cm. At baseline (at admission into the diving
institute), all divers had a number of ULC lower than 5: 76
divers had no ULC, 22 divers had 1 ULC, 4 divers had 2
ULC, 3 divers had 3 ULC, and 1 diver had 4 ULC. These
ULC were mainly recorded on the lower intercostal
spaces on the antero-lateral chest (parasternal,
midclavicular, anterior axillary, and mid-axillary lines) on
the right side (41% of cases) and on the left side (31%).
ULC were more rarely recorded on the posterior chest
(posterior axillary, scapular and paravertebral lines) on
the right (16%) or on the left side (13%).
In total, 263 dives (Table 1) were investigated (from one
to six dives per subject). The water temperature was
16 ± 3 °C (range 10–23 °C). The divers breathed air at
the bottom. They were submitted to a maximal partial
pressure of oxygen (O2) from 0.38 to 1.22 ATA. The
ascent rate was 9 to 15 msw.min−1, with a decompression
stop at 3 and/or 6 msw and sometimes 9 msw, according
to the recommendations of the decompression tables
developed by the French Ministry of Labor in 1992
(MT92 table) [
]. During the decompression period at
3 and 6 msw, the divers breathed 100% O2; they
breathed air at 9 msw. Forty-three successive dives were
Examinations after the dives
Two divers reported one episode of breathlessness
during the decompression stops. In the two cases, these
troubles disappeared before the end of the water
immersion stay and no respiratory trouble was recorded
during the ultrasonographic examinations performed
after the dive. In one diver, a significant increase in ULC
number was recorded by chest ultrasonography, while in
the other diver, no ULC was observed.
Linear relationship between the variables
To assess relationship between the studied variables, a
correlation matrix was performed (Spearman’s rho). The
correlation plot is reported in Fig. 2. High linear
relationships were observed between diving duration
(scaled) and material (SCUBA = 1, HOOKAH = 0), gas
during decompression stops (O2 = 1, Air = 0) and
material, gas during decompression stops and diving
duration (scaled), height (scaled) and weight (scaled)
VAS cold, and material. ULC number has linear
relationship with only two variables: diving and RPE (Borg
Association and potential predictors
When comparing model with all variables plus
interaction between age × Borg scale (RPE) and model with
all variables without interaction, AIC was lower (923.89
vs. 925.64) for the model with no interaction. The
simpler model (without interaction) was kept for the
Results of the generalized linear mixed-effects model
for the negative binomial family are reported in Table 2.
Diving increased ULC number by 3.16 (IRR confidence
interval, 2.01–4.57, z = 6.42, p = 1.37 × 10−10). The only
other predictive variable of an increasing number of
ULC was the exercise intensity (RPE) at depth (z = 3.99,
p = 0.0000659). As seen in 3.4, diving has no significant
linear relationship with other variables than ULCs, and
RPE has weak linear relationship with material, gas
during decompression stops, weight (scaled) VAS cold,
and ULCs restraining confounding of a third variable.
The present study provides several interesting findings
on the risk of increase in EVLW after diving in healthy
Increase in extravascular lung water after diving
First, the occurrence of respiratory troubles in the
context of well-controlled dives is rare. In the population
studied, only 2 divers out of 106 reported one episode of
breathlessness at the end of the dive during the
decompression stop. These respiratory symptoms quickly
disappeared, and the divers were asymptomatic when
they were submitted to the investigations on board. In
our work, chest ultrasonography was used to detect an
increase in EVLW. One diver with breathlessness during
the decompression stop had an elevated ULCs whereas
the other had no ULC. Consequently, the respiratory
troubles could not be related to minor pulmonary
In the whole population, although a significant
increase in ULC number was recorded after the dives,
ULC number remained low. At the baseline
examination, ULC number was lower than 5 in all volunteers.
After the dives, few divers presented a ULC number
greater than 5. According to the criteria chosen to define
a significant increase in ULC number, only 8 divers out
of 106 had ultrasonographic signs of increase in EVLW.
This increase was recorded in 12 dives out of 263
investigated. Consequently, it can be concluded that an
increase in EVLW rarely occurs after a well-controlled
dive in healthy subjects. Given the infrequent occurrence
of the increase in EVLW reported in our population of
divers (104 men and only 2 women), our results should
be interpreted with cautions.
The contribution of exercise in the increase in extravascular lung water
The statistical analysis demonstrated that the intensity
of exercise was the main contributor to the increase in
EVLW. This result, recorded in our population of
healthy subjects, agreed with physiological knowledge
and previous case series of IPE. It is well recognized that
strenuous exercise on land can induce PE in healthy
]. During exercise on land, hemodynamic
changes include an increase in both the cardiac output
and pressure in pulmonary and systemic circulation. The
increase in the pulmonary arterial pressure, leading to
pulmonary capillary leakage and pulmonary capillary
stress failure, is likely to be implicated in the
pathophysiology of exercise pulmonary edema. In some individuals,
such as subjects with left diastolic function impairment,
an increase in both left atrial pressure and pulmonary
capillary pressure can contribute to the pulmonary
In water, supplementary cardio-respiratory strains are
experienced. The high density of water in comparison to
the air density, leads to an increase in ambient pressure.
In SCUBA diving, it has been reported that the wet suit
contributes to the increase in hydrostatic pressure [
This external hydrostatic pressure is transmitted to the
smooth wall peripheral vessels reducing their
compliance and leading to a blood mass transfer towards the
central circulation [
]. In addition, the hydrostatic
pressure also triggers a capillary shift of interstitial fluid
into the blood compartment, which after some delay
significantly increases the plasma volume [
hemodynamic changes have been investigated at depth
using underwater Doppler-echocardiography. A
restrictive transmitral filling pattern, in favor of an impairment
in left ventricular diastolic function leading to an
increase in both left atrial pressure and pulmonary arterial
wedge pressure, has been reported [
Furthermore, the myocardial wall stretch, due to the
blood mass transfer, leads to the release of natriuretic
peptides in the plasma, and is reported both after water
immersion or SCUBA diving [
22, 41, 43, 44
peptides are recognized to increase endothelial
permeability . Lastly, the specific stressors induced by
SCUBA breathing during exercise at depth might
contribute to the development of PE through the increased
workload of breathing, thus leading to great pleural
pressure variations and subsequently to great variations
in cardiac preload and after load [
]. Consequently, the
cardiac, respiratory, and humoral alterations induced by
water immersion could explain that exercising in water
was the main contributor to the increase in EVLW in
The role of other environmental stressors
The other environmental stressors investigated in our
work were not related to the increase in ULCs.
Cold exposure has been implicated in the pathogenesis
of IPE [
]. Nevertheless, PE have also been described
after diving in warm water [
3, 20, 46
]. In our study, the
statistical analysis did not demonstrate a correlation
between ULCs and water temperature or coldness.
Consequently, cold exposure did not seem to be a major
contributor to the increase in EVLW. Nevertheless, our
study was carried out in the Mediterranean Sea and the
water temperature ranged from 10 to 23 °C. One could
object that the cold stressor was not sufficient enough.
Further studies in fresh water would be interesting to
better assess the contribution of cold in the pathogenesis
No correlation between the increase in ULCs and the
parameters of the dive profile was recorded in our work.
The magnitude of the decompression stressor induced
by the dive seemed unrelated to the increase in EVLW.
This result agreed with the previous studies. In a study
performed in 12 divers, Dujic et al. [
] did not observe
any significant increase in ULCs after a no-decompression
dive performed at 33 msw. In this work, a high bubble
grade was recorded in the divers. In the study of Dujic
et al. [
] and the present study, the profiles were
controlled and the decompressions were performed according
to the recommendations of the Norwegian and French
diving tables, respectively. These results did not disagree
with the fact that pulmonary edema can be induced by a
major decompression stress such as pulmonary
decompression sickness, i.e., “chokes” [
Lastly, hyperoxia might be implicated in the
pathogenesis of IPE through an increase in the production of
oxygen-free radicals. Indeed, it has been demonstrated
that a long duration exposure to an increase in oxygen
partial pressure was associated to endothelial damages in
the pulmonary capillaries and alveoli, resulting in
interstitial and alveolar edema [
]. In our work, for short
duration dives (37 ± 16 min at the bottom) performed
by healthy subjects, hyperoxia did not contribute to the
increase in ULCs. This result could be related to the fact
that the risk of IPE in military divers using
oxygenenriched gas mixtures through rebreathers did not seem
to be higher than in swimmers or recreational divers
]. On the other hand, a protective effect of
hyperoxia has been recently evoked. When compared with
normoxia, hyperoxia counteracted the increase in both
central venous pressure and mean arterial pulmonary
pressure recorded at depth (4.7 ATA) during exercise
. In our work, no difference in ULCs was found
after the dives between subjects breathing oxygen and
subjects breathing air at the decompression stop.
Nevertheless, our study was performed in a diving
school of professional divers leading to the use of
specific procedures and materials such as a
HOOKAH. The decompression stops were most frequently
performed using oxygen breathing (88% of cases). A
few SCUBA dives, breathing air for the whole
duration of the dive, were investigated (31 dives).
Although these SCUBA divers did not develop more
numerous ULC after diving than others (232 dives),
the small sample studied led to an impairment of the
power of the statistical analysis. Furthermore, if the
protective effect of hyperoxia is real, the procedure
used in our diving center could explain the scarcity
of the increase in EVLW recorded after the dives,
when compared with the recent study of Castagna
et al. [
]. Consequently, it seems important to
perform further studies on the potential protector effect
of hyperoxia against IPE, in particular among SCUBA
According to the results of our prospective study, the
main contributor to the increase in EVLW in healthy
divers is exercise intensity. The other environmental
stressors investigated such as the temperature of the
water, the dive profile, and the hyperoxic exposure were
not significantly related to the increase in ULC number
assessed by chest ultrasonography. These findings are
important to improve the prevention of IPE in divers.
Indeed, before dive training, it is important to assess the
aerobic fitness of the divers and subsequently adapt the
exercise intensity experienced during the dive to the
aerobic fitness level of the individual diver.
AIC: Akaike information criterion; ATA: Atmosphere absolute; EVLW: Extra
vascular lung water; HOOKAH: Open-circuit umbilical supplying compressed
gas; IPE: Immersion pulmonary edema; msw: meters of sea water;
O2: Oxygen; PE: Pulmonary edema; SCUBA: Self-contained underwater
breathing apparatus; ULC: ultrasound lung comets; ULCs: ULC score
The authors gratefully acknowledge the volunteers. This work was made
possible through the support of the French National Institute of Professional
Diving and its director, Commandant Paul Gavarry.
This study was supported by a French Ministry of Defense research grant
(Direction Générale de l’Armement, PDH-1-SMO-2-0701). The funds were
used to pay the rent of the ultrasound machine.
Availability of data and materials
AB and OG conceived and designed the study. AK, EA, and MB assisted with
the technical aspects of the protocol, recruited all the participants, and were
involved in the acquisition of the data. AB and KA analyzed the data and GC
performed the statistical analysis. AB, GC, and OG have drafted the article
while MB and EA revised it critically for important intellectual content. All the
authors have given final approval of the version to be published.
Ethics approval and consent to participate
All experimental procedures were conducted in accordance with the
declaration of Helsinki and were approved by the local ethical committee
(Aix-Marseille University, n°2011-A01299-32). Each method and potential risk
were explained to the participants in detail and they in turn gave written
informed consent before the experiment.
Consent for publication
Consent to publish has been obtained from the participants to report
Alain Boussuges, Karine Ayme, Guillaume Chaumet, Eric Albier, Marc
Borgnetta, and Olivier Gavarry declare that they have no conflict of interest.
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
1. Wilmshurst PT , Nuri M , Crowther A , et al. Forearm vascular responses in subjects who develop recurrent pulmonary edema when SCUBA diving : a new syndrome . Br Heart J . 1981 ; 45 : 349 .
2. Boussuges A , Pinet C , Thomas P , et al. Haemoptysis after breath-hold diving . Eur Respir J . 1999 ; 13 : 697 - 9 .
3. Hampson NB , Dunford RG . Pulmonary oedema of SCUBA divers . Undersea Hyperb Med . 1997 ; 24 : 29 - 33 .
4. Koehle MS , Lepawsky M , McKenzie DC . Pulmonary oedema of immersion . Sports Med . 2005 ; 35 : 183 - 90 .
5. Pons M , Blickenstorfer D , Oechslin E , et al. Pulmonary oedema in healthy persons during scuba-diving and swimming . Eur Respir J . 1995 ; 8 : 762 - 7 .
6. Cochard G , Arvieux J , Lacour JM , et al. Pulmonary edema in scuba divers: recurrence and fatal outcome . Undersea Hyperb Med . 2005 ; 32 : 39 - 44 .
7. Dwyer N , Smart D , Reid DW . SCUBA diving, swimming and pulmonary oedema . Intern Med J. 2007 ; 37 : 345 - 7 .
8. Gempp E , Boussuges A , Poyet R , et al. Acute coronary event revealed by diving-related pulmonary edema . Ann Cardiol Angeiol . 2009 ; 58 : 240 - 3 .
9. Kenealy H , Whyte K. Diving-related pulmonary oedema as an unusual presentation of alcoholic cardiomyopathy . Diving Hyperb Med . 2008 ; 38 : 152 - 4 .
10. Ludwig BB , Mahon RT , Schwartzman EL . Cardiopulmonary function after recovery from swimming-induced pulmonary edema . Clin J Sports Med . 2006 ; 16 : 348 - 51 .
11. Slade JB Jr, Hattori T , Ray CS , et al. Pulmonary edema associated with scuba diving: case reports and review . Chest . 2001 ; 120 : 1686 - 94 .
12. Peacher DF , Martina SD , Otteni CE , et al. Immersion pulmonary edema and comorbidities: case series and updated review . Med Sci Sports Exerc . 2015 ; 47 : 1128 - 34 .
13. Coulange M , Rossi P , Gargne O , et al. Pulmonary oedema in healthy SCUBA divers: new physiopathological pathways . Clin Physiol Funct Imaging . 2010 ; 30 : 181 - 6 .
14. Miller CC III, Calder-Becker K , Modave F. Swimming-induced pulmonary edema in triathlete . Am J Emerg Med . 2010 ; 28 : 941 - 6 .
15. Wilmshurst PT , Nuri M , Crowther A , et al. Cold-induced pulmonary edema in scuba divers and swimmers and subsequent development of hypertension . Lancet . 1989 ; 1 ( 8629 ): 62 - 5 .
16. Lichtenstein D , Mézière G , Biderman P , et al. The comet-tail artifact. An ultrasound sign of alveolar-interstitial syndrome . Am J Respir Crit Care Med . 1997 ; 156 : 1640 - 6 .
17. Pratali L , Cavana M , Sicari R , et al. Frequent subclinical high-altitude pulmonary edema detected by chest sonography as ultrasound lung comets in recreational climbers . Crit Care Med . 2010 ; 38 : 1818 - 23 .
18. Pingitore A , Garbella E , Piaggi P , et al. Early subclinical increase in pulmonary water content in athletes performing sustained heavy exercise at sea level: ultrasound lung comet-tail evidence . Am J Physiol Heart Circ Physiol . 2011 ; 301 : H2161 - 7 .
19. Frassi F , Pingitore A , Cialoni D , et al. Chest sonography detects lung water accumulation in healthy elite apnea divers . J Am Soc Echocardiogr . 2008 ; 21 : 1150 - 5 .
20. Boussuges A , Coulange M , Bessereau J , et al. Ultrasound lung comets induced by repeated breath-hold diving: a study in underwater fishermen . Scand J Med Sci Sports . 2011 ; 21 : e384 - 92 .
21. Castagna O , Gempp E , Poyet R , et al. Cardiovascular mechanisms of extravascular lung water accumulation in divers . Am J Cardiol . 2017 ; 119 : 929 - 32 .
22. Ljubkovic M , Gaustad SE , Marinovic J , et al. Ultrasonic evidence of acute interstitial lung edema after SCUBA diving is resolved within 2-3h . Respir Physiol Neurobiol . 2010 ; 171 : 165 - 70 .
23. Agricola E , Bove T , Oppizzi M , et al. Ultrasound comet-tail images: a marker of pulmonary edema . Chest . 2005 ; 127 : 1690 - 5 .
24. Picano E , Frassi F , Agricola E , et al. Ultrasound lung comets: a clinically useful sign of extravascular lung water . J Am Soc Echocardiogr . 2006 ; 19 : 356 - 63 .
25. Gargani L. Lung ultrasound: a new tool for the cardiologist . Cardiovasc Ultrasound . 2011 ; 9 : 6 .
26. Lambrechts K , Germonpré P , Charbel B , et al. Ultrasound lung “comets” increase after breath-hold diving . Eur J Appl Physiol . 2011 ; 111 ( 4 ): 707 - 13 .
27. R Core Team. R: A language and environment for statistical computing . Vienna, Austria; 2015 . URL https://www.r-project. org/ . 2015- 03 -09.
28. Shaw RG , Mitchell-Olds T. ANOVA for unbalanced data: an overview . Ecology . 1993 ; 74 : 1638 - 45 .
29. Akaike H. A new look at the statistical model identification . IEEE Trans Autom Control . 1974 ; 19 : 716 - 23 .
30. Bates D , Maechler M , Bolker B , et al. Fitting linear mixed-effects models using lme4 . J Stat Softw . 2015 ; 67 : 1 - 48 .
31. Imbert JP . Calcul des tables du Ministère du travail MT92 . In: Broussolle B , Méliet JL , editors. Physiologie et Médecine de la Plongée . Paris: Ellipses; 2006 . p. 530 - 40 .
32. Hopkins SR , Schoene RB , Henderson WR , et al. Intense exercise impairs the integrity of the pulmonary blood-gas barrier in elite athletes . Am J Respir Crit Care Med . 1997 ; 155 : 1090 - 4 .
33. Zavorsky GS . Evidence of pulmonary edema triggered by exercise in healthy humans and detected with various imaging techniques . Acta Physiol (Oxf) . 2007 ; 189 : 305 - 17 .
34. Agricola E , Picano E , Oppizzi M , et al. Assessment of stress-induced pulmonary edema by chest ultrasound during exercise echocardiography and its correlation with left ventricular function . J Am Soc Echocardiogr . 2006 ; 19 : 457 - 63 .
35. Castagna O , Blatteau JE , Vallee N , et al. The underestimated compression effect of neoprene wetsuit on divers hydromineral homeostasis . Int J Sports Med . 2013 ; 34 : 1043 - 50 .
36. Arborelius M Jr , Balldin UI , Lilja B , et al. Hemodynamic changes in man during immersion with the head above water . Aerosp Med . 1972 ; 43 : 592 - 8 .
37. Gabrielsen A , Johansen LB , Norsk P . Central cardiovascular pressures during graded water immersion in humans . J Appl Physiol . 1993 ; 75 : 581 - 5 .
38. Risch WD , Koubenec HJ , Beckmann U , et al. The effect of graded immersion on heart volume, central venous pressure, pulmonary blood distribution, and heart rate in man . Pfluggers Arch . 1978 ; 374 : 115 - 8 .
39. Khosla SS , DuBois AB . Osmoregulation and interstitial fluid pressure changes in humans during water immersion . J Appl Physiol Respir Environ Exerc Physiol . 1981 ; 51 : 686 - 92 .
40. Miki K , Hajduczok G , Hong SK , et al. Plasma volume changes during headout water immersion in conscious dogs . Am J Phys . 1986 ; 251 : R582 - 90 .
41. Ayme K , Rossi P , Gavarry O , et al. Cardiorespiratory alterations induced by low-intensity exercise performed in water or on land . Appl Physiol Nutr Metab . 2015 ; 40 : 309 - 15 .
42. Marabotti C , Cialoni D , Pingitore A . Environment-induced pulmonary oedema in healthy individuals . Lancet Respir Med . 2017 ; 5 : 374 - 6 .
43. Norsk P , Bonde-Petersen F , Christensen NJ . Catecholamines, circulation, and the kidney during water immersion in humans . J Appl Physiol . 1990 ; 69 : 479 - 84 .
44. Passino C , Franzino E , Giannoni A , et al. B-type natriuretic peptide secretion following scuba diving . Biomark Med . 2011 ; 5 : 205 - 9 .
45. Curry FR . Atrial natriuretic peptide: an essential physiological regulator of transvascular fluid, protein transport, and plasma volume . J Clin Invest . 2005 ; 115 : 1458 - 61 .
46. Gnadinger CA , Colwell CB , Knaut AL . Scuba diving-induced pulmonary edema in a swimming pool . J Emerg Med . 2001 ; 21 : 419 - 21 .
47. Dujic Z , Marinovic J , Obad A , et al. A no-decompression air dive and ultrasound lung comets . Aviat Space Environ Med . 2011 ; 82 : 40 - 3 .
48. Kondo Y , Shiohira S , Kamizato K , et al. Vascular hyperpermeability in pulmonary decompression illness: 'the chokes . Emerg Med Australas . 2012 ; 24 : 460 - 2 .
49. Crapo JD . Morphologic changes in pulmonary oxygen toxicity . Annu Rev Physiol . 1986 ; 48 : 721 - 31 .
50. Arieli R , Shochat T , Adir Y. CNS toxicity in closed-circuit oxygen diving: symptoms reported from 2527 dives . Aviat Space Environ Med . 2006 ; 77 : 526 - 32 .
51. Gempp E , Louge P , Blatteau JE , et al. Descriptive epidemiology of 153 diving injuries with rebreathers among French military divers from 1979 to 2009 . Mil Med . 2011 ; 176 : 446 - 50 .
52. Peacher DF , Pecorella SR , Freiberger JJ , et al. Effects of hyperoxia on ventilation and pulmonary hemodynamics during immersed prone exercise at 4.7 ATA: possible implications for immersion pulmonary edema . J Appl Physiol . 2010 ; 109 : 68 - 78 .