Chapter 117. Submersion Injury

Submersion injury is a major public health problem and accounts for more than half a million deaths annually worldwide.1 It is the second most common cause of unintentional death in the United States for children between ages 1 and 19 years (Fig. 117-1), second only to motor vehicle crashes. In 2002, there were 3447 unintentional submersion deaths in the United States, averaging 9 people per day.3 Of these, 1158 (34%) were children. It is estimated that for each submersion death, there are up to four children who receive emergency medical care for nonfatal submersion injuries. More than 40% of these children are hospitalized and 20% of the survivors suffer permanent disability.

Submersion injury is not a problem unique to the United States. In Scotland, the rate of submersions in 1999 was 1.6 per 100,000 persons per year, where most events occurred not in private swimming pools, but in cold waters such as the sea or a reservoir.4 In Canada, the overall, unintentional submersion rate declined from 21.7 per 100,000 persons per year in 1979 to 5.8 per 100,000 in 2002.5 Drowning is the third most common cause of death (13.5%) in children. Populations that appear to be at particularly high risk of submersion injury include toddlers ages 1 to 4 years (1.85 deaths per 100,000) and youth ages 15 to 19 years (1.28/100,000). Submersion injury is a particular problem in China, where it is estimated that more than one fifth of worldwide submersion deaths occur.6 In Guangxi province of China, the mortality rate for submersions in male and female children ages 1 to 4 years was 29.8 per 100,000 and 29.6 per 100,000, respectively.7 Submersion injury is also a serious public health problem in India, where a study in one region showed average unintentional submersion injury rate for ages 1 to 12 years was 25.9 per 100,000.8 There, young children were most likely to drown in vessels containing water, since there are no piped water supplies in most villages. In the 10- to 12-year age group, 80% of the submersion injuries were in large open-irrigation wells without protective walls.

Submersion injury risks vary by gender, age, race, and socioeconomic status. After the first year of age, US males are at greater risk of death from submersion than females (Fig. 117-1). Among females, submersion deaths peak at 1 year and decline thereafter. Male incidence patterns exhibit a bimodal distribution, with peaks in ages 1 to 4 years and in the adolescent age ranges. Between the ages of 10 to 19 years, African American males have a higher submersion injury rate than do Caucasian males.3

Within the pediatric population, submersion circumstances vary by age group. Children under age 1 year are most likely to drown in bathtubs, toilets, or other household water reservoirs such as large buckets.9 Recent literature describes the association with bathing seats or rings or cobathing with an older sibling.10-14 Since 1983, the US Consumer Product Safety Commission (UPSC) is aware of 66 deaths associated with bathtub seats or rings. These seats usually contain 3 or 4 legs with suction cups that attach to the bottom of the tub. The suction cups may suddenly release, allowing the seat or ring and baby to tip over. The baby can also slip between the legs of the seat or ring and become trapped under water. In most cases, bathtub submersions result from a lapse in adult supervision, although not infrequently they are the result of intentional injury.15-18

Industrial buckets, especially the 3- to 5-gallon buckets with a wide base, are an often-overlooked hazard to older infants and toddlers.19-22 Curious children may find themselves headfirst in the bucket, unable to push themselves out of the bucket or tip it over. It is reported that children can drown in as little as 5 cm of water.

Children ages 1 to 4 years are most likely to experience a submersion injury in swimming pools and hot tubs, whereas older children are most likely to drown in natural freshwater bodies such as rivers and lakes. Adolescents often drown in rivers, lakes, and canals. Adolescent males tend to participate in riskier recreational activities and to fall under the influence of alcohol and drugs that impair judgment.23-25

Numerous studies have shown that alcohol is associated with 25% to 50% of submersion events involving adolescents and adults in the United States.26,27 In 2000, 31% of all 701 recreational- boating deaths involved alcohol.28 The odds ratio of dying from submersion for current alcohol consumers was 3.64 (95% chemotherapeutic index [CI] = 2.23, 5.93) compared to those who did not drink alcohol.29 There is a strong positive association of blood alcohol content (BAC) with the relative risk of death among boaters ages 18 years and older even if the BAC is less than 50 mg/dL. In addition, passenger and operator drinking are associated with the same relative risk of death, regardless of whether the boat is docked or underway. In Finland, the association between alcohol and submersion injury is higher than in other countries.30 Alcohol is involved in 63.8% of nonboating-related and 77.8% of boating-related submersion victims ages 15 years and older. Alcohol can affect one’s safety by impairing balance and coordination, judgment and ability to keep the head above the water, and by increasing one’s risk of hypothermia.

In the United States, submersion injury rates for African American children are almost twice that for Caucasian children.31,32An exception to this trend is the 2-fold higher rate of submersion injury among Caucasian children ages 1 to 4 years. Also, Hispanic males ages 15 to 24 years were found to be 2 to 5 times more likely to drown than Caucasian males of the same age.

Submersion occurrences also exhibit temporal variations. Two thirds of children less than 15 years drown during the warmer months of May through August in the United States and November through January in Australia.33 Submersion injury occurs more frequently on the weekends and between noon and 8 pm.34 Across the United States, there are also regional differences in rates of submersion injury. In Harris County, Texas, the number of submersions in swimming pools exceeded those in natural bodies of freshwater for all age groups younger than 15 years.31 In Florida, children older than 11 years of age were most likely to drown in natural bodies of saltwater. Even in Hawaii, more children drown in swimming pools (41%) than in the surf (29%).35

Certain medical conditions, such as autism,36 inherited long QTc syndrome,37,38 or epilepsy,39 that may result in loss of consciousness while bathing or taking part in water-related activities predispose children to submersion injury. In a small California-based study, results suggest that children with autism may be at increased risk for submersions.36 Cardiac arrhythmias associated with inherited long QTc syndrome or subclinical myocarditis may be causes of unexplained submersions.37,38Epilepsy is a predisposing risk factor for submersion injury. Sixty percent of submersions involving epileptic patients occurred during an unsupervised bath at home.39 Therefore, older children with epilepsy are encouraged to take showers instead of baths. Because of the increased risk of submersion injury in the bath, older children with epilepsy are encouraged to take showers instead. In a study of submersion injury in children with epilepsy in King County, Washington, the relative risk of submersion injury was 18.7 in the bathtub, and 23.4 in the swimming pool.40 In a Finnish study, 19% of the deaths in epileptics resulted from submersion injury. Children with seizure disorders had a 4 times greater risk for submersion injury than that of the normal population.41 Similarly in the United Kingdom, children with seizures have a higher incidence of submersion injuries and all were reported to have occurred while bathing unsupervised.42

Historically, a number of definitions have been used to describe the submersion or drowning event, incorporating characteristics of the event, pathophysiology, and outcomes. The terminology was confusing and inconsistent. To solve this problem an international Utstein-style consensus conference convened in 2002 to develop guidelines for the definition and reporting of data related to submersion injury and drowning.2 The following definitions and recommendations were established. Drowning is “a process resulting in primary respiratory impairment from submersion or immersion in a liquid medium.” The drowning process is “the continuum that begins when the victim’s airway lies below the surface of the liquid, usually water,” until the time the victim is rescued and given appropriate resuscitative measures. At this point the drowning process is interrupted. In addition, the consensus recommended the following terms be abandoned: dry drowning; wet drowning; active, passive, or silent drowning; secondarydrowning; and even near-drowning. The word drowned will refer to a person who died from drowning. Immersion is to be covered in water; thus for drowning to occur, at least the face and airway are immersed. Submersion implies that the entire body, including the airway, is under water.2

A number of definitions have been used to describe the submersion or drowning event, incorporating characteristics of the event, pathophysiology, and outcomes. The terminology was confusing and inconsistent. Currently accepted definitions are that drowning is “a process resulting in primary respiratory impairment from submersion or immersion in a liquid medium.” The drowning process is “the continuum that begins when the victim’s airway lies below the surface of the liquid, usually water,” until the time the victim is rescued and given appropriate resuscitative measures. At this point the drowning process is interrupted. Terms including dry or wet drowning; active, passive or silent drowning were abandoned. The word drowned refers to a person who died from drowning. Immersion is to be covered in water; thus for drowning to occur, at least the face and airway are immersed. Submersion implies that the entire body, including the airway, is under water.2 A more detailed discussion of definitions is available electronically and in eTable 117.1 and eFigure 117.1.

In 1972, Menezes and Costa43 proposed a classification dividing the cases of submersion injury into 4 grades of severity: grade 1 (normal pulmonary auscultation), grade 2 (audible rales in both pulmonary bases), grade 3 (evidence of acute pulmonary edema), and grade 4 (presence of cardiopulmonary arrest). Mortality rates assigned to each grade were grade 1 = 0.0%; grade 2 = 0.6%; grade 3 = 10.6%; and grade 4 = 87.1%.

In 1984, Conn and Barker44 developed a neurologic classification system based on the Glasgow Coma Scale (GCS). This system provides a baseline for organizing optimal patient management and a standardized method of categorizing patients to facilitate research (eTable 117.1).

In 1997, Szpilman proposed a submersion injury classification reevaluating the Menezes and Costa classification.45 The classification was studied using a retrospective review of 1831 cases of submersions and categorized patients into 6 subgroups based upon the degree of respiratory compromise determined during the initial evaluation at the submersion site. The grades of severity were grade 1 (normal pulmonary auscultation with coughing; mortality = 0%); grade 2 (abnormal pulmonary auscultation with rales in some pulmonary fields; mortality = 0.6%); grade 3 (auscultatory evidence of acute pulmonary edema without arterial hypotension; mortality = 5.2%); grade 4 (auscultatory evidence of acute pulmonary edema with arterial hypotension; mortality = 19.4%); grade 5 (isolated respiratory arrest; mortality = 44%); and grade 6 (cardiopulmonary arrest; mortality = 93%). The clinical findings of the victims and interventions that were administered at the submersion site were evaluated and extrapolated and a suggested algorithm for treatment was developed to aid lifeguards in the initial care of a victim. It remains unclear how applicable this classification system is for children and infants, and if similar results would be obtained if the scale were validated prospectively (eFig. 117.1).

Respiratory Manifestations

Submersion usually initiates a complex sequence of behavioral and reflex responses, including panic, frantic attempts to surface, and breath holding. Younger victims, however, may not struggle, but simply sink and drown quietly. When fluid comes in contact with the mucosa of the larynx, laryngospasm typically ensues, but this protective reflex is short lived and water can soon be aspirated into the trachea and distal airways. This is contrary to common wisdom that laryngospasm prevents water aspiration in majority of victims, who go on to die of hypoxia without water entering the lungs.46 Determinations of lung weights and volumes from autopsies of drowned victims have shown evidence of death without aspiration of water in about 2% of the cases.47

Seawater salinity generally ranges from 32 to 37 parts per thousand (ppt), which is equivalent to 32 to 37 grams of dissolved salts per kilogram of water.48 The most abundant ion in seawater is chloride, followed by sodium, sulfate, magnesium, calcium, and potassium. Freshwater salinity is usually less than 0.5 ppt.48 The traditional teaching that seawater draws water from the intravascular spaces into the lungs by osmosis, resulting in fluid-filled alveoli and hypovolemia, and freshwater results in water absorption into the vascular space and hypervolemia and hemolysis is based on impeccable physiologic reasoning. It is, however, rarely corroborated by empirical observation, perhaps because most survivors do not aspirate sufficient volumes of liquid to justify these changes. Similarly, it stands to reason that water entry into the lungs should wash out pulmonary surfactant or inactivate surfactant’s surface tension properties.46,49,50 Once again, the extent to which surfactant deficiency contributes to the alveolar collapse and gas exchange abnormalities seen after submersion injuries is not well defined. As in other forms of respiratory disease, ventilation-perfusion mismatch is the primary underlying mechanism of the hypoxemia and hypercapnia found in submersion victims.

Victims may aspirate not only water (fresh- or seawater), but also foreign materials contained in the water and stomach contents.55 Often the victim swallows large amounts of water during the submersion event. A distended stomach makes vomiting common during the resuscitation. Because the victim is unconscious and the airway is unprotected, the results can be catastrophic. Aspiration of gastric contents causes inflammatory damage to the lung parenchyma and compounds gas exchange abnormalities. It may also further predispose the victim to pulmonary infections caused by oropharyngeal organisms.56

Neurologic Manifestations

Brain hypoxia, secondary to hypoxemia, ischemia, or both, is the main mechanism for neurologic dysfunction in submersion victims, but both rescuers and clinicians need to be alert for the possibility of concurrent trauma or neurologic disease (eg, seizures resulting in loss of consciousness before drowning). The manifestations of the hypoxic injury are often immediate and severe, displaying the entire spectrum found in other forms of hypoxic-ischemic encephalopathy (see Chapter 104). Decreased consciousness is the rule in all patients who have suffered a severe insult. Coma, decreased respiratory drive, decorticate and decerebrate responses, and seizures are common. Brain edema and herniation may occur in severe cases. The pessimistic or, at best, uncertain prognosis of the neurologic dysfunction usually dominate the clinical course after the initial resuscitation, when the shock of the accident is replaced by a waiting vigil for days, or even weeks, until the extent and prognosis of the neurologic injuries can be defined.

Cardiovascular Manifestations

Cardiac dysrhythmias, cardiac dysfunction, and hypotension are caused primarily by hypoxemic or ischemic hypoxia of the myocardium. Increases in systemic and pulmonary vascular resistances have been described,52,57 but they tend to be incidental and subside after appropriate ventilation and oxygenation have been restored. Hypothermia can cause hypotension and bradycardia; the latter should not be interpreted automatically as a more ominous sign of brain herniation.

Other Systems

Although much has been written about the fluid shifts produced by freshwater (hypotonic) and seawater (hypertonic) on circulating blood volume and electrolyte balance, clinically significant alterations in blood volume or electrolyte concentrations are rare after submersion events. Modell found that dogs must aspirate more than 11 mL/kg of liquid before significant fluid shifts could occur, and more than 22 mL/kg before significant electrolyte abnormalities develop.46,58-60 There are a few documented unusual circumstances in which significant electrolyte abnormalities occurred, such as submersion in the Dead Sea or in industrially polluted waters. Submersion in the Dead Sea causes hypercalcemia and hypermagnesemia.61,62

Hypoxemia and hypercarbia cause significant alterations in the acid–base status of a submersion victim.60 Airway obstruction with apnea affects the arterial oxygen tension and acid–base balance in dogs. During the acute asphyxia, arterial carbon tension increases by approximately 6 mm Hg per minute, whereas arterial pH decreases by 0.05 U per minute.

Mild reversible renal impairment is common after a submersion event. Acute renal injury is usually mild and self-limited. Conversely, with severe hypoxia and hypothermia, acute tubular necrosis and renal failure, secondary to shock, and multiple organ failure can also occur.

Early animal studies showed that hemoglobin and hematocrit values varied with the type of fluid aspirated.63 However, recent research on humans fails to support this finding. Hemodilution or hemoconcentration seldom occur and, if present, may relate to ingestion of large volumes of water.46,58,64 Anemia should always raise suspicion of an occult hemorrhage from associated trauma.

The primary goal of the resuscitation after a submersion injury is to restore adequate oxygenation, ventilation, and circulation. Treatment should begin at the scene. If possible, rescue breathing should be initiated while subject is still in the water.65,66 In diving or boating accidents and in any other instance when trauma is possible (including unwitnessed events), the cervical spine should be immobilized and a jaw thrust maneuver should be used as the method of choice to open the airway. Clearing the mouth and airways of water is unnecessary before rescue breathing. All other resuscitative measures should be instituted according to current pediatric advanced life support protocols until arrival at the nearest medical facility.

Although the specter of brain injury tends to overpower all the other consequences of severe submersion injury, the fortunate reality is that a majority of victims will arrive at the emergency department awake, alert, and having sustained no apparent injury. Even in those cases, it is appropriate to administer 100% oxygen with a nonrebreathing mask. Wet clothes should be removed and other warming measures should be applied as needed. A thorough physical examination and a detailed history are essential to define the circumstances of the event and to identify traumatic injuries that may not be apparent initially. Depending on the history and whether respiratory manifestations develop, a chest radiograph may provide helpful information. However, routine radiographs and laboratory tests are usually unnecessary in the asymptomatic victim.

For the spontaneously breathing patient, 100% oxygen should be by a nonrebreathing mask, bi-level positive airway pressure (BiPAP), or continuous positive airway pressure (CPAP). Because of atelectasis, pulmonary edema, and intrapulmonary shunting, positive airway pressure can increase the functional residual capacity and improve lung compliance.73 Bronchospasm should be treated with nebulized or inhaled beta-agonists such as albuterol or levalbuterol. Pulmonary edema may respond to a diuretic such as furosemide in addition to positive airway pressure.

If there is difficulty maintaining the natural airway or there is evidence of impending respiratory failure, endotracheal intubation and mechanical ventilation may be necessary. Rapid-sequence induction with medications that minimize fluctuations in intracranial pressure and do not exacerbate cardiovascular depression or vasodilatation are recommended. Depending on the clinical severity, one should start with physiologic positive end expiratory pressure (PEEP) of 4 to 5 cm H2O and increase in increments to minimize the Aao2 gradient. A normal Aao2 gradient is 35 torr. When PEEP requirements exceed 10 to 15 cm H2O, a pulmonary artery catheter may be helpful. Patients who require high PEEP are at risk for barotrauma and may require alternative means of ventilatory support, such as high frequency ventilation or extracorporeal membrane oxygenation.74,75 Complications of mechanical ventilation include lung parenchymal damage, pneumothorax, or pneumomediastinum. And finally, there have been a few case reports describing the benefits of artificial surfactant76,77 and cardiopulmonary bypass78 in the most severely affected children. However, further prospective studies need to be done before recommendations for practice can be made.

Victims who develop signs of respiratory distress should be managed in a manner as with any other child with respiratory disease (see Chapters 102 and 109). For the spontaneously breathing patient, 100% oxygen should be given by a nonrebreathing mask. Bi-level positive airway pressure (BiPAP), or continuous positive airway pressure (CPAP) can be useful in patients with more severe respiratory dysfunction, but endotracheal intubation and mechanical ventilation are indicated whenever there is difficulty maintaining the natural airway or there is evidence of impending respiratory failure. If aspiration of gross particulate material is suspected, removal via bronchoscopy may be indicated once the patient is sufficiently stable to tolerate it. Currently, there is no evidence to support the prophylactic use of corticosteroids or antibiotics.46,56,79 Daily assessment of the tracheal aspirate for leukocytosis may be helpful to identify infection, but the presence of leukocytosis is not a specific indicator for infection after aspiration injury.

Circulatory dysfunction is usually the result of hypoxia and therefore tends to be associated with neurologic injuries. It should be treated aggressively to minimize the compounding effect of ischemia on already injured tissues. Inotropic and vasoactive medications are often required to support blood pressure and improve perfusion, but, unless other injuries are present, in most patients circulatory function improves rapidly once blood volume is restored and the myocardium recovers from hypoxia.

Maintenance of good cardiac output is critical and aggressive fluid support may be necessary. Treat evidence of poor cardiac function, heart failure, and hypotension with inotropic agents. Blood pressure must be maintained at a level adequate to support cerebral perfusion (cerebral perfusion pressure [CPP] = mean arterial pressure [MAP] – intracranial pressure [ICP]).

In the unfortunate children who present with neurologic dysfunction, the prevention of secondary brain injury is perhaps the most important priority. Intubation of the trachea and institution of mechanical ventilation are usually indicated to minimize the risk from an unprotected airway or a depressed respiratory drive. Placement of a nasogastric tube is recommended to minimize the risk of vomiting and decompress the stomach and intestine. Seizures are common after severe injuries and should be treated with anticonvulsants.

Hypothermia is a common problem in children after a submersion event,81 particularly in cold climates. Esophageal or bladder temperature probes are helpful to monitor the rewarming process, which should be started as soon as possible.49,81 Passive rewarming measures (removal of wet clothing, placement of dry blankets) depend on the child’s ability to produce heat and are often inadequate. Active external rewarming with conventional warmed blankets, heating pads, radiant warmers, and heat lamps is more successful, particularly if the circulation is not impaired and the skin is well perfused. Core rewarming methods are the most efficient, and include administration of warmed IV fluids (36�C–40�C); inhalation of warmed humidified gas mixtures; and performance of gastric, bladder, pleural, or peritoneal lavage with warm saline. Institution of cardiopulmonary bypass or ECMO has been used successfully in children with severe hypothermia and circulatory dysfunction.74,75,78 Complications of rewarming include hypotension, shock, and arrhythmias. Warmed isotonic fluids (Ringer’s lactate or normal saline) and inotropic support are recommended to support blood pressure. Conventional advanced life support methods should be used to treat arrhythmias such as ventricular fibrillation. Cardiac medications, such as lidocaine are safe in the hypothermic child. Electrical defibrillation may also be attempted; however, it is usually ineffective until the temperature is at least 28°C to 30°C.

Hypothermia can also cause hemolysis, coagulation abnormalities, or disseminated intravascular coagulopathy (DIC). Hemolysis leading to hemoglobin deposition in the renal tubules can result in renal tubular insufficiency. Restoration of renal perfusion and circulating volume may reverse this problem, however if renal failure occurs, dialysis might be indicated. Hypothermia is also associated with sepsis syndrome, rhabdomyolysis, pancreatitis, and hypo- or hyperglycemia. Vigilant laboratory monitoring is prudent in the hypothermic child after a submersion event.

The neurologic injuries are by far the main outcome determinant after submersion injuries. Many clinical investigators have proposed scoring systems and identified prognostic factors that may improve our ability to make predictions (Table 117-1). Although these efforts are helpful, one overwhelming consideration emerges from all the studies as well as from the collective experience of a majority of clinicians: Unresponsiveness following cardiopulmonary resuscitation carries a very poor prognosis for neurologic recovery.

There are, however, some exceptions, most notably when the loss of consciousness is not the primary result of hypoxic injury (eg, when concomitant head trauma is the cause). In addition, there is firm evidence that the temperature of the water can modify the prognosis of severe submersion injuries. Over the years, there have been numerous reports of children who recover with minimal or even without neurologic deficit after prolonged icey water submersions. These reports, and the underlying implication that rapid cooling is protective against neurologic injury even if it occurs simultaneously with the injury, justify a modified approach to submersion in cold conditions.49 Thus, it is prudent to continue resuscitation efforts in such conditions, at least until every attempt has been made to reestablish normothermia, including the use of extracorporeal circulation when available. It should be clear, however, that this recommendation is based on a generalization that has not been sanctioned empirically and that the clinician must be circumspect about communicating unrealistic expectations to the family.

In the emergency department, a number of variables have been retrospectively studied to delineate between the intact survivors, severely impaired survivors, and nonsurvivors after warm water submersions. No variable or combination of predictors has been shown to be 100% predictive. Variables that have been studied include scene factors (length of submersion, time to resuscitation) and clinical factors (Glasgow Coma Score [GCS], serum glucose, arterial pH, pupillary response, need for cardiac medications during CPR) in the hopes of being able to predict good or poor outcome.82-85

Prediction of neurologic outcome in the intensive care unit (ICU) may be more reliable. A number of investigators have noted that a GCS less than 5 at emergency department (ED) presentation indicates a high risk of poor neurologic outcome. Lavelle and Shaw86 and Allman et al87 showed that a GCS less than 5 on arrival at the ICU was a generally poor prognostic indicator. However, 3 out of 20 patients with a GCS of 5 on arrival at the ICU had good neurologic outcome. Allman et al noted that although no patients admitted to the ICU with a GCS of 3 had a good outcome, 12 out of 24 patients with a GCS of 4 or 5 survived with normal neurologic function. In most studies, a GCS of greater than 5 on arrival at the ED or ICU is highly predictive of good neurologic outcome.

Orlowski developed a scoring system that assigned points to risk factors associated with poor outcome.83 Risk factors such as age under 3 years, submersion time more than 5 minutes, resuscitation attempts delayed more than 10 minutes after the incident, coma upon presentation to the ED, and an arterial pH equal or less than 7.1 were each assigned 1 point. Patients with 2 or fewer prognostic signs had a 90% chance of full recovery, although those with 3 or more had a 5% chance.

Kyriacou et al88 in 1994 found that children with good outcome were 4.75 (95% chemotherapeutic index [CI] = 3.44, 6.06) times more likely to have a history of immediate resuscitation than children with poor outcome. He concluded that immediate resuscitation before arrival of paramedic personnel is associated with a significantly better neurologic outcome in children with submersion injury.

Lavelle and Shaw86 reviewed all pediatric patients admitted to the ICU for warm-water submersion injury over a 6-year period and found that poor neurologic outcome in this sample was associated with unreactive pupils in the emergency department (ED) (OR 374; 95% CI = 17 to 16,000) and a Glasgow Coma Score of 5 or less upon arrival to the ICU (OR 51; 95% CI = 5 to 2200). Again, no independent predictor was absolute. Two patients who arrived to the ED without vital signs received cardiopulmonary resuscitation and had full neurologic recovery.

Bratton et al89 described a retrospective case series of 44 warm-water submersion victims admitted to the ICU and found that all survivors with good neurologic outcome had spontaneous, purposeful movements within 24 hours after submersion.

In addition to clinical factors, ancillary tests and biochemical markers have been examined to determine their predictive abilities in children after a submersion event. Intracranial pressure monitoring is one such modality and is no longer routinely utilized in the submerged child. Increased intracranial pressure (greater than 20 cm H2O) is associated with poor neurologic outcome or death; however, initial ICPs of less than 20 cm H2O do not necessarily indicate a good outcome. Connors et al90 measured global cerebral blood flow, cross-brain oxygen content difference, cerebral metabolic rate for oxygen, ICP, and cerebral perfusion pressure and were not able to find differences between normal intact survivors and those that have poor outcome.

Neuroimaging and neurophysiologic monitoring has also been used to predict outcome. The role of CT imaging of the head in the evaluation of the extent of hypoxic ischemic injury is currently unknown. An abnormal computed tomography may help identify some who have sustained severe injury. However, the majority of patients will have normal initial studies, regardless of outcome. Early use of magnetic resonance imaging and magnetic resonance spectroscopy may be more useful in predicting outcome. Dubowitz et al91 obtained serial brain magnetic resonance imaging (MRI) and magnetic resonance spectroscopy on 22 children admitted to the ICU after a submersion event. They found excellent correlation between the two techniques, and excellent predictive value of MRI when performed on day 3 or 4 (100% positive predictive value).

The role of electroencephalography (EEG) in predicting outcome is limited because of the use of sedatives, analgesics, and muscle relaxants used during resuscitation and ICU care. If the EEG reveals attenuation while the patient is without medications, neurologic outcome has been shown to be very poor.

Brainstem auditory evoked response (BAER) testing has also been looked at as a method of evaluating global cerebral injury after submersions. Fisher et al92 collected serial brainstem auditory evoked response measurements on all children admitted to the ICU over a 10-year period. This exploratory study revealed that BAER correlated quality of outcome as early as 6 to 24 hours after submersion. BAER use is still controversial and not widely used in the submerged victim. Because the BAER can be altered by central and peripheral auditory disorders, the method is used with caution. Some advocate using the somatosensory-evoked potentials that may be more accurate in predicting outcome in children who have suffered a hypoxic ischemic event. Currently, there are no published reports describing its use after the submersion event.

Investigators have sought to find biochemical markers that would aid in prediction of outcome. Checchia et al93 found in a small case series, that cardiac troponin measurements in children suffering cardiac arrest after a submersion injury are associated with mortality.

Verdu Periz et al94 evaluated serum neuron specific enolase (NSE) levels in children after nontraumatic injuries resulting in acute encephalopathy and found that enzyme activities are significantly higher in those with neurologic sequelae compared to those without. Although the sample size was small, they concluded that serum NSE levels in children obtained 24 to 72 hours postevent were predictive of neurologic outcome. Thus far, biochemical markers of organ injury to predict outcome have been unreliable. Currently, the most powerful individual predictor of neurologic outcome after cardiopulmonary arrest is the neurologic examination obtained at serial points in time. Stating the obvious, patients reaching the hospital with spontaneous breathing and circulation have good outcomes.

Prevention remains the mainstay in reducing the mortality and morbidity associated with submersion injury. In a retrospective analysis of 4806 deaths of children under age 18 years between 1995 and 1999, it was estimated that more than 85% of these deaths could have been prevented by adequate supervision and pool fencing.95 A number of strategies have been recommended in attempts to prevent unnecessary injuries. Strategies have been proposed, but few rigorously studied. Experts recommend that a number of prevention measures be implemented, because no one strategy can prevent all submersion events and injuries.96

Lack of supervision is the most common reason for infant and toddler submersions.24 Parents should be instructed never to leave a child unattended near bodies of water such as the bathtub, toilet, pool, or bucket of water and resist the temptation to leave even for the shortest time while a child is in the water, no matter how safe the situation appears to be. Children with special needs such as those suffering from autism or seizure disorders should also be constantly supervised while bathing or enjoying water activities.

Providing swimming instruction to prevent submersion injury in young children is controversial. Currently there are no data to support that early swimming instruction statistically decreases the risk of submersion injury. The American Academy of Pediatrics (AAP) suggests children are not developmentally ready for swimming instruction until after their fourth birthday.97 Parents of children enrolled in swimming lessons were surveyed and the results found that up to 54% of parents surveyed believe that swimming lessons can prevent a young child from submersion injury. Thirty-five percent of parents felt that it was better to develop swimming skills than to rely on adult supervision.98 Health-care providers need to address parental misconceptions of the role of swimming and reiterate the importance of close adult supervision of toddlers and young children around water.

Pool fencing significantly reduces the risk of submersion injury. To be effective, fencing should be at least 4 feet high, with no openings more than 4 inches. Gates should be self-closing, self-latching, open away from the pool, and be checked regularly.99 The odds ratio for the risk of submersion injury in a fenced pool compared to an unfenced pool is 0.27 (95% chemotherapeutic index [CI] = 0.16–0.47). Four-sided fencing that isolates the pool from the house is superior to perimeter fencing (isolates the property and pool) and reduces submersion injury among young children by more than 50%.100 In addition, rigid pool covers, underwater alarms, a poolside telephone, and knowledge of CPR techniques all aid in reducing submersion injury in children. In contrast, nonrigid pool covers, such as solar or plastic covers are associated with an increased risk of submersion injury because children misjudge the strength of the cover and attempt to walk over the pool cover. Personal flotation devices (PFDs), also known as life jackets, appear likely to decrease submersion injury in children and adolescents while boating or playing near streams, rivers, and lakes. PFDs safe for use carry the seal of approval from the US Coast Guard (UCSG). These are graded 1 through 5 for use in different settings.102

Counseling older children and adolescents should include instructions to use personal flotation devices and to abstain from alcohol and recreational drug use while participating in water-related activities.

Pool Alarms and Pool Covers

The US Consumer Product Safety Commission (CPSC) evaluated three types of pool alarms: floating alarms that detect waves on the surface, underwater alarms, and a wristband alarm that is worn by a child and sounds an alarm when it is exposed to water.101 The underwater alarms perform most consistency and reliably, and can be used in conjunction with a pool cover. The floating alarms performed inconsistently and often in response to water movements attributable to environmental factors such as the wind and toys. The wrist alarm worked well, but was found to be impractical because the caregiver must remember to place the device on the child.

Motorized rigid pool covers placed over the water area prevents access to the water by young children. These covers are able to support the weight of a child. However, nonrigid pool covers such as solar or plastic covers are associated with increased risk of submersion injury. Young children often misjudge the strength of the cover, and attempt to walk over the pool cover. The submerged child may be hidden from view or entangled in the plastic, delaying the rescue.

Personal Flotation Devices (PFDs)

Personal flotation devices (PFDs), also known as life jackets, appear likely to decrease submersion injury in children and adolescents while boating or playing near streams, rivers, and lakes. Approximately 84% of boating-related submersion deaths can be prevented through PFD use. There are a number of PFDs available. Those safe for use carry the seal of approval from the US Coast Guard (UCSG). Five types of PFDs are available, labeled I through V. Type I PFDs have the most buoyancy and are best for rough waters, excellent for flotation, and will be likely to turn an unconscious victim’s face out of the water. Type II PFDs are good for calm waters near the shore, and in some individuals, may turn the face up and out of the water. Type III are designed for calm waters, and are unable to turn the face up independently. Type IV PFDs are life vests that are thrown into the water to aid in a rescue. The Type V PFD or Special Use Device is used for specific activities and includes deck suits, work vests, sailing vests, and hybrid PFDs that inflate upon entering the water.102

Alcohol

The use of alcohol and other substance abuse while near water has been related to submersion injury especially in adolescents and young adults. Several mechanisms include increasing the risk of falling, increasing risk-taking behaviors, and reducing the likelihood of survival in the water.26,27,103 There is some evidence that reducing alcohol use in these areas can prevent some submersion deaths.