Pediatrics

A reappraisal of childhood drowning in a pediatric emergency department

a b s t r a c t

Background: In the present study, we aimed to investigate the demographic and clinical features, laboratory and radiologic characteristics, management, and outcomes of pediatric drowning patients in order to identify predic- tors of hospital admission, and to evaluate the need for respiratory support, and prognosis.

Methods: In this retrospective chart review, children aged 0 to 18 years who presented to the pediatric emergency department due to drowning between July 2009 and September 2019 were included. Demographics, initial vital signs, clinical findings, laboratory and radiologic results, and the need for respiratory support or cardiopulmonary resuscitation in the emergency department were recorded. Subjects were divided into 6 groups using the Szpilman classification system.

Results: A total of 89 patients were enrolled. Among the children who were admitted to the hospital, initial Szpilman score, crepitations on lung auscultation, and pathologic Chest x-ray findings were higher and Glasgow Coma Score and oxygen saturation (SpO 2) levels were lower than those of children who were discharged from the emergency department. A Szpilman score of >=4, a lactate level of >2 mmol/L, and pathologic CXR findings were identified as predictors of hospital admission. Of the 89 patients, 22 (24.7%) underwent non- invasive ventilation (NIV) treatment and were classified as grade 3 or 4 according to the Szpilman score. Length of stay in the pediatric intensive care unit (PICU) and in the hospital was lower in patients who underwent NIV. As the Szpilman score increased as of grade 3, a positive correlation was observed with lactate levels (p <0.001, r: 0.552) and the total length of stay in the hospital (p: 0.001, r : 0.491), both of which gradually increased.

Conclusion: The Szpilman score was associated with the duration of hospital stay and the degree of hypoxia, so it could help the physician make rapid decisions on ventilation strategy. Application of NIV in the emergency de- partment shortened the length of stay in the PICU and in the hospital, suggesting that it can be used more often in pediatric emergency settings.

(C) 2020

  1. Introduction

Drowning is still a relevant medical challenge and a leading cause of Accidental death, constituting an important healthcare, social, and eco- nomic burden worldwide. According to the World Health Organization (WHO), approximately 370.000 drowning deaths occur per annum globally. Children, males, and Low- and middle-income countries are

Abbreviations: WHO, World Health Organization; CPR, cardiopulmonary resuscitation; GCS, Glasgow Coma Scale; CXR, chest X-ray; ICD, International Classification of Diseases; NIV, non-invasive ventilation; HFNC, high-flow nasal cannula; BiPAP, bilevel positive airway pressure; IMV, invasive mechanic ventilation; SD, standard deviation; IQR, interquartile range; SpO2, oxygen saturation; OR, odds ratio; ALT, alanine aminotransferase; AST, aspartate aminotransferase; PICU, pediatric intensive care unit.

? This study was presented as an oral presentation in XVI. Pediatric Emergency

Medicine and Intensive Care Congress in 2019, in Turkey.

* Corresponding author.

E-mail address: [email protected] (M. Duman).

disproportionally affected by drowning. More than 90% of deaths due to drowning occur in low- and middle-income countries worldwide [1-3]. Over half of drowning deaths occur in the pediatric population; it is the leading cause of unintentional injury deaths in children 1-4 years of age and the second leading cause of unintentional injury deaths in children aged 5 to 14 years in the United States, with a mortal- ity rate of 3 per 100.000 events [4]. Global age-standardised drowning mortality rates decreased 57.4%, from 9.3 (8.5 to 10.0) in 1990 to

4.0 (3.8 to 4.1) per 100,000 per annum in 2017 due to Prevention efforts and early, adequate, on-scene, bystander resuscitation, but it still remains a significant preventable cause of pediatric morbidity and mortality [5,6].

According to the current definition proposed by the WHO in 2002, “Drowning is the process of experiencing respiratory impairment from submersion/immersion in liquid” [1]. This definition highlights the role of acute respiratory failure in the pathophysiology. neurological complications caused by hypoxia are the most important prognostic

https://doi.org/10.1016/j.ajem.2020.12.059

0735-6757/(C) 2020

factors to predict outcomes. Risk factors such as duration of submersion, the interval between rescue and cardiopulmonary resuscitation (CPR), successful CPR at the scene, the duration of CPR, mental status, sponta- neous breathing, and circulation on arrival to the emergency depart- ment are known to be associated with neurologic outcomes [7]. Thus, preventing hypoxia is crucial.

There have been a few Classification systems proposed to guide in- terventions for the management of the drowning patient, predict the prognosis, and stratify the mortality. An algorithm was proposed by Semple-Hess and Campwala for pediatric drowning patients based on respiratory distress, hypothermia, and the Glasgow Coma Scale score [8]. In 1997, Szpilman developed a classification system based on the level of consciousness, lung auscultation findings, and hemody- namic status of the patient [9,10]. Nevertheless, there are limited pedi- atric data using these algorithms to evaluate drowning patients.

There are also limited data addressing the discharge of asymptom- atic or mildly symptomatic patients from the emergency department. Previous research advocated routine admission of all drowning patients with the concern that initially stable subjects may deteriorate with signs of respiratory distress even several hours after the event [11,12]. Recent studies demonstrated that it is reasonable to discharge a drowning pa- tient who initially presents with normal mental status in whom respira- tory function is normalized if there is no deterioration of respiratory symptoms after 4-6 h of observation in the emergency department [10,13,14]. It was reported that the initial chest x-ray results did not correlate with arterial blood gas measurements or outcomes but were helpful in tracking changes in the clinical status of drowning patients [14]. Routine use of complete blood count or electrolytes is not recommended. Arterial blood gas was found to be indicated to help guide respiratory resuscitation in drowning patients with evidence of hypoxemia or respiratory distress. For patients whose mental status fails to respond to resuscitation or in whom the initial cause of drown- ing remained uncertain, laboratory testing is recommended [15].

In this study, we aimed to investigate the demographic data, clinical

features, laboratory and radiologic characteristics, management, and outcomes of pediatric drowning patients in order to identify predictors of hospital admission, and to evaluate the need for respiratory support, and prognosis using the Szpilman classification system in a pediatric emergency department.

  1. Materials and methods
    1. Study design

This was a single-center retrospective chart review performed in the pediatric emergency department of a tertiary hospital with approxi- mately 100.000 pediatric emergency department admissions per annum in Izmir, a seaside city in the west of Turkey. The study was ap- proved by the Institutional Review Board of the Dokuz Eylul University Faculty of Medicine.

Children aged 0 to 18 years who were arrived at the pediatric emer- gency department by ambulance or self-transport or referred from other medical facilities due to unintentional drowning between July 2009 and September 2019 were included. We used International Classi- fication of Diseases (ICD) codes (Y21, V90, V92, W65-74, T71 and T75.1) for drowning, Near drowning, immersion, and Submersion injuries to identify patients. Intentional drownings were excluded from the study. We obtained information from a computer database, electronic medical records, medical charts, and nursing records. All data were in- vestigated and recorded by a pediatric emergency fellow and a pediatric resident.

Demographics, the site of drowning, and the time, day (weekend or not), and month of the event were recorded. Site of drowning was cat- egorized into two groups according to the water content as fresh water (events taking place in swimming pools, bathtubs, rivers, or canals) or salt water (events that took place at the sea). Time of day of drowning

was recorded as early morning (12:01 am-6:00 am), morning (6:01 am-12:00 pm), afternoon (12:01 pm-6:00 pm), and evening (6:01 pm-12:00 am). Season of the event was coded as spring (March, April, May), summer (June, July, August), autumn (September, October, November), or winter (December, January, February). The need for respiratory support or CPR at the scene, initial vital signs, clin- ical findings, results of laboratory and radiologic investigations, and the need for respiratory support or CPR in the emergency department were recorded. Subjects were divided into 6 groups using the Szpilman clas- sification system as follows: grade 1, normal pulmonary auscultation with coughing; grade 2, abnormal pulmonary auscultation with rales in some pulmonary fields; grade 3, pulmonary auscultation of acute pul- monary edema without arterial hypotension; grade 4, pulmonary aus- cultation of acute pulmonary edema with arterial hypotension; grade 5, isolated respiratory arrest; and grade 6, cardiopulmonary arrest [9]. Variables were ascertained upon arrival to the emergency department. The need for respiratory support was categorized as non-invasive venti- lation (NIV; either high-flow nasal cannula [HFNC] oxygen therapy or bilevel positive airway pressure [BiPAP] therapy] or invasive mechanic ventilation (IMV) and it was applicable to the emergency department. Finally, admission to the ward or pediatric intensive care unit (PICU), total duration time of stay in the hospital, prognosis, and mortality were recorded. The outcome was classified as poor in the event of death or severe encephalopathy.

    1. Statistical analysis

All statistical analyses were performed using SPSS software version

22.0 for Windows. Categorical and continuous variables were reported as frequencies and percentiles and as means with standard deviations (SDs) or medians with interquartile ranges (IQRs). The Mann- Whitney U test was used to compare non-parametric variables and Student’s t-test and one-way analysis of variance (ANOVA) were used for parametric data. Correlations were assessed with Spearman’s rank correlation coefficient. Multivariable analysis was performed using lo- gistic regression to determine predictors of hospital admissions. Values of p < 0.05 were considered statistically significant.

  1. Results

A total of 89 patients were enrolled during the study period. The me- dian age was 6.0 years (IQR: 4.0-12.0). Of the patients, 62 (69.7%) were male and 27 (30.3%) were female; the male-to-female ratio was 2.58:1. Thirty-one (34.8%) of them were aged between 0 and 4 years. Thirty- seven (41.5%) drownings occurred in fresh and 52 (58.5%) in salt water. Most drowning events (n:67, 75.2%) occurred in the summer with a peak incidence in July (n:26, 29.2%), and about half (n:50, 56.2%) occurred on weekend days. Afternoon was the most common time interval of the day; 60 (67.4%) patients experienced drowning be- tween 12:01 pm and 6:00 pm. According to the Szpilman classification, 33 (37.1%) patients were evaluated as grade 1, 9 (10.1%) as grade 2, 28 (31.5%) as grade 3, 6 (6.4%) as grade 4, 11 (12.4%) as grade 5, and 2 (2.2%) as grade 6. Five patients (5.6%) were intubated before arrival and 7 patients (7.8%) were intubated after arrival to the emergency de- partment. Fourteen patients (15.7%) received CPR at any time between rescue (extrication from the water) to hospital arrival. Of them, 2 pa- tients (2.2%) arrived to the emergency department while CPR was being performed and it was continued in the hospital. One of these two patients was intubated and the other underwent bag-mask ventila- tion. No alcohol use was observed, but there was Substance misuse in 1 (1.1%) case. A total of 5 (5.6%) patients had chronic illnesses, 2 (2.2%) of them with epilepsy and 3 (3.3%) with developMental disorders. Of the 89 patients, blood tests were obtained from 83 (93.2%) and CXR were obtained from 86 (96.6%) of them. Antibiotic treatment was started for 5 (5.6%) patients who had clinical and radiologic findings of pneu- monia; among them, 2 drownings had occurred in a polluted canal. Of

the children, 54 (60.6%) were discharged from the emergency depart- ment while 31 (34.8%) were admitted to the hospital, and among the latter, 21 (23.6%) were admitted to the PICU. Four (4.4%) patients were discharged with severe complications (with neurological and re- spiratory complications) and 4 (4.4%) died. Thus, poor outcome was identified in 8 (8.9%) cases.

When we evaluated hospital admissions, the initial Szpilman score,

crepitations on lung auscultation, and pathologic CXR findings were higher (p < 0.001) and the Glasgow Coma Score (p: 0.001) and oxygen saturation (SpO2) (p < 0.001) levels were lower in children who were admitted to the hospital compared with subjects discharged from the emergency department. The rate of patients who received CPR was higher in the group admitted to the hospital (p < 0.001). The admitted children also had higher Serum glucose, alanine aminotransferase (ALT), and lactate levels in blood gas analyses (p < 0.05) and lower pH and bicarbonate measurements (p < 0.01) (Table 1). A Szpilman score of >=4 [odds ratio (OR) = 12.109, 95% CI: 2.327-63.010, p: 0.003],

Table 2

Multivariable analysis to predict hospital admissions

Variable

Odds Ratio

95% Confidence Interval

p

Glasgow Coma Scale Score < 14

0.274

0.039-1.917

0.192

Szpilman score >=4

12.109

2.327-63.010

0.003

Lung crepitation

0.901

0.111-7.308

0.922

pH<7.30

0.454

0.076-2.719

0.387

Lactate >2 mmol/L

4.390

1.365-14.121

0.013

Abnormal CXR

19.500

3.761-101.112

<0.001

CXR: chest X-ray. While evaluating hospital admissions, the 4 children who had died in the emergency department were excluded.

Table 3

Comparison of NIV (+) and NIV (-) patients.

Variable NIV (+) (n: 22) NIV (-) (n: 54) p value

Male gender, n (%) 18 (81.8) 37 (64.9) 0.155

a lactate level of >2 mmol/L (OR = 4.390, 95% CI: 1.365-14.121, p: 0.013), and pathologic CXR findings (OR = 19.500, 95% CI: 3.761-101.112, p < 0.001) were identified as predictors of hospital ad- missions (Table 2). While comparing hospital admission, we excluded

Age in years, median (IQR) 7.00

H R

(4.00-10.50)

(4.00-13.25)

eart rate (beats/min), mean+-SD 124.00+-24.03 114.87+-20.24 0.187

espiratory rate (breaths/min), 42.00+-11.2042 31.00+-7.61,0 0.001

mean+-SD

7.00

0.787

the 4 children who had died in the emergency department to prevent confusion among the findings of the children who were discharged from the emergency department.

Of the 89 patients, 22 (24.7%) underwent NIV and were classified as grade 3 or 4 according to the Szpilman score. Twenty (22.4%) of them received BiPAP and 2 (2.3%) received HFNC oxygen therapy. Applica- tions of NIV were started between 1 and 6 h after arrival to the pediatric emergency department. Of the 22 patients, 10 (45.4%) were admitted to the PICU, and 12 (54.6%) of them were observed in the pediatric emer- gency department and then discharged. Of all NIV applications, no pa- tient deteriorated to IMV. After excluding children who received IMV, the patients who were treated with NIV had higher respiratory rate (p: 0.001), Szpilman score (p < 0.001), lung crepitation (p < 0.001), pathologic CXR findings (p < 0.001), blood glucose (p < 0.05), urea (p

< 0.05), and aspartate aminotransferase (p < 0.05) and lower

Glasgow Coma Scale Score, mean

+-SD

SpO2, mean+-SD

91.10+-4.80

94.20+-6.50

0.001

Lung crepitation, n (%)

22 (100)

27 (46.5)

<0.001

Szpilman score, mean+-SD

3.20+-0.40

1.90+-1.20

<0.001

Abnormal CXR, n (%)

15 (68.1)

16 (29.6)

<0.001

pH, mean+-SD

7.28+-0.07

7.32+-0.08

0.004

Lactate (mmol/L), mean+-SD

2.70+-1.95

2.39+-1.20

0.995

Bicarbonate (mmol/L), mean+-SD

19.76+-2.61

22.39+-13.62

0.303

Glucose (mg/dL), mean+-SD

136.00+-37.50

120.00+-31.90

0.025

Sodium (mmol/L), mean+-SD

141.36+-5.91

141.32+-6.03

0.635

Urea (mg/dL), mean+-SD

12.90+-1.90

11.70+-3.1

0.032

Creatinine (mg/dL), mean+-SD

0.49+-0.21

0.55+-0.17

0.149

AST (U/L), median (IQR)

36.00

30.00

0.041

(29.00-70.70)

(23.00-39.00)

ALT (U/L), median (IQR)

20.00

16.00

0.220

(13.00-36.25)

(14.00-19.00)

Length of stay in PICU (hours), mean

23.90+-4.60

52.00+-15.00

0.026

+-SD

length of stay in hospital (hours),

48.00

54.00

0.001

median (IQR)

(35.00-56.00)

(46.00-96.00)

14.45+-0.91 14.49+-1.46 0.154

Table 1

Comparison of patients who were discharged from the emergency department and those who were admitted to the hospital.

Seven patients were intubated before arrival and 5 patients were intubated after arrival to the emergency department, and one patient underwent Bag-mask ventilation, so all these patients were excluded. NIV: Non-invasive ventilation, SD: standard deviation, IQR: inter-

quartile range, SpO2: oxygen saturation, CXR: chest X-ray, AST: aspartate aminotransfer-

ase, ALT: alanine aminotransferase, CPR: cardiopulmonary resuscitation, PICU: pediatric intensive care unit.

Variable

Discharged

Admitted

p value

(n: 54)

(n: 31)

Male gender, n (%)

37 (68.5)

22 (70.9)

0.380

Age in years, median (IQR)

8.00

5.5 (3.39-11.25)

0.301

(4.00-12.00)

Heart rate (beats/min), mean+-SD

117.80+-20.41

119.74+-25.73

0.890

Respiratory rate (breaths/min),

33.60+-9.20

36.25+-11.61

0.398

mean+-SD

Glasgow Coma Scale Score, mean

14.80+-0.70

12.6+-3.20

0.001

+-SD

SpO2, mean+-SD

94.20+-5.70

91.30+-6.90

<0.001

Lung crepitation, n (%)

26 (48.1)

28 (90.3)

<0.001

Szpilman score, mean+-SD

1.80+-1.00

3.41+-1.23

<0.001

Abnormal CXR, n (%)

10 (18.5)

26 (83.8)

<0.001

pH, mean+-SD

7.33+-0.06

7.25+-1.00

0.004

Lactate (mmol/L), mean+-SD

2.14+-1.30

3.38+-1.80

0.001

Bicarbonate (mmol/L), mean+-SD

22.78+-13.90

18.83+-3.00

0.008

Glucose (mg/dL), mean+-SD

139.00+-61.70

154.70+-55.30

<0.001

Sodium (mmol/L), mean+-SD

141.07+-5.70

141.45+-6.41

0.759

Urea (mg/dL), mean+-SD

11.94+-2.90

12.74+-3.41

0.318

Creatinine (mg/dL), mean+-SD

0.50+-0.17

0.57+-0.19

0.141

AST (U/L), mean+-SD

31.00

37.00

0.129

(24.00-39.00)

(24.00-71.00)

Received CPR (n, %)

9 (29.0%)

1 (1.8%)

<0.001

SpO2 (p: 0.001) and pH (p < 0.05) measurements than those treated without NIV. Additionally, total length of stay in the PICU and in the hos- pital was shorter in patients who underwent NIV treatment (p: 0.026, p: 0.001) (Table 3). The Szpilman score had positive correlations with heart rate (p: 0.039, r: 0.239), respiratory rate (p < 0.001, r: 0.477),

blood glucose (p < 0.001, r: 0.534), urea (p: 0.008, r: 0.289), AST

(p: 0.002, r: 0.349), ALT (p: 0.002, r: 0.339), and lactate (p < 0.001, r:

0.552) levels and negative correlations with SpO2 (p < 0.001, r:

-0.483), body temperature (p: 0.009, r: -0.309), GCS score (p < 0.001, r: -0.758), and pH levels (p < 0.001, r: -0.613). As the Szpilman score increased as of grade 3, a positive correlation was observed with lactate levels (p < 0.001, r: 0.552) and the total length of stay in the hos- pital (p: 0.001, r: 0.491), both of which gradually increased (Figs. 1 and 2).

While comparing hospital admissions, the 4 children who had died in the emergency department were excluded. SD: Standard deviation, IQR: interquartile range, SpO2: oxygen saturation, CXR: chest X-ray, AST: aspartate aminotransferase, ALT: alanine aminotransferase.

Evaluating the 8 patients with poor outcomes, they had lower body temperature (p: 0.015), Glasgow Coma Score (p < 0.001), pH (p: 0.012), and bicarbonate (p: 0.016) levels and higher Szpilman score (p < 0.001), AST (p: 0.009), ALT (p: 0.011), and lactate (p: 0.003) levels, with longer duration time of CPR (p: 0.03). The rate of patients who

Fig. 1. Mean length of stay in the hospital according to the Szpilman classification.

160

140

120

100

80

60

40

20

0

Grade 1

Grade 2

Grade 3

Szpilman score

Grade 4

Grade 5

Length of stay (hours)

received CPR was also higher in the group with poor outcomes (p < 0.001) (Table 4).

  1. Discussion

Drowning still remains an important cause of mortality. Loux et al. evaluated admissions to a pediatric trauma center and reported that the risks of mortality, severe outcome, ventilator assistance, and PICU stay were higher for drowning versus other causes of trauma [16]. Al- though age distribution was not discriminative, drownings occurred more commonly in males and in summer, with a peak incidence of July in our study, in accordance with the literature [2,3,17-19].

Although some authors recommended routine hospital admission for all drowning patients in case of deterioration, subsequent studies have demonstrated that discharge is possible for a drowning patient who initially presents with normal mental status and if respiratory func- tion is normalized, and if there is no deterioration of respiratory symp- toms after 4-6 h of observation in the emergency department [10-14]. Cohen et al. assessed asymptomatic to moderately symptomatic pediat- ric drowning patients and found that respiratory distress and lung

Lactate (mmol/L)

crepitations were independent factors of hospital admission; there were no admissions of children who were discharged from the emer- gency department after an observation period [20]. In our study, a Szpilman score of >=4, a lactate level of >2 mmol/L, and pathologic CXR findings were valuable for predicting hospital admissions. In another pediatric study, Brennan et al. evaluated predictors for safe discharge and found that children with abnormal results for body temperature, SpO2, respiratory rate, cardiopulmonary physical examination, mental status, or pathologic CXR findings were more likely to require hospital admission, in accordance with our results. Additionally, patients with initial abnormal SpO2 levels were more likely to develop complications [21]. We also reported higher blood glucose and lower pH values in our hospitalized group. Meanwhile, patients who were admitted to the PICU had higher ALT levels in addition to all these parameters. The fact that we observed higher levels of AST, ALT, and lactate and lower pH values demonstrates the importance of tissue damage and hypoxia.

Evaluating the 8 patients who had poor outcomes in our study, they had lower body temperature, Glasgow Coma Score, pH, and bicarbonate levels and higher Szpilman score, AST, ALT, and lactate levels and longer duration time of CPR. A PICU trial reported that the need for advanced

Fig. 2. Mean lactate levels of the patients according to the Szpilman classification.

26

24

22

20

18

16

14

12

10

8

6

4

2

0

Grade 1

Grade 2

Grade 3

Grade 4

Grade 5

Grade 6

Szpilman score

Table 4

Comparison of patients with poor and good prognosis.

Variable Poor prognosis

(n: 8)

Male gender, n (%)

5 (62.5)

57 (70.4)

0.891

Age in years, median (IQR)

7.4 (5.0-8.7)

6.0 (4.0-12.0)

0.646

Heart rate (beats/min), mean+-SD

118.83+-24.77

118.98+-22.80

0.974

Respiratory rate (breaths/min),

38.00+-2.82

34.62+-10.40

0.486

mean+-SD

Glasgow Coma Scale Score, mean

+-SD

5.25+-0.46

14.35+-1.73

<0.001

SpO2, mean+-SD

85.50+-4.94

93.36+-6.28

0.071

Body temperature, mean+-SD

35.28+-1.07

36.38+-0.88

0.015

Szpilman score, mean+-SD

5.25+-0.46

2.27+-1.24

<0.001

pH, mean+-SD

7.00+-0.30

7.31+-0.08

0.012

Lactate (mmol/L), mean+-SD

11.35+-9.58

2.54+-1.59

0.003

Bicarbonate (mmol/L), mean+-SD

13.64+-8.17

21.38+-11.53

0.016

Glucose (mg/dL), mean+-SD

162.37+-69.16

128.58+-42.50

0.057

Sodium (mmol/L), mean+-SD

141.57+-7.34

141.41+-5.94

0.994

Urea (mg/dL), mean+-SD

14.83+-6.05

12.12+-2.74

0.169

Creatinine (mg/dL), mean+-SD

0.72+-0.33

0.53+-0.17

0.311

AST (U/L), median (IQR)

203.50

32.00

0.009

(50.50-904.25)

(24.00-40.00)

ALT (U/L), median (IQR)

121.50

17.00

0.011

(26.00-725.75)

(14.00-23.00)

Received CPR (n, %)

7 (87.5%)

7 (8.6%)

<0.001

Duration of CPR (min), median

55.00

3.00 (2.00-)

0.030

(IQR)

(15.00-75.00)

Good prognosis (n: 81)

p value

to the Szpilman classification. Of the 25 patients who underwent NIV, 4 were intubated due to respiratory or Neurological deterioration [27]. In our study, children were successfully treated by NIV, using HFNC or BiPAP, and none of them deteriorated enough to require IMV. The Szpilman classification was grade 3 or 4 for these patients. The decision to apply NIV was made on the basis of confidence that the early applica- tion of NIV could be a preventive method for reducing the morbidity of nonfatal cases with the reversible nature of drowning [28]. At this point, the Szpilman score is useful for the choice of ventilation strategy. Szpilman et al. proposed administering High-flow oxygen therapy and/or mechanical ventilation for patients of grades 3 and 4 [10]. The improvement of neurological status as well as the lower incidence of he- modynamic instability could suggest alternative treatment strategies such as NIV. The existence of neurological deterioration is an indication for invasive mechanic ventilation rather than NIV use. Gregorakos et al. evaluated drowning adults with acute respiratory failure in the emer- gency department and concluded that intubation could be avoided in non-comatose patients [29]. Dottorini et al. reported successful treat- ment with nasal CPAP in 2 drowning cases without loss of conscious- ness, the patients being 13 and 19 years old [30].

There were some limitations related to the retrospective nature of

our study. We used ICD codes to identify patients, but missing data may lead to the underestimation of the number of submersion injuries, as previously reported by Peden et al. [31]. Our information was ob-

SD: Standard deviation, IQR: interquartile range, SpO2: oxygen saturation, AST: aspartate aminotransferase, ALT: alanine aminotransferase, CPR: cardiopulmonary resuscitation.

CPR with epinephrine administration on the scene predicted poor out- come, and patients who had advanced CPR with epinephrine admission had lower Glasgow Coma Score, body temperature, pH, and bicarbonate and higher glucose levels at admission [22]. In an adult study, drowning-related cardiac arrest still remained the cornerstone of prog- nosis [23]. Another pediatric study found that immediate intervention after discovery was associated with approximately 70% lower odds of a severe outcome [16]. Submersion time was a well-known risk factor; in one review, it was considered as the strongest predictor for outcomes [24]. Unfortunately, we could not obtain sufficient information about submersion times in our study. There was no difference with the victim’s age, water temperature, or events being witnessed or unwitnessed; we also found no difference regarding age, but we had in- sufficient data about water temperature and whether the event was witnessed or not.

There are few pediatric studies using the Szpilman classification sys- tem for drowning patients. Son et al. evaluated 29 children who experi- enced drowning and found that the Szpilman score correlated with consciousness level and serum AST and ALT levels, and patients with poor prognosis had higher Szpilman scores, consistent with our findings [25]. They concluded that they could identify a more accurate prognosis by using the Szpilman classification system together with the level of consciousness. Szpilman declared hospital mortality of 19% for grades 2 to 6, compared with 7.1% in our study [9]. In another adult trial, the mortality was reported as 18.5% for the group of patients of Szpilman grades 2-6 [23]. This discrepancy may be related to the fact that chil- dren had far fewer comorbidities than adults, which could contribute to the worsening of the clinical condition and increased mortality. In- deed, children might have a rapid and better response to resuscitation attempts and treatment of hypoxia. In two pediatric studies, the mortal- ity was calculated as 35% and 9.5% for grades 2-6, higher than our rates [25,26]. However, the number of such patients was lower in those stud- ies, with 31 and 26 versus 56 children. Thus, studies with larger case se- ries are required to obtain more accurate pediatric mortality rates.

There are limited data about the safety and effectiveness of NIV use for drowning patients. Michelet et al. used NIV (continuous positive air- way pressure [CPAP] or BiPAP therapy) for drowning adults in intensive care units and declared that their patients were grades 3 to 5 according

tained from a computer database, electronic medical records, medical charts, and nursing records, but there were missing data. We did not know whether patients had previously learned to swim or not, and we found very little information about the characteristics of the res- cuers. Additionally, data about submersion time or time of the interval between the event and arrival to the emergency department were miss- ing, so we could not accurately evaluate their role in the clinical features or outcomes. Providing all these data could inform clinical practice and public health priorities such as health care planning to ensure availabil- ity of resources needed for acute care of drowning children. So, future work is needed.

  1. Conclusion

A Szpilman score of >=4, a lactate level of >2 mmol/L, and pathologic CXR findings were valuable for predicting hospital admissions in pediat- ric patients. The Szpilman score was associated with the length of stay in the hospital and the degree of hypoxia, so it could help the physician make rapid decisions about ventilation strategy. Application of NIV in the emergency department shortened the length of stay in the PICU and in the hospital, suggesting that it can be used more often in pediat- ric emergency settings.

Funding

This research did not receive any specific grant from funding agen- cies in the public, commercial, or not-for-profit sectors.

Declaration of Competing Interest

The authors declare no conflict of interest.

Acknowledgements

None.

References

  1. World Health Organization. Global report on drowning: Preventing a leading killer. Geneva: World Health Organization; 2014 Available from: https://www.who.int/ violence_injury_prevention/global_report_drowning/en/.
  2. Tyler MD, Richards DB, Reske-Nielsen C, Saghafi O, Morse EA, Carey R, et al. The ep- idemiology of drowning in low- and middle-income countries: a systematic review. BMC Public Health. 2017;17(1):413.
  3. Rahman A, Mashreky SR, Chowdhury SM, Giashuddin MS, Uhaa IJ, Shafinaz S, et al. Analysis of the childhood fatal drowning situation in Bangladesh: exploring preven- tion measures for low-income countries. Inj Prev. 2009;15(2):75-9.
  4. Centers for Disease Control and Prevention. Web-Based Injury Statistics Query and Reporting System (WISQARS). Atlanta: CDC; 2017 s http://www.cdc.gov/injury/ wisquars/leading_causes_death.html.
  5. Roth GA, Abate D, Abate KH, Abory SM, Abbafati C, Abbasi N, et al. Global, regional, and national age-sex-specific mortality for 282 causes of death in 195 countries and territories, 1980-2017: a systematic analysis for the global burden of disease study 2017. Lancet. 2018;392(10159):1736-88.
  6. Franklin RC, Peden AE, Hamilton EB, Bisignano C, Castle CD, Dingels ZV, et al. The burden of unintentional drowning: global, regional and national estimates of mor- tality from the global burden of disease 2017 study. Inj Prev. 2020 Oct;26(Supp 1):i83-95.
  7. Suominen PK, Vahatalo R. Neurologic long term outcome after drowning in children. Scand J Trauma Resusc Emerg Med. 2012;20:55.
  8. Semple-Hess J, Campwala R. Pediatric submersion injuries: emergency care and re- suscitation. Pediatr Emerg Med Pract. 2014;11(6):1-21.
  9. Szpilman D. Near-drowning and drowning classification: a proposal to stratify mor- tality based on the analysis of 1,831 cases. Chest. 1997;112(3):660-5.
  10. Szpilman D, Bierens JJ, Handley AJ, Orlowski JP. Drowning. N Engl J Med. 2012;366 (22):2102-10.
  11. Pearn JH. Secondary drowning in children. Br Med J. 1980;281(6248):1103-5.
  12. Pratt FD, Haynes BE. Incidence of “secondary drowning” after saltwater submersion. Ann Emerg Med. 1986;15(9):1084-7.
  13. Causey AL, Tilelli JA, Swanson ME. Predicting discharge in uncomplicated near- drowning. Am J Emerg Med. 2000;18(1):9-11.
  14. Noonan L, Howrey R, Ginsburg CM. Freshwater submersion injuries in children: a retrospective review of seventy-five hospitalized patients. Pediatrics. 1996;98(3 Pt 1):368-71.
  15. Schmidt AC, Sempsrott JR, Hawkins SC, Arastu AS, Cushing TA, Auerbach PS. Wilder- ness medical society clinical practice guidelines for the treatment and prevention of drowning: 2019 update. Wilderness Environ Med. 2019;30(4S):S70-86.
  16. Loux T, Mansuri F, Brooks SE, Slye N, Lewis B, Lu Y, et al. Factors associated with pe- diatric drowning admissions and outcomes at a trauma center, 2010-2017. Am J Emerg Med. 2021(39):86-91. https://doi.org/10.1016/j.ajem.2020.01.025.
  17. Chauvin M, Kosatsky T, Bilodeau-Bertrand M, Gamache P, Smargiassi A, Auger N. Hot weather and risk of drowning in children: opportunity for prevention. Prev Med. 2020;130:105885.
  18. Chang SSM, Ozanne-Smith J. Drowning mortality in children aged 0-14 years in Victoria, Australia: detailed Epidemiological study 2001-2016. Inj Prev. 2019;6 (26):593-8. https://doi.org/10.1136/injuryprev-2019-043307 (in press). pii: injuryprev-2019-043307.
  19. El Sibai R, Bachir R, El Sayed M. Submersion injuries in the United States: patients characteristics and predictors of mortality and morbidity. Injury. 2018;49(3):543-8.
  20. Cohen N, Scolnik D, Rimon A, Balla U, Glatstein M. Childhood drowning: review of patients presenting to the emergency departments of 2 large tertiary care pediatric hospitals near and distant from the sea coast. Pediatr Emerg Care. 2020;36(5): e258-62.
  21. Brennan CE, Hong TKF, Wang VJ. Predictors of safe discharge for pediatric drowning patients in the emergency department. Am J Emerg Med. 2018;36(9):1619-23.
  22. Salas Ballestin A, de Carlos Vicente JC, Frontera Juan G, Sharluyan Petrosyan A, Reina Ferragut CM, Gonzalez Calvar A, et al. Prognostic factors of children admitted to a pe- diatric intensive care unit after an episode of drowning. Pediatr Emerg Care. 2018. https://doi.org/10.1097/PEC.0000000000001554.
  23. Markarian T, Loundou A, Heyer V, Marimoutou C, Borghese L, Coulange M, et al. Drowning classification: a reappraisal of clinical presentation and prognosis for severe cases. Chest. 2020;158(2):596-602.
  24. Quan L, Bierens JJ, Lis R, Rowhani-Rahbar A, Morley P, Perkins GD. Predicting out- come of drowning at the scene: a systematic review and meta-analyses. Resuscita- tion. 2016;104:63-75.
  25. Son KL, Hwang SK, Choi HJ. Clinical features and prognostic factors in drowning chil- dren: a regional experience. Korean J Pediatr. 2016;59(5):212-7.
  26. Guzel A, Duran L, Paksu S, Akdemir HU, Paksu MS, Kati C, et al. Drowning and near- drowning: experience of a university hospital in the Black Sea region. Turk J Pediatr. 2013;55(6):620-7.
  27. Michelet P, Bouzana F, Charmensat O, Tiger F, Durand-Gasselin J, Hraiech S, et al. Acute respiratory failure after drowning: a retrospective multicenter survey. Eur J Emerg Med. 2017;24(4):295-300.
  28. Caglar A, Er A, Ozden O, Karaarslan U, Akgul F, Koroglu TF, et al. Efficacy of early non- invasive ventilation in three cases of nonfatal drowning with pulmonary oedema in the pediatric emergency department. Hong Kong J Emerg Med. 2017;23(2):42-6.
  29. Gregorakos L, Markou N, Psalida V, Kanakaki M, Alexopoulou A, Sotiriou E, et al. Near-drowning: clinical course of lung injury in adults. Lung. 2009;187(2):93-7.
  30. Dottorini M, Eslami A, Baglioni S, Fiorenzano G, Todisco T. Nasal-continuous positive airway pressure in the treatment of near-drowning in freshwater. Chest. 1996;110 (4):1122-4.
  31. Peden AE, Franklin RC, Mahony AJ, Scarr J, Barnsley PD. Using a retrospective cross- sectional study to analyse unintentional fatal drowning in Australia: ICD-10 code- based methodologies versus actual deaths. BMJ Open. 2017 Dec 21;7(12):e019407.