Abstract
-
Objective
We evaluated prognostic factors for pediatric drowning patients. The association between functional outcomes and clinical factors was investigated.
-
Methods
A retrospective cohort study was conducted using data for pediatric drowning patients from the Korean Community-based Severe Trauma Survey from 2016 to 2020. The primary outcome was a good prognosis at discharge, defined as a Glasgow Outcome Scale score of 5. A multivariable logistic regression analysis was performed to evaluate independent factors associated with the primary outcome.
-
Results
From 237,616 patients, we identified 406 drowning patients aged ≤19 years (mean age, 8.8 years). At discharge, 41.0% of those patients had a good recovery. The absence of prehospital cardiac arrest (adjusted odds ratio [aOR], 98.7; 95% confidence interval [CI], 32.9–295.8), indoor location (aOR, 4.0; 95% CI, 1.7–9.3), and transfer to a high-volume hospital (aOR, 2.5; 95% CI, 1.1–5.8) were significant independent factors associated with a good outcome. Age, sex, the intent of injury, and prehospital time were not associated with the outcome.
-
Conclusion
Our study identified independent prognostic factors for drowning patients, highlighting the importance of prehospital conditions and hospital care settings in determining outcomes. These findings could be useful in developing clinical strategies for managing such patients.
-
Keywords: Cardiac arrest; Drowning; Hospitals, high-volume; Prognosis
Capsule Summary
What is already known
Previously established risk factors for pediatric drowning include being male, low income, living in rural areas, outdoor place of occurrence, and neglect of parental supervision.
What is new in the current study
In this study, the absence of prehospital cardiac arrest, indoor location, and transfer to a high-volume hospital were associated with good outcomes after pediatric drowning. Age, sex, and intent of injury were not associated with outcomes.
INTRODUCTION
Drowning is a leading cause of mortality from injury worldwide [
1]. In Korea, drowning is the third most common cause of accidental deaths among children aged 1 to 9 years [
2]. Despite a decreasing number of drowning-related deaths [
3], the death to survival ratio is reported to be 1:10, with two out of three of those who survive having been admitted to a hospital [
4].
Drowning is the process of experiencing respiratory impairment from submersion or immersion in liquid. Male sex [
5], age, low income [
6], rural residence, location of occurrence, climatic conditions, safety equipment, safety policies, and parental supervision affect the prognosis of drowning patients [
7,
8]. Drowning-related studies have mainly analyzed risk factors and the final prognoses of patients treated in hospitals [
9]. Although some studies have used national data on deaths by drowning, sequelae other than death as the final prognosis have not been described [
10].
Despite advances in medical science, the prognosis for pediatric drowning patients has not improved significantly, prompting research to identify factors associated with better outcomes. In particular, factors such as hospital volume, prehospital time, and the patient's prehospital condition, in addition to environmental factors, need to be considered.
Our goal in this study was to evaluate prognostic factors in pediatric drowning patients. We investigated the association between functional outcomes and clinical factors, including demographic, prehospital, and hospital factors. We evaluated the variables from prehospital to emergency department (ED) treatment using national data from Korea.
METHODS
Ethics statement
This study was approved by the Institutional Review Board of Kangwon University Hospital (No. KNUH-2023-10-014). The requirement for informed consent was waived because of the study’s retrospective design.
Study design and data source
We conducted a retrospective cohort study using data from the Korean Community-based Severe Trauma Survey, managed by the Korean Disease Control and Prevention Agency. Raw data from the Community-based Severe Trauma Survey were collected for patients with severe injuries and multiple casualties who were transferred by emergency medical services (EMS) between January 1, 2016, and December 31, 2020 [
11,
12]. The data comprise records from the National Fire Agency, EMS, and medical records acquired from hospitals, and they were de-identified in accordance with the Personal Information Protection Act and the Statistics Act. A community-based severe trauma investigation specialist visited hospitals involved with the EMS transfer information and reviewed medical records to confirm the data related to patients with severe injuries and multiple casualties.
Our study population comprised patients included in the 2016–2020 Korean Community-based Severe Trauma Survey whose mechanism of injury was respiratory failure and drowning and who were aged ≤19 years at diagnosis. Those with missing baseline data and cases in which the location of drowning was difficult to determine or unknown were excluded.
Variables
Age was calculated from the date of birth confirmed in the medical record to the reporting date. Hospitals that provided treatment were classified as follows: regional trauma centers (level I trauma centers) [
13], regional emergency medical centers (level I), local emergency medical centers (level II), and local emergency medical institutions (level III) (
Supplementary Table 1). We defined regional trauma centers, regional emergency medical centers, and local emergency medical centers as high-volume hospitals and local emergency medical institutions as low-volume hospitals. Prehospital cardiac arrest was defined as a loss of consciousness, breathing, or pulse, or the absence of breathing during the rescue phase. The rescue phase included the entire process from the moment EMS reached a patient until their arrival at the hospital. We also defined cases of cardiac arrest during transfer to the hospital as prehospital cardiac arrest. In the study data, prehospital cardiac arrest was defined as out-of-hospital cardiac arrest or prehospital death. Among patients with out-of-hospital cardiac arrest, data on return of spontaneous circulation status before and after transport were not included.
The locations where drowning occurred were classified as follows: (1) house; (2) residential facilities other than private homes (nursing homes, shelters); (3) medical facilities or schools (regular curriculum); (4) educational facilities (nonformal curriculum); (5) exercise facilities; (6) road or off-road traffic areas; (7) factories or construction facilities; (8) farms and other primary industries; (9) public entertainment and culture facilities; (10) other commercial facilities; (11) outdoor areas (sea, river); (12) a location difficult to define; and (13) an unknown location. Locations difficult to define and unknown locations were excluded from location of occurrence analysis. We classified locations numbered 11 (outdoor area, sea, river) as outdoor and those numbered 1 to 10 as indoor locations.
Intention in relation to drowning was classified as unintentional, suicide, or as a consequence of violence. The prehospital time was the interval from reporting to EMS until arrival at the hospital, recorded in minutes.
The degree of disability before and after injury was assessed using the Glasgow Outcome Scale (GOS)
(Supplementary Table 2). A GOS score of 2 to 4 after discharge from the hospital was defined as having sequelae after drowning. The primary outcome was defined as a good prognosis at discharge, indicated by a GOS score of 5 (good recovery).
Categorical variables are expressed as numbers (percentages), and continuous variables are expressed as means±standard deviations. Each variable was analyzed using chi-square or Fisher exact tests for categorical variables and Student t-test for continuous variables. To identify independent factors associated with the primary outcome, variables with a P-value 0.05 in the univariable logistic regression analyses were taken as candidates for the multivariable logistic regression analysis. After collinearity was tested for each univariable predictor, a logistic regression model of the eligible variables was constructed using the backward elimination method. Adjusted odds ratios (aORs) and 95% confidence intervals (CIs) were then calculated for each variable in the multivariable logistic regression analysis. For the statistical analysis, IBM SPSS ver. 25.0 (IBM Corp) was used. Statistical significance was set at a P<0.05.
RESULTS
Of 237,616 individuals in the survey results, 24,897 children and adolescents aged ≤19 years were selected. Among them, 406 individuals (mean age, 8.8±6.8 years; 279 male patients, 68.7%) had drowning-related injuries (
Fig. 1), and a bimodal peak was observed (
Fig. 2). The most frequently used medical institution was the local emergency medical center (43.3%), followed by the local emergency medical institution (25.6%). A large number of childhood drownings occurred outdoors (55.4%). Prehospital cardiac arrest occurred in 188 patients (46.3%). Intention in relation to drowning was unintentional in 88.8% of cases, suicide in 10.6%, and subsequent to violence in 0.5%. Before drowning, only three cases (0.7%) had moderate disability. Most of the drowning children (77.8%) arrived at the ED within 60 minutes of the report to EMS. The median prehospital period was 40 minutes. The median prehospital period when death was the final result was 54.5 minutes (interquartile range, 16–1,214 minutes). The median prehospital period when a recovery state was the final result was 30.5 minutes (interquartile range, 8–92 minutes). The GOS at discharge was as follows: good recovery in 166 patients (41.0%), death in 156 (38.5%), severe disability in 37 (9.1%), moderate disability in 33 (8.1%), and vegetative state in 13 (3.2%) (
Table 1). The rate of moderate disability reported in patients before drowning was 0.7% (3 of 406); the final rate of death after drowning among patients with moderate disability was 33.3%, and the final death rate among patients without underlying diseases was 38.5% (P=0.853).
In the univariable logistic analyses, younger age, female sex, high-volume hospital, shorter prehospital time, indoor location, and absence of prehospital cardiac arrest were significant factors associated with a good outcome in drowning children (all P<0.001). In the multivariable logistic analysis, the absence of prehospital cardiac arrest (aOR, 98.7; 95% CI, 32.9–295.8; P<0.001), indoor location (aOR, 4.025; 95% CI, 1.743–9.297; P=0.001), and high-volume hospital (aOR, 2.498; 95% CI, 1.069–5.834; P=0.034) were significant factors associated with a good outcome (
Table 2). In cases without prehospital cardiac arrest, high-volume hospital (aOR, 2.868; 95% CI, 1.005–5.978; P=0.047) and indoor location (aOR, 3.215,95% CI, 1.225–7.884; P=0.008) were also significant factors associated with a good outcome (
Table 3).
Among drowning patients without prehospital cardiac arrest, those treated in high-volume hospitals showed lower rates of death or sequelae compared with those treated in low-volume hospitals (23.5% vs. 40.0%, P<0.01) (
Table 4). However, in cases with prehospital cardiac arrest, no difference in the final outcomes of drowning children was observed based on the hospital volume (
Table 4). The hospital mortality rates in patients with prehospital cardiac arrest were 79.2% in 2016, 78.6% in 2017, 84.8% in 2018, 86.0% in 2019, and 83.3% in 2020. Among the 218 patients without prehospital cardiac arrest, only 1 patient (0.5%) died, and 56 patients (25.7%) had a poor prognosis other than death. Conversely, among the 188 patients who experienced prehospital cardiac arrest, only 5 patients (2.7%) had a good recovery as the final prognosis (
Fig. 3).
DISCUSSION
For this study, we analyzed various factors affecting the final prognosis in pediatric drowning. The multivariate analysis results show that receiving treatment at a high-volume hospital and drowning occurring in an indoor location were associated with significantly better outcomes.
Among the prognostic factors for drowning children, prehospital cardiac arrest and the location where drowning occurred were variables determined before the arrival of EMS. Otherwise, varying hospital types could be involved. The number of deaths or sequelae for the final prognosis was higher in low-volume hospitals than in high-volume hospitals. We attribute this finding to the specialized critical care using a multidisciplinary approach provided in high-volume hospitals, where numerous medical devices are available to treat drowning-related complications such as pulmonary edema, aspiration pneumonia, respiratory failure, cerebral edema, and metabolic disorders.
Emergency and critical care treatment for children is often particularly challenging in low-volume hospitals; therefore, this discrepancy can be expected to be greater than that found in adults. In the case of severe trauma patients, it has been shown that patient prognosis is better when they are transferred directly to a trauma center for treatment rather than being transferred to the nearest hospital [
14]. We recommend that children in prehospital noncardiac arrest be treated in a high-volume hospital, even considering the additional prehospital time. In the case of prehospital cardiac arrest, most patients died or had sequelae, which was not related to the volume capacity of the hospital (
Table 4); therefore, we consider this to be an important factor in preventing deaths related to drowning. Several studies have reported a lack of parental attention as a risk factor for childhood drowning. Additionally, the absence of barriers around swimming pools has been reported to be a risk factor for drowning. Therefore, prevention strategies are important in cases of drowning, such as improving environments where drowning is a risk and educating parents and guardians.
Although the quality of intensive care at high-volume hospitals could be a contributing factor, it is also possible that high-volume hospitals are located in more accessible areas. We analyzed the effect of this factor and the differences in prehospital times between high- and low-volume hospitals (
Supplementary Table 3).
Many other factors have been associated with pediatric drowning. Age is an important factor in pediatric drowning. Previous studies have reported that the incidence of drowning peaks in children aged <4 and >10 years [
15,
16]. Young men have been reported to have a high rate of drowning [
17]. In this study, men (68.7%) were overrepresented; however, this was not a significant factor in the final outcome for drowning after adjusting for various risk factors (
Table 2). In deaths resulting from childhood injuries, sex differences have also been confirmed. In children aged <5 years, the male to female ratio is similar, but in children aged ≥5 years, the number of deaths among boys is approximately three times higher than that among girls [
18]. Drowning-related injuries were also prevalent in boys, which could be related to boys' high activity levels and proactiveness. Gender differences relating to styles of play have been reported. These differences in playing styles result in play polarization between boys and girls aged 10 to 15 years, leading to differences in activity characteristics and areas of play [
19].
A longer time for a patient to arrive at the ED is considered to be a poor prognostic factor for pediatric drowning. However, in this study, the univariate analysis showed an OR of 1.025 (95% CI, 1.016–1.035; P<0.001) in relation to this factor, whereas the multivariate analysis did not show statistical significance. The time between a patient falling into water and being reported at the time of patient transfer and the time from being reported to being rescued were not in our data; therefore, this factor was analyzed as the time from reporting to arrival. In Korea, many hospitals are available in close proximity, and the national EMS system is well-equipped [
20], so delays in transfer are rare. We considered that the time taken for a drowning patient to be transferred to the ED had little effect on prognosis. Drowning at an indoor location was associated with a good final prognosis, even after adjusting for age. This might be because the time to discovery and rescue from drowning that occurs indoors is shorter than that involved in cases of drowning that occur outdoors. Medical conditions such as epilepsy are also considered to increase the risk of drowning [
21].
This study had some limitations. The lack of precise data concerning the drowning time is an important factor affecting prognosis according to previous studies. No data were available concerning the patients’ drowning occurrence times, drowning report times, or EMS arrival times in this study. However, we obtained data concerning the patients’ EMS report times and arrival times at the ED. In the study data, prehospital cardiac arrest was defined as out-of-hospital cardiac arrest or prehospital death. Among patients with out-of-hospital cardiac arrest, data concerning return of spontaneous circulation status before and after transport were not included.
This is the first study to analyze the association between the presence of prehospital cardiac arrest and the volume capacity of the hospital where treatment was undertaken and the prognosis of pediatric drowning patients.
NOTES
-
Author contributions
Conceptualization: HC, JHC; Data curation: HC, SHL; Formal analysis: HC, SHL; Methodology: HC, SHL; Visualization: all authors; Writing–original draft: HC, SHL; Writing–review & editing: all authors. All authors read and approved the final manuscript.
-
Conflicts of interest
Jun Hwi Cho is an editorial board member of this journal, but was not involved in the peer reviewer selection, evaluation, or decision process of this article. The authors have no other conflicts of interest to declare.
-
Funding
This study was supported by a research grant from Kangwon National University School of Medicine in 2024.
-
Data availability
Data analyzed in this study are available from the corresponding author upon reasonable request.
Supplementary materials
Fig. 1.Study flowchart. KCBSTS, Korean Community-based Severe Trauma Survey.
Fig. 2.Prognosis of pediatric drowning patients according to age.
Fig. 3.Prognosis of pediatric drowning patients according to the occurrence of prehospital cardiac arrest. ED, emergency department; CPR, cardiopulmonary resuscitation; ICU, intensive care unit.
Table 1.Characteristics of the study population
Table 1.
|
Characteristic |
Total (n=406) |
Good outcome (n=166) |
Poor outcome (n=240) |
P-value |
|
Age (yr) |
8.8±6.8 |
7.2±7.0 |
9.9±6.5 |
0.299 |
|
Sex |
|
|
|
<0.001 |
|
Male |
279 (68.7) |
96 (57.8) |
183 (76.3) |
|
|
Female |
127 (31.3) |
70 (42.2) |
57 (23.7) |
|
|
Hospital type |
|
|
|
<0.001 |
|
Regional trauma center |
48 (11.8) |
25 (15.1) |
23 (9.6) |
|
|
Regional emergency medical center |
78 (19.2) |
43 (25.9) |
34 (14.2) |
|
Local emergency medical center |
176 (43.4) |
77 (46.4) |
99 (41.4) |
|
Local emergency medical institution |
104 (25.6) |
21 (12.6) |
83 (34.8) |
|
Location of drowning (outdoor) |
225 (55.4) |
65 (39.2) |
160 (66.7) |
<0.001 |
|
Prehospital cardiac arrest |
188 (46.3) |
5 (3.0) |
183 (76.3) |
<0.001 |
|
Intention of drowning |
|
|
|
<0.001 |
|
Unintentional |
325 (80.0) |
144 (86.4) |
181 (75.7) |
|
|
Suicidal |
39 (10.6) |
19 (11.4) |
20 (8.3) |
|
Violence |
2 (0.5) |
0 (0) |
2 (0.8) |
|
GOS score before injury |
|
|
|
0.147 |
|
1 (Death) |
0 (0) |
0 (0) |
0 (0) |
|
|
2 (Vegetative state) |
0 (0) |
0 (0) |
0 (0) |
|
3 (Severe disability) |
0 (0) |
0 (0) |
0 (0) |
|
4 (Moderate disability) |
3 (0.7) |
0 (0) |
3 (1.3) |
|
No disabilities mentioned (healthy) |
403 (99.3) |
166 (100) |
237 (98.7) |
|
GOS score at hospital dischargea)
|
|
|
|
<0.001 |
|
1 (Death) |
156 (38.5) |
0 (0) |
156 (65.3) |
|
|
2 (Vegetative state) |
13 (3.2) |
0 (0) |
13 (5.4) |
|
3 (Severe disability) |
37 (9.1) |
0 (0) |
37 (15.5) |
|
4 (Moderate disability) |
33 (8.1) |
0 (0) |
33 (13.8) |
|
5 (Good recovery) |
166 (41.0) |
166 (100) |
0 (0) |
|
Prehospital time (min) |
40 (2–1,214) |
30 (8–92) |
45 (2–1,214) |
0.001 |
Table 2.Logistic regression analysis for good recovery at hospital discharge in pediatric drowning patients
Table 2.
|
Variable |
Univariable analysis
|
Multivariable analysis
|
|
OR |
95% CI |
P-value |
aOR |
95% CI |
P-value |
|
Age (yr) |
0.942 |
0.914–0.971 |
<0.001 |
1.054 |
0.984–1.128 |
0.131 |
|
Female sex |
2.328 |
1.517–3.573 |
<0.001 |
0.950 |
0.489–1.845 |
0.879 |
|
High-volume hospital |
3.674 |
2.164–6.238 |
<0.001 |
2.498 |
1.069–5.834 |
0.034 |
|
Prehospital time (min) |
0.975 |
0.966–0.985 |
<0.001 |
0.999 |
0.987–1.012 |
0.932 |
|
Intent (unintentional) |
1.086 |
0.566–2.083 |
0.805 |
1.712 |
0.536–5.471 |
0.364 |
|
Location (indoor) |
3.257 |
2.114–5.018 |
<0.001 |
4.025 |
1.743–9.297 |
0.001 |
|
No prehospital cardiac arrest |
102.8 |
40.2–262.8 |
<0.001 |
98.7 |
32.9–295.8 |
<0.001 |
Table 3.Logistic regression analysis for good recovery at hospital discharge in pediatric drowning patients without prehospital cardiac arrest
Table 3.
|
Variable |
Univariable analysis
|
Multivariable analysis
|
|
OR |
95% CI |
P-value |
aOR |
95% CI |
P-value |
|
Age (yr) |
1.020 |
0.955–1.062 |
0.580 |
0.956 |
0.890–1.040 |
0.244 |
|
Female sex |
1.435 |
0.754–2.696 |
0.280 |
1.077 |
0.542–2.139 |
0.821 |
|
High-volume hospital |
2.215 |
1.125–4.874 |
0.043 |
2.868 |
1.005–5.978 |
0.047 |
|
Prehospital time (min) |
0.998 |
0.980–1.015 |
0.748 |
0.992 |
0.968–1.021 |
0.465 |
|
Intent (unintentional) |
1.810 |
0.587–5.588 |
0.280 |
2.312 |
0.602–8.892 |
0.221 |
|
Location (indoor) |
1.980 |
1.045–3.766 |
0.025 |
3.215 |
1.225–7.884 |
0.008 |
Table 4.Final outcomes in pediatric drowning patients depending on the hospital volume
Table 4.
|
Glasgow Outcome Scale score |
No. of patients (%)
|
P-value |
|
Low-volume hospital (n=104) |
High-volume hospital (n=301) |
|
|
Without prehospital cardiac arrest |
35 (33.7) |
183 (60.8) |
|
|
5 (Good recovery) |
21 (60.0) |
140 (76.5) |
0.004 |
|
1–4 (Death and sequelae) |
14 (40.0) |
43 (23.5) |
<0.001 |
|
1 (Death) |
0 (0) |
1 (0.5) |
0.234 |
|
With prehospital cardiac arrest |
69 (66.3) |
118 (39.2) |
|
|
5 (Good recovery) |
0 (0) |
5 (4.2) |
0.083 |
|
1–4 (Death and sequelae) |
69 (100) |
113 (95.8) |
<0.001 |
|
1 (Death) |
59 (85.5) |
96 (81.4) |
0.389 |
REFERENCES
- 1. Brenner RA. Childhood drowning is a global concern. BMJ 2002;324:1049-50.
- 2. Park CH, Yang DH, Kim JW, et al. Clinical practice guideline for endoscopic resection of early gastrointestinal cancer. Korean J Gastroenterol 2020;75:264-91.
- 3. Hong J, Lee WK, Park H. Change in causes of injury-related deaths in South Korea, 1996-2006. J Epidemiol 2011;21:500-6.
- 4. Wallis BA, Watt K, Franklin RC, Nixon JW, Kimble RM. Drowning mortality and morbidity rates in children and adolescents 0-19 yrs: a population-based study in Queensland, Australia. PLoS One 2015;10:e0117948.
- 5. Simons A, Govender R, Saunders CJ, Singh-Adriaanse R, Van Niekerk A. Childhood vulnerability to drowning in the Western Cape, South Africa: risk differences across age and sex. Child Care Health Dev 2020;46:607-16.
- 6. Hossain M, Mani KK, Sidik SM, Hayati KS, Rahman AK. Socio-demographic, environmental and caring risk factors for childhood drowning deaths in Bangladesh. BMC Pediatr 2015;15:114.
- 7. Peden MM, McGee K. The epidemiology of drowning worldwide. Inj Control Saf Promot 2003;10:195-9.
- 8. Salomez F, Vincent JL. Drowning: a review of epidemiology, pathophysiology, treatment and prevention. Resuscitation 2004;63:261-8.
- 9. Vahatalo R, Lunetta P, Olkkola KT, Suominen PK. Drowning in children: Utstein style reporting and outcome. Acta Anaesthesiol Scand 2014;58:604-10.
- 10. Armstrong EJ, Erskine KL. Investigation of drowning deaths: a practical review. Acad Forensic Pathol 2018;8:8-43.
- 11. Seo W. Health system factors associated with community-level incidence of self-harm and suicide: using 「Community-based Severe Trauma Surveillance」 data in 2016-2020 [Dissertation]. Seoul National University; 2024.
- 12. Hwang J, Lee H. Influencing factors of the mortality in patients with major trauma: 2016-2022 Community-based Severe Trauma Survey in Korea. J Korea Acad Ind Coop Soc 2024;25:564-71.
- 13. Im J, Seo EW, Jung K, Kwon J. Understanding regional trauma centers and managing a trauma care system in South Korea: a systematic review. Ann Surg Treat Res 2023;104:61-70.
- 14. Sampalis JS, Denis R, Frechette P, Brown R, Fleiszer D, Mulder D. Direct transport to tertiary trauma centers versus transfer from lower level facilities: impact on mortality and morbidity among patients with major trauma. J Trauma 1997;43:288-96.
- 15. Zuckerman GB, Conway EE. Drowning and near drowning: a pediatric epidemic. Pediatr Ann 2000;29:360-6.
- 16. Kim KS, Joung JI, Han MK, et al. Clinical aspects and prognostic factors of near drowning children. J Korean Child Neurol Soc 2003;11:322-7.
- 17. Yang L, Nong QQ, Li CL, Feng QM, Lo SK. Risk factors for childhood drowning in rural regions of a developing country: a case-control study. Inj Prev 2007;13:178-82.
- 18. World Health Organization. World health statistics 2008. World Health Organization; 2008.
- 19. You JY, Lee JI, Ryu JY. A comparison of characteristics in pediatric trauma patients under 7 years. J Korean Soc Traumatol 2004;17:197-205.
- 20. Jang K, Kang K, Jang Y, Hahn K. Analysis of the time intervals in 119 ambulance services. Fire Sci Eng 2016;30:128-34.
- 21. Franklin RC, Pearn JH, Peden AE. Drowning fatalities in childhood: the role of pre-existing medical conditions. Arch Dis Child 2017;102:888-93.
