The Management of Drowning: Prognosis

It is difficult to prognosticate in individual cases because data from the literature, much of it in the paediatric age group, arise from widely different situations. These situations range from childhood bath and pool incidents in fresh water to boating, swimming and diving activities in the open sea. Nevertheless, several reasonably consistent observations emerge.

Factors that negatively affect outcome include submersion time, time to initiation of effective CPR, severe metabolic acidosis, Glasgow Coma Scale (GCS <5), cardiac arrest and the presence of fixed dilated pupils. However, complete recoveries have been reported despite the presence of one or more of these adverse predictors.

Poor prognostic factors

Prolonged submersion.Prolonged time to effective cardiopulmonary resuscitation.Cardiac arrest.Absence of spontaneous respiration.Prolonged coma.

Several pooled series indicate that 90 per cent of patients who arrive at least arousable with spontaneous respiration and purposeful response to pain will survive neurologically intact. In contrast, of the patients in this series who arrived comatose, 34 per cent died, and 20 per cent of the survivors had neurological damage18.

Unresponsive coma, decorticate and decerebrate rigidity, areflexia and fixed dilated pupils are not in themselves diagnostic signs of death, although they are, of course, signs of a poor prognosis. Patients who arrive in hospital in asystole usually have a poor prognosis, and one series reported a 93 per cent mortality rate after cardiac arrest. Current treatment regimens do not alter the outcome.

The rapid development of severe hypothermia, either before or during the final submersion, is probably protective and helps to explain some spectacular recoveries after prolonged periods of submersion. The role of the diving reflex remains controversial.

The Management of Drowning: Hospital (Emergency department/Intensive care)

Hospital management is subdivided into initial emergency department management and continuing therapy in the intensive care unit. All patients should receive oxygen (see Chapter 49) while undergoing evaluation.

Emergency department

On arrival, the emphasis is on evaluation and resuscitation of respiratory failure. Preliminary assessment includes airway, circulation and level of consciousness re-evaluation. Continuous monitoring of pulse, blood pressure, pulse oximetry and electrocardiography are commenced.

The severity of the case determines the appropriate care. If submersion victims show no signs of aspiration on arrival in the emergency department, there may be no need for hospital admission. Patients who are asymptomatic and have normal chest auscultation, chest x-ray findings and arterial blood gases will not subsequently deteriorate7–9. They may safely be discharged after 6 hours. In contrast, patients with mild hypoxaemia, auscultatory rales or rhonchi or chest x-ray changes should be admitted for observation because they may deteriorate.

In moderately hypoxaemic patients who have not lost consciousness and who are breathing spontaneously, the use of non-invasive ventilatory support with face mask or nasal continuous positive airway pressure (CPAP) may be an alternative to sedation and endotracheal intubation, provided adequate gas exchange is achieved.

Seriously affected patients should be admitted to an intensive care unit or a high-dependency unit. Patients with symptomatic hypoxaemia or disturbed consciousness may rapidly deteriorate further as a result of progressive hypoxaemia. Patients who have had a cardiac arrest or are unconscious and/or severely hypoxaemic require ventilatory support and should be intubated using a rapid sequence induction technique. A nasogastric tube should be inserted and the stomach emptied before induction, if possible.

Concurrent with resuscitation measures, a careful search for any other injuries, such as cranial or spinal trauma, internal injuries and long bone fractures, should be undertaken. Initial x-ray studies should include the chest and cervical spine. Cerebrovascular accident, myocardial infarction, seizure or drug abuse should be suspected if the cause of the incident is not readily apparent. One study revealed that 27 per cent of recreational diving deaths appeared to have a cardiac event as the disabling injury11, so a high index of suspicion for myocardial ischaemia should be maintained. A 12-lead electrocardiogram would be advisable early in the evaluation.

Similarly, in the scuba diver, other disorders such as pulmonary barotrauma and cerebral arterial gas embolism may have initiated or complicated the drowning. These conditions may require specific treatment such as recompression, hyperbaric oxygen or drainage of a pneumothorax. Nevertheless, it should never simply be assumed that a diver ‘must have’ suffered pulmonary barotrauma and arterial gas embolism just because the diver was brought to the surface rapidly or unconscious. There have been many unconscious ascents from significant depths in which it appears that pulmonary barotrauma did not occur. Moreover, such an assumption would indicate hyperbaric therapy. Although recompression should not be withheld from a patient who clearly warrants it, it is logistically difficult and more hazardous for an intensive care patient, and recompression should not be used speculatively.

The main goal of therapy is to overcome major derangement of hypoxaemia with its subsequent acidosis. The benchmark should be an arterial oxygen tension (PaO2) of at least 60 mm Hg. This may be achieved by administration of oxygen by mask in milder cases, possibly with CPAP, but some patients will require more aggressive therapy, employing intermittent positive pressure ventilation (IPPV) with a high fractional inspired oxygen concentration (FIO2).

High ventilatory pressures may be required to obtain adequate tidal volumes. Progress should be monitored by serial arterial blood gas determinations and continuous pulse oximetry.

The institution of continuous positive end-expiratory pressure (PEEP) with either IPPV or spontaneous ventilation, (i.e. CPAP) will decrease the pulmonary shunting and ventilation-perfusion inequality and increase the functional residual capacity, thus resulting in a higher PaO2. Nebulized bronchodilator aerosols may be used to control bronchospasm. In sedated patients, fiberoptic bronchoscopy can be used to remove suspected particulate matter, and repeated gentle endotracheal suction will assist in the removal of fluid from the airway.

Early on in the resuscitation sequence, large-bore intravenous access should be established and warmed crystalloid fluids commenced. Moderate volumes may be required initially because of tissue losses, immersion diuresis and dehydration, but care should be taken not to overhydrate. Simultaneously, blood for haematology and biochemistry laboratory work can be drawn for baseline assessment. Testing should include cardiac enzymes. Intra-arterial pressure monitoring is useful and allows frequent arterial blood gas determinations to guide ventilation and acid-base management.

If cardiac arrest is diagnosed, the rhythm should be rapidly determined, and defibrillation and/or intravenous adrenaline (epinephrine) should be administered according to advanced life support protocols (see Figure 23.3). Other arrhythmias should also be appropriately treated if they have not responded to correction of hypoxaemia and restoration of adequate tissue perfusion.

In the past, it was common to attempt to correct metabolic acidosis by giving bicarbonate. The use of bicarbonate in this setting is controversial, and some clinicians may prefer to hyperventilate a patient to create a respiratory compensation for the metabolic acidosis. A more modern alternative intravenous alkalinizing agent is tromethamine acetate (tris-hydroxymethyl aminomethane [THAM]). THAM has the apparent advantage that its action does not result in the generation of carbon dioxide (as occurs with sodium bicarbonate), and as such it may be a better choice in hypercapnic patients who have mixed acidaemia or in patients who are difficult to ventilate. These descriptors may apply to drowning victims.

Some drowning victims have been noted to be markedly hypoglycaemic, and an association with alcohol intoxication, physical exhaustion and hypothermia is relevant. Blood glucose concentrations should be rapidly determined along with blood gases on arrival at hospital, and intravenous glucose therapy should be instituted if appropriate. Untreated hypoglycaemia may aggravate hypoxic brain lesions. However, intravenous glucose must be used with care because hyperglycaemia is also potentially harmful to injured neurons. Hyperglycaemia resulting from massive catecholamine release or other causes may require insulin infusion.

Hypothermia may complicate drowning and pose difficulties with resuscitation end points in the event of cardiac arrest (see Chapter 28). The emergency department management of hypothermia depends on severity, and low-reading thermometers are required because severe hypothermia may otherwise be overlooked. Warmed intravenous fluids and inspired gases, insulation, forced-air warming blankets and radiant heat may be sufficient, but in severe cases, gastric lavage, peritoneal lavage or even cardiopulmonary bypass has been employed. Resuscitation should continue at least until core temperature approaches normal. Care should be taken to avoid hyperthermia because even mild degrees of cerebral hyperthermia can be profoundly disadvantageous to the injured brain.

Hypovolaemia may result from the combined effects of immersion diuresis and pulmonary and tissue oedema. Circulatory support maybe required to provide adequate perfusion of vital tissues. The maintenance of effective cardiac output may require the correction of hypovolaemia, which may be unmasked by the instigation of body rewarming. PPV also decreases venous return to the heart and thus lowers cardiac output. This can usually be overcome by volume restoration or even augmentation. If after volume replacement the patient does not rapidly regain adequate cardiac output, then inotropic support will be required.

Although not studied specifically in the context of drowning, there is compelling evidence from large randomized studies suggesting that resuscitation of intensive care patients with crystalloid fluids results in better outcome (and less requirement for renal replacement therapy) than does resuscitation using colloids. If colloids are used, small volumes of concentrated albumin may be optimal. Care must be taken not to overhydrate drowning patients.

Once intravascular volume is normalized and adequate cardiac output is established, fluid administration should be parsimonious. Diuretics (e.g. frusemide) have also been employed where overhydration is suspected. A urinary catheter with hourly output measurements is essential to determine renal perfusion and function and is a good indication of adequate volume. Electrolyte disturbances are usually not a significant problem in the initial phases, but any abnormality should be corrected.

Antibiotics given prophylactically are of dubious benefit. Antibiotics should be employed only where clearly clinically indicated, guided by sputum and blood cultures. Routine use may encourage colonization by resistant organisms.

Intensive care

The general principles of intensive therapy are followed, but with special emphasis on respiratory function because near drowning is a common cause of the acute respiratory distress syndrome (ARDS).

The following clinical parameters, established on arrival in the emergency department, should be regularly monitored:

  1. Routine clinical observations such as pulse, blood pressure, temperature, respiratory rate, minute volume, inspiratory and PEEP pressures, electrocardiography, pulse oximetry and urine output.
  2. Where indicated, invasive monitoring such as central venous pressure, pulmonary artery wedge pressure and cardiac output by a pulmonary artery catheter.
  3. Blood tests such as arterial blood gas and acid-base status, haemoglobin, packed cell volume, white cell count, serum and urinary electrolytes, serum and urinary haemoglobin and myoglobin levels, serum creatinine, urea, glucose, protein and coagulation status.
  4. Regular chest x-ray examination to detect atelectasis, infection, pneumothorax, pulmonary oedema, pleural effusions and other disorders.
  5. Serial measurement of pulmonary mechanics, by measurement of airway pressures and compliance, which are also useful in monitoring progress. In the less severely ill patient, simple spirometry is a useful guide to recovery.

The optimal method of ventilation aims to produce an adequate oxygen tension at the lowest FIO2 (preferably FIO2 of 0.6 or less to avoid pulmonary oxygen toxicity; see Chapter 17) with the least haemodynamic disturbance and the least harm to the lung. CPAP can be dramatic in improving oxygenation by reducing intrapulmonary shunting. Pressures of 5 to 10 centimetres of water are usual, but greater pressures may be required. Patients receiving PPV tend to retain salt and water, so fluid intake should be reduced to about 1500 ml/day with low sodium content. Fluid overload may have a deleterious effect on pulmonary function.

Various modes of ventilation have been employed. These include spontaneous respiration with CPAP, IPPV with and without PEEP, synchronous intermittent mandatory ventilation, pressure support and high-frequency ventilation. Increasing experience in the management of ARDS has led to the development of so-called ‘lung protective ventilator strategies’ characterized by relatively long inspiratory times, high end-expiratory pressures and relatively small tidal volumes. This may require acceptance of both a degree of ‘permissive hypercapnia’ and mild respiratory acidosis that in the short term may be treated with bicarbonate or THAM, before medium-term compensation through renal retention of bicarbonate. Postural changes (e.g. prone ventilation) are sometimes experimented with in individual patients. There is no universally applicable formula for ventilating these patients, and much individualization of regimens occurs in high-level intensive care units. These types of strategy may be necessary in severely affected drowning victims.

Femoro-femoral or full cardiopulmonary bypass has been employed both for rewarming hypothermic patients and for establishing adequate oxygenation in severe cases12. Severe ARDS has been successfully managed with variable periods (up to weeks) of extracorporeal membrane oxygenation (ECMO).

Because of improvements in cardiorespiratory support, preservation of the central nervous system is now the major therapeutic challenge13. The application of various brain protection techniques including deliberate hypothermia, hyperventilation, barbiturate coma and corticosteroids has not altered cerebral salvage rates in the specific context of drowning14,15. Studies in community cardiac arrest do provide circumstantial support for the use of hypothermia after prolonged resuscitation. Although some intensive care units may try this in drowning victims, it is certainly not considered a standard of care. It is notable that corticosteroids have not been shown to reduce cerebral oedema or intracranial pressure (ICP) and are not recommended.

Central nervous system function is assessed clinically and potentially by electroencephalography. ICP monitoring has been advocated where intracranial hypertension is suspected, with prompt therapy for any sudden elevation.

Serial creatinine estimations often reveal mild renal impairment in patients requiring intensive care. Severe acute renal failure16 requiring dialysis is less common, but it may develop in patients who presented with severe metabolic acidosis and elevated initial serum creatinine levels. Occasional cases of rhabdomyolysis have also been reported.

Hyperpyrexia commonly follows drowning, and its effect may be deleterious, especially to the injured brain. External cooling and antipyretic drugs to prevent shivering and to keep the temperature lower than 37°C may be indicated.

Continuing hyperexcitability and rigidity may require the use of sedative and relaxant drugs.

The Management of Drowning: Advanced Life Support and Transport

A regional organized emergency medical service (e.g. paramedics) that carries specialized apparatus such as oxygen, endotracheal tubes, suction and intravenous equipment should be activated, if available. In any case, the patient should be transferred to hospital as soon as possible. The early administration of oxygen by suitable positive pressure apparatus, by personnel trained in its use, may be the critical factor in saving lives. For this reason, oxygen administration equipment should be carried on all dive boats. Patients who regain consciousness or who remain conscious after drowning events may have significant pulmonary venous admixture with resultant hypoxaemia. All such patients should receive supplementary oxygen and be further assessed in hospital. Respiratory and cardiac arrests have occurred after apparently successful rescue.

Although endotracheal intubation remains the best method for securing an airway and achieving adequate ventilation, the necessary expertise may not be available until the victim is transferred to hospital. In such cases, the use of airway devices such as the laryngeal mask airway may improve ventilation while the patient is being transported to hospital. Other airways such as the pharyngo-tracheal lumen airway and the Combitube tube are alternatives, but they require more training and have their own problems. One potential problem with all supraglottic devices, and with mouth-to-mouth and bag-mask ventilation techniques for that matter, is that the airway pressures required to inflate a ‘wet’, non-compliant lung may be very high and not easily achieved with these devices or methods. Endotracheal intubation may the only way to achieve adequate tidal volumes in such patients.

Properly trained and equipped personnel attending a case in the field may be able to invoke advanced resuscitation techniques such as the airway interventions mentioned earlier and the monitoring methods, drug administration strategies and arrhythmia treatments specified in Figure 23.3.

Advanced life support algorithm. CPR, cardiopulmonary resuscitation; ECG, electrocardiogram; ETT, endotracheal tube; IV, intravenous; IO, intra-osseous; LMA, laryngeal mask airway. (From the Australian Resuscitation Council.)
Advanced life support algorithm. CPR, cardiopulmonary resuscitation; ECG, electrocardiogram; ETT, endotracheal tube; IV, intravenous; IO, intra-osseous; LMA, laryngeal mask airway. (From the Australian Resuscitation Council.)

Advanced life support algorithm. CPR, cardiopulmonary resuscitation; ECG, electrocardiogram; ETT, endotracheal tube; IV, intravenous; IO, intra-osseous; LMA, laryngeal mask airway. (From the Australian Resuscitation Council.)

The Management of Drowning: Rescue and initial resuscitation

In the diving setting, the management of a drowning situation often begins with witnessing a diver become unconscious underwater. Before resuscitation efforts can begin, the victim must be retrieved to the surface. Related considerations were reviewed by the Undersea and Hyperbaric Medical Society (UHMS) Diving Committee1, and their findings are outlined here.

The overarching goal of this initial phase of the rescue is to retrieve the diver to the surface as quickly as possible, even if the victim has a mouthpiece in place and appears to be breathing (which would be a most unusual circumstance). More typically, the victim is found unconscious with the mouthpiece out. No attempt should be made to replace it; however, if the mouthpiece is retained in the mouth, then the rescuer should make an attempt to hold it in place during the ascent. An ascent should be initiated immediately. If there is significant risk to the rescuer in ascending (if the rescuer has a significant decompression obligation), then making the victim buoyant and sending him or her to the surface may be the only option, depending on the degree to which the rescuer wishes to avoid endangering himself or herself.

The committee flagged one exception to the advice to surface immediately. In the situation where a diver is in the clonic phase of a seizure with the mouthpiece retained, then the mouthpiece should be held in place and ascent delayed until the seizure abates. To be clear, however, this does not apply to the more common situation of the seizing diver whose mouthpiece is out. In the latter situation, the ascent should be initiated while the diver is still seizing. This dichotomy arises because of the committee’s perception of the shifting balance of risk between pulmonary barotrauma and drowning in situations where the airway is at least partially protected or not. Thus, where the airway is completely unprotected (mouthpiece not retained), the risk of drowning outweighs the risk of barotrauma imposed by seizure-induced apposition of the glottis tissues. Where the airway is partly protected (mouthpiece retained and held in place), the opposite holds true. This matter is discussed in more detail in the committee report.

At the surface, the victim should be made positively buoyant face-up, and a trained rescuer should attempt to give two mouth-to-mouth rescue breaths. Experience has shown that this is often all that is required to stimulate the victim to breathe. Pausing to give rescue breaths will slightly delay removal from the water for definitive cardiopulmonary resuscitation (CPR) and is therefore a gamble that the victim has not yet having suffered cardiac arrest. However, given the extremely poor outcome expected if a drowning victim suffers a hypoxic cardiac arrest and the time it usually takes to remove a diver from the water, the committee determined that this was a gamble worth taking. The best chance of survival lies in preventing hypoxic cardiac arrest, and establishing oxygenation is the means of such prevention. If the diver has already had a cardiac arrest, then a small extra delay in initiating CPR imposed by performing in-water rescue breaths is not likely to alter the outcome. There is some human evidence suggesting a survival advantage for in-water rescue breathing in non-diving drowning situations2.

Once at the surface and in a situation where the surface support is not immediately to hand, a choice must be made whether to wait for rescue or initiate a tow to shore or nearest surface support. The committee determined that if surface support is less than a 5-minute tow away, then a tow should be commenced with intermittent rescue breaths administered if possible. If surface support or the shore is more than a 5-minute tow away, then the rescuer should remain in place, continuing to administer rescue breaths for 1 minute. If there is no response in this time, then a tow toward the nearest surface support should be initiated without ongoing rescue breaths. These guidelines are summarized in Figure 23.1.

Undersea and Hyperbaric Medical Society Diving Committee guidelines for rescue of an unresponsive diver from depth.
Figure 23.1 Undersea and Hyperbaric Medical Society Diving Committee guidelines for rescue of an unresponsive diver from depth. It is recommended that the interested diver read the original paper which contextualizes these recommendations more thoroughly. CPR, cardiopulmonary resuscitation. (From Mitchell SJ, Bennett MH, Bird N, Doolette DJ, et al. Recommendations for rescue of a submerged unconscious compressed gas diver. Undersea and Hyperbaric Medicine 2012;39:1099–1108.)

It is notable that these guidelines contain no reference to in-water chest compressions. Although techniques for in-water chest compressions have been described3,4, the committee did not consider there was adequate evidence of efficacy to justify the extra difficulty and stress to the rescuer for their inclusion in the rescue protocol.

The victim should be kept horizontal as much as possible during and after removal from the water. The patient should be moved with the head in the neutral position if cervical spine injury is suspected. Scuba divers are most unlikely to have suffered cervical spine trauma. A basic life support algorithm should be initiated immediately, beginning with assessment of the airway (Figure 23.2).

Basic life support algorithm. AED, automatic external defibrillator; CPR, cardiopulmonary resuscitation. (From the Australian Resuscitation Council.)
Figure 23.2 Basic life support algorithm. AED, automatic external defibrillator; CPR, cardiopulmonary resuscitation. (From the Australian Resuscitation Council.)


Vomiting and regurgitation frequently follow a submersion incident. Foreign particulate matter causing upper airway obstruction should be removed manually or later by suction. Obstruction of the upper airway by the tongue is common in the unconscious patient.

Two methods are used to overcome the obstruction:

Head-tilt/chin-lift is accomplished by pushing firmly back on the patient’s forehead and lifting the chin forward by using two fingers under the jaw at the chin. The soft tissues under the chin should not be compressed, and unless mouth-to-nose breathing is to be employed, the mouth should not be completely closed. This technique should be avoided if cervical spine injury is suspected.

Jaw-thrust describes the technique of forward displacement of the lower jaw by lifting it with one hand on either side of the angle of the mandible. Unless cervical spine injury is suspected, this technique is often combined with head-tilt.

Time should not be wasted in trying to clear water from the lower airways. If airway obstruction is encountered and has not responded to normal airway management, the Heimlich manoeuvre (sub-diaphragmatic thrust) has been suggested5. This manoeuvre, which was proposed as a routine step to clear water from the airway, has not received the widespread endorsement of resuscitation councils around the world. It should be used with caution and only as a last resort because of the risks of regurgitation of gastric contents, rupture of the stomach and causing delay in initiating effective ventilation. Persistent airway obstruction may result from a foreign body, but other causes include laryngeal oedema or trauma, bronchospasm and pulmonary oedema.


Respiration can be assessed by placing one’s ear over the victim’s mouth while looking for chest movement, listening for air sounds and feeling for the flow of expired air. If breathing is detected, oxygen should be administered and the victim maintained in the ‘recovery’ position to avoid aspiration of fluid or vomitus.

If breathing is absent, mouth-to-mouth or mouth-to-nose breathing is instituted. Initially, two full breaths of air, with an inspiratory time (for the victim) of 1 to 1.5 seconds, are recommended. For adults, an adequate volume to observe chest movement is about 800 ml. If no chest movement is seen and no air is detected in the exhalation phase, then head-tilt or jaw-thrust manoeuvres should be revised. Failing that, further attempts at clearing the airway with the fingers (only if the victim is unconscious!) should be undertaken. With mouth-to-mouth respiration, the rescuer pinches the victim’s nose and closes it gently between finger and thumb. Mouth-to-nose rescue breathing may be more suitable in certain situations, such as when marked trismus is present or when it is difficult to obtain an effective seal (e.g. injury to mouth, dentures).

Paramedics or other practitioners with advanced skills are likely to use a bag-mask-reservoir device connected to an oxygen source for manual positive pressure ventilation in the field. Useful adjuncts in resolving upper airway obstruction may include a nasopharyngeal airway, oropharyngeal airway or supraglottic airway device such as a laryngeal mask. Endotracheal intubation in the field should be undertaken only by highly trained and experienced practitioners.

The rate of chest inflation should be about 12 per minute (one every 5 seconds) with increased rate and decreased volume in young children.

It must be made clear that the recent advocacy for ‘compression-only CPR’ in which rescue breathing is omitted and first responders provide only chest compressions to victims of community cardiac arrest is not relevant to CPR in the context of drowning.The cause of cardiac arrest in the community is usually some sort of cardiac disease, whereas it is hypoxia in drowning. Compression-only CPR works in community cardiac arrest because the victim is not hypoxic at the onset of cardiac standstill, and the lungs are filled with air to functional residual capacity. In contrast, hypoxia is usually the cause of cardiac arrest in drowning, and the lungs are frequently compromised by aspirated fluid and alveolar collapse. Failing to ventilate the lungs during resuscitation of a drowning victim is likely to bias against a good outcome.


The presence of a carotid or femoral pulse should be sought in the unconscious non-breathing victim. This is often difficult because the patient is usually cold and peripherally vasoconstricted. Although it is possible that external cardiac compression (ECC) could precipitate ventricular fibrillation in a hypothermic patient, if in doubt it is safer to commence ECC than not.

If no carotid pulse is detected, ECC should be commenced after two initial breaths. Higher rates of compression are now recommended, with greater outputs achieved at 100/minute compared with the traditional 60/minute standard. Controversy still exists over the mechanism of flow in external compression, with the evidence for the older ‘direct compression’ model being challenged by the ‘thoracic pump’ theory.

Cardiac compression should be performed with the patient supine on a firm surface. The legs may be elevated to improved venous return. The rescuer kneels to the side of the patient. The heel of the rescuer’s hand should be placed in line with the patient’s sternum. The lower edge of the hand should be about two fingers above the xiphisternum (i.e. compression is of the lower half of the sternum). The second hand should be placed over the first, and the compression of the sternum should be about 4 to 5 centimetres in adults in the vertical plane. To achieve this, the rescuer’s elbow should be straight, with the shoulders directly over the sternum. A single rescuer may be able to achieve rates of only 80/minute because of fatigue, but if several rescuers are present, it may be possible to maintain high rates.

Further help should be sought immediately, by a third person, if possible, without compromising resuscitation efforts.

Survival from Drowning

Treatment at the scene of an accident is sometimes of little ultimate consequence with many disorders, but in drowning it often determines whether the victim lives or dies. The standard of first aid and resuscitation training of the rescuers therefore influences outcome.

In human drowning, deterioration after initial resuscitation is frequently recorded, and this influences management (see Chapter 23).

The temperature of the water and thus the degree of hypothermia may also be factors. Poorer results are achieved in warm water drowning.

In what was previously referred to as ‘dry’ drowning (in which the distal airway remains relatively dry because of early laryngospasm), the patient is hypoxic and, if rescued in time, may– make a rapid recovery. However, when laryngospasm relaxes and fluid aspiration occurs as it eventually does if the victim remains immersed, the result is drowning.

Other factors that influence outcome include the following: the presence of chlorine, other chemicals and foreign bodies; the aspiration of stomach contents; and the subsequent development of pneumonitis, respiratory infection and multi-organ failure.

One likely cause for delayed death is progressive lung injury2. ARDS develops in a significant proportion of drowning cases; usually hours or days after the aspiration. Other causes of death in the days after the event include cerebral hypoxia, secondary infections (usually of the lungs), renal failure and iatrogenic events.

Factors that negatively influence survival have been well documented by Modell:

  • Prolonged immersion.
  • Delay in effective cardiopulmonary resuscitation.
  • Severe metabolic acidosis (pH <7.1).
  • Asystole on admission to hospital.
  • Fixed dilated pupils.
  • Low Glasgow Coma Scale score (<5).

Nevertheless, none of these predictors is infallible, and survival with normal cerebral function has been reported with all the foregoing factors.

Claims of survival after extended duration underwater without ventilation of the lungs have been used to encourage rescuers to persevere with resuscitation efforts. There have been cases reported in victims who have been submerged for between 15 and 45 minutes4–7 and who have survived without neurological sequelae. The explanations given for such prolonged durations of survival are as follows:

1. Hypothermia is protective and develops very rapidly with aspiration of water. In swimmers and divers, hypothermia may be present before the incident.

2. The ‘diving reflex’ is a possible, but contentious, explanation. Within seconds of submersion, the diving reflex may be triggered by sensory stimulation of the trigeminal nerve and by reflex or voluntary inhibition of the respiratory centre in the medulla. This produces bradycardia and shunting of the blood to the areas more sensitive to hypoxia – the brain and coronary circulations. It is independent of baroreceptor or chemoreceptor inputs. The diving reflex is more intense in the frightened or startled animal, compared with animals which dive or submerge voluntarily, but it is not known whether this finding is applicable to humans. Water temperatures higher than 20°C do not inhibit the diving reflex, but progressively lower temperatures augment it.

3. Gas exchange in the lungs can continue after submersion. With or without the effects of laryngospasm, there may be several litres of air remaining within the lungs, thus allowing for continued exchange of respiratory gases. Increased pressure (depth) transiently enhances oxygen uptake by increasing the PO2 in compressed lungs. In an unconscious state, with low oxygen use and the effects of hypothermia, a retained respiratory gas volume could add considerably to the survival time, although it is not often considered in the literature on drowning.

Whether fluid enters the lungs in an unconscious victim depends on many factors, including the spatial orientation of the body. For example, a dependent position of the nose and mouth, facing downward, is not conducive to fluid replacement of the air in the lungs.

Even though spectacular and successful rescue can be achieved after prolonged submersion, it is more frequent that this is not so. Many victims lose consciousness and die after only a few minutes of submersion.

Clinical Features of Drowning

The respiratory manifestations of drowning include the following:

  • Dyspnoea.
  • Retrosternal chest pain.
  • Blood-stained, frothy sputum.
  • Tachypnoea.
  • Cyanosis.
  • Pulmonary crepitations and rhonchi.
  • Hypoxaemia.

Pulse oximetry typically reveals low oxygen saturations, but a pulse oximeter may not read at all on a cold, peripherally shut-down victim. An arterial blood gas determination reveals hypoxaemia (lower limit of the ‘normal’ range for arterial oxygen tension [PO2] is 80 mm Hg [10.5 kPa]). There is often acidaemia that usually has a metabolic component, but that may be mixed and very severe in a respiratory peri-arrest situation. Carbon dioxide levels are frequently elevated in a peri-arrest condition, but they may be normal or even low during spontaneous breathing or manual ventilation.

Initial chest x-ray studies may be normal, or they may show patchy opacities or pulmonary oedema. Significant hypoxia may be present even when chest x-ray changes are subtle or even absent.

Complications may include pneumonitis, pulmonary oedema, bronchopneumonia, pulmonary abscess and empyema. Severe pulmonary infections with unusual organisms leading to long-term morbidity have been reported. Progression to the acute respiratory distress syndrome (ARDS) is not uncommon in drowning situations.

Central nervous system effects of hypoxia include variable impairment of consciousness, ranging from awake to comatose, with decorticate or decerebrate responses. If hypoxia is prolonged, a global hypoxic brain injury can result with cerebral oedema, raised intracranial pressure and sustained coma. Seizure activity is common in this setting.

Cardiovascular manifestations are largely the result of the effects of hypoxaemia on the heart. Progressive bradycardia leading to asystolic cardiac arrest is not uncommon. After rescue and resuscitation, supraventricular tachycardias are frequent, but various other dysrhythmias may occur. When the hypoxic acidotic insult has been severe, hypotension and shock may persist despite re-establishment of a perfusing rhythm. The central venous pressure may be elevated as a result of right-sided heart failure exacerbated by elevated pulmonary vascular resistance, rather than by volume overload. Mixed venous oxygen tension may also be low, indicating tissue hypoperfusion.

Multi-system organ failure may develop secondary to the hypoxaemia, acidosis and resultant hypoperfusion. Decreased urinary output occurs initially and occasionally progresses to acute tubular necrosis and renal failure. Haemoglobinaemia, coagulation disorders and even disseminated intravascular coagulation may complicate the clinical picture.

Laboratory findings include decreased arterial oxygen with variable PaCO2 values, metabolic and respiratory acidosis, haemoconcentration, leucocytosis, increased lactic dehydrogenase, occasional elevated creatinine levels and haemolysis as indicated by elevated free haemoglobin. Serum electrolytes are usually within the normal range.

The arterial oxygen tension is always low, but the carbon dioxide tension may be low, normal or elevated.

Recovery from drowning is often complete in survivors. However, residual neurological deficiencies may persist in the form of either cognitive impairment or extrapyramidal disorders.

Pathophysiological of Drowning

The effects of drowning are multiple, but the initial and primary insult is to the respiratory system, with hypoxaemia being the inevitable result (Case Report 22.1).
The sequence of events that occur with drowning includes the following:

Initial submersion in water preventing air breathing. This is usually followed by voluntary breath-holding. Duration of the breath-holding depends on several factors, which include general physical condition, exercise, prior hyperventilation and psychological factors (see Chapter 61). This is frequently a period when the victim swallows substantial amounts of water.

Fluid aspiration into the airway at the point of breaking the breath-hold. Eventually, the rising arterial carbon dioxide tension (PaCO2) compels inspiration, and fluid is aspirated. Laryngeal spasm may follow the first contact of the glottis with water. While laryngospasm is maintained, the lungs may remain dry; however, the inevitable result of the associated hypoxaemia is that the spasm will eventually also break, and if the victim remains immersed, then aspiration of water into the lungs will follow. Vomiting of swallowed liquid may occur, and this may also be aspirated into the lungs.

Progressive hypoxaemia. This may initially result from oxygen use during voluntary breath-holding and any subsequent laryngospasm, but ultimately it is aspiration of water or regurgitated stomach contents into the gas-exchanging segments of the lungs that provokes persistent and progressive hypoxaemia. The inhalation of water can occur through involuntary diaphragmatic contractions even if the victim is not breathing per se. The presence of water instead of air and the dilution of surfactant function with consequent alveolar atelectasis result in a ventilation-perfusion (V/Q) mismatch with a preponderance of low V/Q units and extensive venous admixture. The resulting hypoxaemia leads to unconsciousness, bradycardia and ultimately asystolic cardiac arrest. Hypoxic brain damage follows within a very short space of time.


Ernie Hazard, age 35: ‘I was thinking “This is it. Just take a mouthful of water and it’s over.” It was very matter of fact. I was at a fork in the road and there was work to do – swim or die. It didn’t scare me. I didn’t think about my family or anything. It was more businesslike. People think you always have to go for life, but you don’t. You can quit….’

The instinct to breathe underwater is so strong that it overcomes the agony of running out of air. No matter how desperate the drowning person is, he or she does not inhale until on the verge of losing consciousness. That is called the ‘break point’.
The process is filled with desperation and awkwardness: ‘So this is drowning…so this is how my life finally ends…. I can’t die, I have tickets for next week’s game’…. The drowning person may even feel embarrassed, as if he or she has squandered a great fortune. He or she has an image of people shaking their heads over this dying so senselessly. The drowning may feel as if it is the last, greatest act of stupidity in his or her life. The thought shrieks through the mind during a minute or so that it takes the panicky person to run out of air.

Occasionally, someone makes it back from this dark world. In 1892, a Scottish doctor, James Lowson, was on a steamship bound for Colombo. Most of the 180 people on board sank with the ship, but Lowson managed to fight his way out of the hold and over the side:

‘I struck out to reach the surface, only to go further down. Exertion was a serious waste of breath and after 10 or 15 seconds the effort of inspiration could no longer be restrained. It seems as if I was in a vice which was gradually being screwed up tight until it felt as if the sternum of the spinal column must break. Many years ago my old teacher used to describe how painless and easy death by drowning was – “like falling about a green field in early summer” – and this flashed across my brain at the time. The “gulping” efforts became less frequent and the pressure seemed unbearable, but gradually the pain seemed to ease up. I appeared to be in a pleasant dream, although I had enough willpower to think of friends at home and the site of the Grampians, familiar to me as a boy, that was brought into my view. Before losing consciousness the chest pain had completely disappeared and the sensation was actually pleasant.

‘When consciousness returned I found myself on the surface. I managed to get a dozen good inspirations. Land was 400 yards distant and I used a veil of silk and then a long wooden plank to assist me to shore. On landing and getting on a sheltered rock, no effort was required to produce copious emesis. After the excitement, sound sleep set in and this lasted three hours, when a profuse diarrhoea came on, evidently brought on by the sea water ingested. Until morning break, all my muscles were in a constant tremor which could not be controlled’.

From Junger S. The Perfect Storm. London: Fourth Estate; 1997, with quotes from James Lowson in The Edinburgh Medical Journal.

Drowning: Radiology and Pathology

Death may occur during immersion, soon after or from delayed complications.

In recreational scuba deaths, drowning is the most common cause – but it is usually a secondary effect, with the primary cause leading to loss of consciousness. Drowning reflects the fact that unconsciousness occurred in a watery environment.

Accidents (e.g. hypoxia, gas toxicities, immersion pulmonary oedema, dysbarism, medical illnesses, trauma) that occur while someone is immersed or submerged may result in the secondary complication of drowning, with all its pathological sequelae. Drowning then complicates the interpretation of the diving accident and contributes to a combined disorder.

Certain external characteristics of drowned victims are common, although these are more specific for immersion than for drowning (see Chapter 51). These include the following: pale, wrinkled, ‘washerwoman’s’ skin; post-mortem decomposition; lacerations and abrasions from impact with rocks, coral, shells, motor boats and their propellers; post-mortem injuries from aquatic animals, varying from the nibbling of protuberances (fingers, ears, nose and lips) by crustaceans and fish to the large tearing wounds of sharks and crocodiles.

Radiology, and especially computed tomography scans, may be of value in drowning cases. These studies are likely to detect pulmonary interstitial oedema, sub-glottic fluid in the trachea and main bronchi, frothy airway fluid, hyper-expanded lungs, high-attenuation particles (indicating sand or other sediment), pleural effusion, haemorrhage or effusion in the middle ear and para-nasal or mastoid sinuses, gas and fluid in the stomach and other contributory causes of death.

The theory and the procedures for autopsies of drowning victims are surprisingly contentious for such a common disorder. An autopsy can imply, but not reliably prove, that death is the result of drowning. There is no pathological feature that is pathognomonic for drowning.

Respiratory system

At autopsy, the main macromorphological changes associated with drowning are caused by the penetration of the liquid into the airways. These are external foam from the mouth and nose, frothy fluid in the airways and overexpansion of the lung. They are not specific to drowning and can be found, usually to a lesser degree, in other cardiorespiratory disorders.

The mushroom-like foam or ‘plume’ from the mouth and nostrils, often exacerbated by resuscitation efforts, is composed of the drowning aspirant (usually sea water), pulmonary oedema, mucus and pulmonary surfactant, together with fine air bubbles. It extends into the lower airways, is relatively resistant to collapse and may last some days. It is particularly common with salt water drowning. It is thought to develop only if there is some inspiratory action (i.e. it is not a passive, post-mortem event). Respiratory epithelial cells and CD68+ alveolar macrophages have been detected in the foam.

The lungs are water-logged, often heavier than normal, and overdistended (emphysema aquosum). The lung weights reduce over the subsequent few days, as fluid is redistributed, and sometimes fluid accumulates in the pleural cavity. This is another reason for performing early autopsies. The overdistension may cause the ribs to be imprinted on the pleural surfaces, and the lungs may extend over the mediastinal midline. The lung surfaces may be pale and mottled, and areas of distended alveoli or bullae may be evident, as may subpleural haemorrhages (Paltauf’s spots). The lungs retain their shape and size on sectioning.

Pleural effusions are common and increase with the duration between death and autopsy.

When non-breathing bodies are immersed, significant quantities of fluid usually do not enter the lower respiratory system, but some may enter with the descent of the body as a result of replacement of the contracting gas space (Boyles’ Law). This is unlike the foam referred to earlier.

Histological evidence of focal alveolar damage and emphysema is frequent. Microscopic changes may demonstrate toxic effects both of chemicals and of the specific aspirant. The surfactant changes, including denaturation, can progress even after apparent clinical improvement. Epithelial and endothelial changes, with detachment of the basilar membrane and cellular disruption, have been described. Sand, marine organisms, algae and diatoms may be observed in the lungs.

The finding of water, effusion or blood in the middle ears, mastoid or para-nasal sinuses is not evidence of drowning per se, but it may indicate descent of the body while still alive (see Chapter 51). Other explanations include the inflow of water after death and the effects of venous congestion during the agonal struggle.

Usually the death results from hypoxia from the acute pulmonary damage and the shunting of blood through non-aerated tissue. Sometimes there is progressive or irreversible lung damage, for various reasons. They include progressive surfactant damage, pneumonitis from the aspirant, vomitus and foreign bodies. Even victims who appear normal on arrival at hospital can deteriorate over the next 6 to 12 hours. Respiratory infections and abscesses are not infrequent if death is delayed. Pulmonary oxygen toxicity, associated with prolonged resuscitation attempts, may also be present (see Chapter 17).

Other organs

The stomach often contains free fluid, water inhaled during the incident, together with debris and organisms. Wydler’s sign is a three-layer separation of foam, fluid and solids in the stomach.

Local haemorrhages in the upper torso musculature are sometimes claimed to be drowning induced, but this finding is controversial and non-specific.

The major effects on the neurological system are those of hypoxic brain damage with petechial haemorrhages and subsequent cerebral oedema and raised intracranial pressure.

Autopsies on drowning victims who submerged while still alive, although unconscious, may also show other cranial haemorrhages, which are sometimes misinterpreted as a cause of the accident. Meningeal haemorrhages, both dural and arachnoid, may be observed. These are usually not extensive and are quite different from the brain haemorrhages of arterial gas embolism or decompression sickness or from the petechial haemorrhages of asphyxia. They are probably derived from the bleeding of descent sinus barotrauma, which ruptures into the cranial cavity when the enclosed gas spaces expand during ascent.

There is often considerable venous congestion of the viscera, especially the brain, kidneys and other abdominal organs. Hypoxic cerebral necrosis and acute renal tubular necrosis with blood pigment casts are both described.
Because of the relatively small amount of fluid usually aspirated in drowning, it is considered unlikely that victims die acutely of electrolyte imbalance and/or associated ventricular fibrillation.

Cardiac arrhythmias may be initiated by hypoxia, but this is not demonstrable at autopsy. Cases of prolonged QT interval causing death from immersion have been postulated, but support for this as a common cause of death is lacking in drowning surveys.

The possibility of vagally induced death immediately following immersion has also been proposed when dealing with water colder than 15°C.

In the event of delayed drowning deaths, the lungs, brain or kidneys may all be involved.

Laboratory findings

A series of biochemical tests has been designed to verify drowning as the cause of death. The rationale is that the inhaled fluid, because it has different osmotic pressures, electrolytes and particulates compared with the pulmonary blood flowing past it, will change the latter and alter the character of the blood in the pulmonary veins, the left side of the heart and the systemic arterial system.

Thus, one can compare the left-sided heart blood with the right-sided heart blood and deduce the nature of the aspirate. In fresh water drowning, the left-sided heart blood should have a lower osmotic pressure and a dilution of most electrolytes. The hypo-osmolarity could result in haemolysis with a raised serum haemoglobin and potassium levels. Sea water aspirate should draw fluid from the circulating blood, thereby causing a rise in specific gravity and most blood constituents in the arterial blood. Such changes can be verified in animal experiments when the lungs are flooded with large volumes of fluid and the parameters are measured immediately post mortem.

Neither situation is likely in human drowning, in which the volume of aspirate is relatively small. There are usually many hours between death and autopsy, thus allowing blood constituents and electrolytes to equilibrate. Effective resuscitation is also likely to diminish any variations in venous and arterial blood.

For the foregoing reasons, the Gettler test (chloride variation) and sodium, magnesium, calcium, strontium, haematocrit, haemoglobin, pulmonary surfactant protein and specific gravity levels are unlikely to contribute to the autopsy diagnosis of drowning, although all have had their proponents. Some pathologists use the measurement of vitreous electrolytes to support the diagnosis. Atrial natriuretic peptide levels may increase in drowning, but also in immersion per se, in cardiac disease and in any hypervolaemic state. They also only persist for short periods.

Identification and comparison of environmental and systemic diatoms and algae in the lungs, blood, kidneys and vertebrae have been recommended. The single-celled diatoms, usually 10 to 80 micrometres long, are ubiquitous with about 15 000 different species – some inhabiting most waterways. They do not enter the tissues from the lungs unless there is an active circulation. Their presence in both the water environment and the body tissues does not prove drowning, merely the aspiration of that water while the body’s circulation is still functional. The silica shell makes diatoms stable and thus detectable by complex autopsy procedures. Despite its potential, the detection of diatoms is not often employed in pathological laboratories because of its complexity and the possibility of contamination. In addition, pollution of waterways reduces the presence of diatoms.

Further discussion relevant to drowning is found in Chapters 22 to 25.

‘Dry’ Drowning

This misnomer continues to be reported, despite the absence of experimental and clinical support for drowning without aspiration. The proponents originally quoted the observations of Charles Cot7, a Belgian doctor who reported in the French literature in 1931. He observed the ‘dry’ lungs of dogs fished from the Seine. Because there was no reason to believe that the dogs had drowned, as opposed to being disposed of in the water after death, this support was dubious.

The colourful terminology ensured the popularity of ‘dry drowning’, and many clinicians observed that persons who drowned may not have had obvious water in the lungs at autopsy. This was attributed to laryngospasm caused by asphyxia, continuing until death. Virtually every review of drowning over the rest of that century acknowledged this concept, without question, although the incidence was often increased to 20 to 40 per cent. It did conflict with the findings of earlier animal experiments.

Clinicians who dealt with sea water drownings, such as in scuba divers, never witnessed this paradox – indeed, they marveled at the degree of foam in the lungs and airways, whereas clinicians who dealt with fresh water drownings were much more enthusiastic about the ‘dry drowning’ pathological observation. Now the incidence of presumed ‘dry’ drownings has sunk to less than 2 per cent, even among the earlier proponents of this concept. It is a pathological entity, not a clinical one.

As it has been stated, fresh water is absorbed very rapidly from the lungs after death, and therefore autopsy findings cannot be used to imply (let alone prove) the absence of a previous aspirant. This is especially so when these investigations are performed sometime after the event or after cardiopulmonary resuscitation. ‘Dry drowning’ is probably an artefact of fluid absorption from the lungs, or it may indicate death from other causes. The deleterious effects of the aspiration can proceed even after the absorption of the fluid.

In the absence of more information, it would be prudent to presume that all victims of near drowning or drowning have aspirated and base one’s first aid and management on this presumption. This is supported by the knowledge that laryngospasm does not usually continue until death, and thus even if it does occur during the drowning process, it will not prevent aspiration as hypoxic death is approached.

Drowning: Animal Experiments

Biochemical and circulatory changes after flooding animals’ lungs with fresh water and sea water. Cl, chloride; K, potassium; Mg, magnesium; Na, sodium.

In the early 1900s many animal experiments conducted both in Europe and North America demonstrated that if an animal was immersed and drowned in water containing chemical traces or dyes, these would spread through the tracheo-bronchial tree to the alveolar surfaces. In the case of fresh water, this was also absorbed into the bloodstream.

A consistent fall in arterial oxygen content was observed, followed by a rise in arterial carbon dioxide and sometimes cardiac arrhythmias.

Swann and his colleagues from Texas3,4, in a series of accurate but misleading experiments, flooded animals’ lungs with fresh or salt water and demonstrated the significant differences between the two, attributable to osmotic pressures. In both cases, flooding of the lungs produced a reduction in PaO2 and pH, with a rise in the arterial partial pressure of carbon dioxide (PaCO2).

Because fresh water was osmotically much weaker than blood, it moved into the bloodstream and produced haemodilution – reducing blood concentrations of proteins, sodium, chloride and so forth. The subsequent reduction in the osmotic pressure of the blood caused haemolysis and a liberation of both haemoglobin and potassium, with resultant metabolic and renal complications, aggravated by hypoxia. Deaths were often cardiac and resulted from ventricular fibrillation.

When, however, the animals’ lungs were flooded with sea water – which has a higher osmotic concentration than blood – water was drawn from the bloodstream into the lungs, thereby producing pulmonary oedema and haemoconcentration. This caused an increase in the haematocrit, blood proteins and electrolytes.

For many years physicians attempted to correct these presumed electrolyte, metabolic and cardiac abnormalities in human drownings, but their cases did not replicate the animal model (Figure 21.1).

Biochemical and circulatory changes after flooding animals’ lungs with fresh water and sea water. Cl, chloride; K, potassium; Mg, magnesium; Na, sodium.
Figure 21.1 Biochemical and circulatory changes after flooding animals’ lungs with fresh water and sea water. Cl, chloride; K, potassium; Mg, magnesium; Na, sodium.

Earlier workers had shown that in dogs that drowned, there were still large volumes of air in the lungs, as there are in humans.

Colebatch and Halmagyi5, working in Australia in 1961, produced an animal model more relevant to the clinical management of patients, by aspiration of only 1 to 3 ml/kg body weight. By using these smaller volumes, these researchers demonstrated the sudden arterial hypoxia, not directly proportional to the amount of fluid inhaled. Pulmonary hypertension, vagal inhibition and reduced compliance were also observed. These investigators demonstrated that the weight of the lungs increased threefold the weight of the instilled sea water. Sea water aspiration usually caused significant pulmonary oedema, but aspirated fresh water was often absorbed from the lungs within 2 to 3 minutes.

Subsequent animal experiments by Modell6 and others, using intermediate volumes of aspirant, demonstrated that shunting of blood was the predominant cause of persistent arterial hypoxaemia, as a result of perfusion of blood through non-ventilated areas of lung. Destruction of lung surfactant in fresh water installation also resulted in alveolar wall damage and pulmonary oedema.