Chapter 112. Diving Medicine and Near Drowning

• Immersion and diving produce physiologic effects from increased hydrostatic pressure and its effects on the behavior of gases.
• When diving with compressed air or other breathing gases, the body takes up extra nitrogen or other inert gas, which must be allowed to leave through the lungs gradually to avoid decompression sickness (DCS). DCS occurs when gas comes out of solution and forms bubbles in body tissues, which causes pain or neurologic symptoms by obstruction of blood flow and initiation of secondary inflammatory events.
• DCS is treated by prompt administration of high-flow O2 and hydration. Oxygen recompression therapy is definitive treatment for symptom resolution and avoiding recurrence, even if delayed 1 or 2 days.
• Arterial gas embolism is usually the result of pulmonary overpressurization during ascent from diving with compressed gas and may occur at shallow depths. Arterial gas embolism is the second leading cause, after drowning, of fatalities in recreational diving.
• Near drowning (ND) is caused by asphyxia underwater, which may or may not be associated with aspiration. The primary injuries are to the brain, heart, lungs, and kidneys.
• In young adults, ND is often associated with drug or alcohol ingestion and may be accompanied by traumatic injuries.
• ND may be complicated by acute respiratory distress syndrome; its onset may be delayed or aggravated by aspiration of gastric contents or other foreign debris and pneumonia.
• ND may be complicated by pneumonia (or sepsis) caused by unusual pathogens present in contaminated water. Prophylactic antibiotic treatment is not indicated.

Recreational activities involving water are enjoyed by millions of swimmers, boaters, and divers of all ages with various degrees of physical skill and judgment. However, the water environment is deceptively hazardous, and too often, swimmers or divers venture into peril with deadly results. In many situations, they ignore their physical limitations or impair their faculties with alcohol or other drugs. In other situations, such as with young children, the encounter with water is unsupervised or unexpected and frequently results in death by drowning. There are millions of recreational divers and many more recreational swimmers in the United States, which translates to thousands of diving accidents, drowning, and near drowning (ND) episodes each year. The exact number of such aquatic incidents and their effect on the health care system are difficult to estimate, but there may be more than 80,000 episodes per year in the United States. Victims often survive the incident only to die hours or days later in the hospital. Thus, the intensive care specialist should be knowledgeable about the clinical consequences and management of diving accidents and ND.

Underwater environments produce physiologic changes as a result of the direct effects of increased hydrostatic pressure and its effects on the physical behavior of gases. Pressure is measured in units of force per area, which is expressed commonly in several convenient forms (Table 112-1). The pressure of the atmosphere at sea level is approximately 760 mm Hg (14.7 pounds per square inch). Underwater, the pressure of the water changes directly with depth, and this pressure must be added to atmospheric pressure to obtain total pressure, usually expressed in atmospheres absolute (ATA). Table 112-1 shows that a seawater column 33 ft deep (fsw) exerts a pressure equivalent to 1 atmosphere at sea level. Thus, a diver immersed at 33 fsw is exposed to a total pressure of 2 ATA.

When diving with compressed breathing gases, for the diver to inhale, the inspired gas pressure must be increased in proportion to the absolute pressure. Therefore, a larger number of gas molecules must be contained in the lungs (or other gas-filled body cavities) to maintain constant volume at a given temperature. The relation of pressure (P), volume (V), temperature (T) to number of moles of gas (n) is described by the ideal gas law:

where R is the universal gas constant. The ideal gas law gives rise to the special gas laws important in diving. Three relevant special gas laws in which one of the three variables is held constant are shown in Table 112-1. In diving the most important of these is the Boyle law, which explains the need to equilibrate pressure inside gas-containing spaces in the body, such as the middle ear space and lungs, to avoid barotrauma of descent (e.g., ear squeeze) or ascent (e.g., gas embolism).

Air and other breathable gases are mixtures of different molecules in which the total pressure is equal to the sum of the partial pressures of each gas. This expresses the Dalton law of partial pressures, which means that each gas in a mixture behaves as though it alone occupies the entire space. The uptake of gases by tissue is determined primarily by diffusion of gas from the alveolar spaces into blood and by transport of gas to tissues by the circulation (perfusion). The amount of a gas dissolved in liquid at any temperature, such as blood or tissues of the body at 37°C (98.6°F), is also proportional to its partial pressure (Henry law). The gas concentration in tissue at equilibrium is related to the partial pressure of the gas multiplied by its solubility coefficient. The physiologic effects of diving gases, such as nitrogen narcosis or the risk of decompression sickness (DCS), generally correlate with the partial pressure of the gas in the tissues of the body.

Water immersion produces adjustments by the respiratory, cardiovascular, renal, and endocrine systems.1 Immersion produces three main physiologic effects: a decrease in thoracic gas volume, an increase in cardiac output, and a diuresis, which may lead to dehydration. These responses are caused by hydrostatic pressure and the high density of water relative to air. During immersion, the blood vessels outside the thorax are supported by water, and an upright body is exposed vertically to a hydrostatic pressure gradient. The pressure around the body compresses the abdomen relative to the thorax, thereby causing negative pressure breathing (approximately –20 cm H2O). The diaphragm is displaced upward, which decreases thoracic gas volume and expiratory reserve volume. The pressure gradient across the diaphragm, coupled to a hydrostatic decrease in venous capacitance in the legs, increases the volume of blood in the thorax, including the heart, by about 20%. Blood volume may be increased further by arterial vasoconstriction if the water is cooler than the thermoneutral point (∼34°C, 93.2°F).

During immersion, cardiovascular distention activates mechanoreceptors that normally respond to hypervolemia. The apparent hypervolemia is detected in the hypothalamus via vagal afferents and leads to an immersion response. This response consists of diuresis and natriuresis, which have time profiles that suggest they operate by different mechanisms. Peak diuresis occurs rapidly (1 to 2 hours), whereas peak natriuresis occurs later (4 to 5 hours). Immersion diuresis but not natriuresis can be prevented by fluid restriction and vasopressin administration. The distention of the heart correlates with urinary sodium excretion. The mechanism of the immersion response is suppression of antidiuretic hormone release, also known as the Gauer-Henry response. The mechanism of natriuresis is related to a decrease in tubular reabsorption of sodium and not to an increase in filtered sodium. Natriuresis involves aldosterone suppression via decreased renin-angiotensin activity, increased release of atrial natriuretic factor, release of renal prostaglandins, and decreased renal sympathetic activity.

The increase in thoracic blood volume in immersion distends the heart and enhances ventricular diastolic filling (preload), which increases cardiac output.2 Cardiac output increases almost entirely due to an increase in stroke volume, which may double. The elevated cardiac output is sustained and not accompanied by an increase in oxygen consumption.

Immersion has important implications for the physiology of the breath-hold.3 In a breath-hold maneuver, the lungs are a reservoir for the exchange of O2 for CO2. During breath-hold in air, mean alveolar partial pressure of O2 (PO2) decreases linearly as a function of the decrease in the mixed venous PO2. As alveolar PO2 decreases, body O2 consumption eventually decreases and anaerobic metabolism increases. CO2 enters the lungs in proportion to pulmonary blood flow and the gradient for CO2 diffusion between mixed venous and alveolar partial pressure of CO2 (PCO2). The CO2 transfer rate is high initially but then decreases because the diffusion gradient decreases as alveolar PCO2 approaches the mixed venous PCO2. Further CO2 production increases mixed venous PCO2, which causes alveolar PCO2 to increase again. When the rising PCO2 causes breathing to resume, the break point is reached. The time to the break point can be extended by maneuvers that lower PCO2 or raise PO2 such as hyperventilation or O2 breathing. Hyperventilation does not increase the body's O2 store appreciably because the increase in alveolar PO2 resulting from a decrease in alveolar PCO2 increases blood O2 content only slightly. Thus, hyperventilation extends breath-hold time, but profound hypoxia may develop before the CO2 reaches the break point.

Alveolar gas exchange during breath-holding is altered by underwater descent when thoracic compression decreases lung volume and the partial pressures of O2, CO2, and N2 increase in the lungs (Fig. 112-1). On descent, alveolar O2 and CO2 concentrations decrease because of greater transfer of those gases to pulmonary capillary blood in relation to inert gas (N2). Alveolar PO2 increases during breath-hold diving compared with simple breath-holding (due to the increase in ambient pressure) and the transfer of CO2 during early descent is opposite normal, that is, CO2 moves from alveoli to pulmonary capillary blood. During ascent, the lung re-expands and alveolar PO2 and PCO2 decrease (as pressure decreases). Near the surface, the alveolar PO2 may be almost as low as mixed venous PO2. In strenuous dives, expansion of hypoxic alveoli during ascent may result in reverse O2 transfer from mixed venous blood to alveoli (see Fig. 112-1). Carbon dioxide retained in the blood during the dive also leaves during ascent as alveolar PCO2 decreases. Carbon dioxide elimination continues after the dive, in addition to the small amount of N2 that entered the blood during the dive.

Hyperventilation before breath-hold diving is a dangerous way to extend the dive. During the dive, compression of the lungs causes the alveolar PO2 to increase. Thus, the primary signal to return to the surface is the PCO2. When the diver hyperventilates before the dive, the arterial PCO2 begins at a lower level, and the time to the breath-hold break point is extended. However, alveolar O2 concentration decreases to lower levels during the longer dive, and life-threatening hypoxia may occur as the diver approaches the surface. Loss of consciousness by this mechanism, known as shallow water blackout, causes many drowning episodes.

The body's response to breath-hold diving can be modified by a diving response induced by apnea and facial immersion, particularly in cold water. This diving response, manifested primarily as bradycardia, is most pronounced in young children. The diving response has been interpreted as an O2-conserving response that redistributes blood flow from organs resistant to hypoxia to organs with continuous O2 needs such as heart and brain. This interpretation is controversial for humans, although a diving response may contribute to survival of children resuscitated successfully after submersion in cold water.

The limitations of breath-hold diving are precluded by underwater breathing apparatuses that provide the diver with a continuous supply of breathing gas at any depth. In shallow water, for instance, 0 to 135 fsw, diving is usually carried out by using compressed air because it is inexpensive and readily available, even at remote locations. Air divers are usually free swimming, that is, they use a self-contained underwater breathing apparatus (SCUBA). The maximum safe depths are considered to be 135 fsw for SCUBA divers and approximately 200 fsw for divers tethered to a safety line.4 Safety concerns are brought about by three factors: nitrogen narcosis, safe decompression, and oxygen toxicity. The problems of greatest importance to the intensive care specialist are caused by decompression and are discussed below. The other problems are covered in standard textbooks on diving medicine.

To dive beyond the practical range of compressed air, special gas mixtures, such as helium-oxygen (heliox) or oxygen-enriched air (nitrox) must be supplied. In heliox operations, inspired oxygen pressure is usually maintained at a constant 0.4 to 1.0 ATA, and the helium may be recycled by gas reconditioning equipment. Surface-supplied heliox diving is expensive but effective and relatively safe to depths of 400 fsw. Most dives deeper than 400 fsw require saturation of the diver with inert gas at the approximate working depth of the dive. Divers can live and work for weeks after saturation and then undergo a single slow decompression.

Principles of Decompression

The use of compressed gases in diving produces physiologic problems during ascent because of the escape of inert gas from tissues. Elimination of inert gas from the body as the pressure decreases during ascent from diving or to altitude is known as decompression. The rate of uptake or elimination of inert gases from the body after a change in pressure is determined primarily by gas solubility in blood and tissue, blood flow, and the volume of tissue. Inert gas exchange usually follows an exponential function with respect to time. Tissues that behave this way are perfusion limited, and the characteristics of their inert gas exchange are defined by a half-time. Because the body tissues receive different amounts of blood flow and nitrogen is more soluble in fatty tissues, half-times for different tissues vary considerably. The principle of multiple tissues was used to calculate the first safe decompression tables, with the assumption that gas bubbles and, hence, DCS would occur only if the tissues were supersaturated to allow a nitrogen partial pressure of about twice the absolute pressure. Modern decompression tables are still based on multiple parallel exponential models, but other concepts have been used with success.5 The variable that most affects inert gas uptake and elimination is tissue perfusion, but diffusion may limit tissue gas exchange under some conditions. Diffusion limitation may occur if the rate of inert gas exchange by perfusion is extremely rapid, if tissue perfusion is very low, or if a tissue barrier such as capillary endothelium has a high diffusion resistance.

Diffusion also may be important when two adjacent tissues have very different rates of perfusion. In such conditions, elimination of inert gas from the faster tissue may allow more inert gas to diffuse from the slower tissue. Thus, a faster tissue may remain supersaturated for longer than expected. Diffusion is also important after gas bubbles have formed in a tissue during decompression. Gas bubbles contain large amounts of N2 gas that can be removed by perfusion only after the N2 diffuses back into the tissue. The rate at which N2 diffuses away from a gas bubble is determined by the bubble surface area, the intrabubble pressure, and the difference in partial pressure between bubble and tissue.

During decompression, bubbles form in specific places in the body called nucleation sites. Microscopic gas nuclei remain stable for a long time at hydrophobic sites in the body and grow into bubbles during decompression. The number of nucleation sites and their location and propensity to form macroscopic bubbles differ according to physiologic conditions. For example, exercise may increase the number of bubbles formed by tribonucleation, a mechanism whereby large negative pressures can generate bubbles by traction between surfaces lubricated by a liquid, such as joints. Bubbles created by tribonucleation may arise from preexisting gas nuclei or by de novo formation.

Decompression Sickness

There is little doubt that gas bubbles cause DCS, but some bubbles are clinically silent. However, there is uncertainty about the exact mechanisms of bubble growth and precisely how bubbles trigger the diverse features of DCS. The factors that govern bubble formation in “aqueous” tissues have been used to develop tables for safe decompression. Decompression tables usually assume that DCS will not occur unless the inert gas tension exceeds some threshold of critical supersaturation. The assumption of such a threshold conflicts with data indicating that DCS occurs as a stochastic event.6 Therefore, even appropriate use of well-tested decompression tables or a decompression computer is associated with a finite risk of DCS for dives of approximately 25 fsw or deeper. However, most serious cases of DCS are due to omitted decompression.

The clinical manifestations of DCS are attributed to growth of bubbles in body tissues. Primary sites of bubble growth are the joint spaces, tendon sheaths, and periarticular tissues, including peripheral nerves. Bubbles in these locations produce pain that is known as type I DCS, or bends. Bubbles also form in the spinal cord and other areas of the central nervous system. This is called type II or serious DCS. Other serious forms of DCS involving the audiovestibular system (staggers) and pulmonary system (chokes), although relatively rare, also occur. The common clinical manifestations of DCS are presented in Table 112-2.

DCS in military and commercial divers occurs with a predominance of type I symptoms. Altitude bends, or dysbarism, is generally accompanied by mild to moderate type I DCS. Tunnel workers have a reported incidence of DCS of 0.7% to 1.5%, primarily consisting of type I symptoms of the knee and lower leg.7 In contrast, some survey data have suggested a predominance of type II symptoms in recreational divers.8 The reasons for this difference in DCS in recreational diving are not clear, although underreporting of type I DCS and overdiagnosis of type II DCS are likely. In some cases, long delays in the diagnosis and treatment of type I DCS may allow more severe manifestations to evolve. Recreational divers are more likely to omit decompression than are professional divers, thus increasing the probability of serious DCS.

One of the most serious forms of DCS is spinal paralysis.9 Spinal cord DCS is not fully understood, but in some animals the injury has been related to intravascular bubbles that form in the low-pressure, epidural venous plexus of the spinal cord.10 Because of its slow blood flow, the plexus is susceptible to bubble formation. Bubble-induced thrombi can obstruct venous outflow, leading to ischemic injury of the spinal cord. Despite evidence for bubble formation in the spinal venous plexus, intravascular formation of bubbles appears to be otherwise uncommon. Most intravascular bubbles probably originate at the tissue-blood interface and stream into the circulation to be absorbed by the lungs. Bubbles arising in the body of the spinal cord (autochthonous) have also been implicated in the etiology of spinal DCS.11

Once bubbles enter the circulation, surface activity at the blood-to-bubble interface produces changes in blood proteins. These include complement activation, activation of coagulation and fibrinolysis, platelet and neutrophil aggregation and destruction, and release of inflammatory mediators. Precise roles for these events have not been determined in DCS pathogenesis.

If a large number of bubbles is released into the venous system during decompression, they may obstruct the pulmonary circulation and cause sore throat, cough, chest pain, and shortness of breath. This syndrome, known as the chokes, if untreated, may lead to cardiovascular collapse and death. Some bubbles may also cross pulmonary capillaries as small gas emboli or pass into the arterial circulation through right-to-left cardiac shunts, for example, patent foramen ovale (PFO).12 However, PFO is present in about 20% of normal individuals, and its presence probably leads to more severe rather than more frequent episodes of DCS.13

Divers with the chokes develop symptoms minutes to hours after decompression, which may be progressive. Physical examination reveals tachypnea, tachycardia, crackles, wheezing, cyanosis, and gasping in severe cases. The chest radiograph may show diffuse pulmonary infiltrates similar to acute respiratory distress syndrome (ARDS). Should bubbles pass into the arterial circulation, neurologic findings may appear, primarily involving cerebral symptoms and signs. Arterial blood-gas determinations often show hypoxemia and respiratory alkalosis.

Pulmonary Barotrauma

Pulmonary barotrauma of ascent, also known as pulmonary overpressurization, is a potentially serious consequence of failure of expanding gas in the lung to escape from alveoli. Overstretching of lung regions may rupture acini or alveoli and cause pulmonary interstitial emphysema. Disruption of the pulmonary parenchyma may cause mediastinal or soft tissue emphysema, pneumopericardium, pneumothorax, or arterial gas emboli (AGE). Pulmonary overinflation occurs during ascent while diving with compressed breathing gases. It is most likely to occur with breath-holding, loss of consciousness underwater, or airway obstruction that traps gas. Pulmonary overinflation also may occur after explosive decompression of aircraft at altitude.

Lung rupture during ascent also depends on physiologic factors such as pulmonary compliance, transpulmonary pressure, and lung volume. Airway closure and air trapping induced by immersion in the upright position may increase the risk of lung overstretching during ascent.14 Once alveolar pressure becomes positive by about 100 cm H2O to that at the mouth, the lung will rupture. During breath-holding at total lung capacity (TLC), the difference between alveolar and ambient pressure is approximately 50 cm H2O. Thus, hydrostatic pressure outside the body during ascent must decrease another 50 cm H2O for the lung to rupture. Assuming a compliance of the lung and chest wall at a TLC of 15 mL/cm H2O, lung volume during ascent must increase by 15 mL × 50 cm H2O, or 750 mL, before the lung ruptures. Using the Boyle law (see Table 112-1), an approximate depth can be determined from which a diver must ascend for pulmonary rupture during a breath-hold at TLC. At the surface, P1 = 1.0 ATA; suppose V1 = 7000 mL. Then, V2 = 7000 − 750, or 6250 mL. Therefore,

and

This calculation illustrates why pulmonary overinflation and AGE occur in shallow water during rapid ascent. The calculation also points out that relative volume changes are greatest at low hydrostatic pressures.

Venous gas emboli (VGE) occur fairly often in compressed gas diving and in diverse clinical settings (see Chap. 27). Asymptomatic venous bubbles occur often in divers during and after decompression. In clinical settings, even careful intravenous infusion of solutions or contrast materials may be accompanied by VGE. If a small volume of gas enters the circulation, such emboli are usually clinically silent. VGE thus create injurious consequences in four situations: (1) obstruction of the heart or major vessels by gas, (2) arterialization of bubbles across the pulmonary vasculature, (3) arterialization of bubbles across a PFO, and (4) physical interactions between air and blood. When venous gas bubbles enter the arterial system, even a small amount of gas can produce substantial morbidity or death.15,16

The lungs effectively filter VGE larger than about 20 μm in diameter;17 this size barrier is not absolute because bubbles of 500 μm may cross the lung in some cases. VGE spillover through the pulmonary system has been demonstrated in animals such as dogs, in which spillover to the pulmonary veins occurred at an intravenous air bubbling rate of 0.3 mL/kg per minute.15 The VGE crossover rate and passage size increases in the setting of a large difference between pulmonary arterial and pulmonary venous pressures. Pulmonary arterioles constrict in response to VGE and pulmonary arterial pressures increases.18 As more VGE mechanically obstruct the vasculature, the gradient of pulmonary arterial versus venous pressure increases, thereby decreasing the filtering effectiveness of the pulmonary capillaries and allowing VGE to pass.19 Anatomic intrapulmonary shunting also decreases the filtering efficiency of the lungs.

Paradoxical air embolism may occur when VGE are arterialized via a PFO in divers or patients.11 Detection of PFO by echocardiography relies on gas microbubbles for ultrasonic contrast. Right-to-left atrial crossover of these bubbles is variable and may require Valsalva or other special maneuver to increase right atrial pressure. Bubble crossover also may occur spontaneously during some phases of the cardiac cycle.20 The probability of paradoxical gas embolism increases in patients who develop pulmonary hypertension and vascular obstruction from VGE.

VGE also become physiologically significant when a large quantity of gas obstructs major vessels at the level of the pulmonary vasculature or the heart. In the heart, gas churned to foam obstructs inflow and outflow and diminishes cardiac output. In addition to obstructing the pulmonary vasculature, venous gas triggers pulmonary arterial constriction, bronchospasm, and dyspnea.21

Physical interactions at the blood-bubble interface complicate mechanical obstruction of the circulation and amplify the physiologic effects of small volumes of intravenous gas. The blood-bubble interface stimulates biochemical events associated with release of multiple inflammatory mediators with subsequent damage to the vascular endothelium. Noncardiogenic pulmonary edema may develop quickly as extravasated fluid floods alveoli and may progress to ARDS.22 Compressed gas divers have long recognized a similar sequence of events as the syndrome of chokes.23

AGE may occur in diving by VGE that cross into the arterial circulation or by gas entering the left heart after pulmonary overpressurization. AGE is second to drowning as the leading cause of fatal accidents in SCUBA diving. The most common source of AGE in divers is pulmonary barotrauma. Similar events may occur in patients on mechanical ventilation who require high airway pressures for ventilation. Pulmonary overdistention, which produces AGE in divers or ventilated patients, may not produce other evidence of barotrauma such as pneumothorax, pneumomediastinum, or pneumopericardium. Gas in vessels is not detected reliably by radiography, and the diagnosis is suggested by clinical evidence of end-organ damage, primarily ischemia of the brain or heart.24 Therefore, procedures to document the presence of air may delay treatment of a critically ill patient. Appropriate management of AGE begins with heightened suspicion of the problem in an appropriate setting.

Like VGE, AGE obstruct, induce vasoconstriction, activate coagulation, complement, and neutrophils and aggregate platelets. This leads to release of mediators of inflammation that activate and may subsequently damage vascular endothelium. Even minor AGE have the potential for permanent injury and should never be regarded as benign, but the severity of injury in general appears to be related to the volume of intravascular gas. In the heart, the myocardial response to AGE is similar to coronary insufficiency of any etiology. Ventricular arrhythmias and ST-segment elevation or depression are common and may lead to infarction.25,26 Cerebral manifestations of AGE are similar to those of thromboembolism and include focal neurologic deficits, loss of consciousness, seizures, and death. The distinction between AGE and VGE may become blurred clinically because the possibility of arterial spillover is difficult to exclude, and either event may be heralded by cardiovascular collapse.

Divers who have AGE are most often inexperienced and frequently arrive suddenly at the surface after diving. They may cry out, indicating a rapid ascent with a closed glottis, or they may surface unconscious. AGE are complicated occasionally by pneumothorax or pneumomediastinum and accompanied by clinical signs of pulmonary barotrauma, such as mediastinal crunch (Hamman sign), subcutaneous air, or tension pneumothorax. The onset of AGE is quite sudden, virtually always within 15 minutes of surfacing, and presents with symptoms of brain ischemia, which may be progressive. Initial symptoms of AGE appearing more than 1 hour after ascent should not be attributed to AGE but to serious DCS, ND, or stroke. Common clinical findings of AGE include headache, confusion, nausea, vomiting, blindness, hemiparesis, seizures, and unconsciousness. Chest pain, hemoptysis, and shortness of breath may also occur. Laboratory abnormalities may include elevated serum creatine kinase levels, signifying injury to skeletal and cardiac muscle.27

Adjunctive Therapy

The management of VGE or AGE begins with identifying the problem and preventing additional emboli. In VGE of any etiology, a “mill wheel” murmur produced by air in the right ventricle may be audible with a stethoscope.28 Positioning the patient in the left lateral decubitus and Trendelenburg position in an effort to avoid PFO crossover and brain emboli is traditional clinical practice for AGE and serious DCS.28,29 The Trendelenburg position has been proposed to reduce cerebral bubble diameter by increasing hydrostatic pressure in the cranium; however, animal studies have suggested that positioning efforts are not efficacious because blood flow rather than buoyancy of the bubble determines the course of air emboli.29,30 Further, Trendelenburg positioning has been associated in animals with increased cerebral edema. Ventilation with 100% O2 helps correct hypoxemia and increases the diffusion gradient for nitrogen out of the bubbles, causing them to shrink.31,32 Vasopressors and antiarrhythmic agents may be required for hypotension and ventricular arrhythmias associated with AGE. Lidocaine at 2 to 4 mg/min intravenously after a 1-mg/kg loading dose ameliorates cerebral injury after experimental cerebral AGE.33

Adjunctive therapy for AGE and serious DCS, other than oxygen and lidocaine, consists of correction of hemoconcentration with intravenous fluids and management of complications.34,35 Parenteral corticosteroids such as dexamethasone, 4 mg every 6 hours, have been recommended to decrease cerebral or spinal cord edema after AGE and serious DCS, but evidence demonstrating efficacy in either situation is lacking. Stress ulceration may occur after AGE, and prophylaxis for this complication should be used routinely. Because of blood-bubble physical interactions discussed earlier, treatment of cerebral AGE with heparin was once recommended, but has not proved efficacious, and is associated with hemorrhage into areas of cerebral infarction induced by the air embolus.33 However, heparin prophylaxis of deep venous thrombosis and pulmonary emboli should be given for spinal cord DCS because of immobility of the patient.

Recompression Therapy

Gas bubbles in tissue or circulation slowly resolve spontaneously, but the rate at which they are removed can be enhanced greatly by oxygen breathing and recompression. Recompression and hyperbaric oxygen (HBO) administration are primary therapy for DCS and AGE.34 Bubble resolution is related to size and the partial pressure difference between the bubble and tissue. This partial pressure difference is due to the inherent unsaturation of venous blood that results from the difference in solubility of CO2 and O2 in body tissues. CO2 is 20 times more soluble than is O2 in tissues. As O2 is consumed, it is replaced by CO2 from substrate oxidation at a ratio of about 0.8 mol of CO2 for each mole of O2. Thus, O2 entering tissue at an arterial PO2 of 100 mm Hg leaves the venous capillary at a PO2 of 40 mm Hg. In contrast, CO2 enters at 40 mm Hg and leaves at only 46 mm Hg. The remaining gas pressure in tissues is primarily N2, which is inert and has the same partial pressure at equilibrium, 573 mm Hg, on both sides of the circulation. Therefore, the sum of partial pressures in the venous system is 54 mm Hg less than in the arterial system. This “oxygen window” provides a pressure gradient to eliminate N2. As a gas pocket or bubble collapses from removal of O2 by metabolism, the internal partial pressure of N2 (PN2) increases above tissue PN2 because total pressure in the bubble remains in equilibrium with ambient pressure. In this way, N2 is gradually absorbed. The oxygen window can be expanded during and after decompression and during recompression by administration of high values of PO2, which decreases tissue PN2 and increases the partial pressure gradient for N2 between bubble and tissue. Immediate O2 administration at the scene of a diving accident before recompression therapy is clearly beneficial in the management of DCS and AGE.

Different recompression schedules are effective for AGE and DCS if treatment is begun promptly. Recompression tables that use HBO and minimal recompression are the treatment of choice. U.S. Navy Treatment (USNT) Table 6, which uses intermittent HBO at a maximum depth of 2.8 ATA (60 fsw), is a standard for primary DCS and for recurrent symptoms.4 Recurrent or persisting symptoms should be retreated with USNT Table 6 once a day or with USNT Table 5 twice a day; in general, a point of diminishing returns is reached after a small number (approximately six to eight) treatments. Persistent bladder and bowel dysfunction or neuromuscular weakness also may respond to saturation therapy (e.g., USNT Table 7), if resources and expertise are available for conducting prolonged treatments.4 In some cases of DCS, oxygen recompression is not available, and air recompression tables must be used. Air treatment tables are longer and less effective than HBO tables and are more likely to produce DCS in the chamber attendants. Recompression presents several logistical challenges for the intensive care team.36

HBO is also used to treat AGE and clinically serious VGE on the basis of the following rationale. First, an increase in ambient pressure compresses the gas and increases the ratio of surface area to volume of the bubble, thus improving nitrogen diffusion from the bubble. Second, greater oxygen tension in the blood increases the nitrogen gradient across the blood-bubble interface. Third, the amount of O2 physically dissolved in plasma during HBO (∼6 mL O2/dL of plasma at 2.8 ATA) may avert or reverse ischemia if plasma streaming occurs around emboli, which obstruct erythrocyte flow.

The optimal depth and duration of HBO treatment of AGE are still debatable. Experience by the U.S. Navy with air embolism after military diving accidents led to the development of USNT Table 6A, which uses initial recompression of the patient breathing air to an equivalent pressure of 165 fsw (6 ATA) followed by prompt decompression to 60 fsw (2.8 ATA), when the patient breathes 100% O2. The chamber is then decompressed slowly to the surface while the patient breathes O2 in cycles alternating with air. The intent of the initial, deep compression in USNT Table 6A is reduction of emboli size to 55% of their original diameter (assuming spherical geometry). This approach may be appropriate for very recent AGE or when the quantity of intravascular gas is very large. Deep compression on air, however, may actually allow nitrogen into the emboli so that they are larger after decompression than before compression. This problem can be circumvented in part by having the patient breathe a mixture of 50% O2 and 50% N2 during compression to decrease the driving force of N2 into the emboli. Of note, recent experience using USNT Table 6 to treat air embolism suggests that equally good results are obtained when using USNT Table 6 at 2.8 ATA. This compresses the emboli to about 70% of their original diameter and expedites release of nitrogen. USNT Table 6, originally designed for the treatment of type II DCS, is more amenable to hospital-based chambers and is particularly well suited to critically ill patients with AGE at sea level pressure. Like DCS, recurrent or persisting symptoms of gas embolism may respond to daily USNT Table 6 treatments until no further improvement is noted.

Many hospitals may not have the experience or equipment to manage critically ill patients in a hyperbaric chamber. In that case, the patient may require transfer to a suitably equipped facility. This can be accomplished by ground or air transportation as long as the aircraft can be pressurized to 1 ATA or safely fly below 1000 ft of altitude to avoid further enlargement of gas emboli.

The best responses to therapy for AGE and serious DCS occur with prompt treatment. However, even recompression therapy delayed as long as 12 to 24 hours may be useful for both conditions.37 Advice from a diving medicine physician concerning diving-related injuries is available 24 hours a day by calling the Divers Alert Network (919- 684-8111. The Divers Alert Network is a nonprofit organization that collects and disseminates information on diving safety and related injuries and is associated with the Center for Hyperbaric Medicine at Duke University Medical Center.

Pathophysiology

The consequences of ND are primarily those of asphyxia. When effective gas exchange ceases, the individual presents the consequences of hypoxia and hypercapnia. Excellent summaries of the demographics and emergency response to ND are available and are not covered here.38–40 This section summarizes the pathogenesis of ND injury of the frequently injured systems supported in the intensive care unit (ICU): lungs, brain, heart, and kidneys.

Lung

The pathogenesis of lung injury in ND is initiated by upper airway obstruction when fresh water (FW) or sea water (SW) contacts the respiratory tract mucosa and provokes laryngospasm. Laryngospasm is a protective response provided the duration of hypoxemia is limited by a brief immersion time. Approximately 10% to 15% of individuals with ND aspirate trivial amounts of water, but some of these individuals develop sufficient O2 deprivation to produce hypoxic encephalopathy or ventricular arrhythmia due directly to laryngospasm.41 Aspiration of SW and FW also induces mechanical airway obstruction with a small airway component. Small airway obstruction is aggravated by bronchoconstriction, mucosal edema, and plugging by water and suspended debris such as algae, diatoms, sand, mud or by teeth and gastric contents.42

Aspiration of even small quantities of water immediately decreases lung compliance and creates persistent areas of low ventilation-perfusion ratio and shunt.42,43 Therefore, aspiration of water may produce a longer lasting hypoxemia than laryngospasm alone. Some of the mechanisms of the early changes in pulmonary gas exchange have been elucidated in animals and are attributable, in general, to loss of surfactant or its activity, damage to the alveolar epithelium and capillary endothelium, and alveolar flooding. In human beings, vomiting and aspiration of stomach contents during ND is common and aggravates mucosal and alveolar epithelial injury.

The combination of alveolar flooding, loss of surfactant or its function, atelectasis, and alveolar damage may give rise to progressive hypoxemia from intrapulmonary shunting, which in severe cases may reach 70% of the cardiac output.42,44 In about 40% of individuals with ND, the injury culminates in ARDS hours to days after the episode.45,46 Hypoxemia usually necessitates treatment with supplemental oxygen at high inspired O2 fraction, which may superimpose pulmonary oxygen toxicity on ARDS. Fortunately, ARDS, which develops after ND, is more likely to be reversible than ARDS from other causes.46

Brain

ND is the second leading cause of brain death, after trauma, in children admitted to a pediatric ICU.47 Brain injury pathogenesis in ND is that of global anoxia or severe hypoxia. Prolonged anoxia or hypoxia produces diffuse neuronal damage that, if severe, compromises the function of the blood-brain barrier, leading to cerebral edema. As edema develops, intracranial pressure (ICP) may increase, further decreasing brain perfusion, exacerbating intracellular hypoxia, and, in severe cases, causing uncal herniation. Profound increases in ICP are infrequent after ND but tend to appear more than 24 hours after resuscitation of patients who already show evidence of neurologic dysfunction.46 There is some evidence that the increase in ICP indicates severity of neuronal injury rather than being a major contributor to the insult.46,48

The difference between ND and anoxic brain injury of other etiologies involves the potential mitigating effects of the diving reflex and hypothermia.49 In human beings, the diving reflex is not well developed and most apparent in young children exposed to cold water.49,50 ND in cold water leading to rapid hypothermia slows cerebral metabolism, thereby postponing the deleterious effects of anoxia.51 These factors are associated with a better prognosis after an apparently severe ND episode.50–52

Heart

The most important cardiac effects of ND are atrial and ventricular arrhythmias, in particular ventricular fibrillation.53 Studies of drowning in animals have demonstrated hemolysis and rapid shifts in blood electrolyte composition after instillation of FW and SW into the lungs. These responses correlated with the occurrence of ventricular arrhythmias. However, human studies have not confirmed significant electrolyte changes even in patients with ventricular fibrillation,54 except for drowning in Dead Sea waters, which have a much higher mineral content than other SW. Individuals with ND in the Dead Sea develop hypernatremia, hyperchloremia, hypermagnesemia, and hypercalcemia more than 24 hours after exposure because electrolytes are absorbed from the gastrointestinal tract after swallowing large amounts of water during the episode.52

ND in human beings differs from that in animal studies primarily because humans rarely aspirate enough water to produce significant electrolyte changes.54 Human pathologic studies after SW or FW drowning have demonstrated cardiac myocyte hypercontraction and hypereosinophilic sarcomeres. These pathologic changes are characteristic of catecholamine excess and suggest that intense adrenergic stimulation contributes to the arrhythmias after ND.55 Thus, the etiology of ventricular fibrillation in human beings is most likely related to hypoxia, respiratory and metabolic acidosis, and catecholamine excess. A review of cases of children with brain death has demonstrated myocardial infarction to be commonly associated with ND.56

Kidney

The incidence of renal insufficiency after ND is unknown but is reported far less frequently than lung, brain, or cardiac injury. The renal complication cited most often is oliguria attributable to acute tubular necrosis,57 probably caused by hypoxemia and hypotension. Infrequently, ND may be complicated by rhabdomyolysis and hemolysis with disseminated intravascular coagulation.58,59 Hemolysis and disseminated intravascular coagulation also may contribute to acute tubular necrosis. Although patients with acute renal failure after ND may require transient dialysis, recovery of adequate renal function can be expected in most patients.

Management

General Measures

ND in adults and adolescents is associated with several predisposing factors and complicating injuries, which are easily overlooked in the unconscious, critically ill ND patient. They must be considered in each case because they may affect the treatment and prognosis of the patient.

Alcohol and other drugs that alter the central nervous system are commonly implicated in adult ND.60 Sedatives and alcohol in particular may complicate the patient's initial ICU course by exacerbating hypothermia and hypotension and impairing mental status and respiratory drive. Levels of toxic drugs and blood alcohol should be measured in complicated patients admitted to the ICU.

Other predispositions to ND include unrelated pathologic events in otherwise unimpaired adult swimmers. Myocardial infarction, cardiac arrhythmias, seizures, subarachnoid hemorrhage, and AGE in the SCUBA diver have been implicated in many ND episodes.60 These events may require extraordinary diagnostic efforts in the immediate postresuscitation environment of the ICU. Electrocardiography should be obtained routinely because the heart is a target organ for hypoxemia in patients with ND. Serial measurements of cardiac enzymes in ND are useful for confirming the diagnosis of myocardial infarction when electrocardiographic changes are nondiagnostic. Acute intracranial hemorrhage or status epilepticus may need to be ruled out in patients whose course is complicated by neurologic dysfunction. Recompression and HBO therapy should be considered for ND in SCUBA divers with unexplained obtundation, coma, or other neurologic deficit (see section on VGE and AGE).

Injuries to the spine and skull are common in patients with ND. These occur most often when a swimmer dives into shallow water and strikes the head on the bottom or on a submerged object.60 A common scenario involves a motor vehicle accident that leaves the passenger submerged in a body of water. Burst fractures of the cervical vertebrae resulting in tetraplegia have been reported in these settings. In addition, skin, middle ear, or sinus trauma sustained during ND may serve as portals of entry for infection.61

ICU Management

The ND patient should be placed in an ICU for respiratory insufficiency, after cardiac arrest or arrhythmia, and for altered mental status. Clinically, the patient may exhibit cyanosis, tachycardia, hypo- or hypertension, hypothermia, respiratory distress with frothy, blood-tinged sputum, diffuse crackles, and wheezing on examination. Initial laboratory evaluation often shows a metabolic acidosis (caused by lactic acid) and hypoxemia on arterial blood-gas analysis. Serum electrolytes, with the exception of a decreased bicarbonate concentration, are rarely perturbed significantly.62 However, ND in industrial fluids or the Dead Sea can perturb serum electrolytes. As an example, a 28-year-old male developed severe hypercalcemia after a ND incident in an industrial vessel with oil-well drilling fluid that contained calcium salts.63 Hypoglycemia is common.64 Hemolysis and rhabdomyolysis, if seen, tend to occur early in the clinical course unless they are the result of late sepsis. Electrocardiographic abnormalities include evidence of ischemia or injury and ventricular and atrial arrhythmias. Initial chest radiographic findings range from patchy infiltrates to diffuse airspace disease (Fig. 112-2). Progressive increase in parenchymal infiltrates over hours to days is not unusual.

Respiratory Care

Mechanical ventilation can be a challenge in the severely injured ND patient. Atelectasis and pulmonary edema with intrapulmonary shunting are encountered in all types of ND. ARDS can develop with poor lung compliance that should be treated, like all cases of ARDS, with normal tidal volume ventilation (6mL/kg ideal body weight; see Chap. 38). These conditions can be further complicated by the presence of foreign bodies in the airways. Heavy sedation or muscle relaxants are best avoided in ND patients because the drugs impair the clinician's ability to follow the neurologic examination. However, careful use of these agents may improve mechanical ventilation by synchronizing the patient with the ventilator or decreasing airway pressures and the risk of barotrauma. The advantages of positive end-expiratory pressure in decreasing intrapulmonary shunt can be dramatic in severe hypoxemia after ND. Treatment of children with ARDS after ND with artificial surfactant did not improve outcome but did improve pulmonary function modestly.65

Respiratory insufficiency in ND may be complicated by factors other than atelectasis and intrapulmonary shunt. Airway obstruction may occur as bronchospasm or from a foreign body. Bronchodilator therapy may benefit the patient with diffuse wheezing. Patients with localized atelectasis that fails to improve with effective ventilation or those who exhibit localized wheezing should undergo fiberoptic bronchoscopy to exclude or remove a foreign body.

Many ND accidents occur in water contaminated with human or animal waste or naturally containing pathogenic bacteria or fungi. The lung is the most common portal of entry for a wide variety of these organisms. Infection is heralded by fever 2 to 7 days after the event and should prompt sputum and blood cultures before initiation of antibiotic therapy.66 Prophylactic antibiotic coverage has not improved outcome after ND, and routine use of antibiotics is not indicated. Reported infections are presented in Table 112-3. Awareness of infection by more unusual organisms is crucial because they may have specific culture requirements not offered routinely in many hospital microbiology laboratories.61

Brain Resuscitation

Brain resuscitation measurements after ND are controversial. The debate is complicated by the diverse parameters that influence brain injury and recovery, including age, water temperature and submersion time (or period of asphyxia), coexisting injuries, and preexisting disease.67 The matter is complicated further by anecdotes of complete or nearly complete neurologic recovery in association with specific therapeutic modes after prolonged immersion. The initial prognostic uncertainties about brain recovery after ND mandate a full effort at cardiopulmonary resuscitation, including correction of hypothermia.68

Classification of neurologic status 1 to 2 hours after resuscitation allows assessment of prognosis. A system with reasonable discrimination for outcome in humans classifies patients after resuscitation into three categories, as listed in Table 112-4.50,51

The best discrimination for outcome using this system has been found in children in whom all category A and B patients (n = 57) recovered completely. Level C patients (n = 39) had 33.3% and 23.9% cerebral morbidity rates (mortality rate + morbidity rate = 56.2%), with the lowest survival rate in the C.3 group.51 In another series of patients that included 52 adults, the category A patients recovered completely, and two adults and one child in level B eventually succumbed to barotrauma or other complications.50 Eight of 11 (73%) adult and 8 of 18 (44%) child category C patients recovered completely in that series.

Use of a brain resuscitation regimen known as HYPER (hyperhydration, hyperpyrexia, hyperexcitability, and hyper-rigidity) suggested that a larger fraction of seriously injured ND children had complete recovery.51 The acronym addressed the overhydration, fever, excitability, and muscular rigidity noted in some patients. These findings were thought to affect outcome adversely, and aggressive therapy was recommended to minimize them.

HYPER therapy consists of systemic corticosteroids, osmotic diuretics, hyperventilation, barbiturate coma, and muscle relaxants administered to minimize cerebral edema and decrease ICP. Controlled hypothermia (32°C, 89.6°F) to decrease neuronal metabolism also was advocated.52 ICP monitoring is necessary to guide such aggressive therapy. The rationale for HYPER therapy was based on the idea that control of ICP would minimize neuronal damage after diffuse anoxia. However, the increase in ICP that develops 24 to 48 hours after ND may be a result of severe neuronal injury rather than its cause.

Critical reviews and subsequent experience with HYPER therapy failed to confirm its efficacy and highlighted its potentially detrimental effects.41,46,68 A retrospective review of 40 ND patients from the same institution that reported the original experience with HYPER found increased incidences of sepsis and multiple organ failure in patients treated with hypothermia.67 This may result from cold-induced immune suppression (including neutropenia) complicated by cold-induced bronchorrhea and decreased mucociliary clearance.67

Corticosteroids have no proved benefit in decreasing brain edema associated with trauma, intracerebral hemorrhage, or stroke. In the absence of convincing evidence that corticosteroids increase edema and, hence, ICP and that decreased ICP improves neurologic outcome after ND, these agents should be avoided because they are immunosuppressive and predispose to infection and gastric ulceration.69 Although hypothermia and barbiturates can decrease ICP in some circumstances, their use has not been demonstrated to improve neurologic outcome. Attempts to decrease ICP by use of osmotic agents also have not been shown to improve neurologic outcome after ND and may induce hyperosmolarity and renal insufficiency. Mild hyperventilation is a comparably benign intervention to decrease ICP temporarily, and ICP monitoring may help direct therapy in the subset of ND patients with increased ICP and poor prognosis. More recent experience with aggressive cerebral monitoring and resuscitation of the ND patient has shown no benefit.70

Prognosis

Overall, about 80% of child and adult ND victims recover completely, 8% to 10% survive but with brain damage, and 10% to 12% die. About 90% of category A and B and approximately 50% of category C patients survive with full recovery, whereas 10% to 23% of the later group survive but have permanent neurologic sequelae.45,46,50,54 Thus, respiratory insufficiency in the absence of sepsis or infection is seldom the cause of death in ND patients in hospitals with modern intensive care capabilities.

Many parameters such as serum electrolytes, arterial blood-gas and pH values, electroencephalographic findings or clinical features (body temperature, absence of pupillary response, cardiac arrest, duration of submersion, and resuscitative efforts), and cross-brain oxygen content differences71 have been examined for use as indicators of prognosis. None is sufficiently discriminating to guide early therapy. Conversely, the presence of cardiac arrest and absence of spontaneous respirations after resuscitation are ominous signs associated with permanent neurologic impairment or death.68 In a retrospective review of 44 ND children, all survivors who regained good neurologic function, were awake with purposeful motion 24 hours after the incident.72

Therapeutic flooding of the whole lung is a form of controlled ND used for many years to treat alveolar filling diseases including lipoid pneumonia, cystic fibrosis, and pulmonary alveolar proteinosis (PAP). Whole lung lavage is currently used primarily for PAP, a rare diffuse lung disease of unknown etiology.73 In PAP, a combination of lipoproteins and noninflammatory cellular debris accumulates in the alveoli and creates a physiologic picture of low ventilation-perfusion ratio that may progress to severe shunting and refractory hypoxemia. The disease affects men twice as often as women and begins commonly in the fourth to sixth decade of life. Occasional onset in childhood is also well documented. Approximately 33% of patients with PAP spontaneously remit, 33% do not progress, and 33% progress to pulmonary fibrosis that may be fatal.

Indications for whole lung lavage include resting hypoxemia, for instance, less than 90% hemoglobin oxygen saturation or arterial desaturation with routine activities. The frequency of need for lavage is variable; some patients require only one, whereas others require the procedure every few months for extended periods. Therapeutic whole lung lavage with warm saline is more effective for physical removal of the debris than segmental lavage via a bronchoscope in patients with PAP. The efficiency of protein clearance is increased by manual percussion of the chest during the emptying phases of the lavage but not by additives such as proteolytic enzymes or heparin in the lavage fluid.74 Complications of the procedure include inadvertent flooding of the contralateral lung, hypoxemia, arrhythmias, hydropneumothorax (<5%), and bronchitis or pneumonia (<5%).75,76

Technique

Whole lung lavage should be performed under general anesthesia by an experienced team. Standard monitoring during the procedure includes pulse oximetry, electrocardiography, blood pressure by indwelling radial artery catheter, core temperature, and volume of exchange in and out of the lung.74,77 The trachea and left main bronchus are intubated with a dual-lumen endotracheal tube (Fig. 112-3). Although intubation of the left main bronchus is technically more difficult than that of the right main bronchus, the longer left mainstem before the first bifurcation provides better mechanical stability of the dual-lumen tube and less chance of bronchial obstruction. Proper placement of the tube is crucial for success in separating the lungs for independent lavage and ventilation. Tube placement is verified anatomically by fiberoptic bronchoscopy and functionally by auscultation and inflation of one lung to 50 cm H2O to ensure that the contralateral main bronchus has been isolated. This is accomplished by connecting the appropriate bronchial tube to a connecting tube to a beaker of saline and observing for bubbles. Although a few bubbles are seen initially from compression by the pressurized lung, persisting bubbles indicate a leak and incomplete separation of the lungs.

Some patients with PAP are referred for lavage with advanced lung disease and severe hypoxemia despite supplemental O2. In these cases, the procedure may be performed in a large hyperbaric chamber by a team composed of an anesthesiologist, a pulmonologist, a nurse and a physical therapist. This permits HBO to be used to maintain adequate arterial oxygenation during the procedure, if necessary.78

In many patients with PAP, hypoxemia worsens during the lavage procedure, particularly during the emptying phases of the lavage. This occurs because pulmonary blood flow to the lavaged lung increases when the hydrostatic compression of the pulmonary capillaries is alleviated by lung emptying. Therefore, it is useful to simulate the emptying phase of the lavage before the procedure by ventilating the lung to be lavaged with 6% O2 (approximate mixed venous PO2) and the other lung with 100% O2 for 5 minutes. Hemoglobin saturation is monitored by pulse oximetry, and an arterial PO2 at the nadir of the response is used to confirm the level of hypoxemia. Hemoglobin saturations below 85% (PO2 <55 mm Hg) during this maneuver indicate that lavage will be more safely performed under hyperbaric conditions (1.2 to 2.8 ATA).

The first phase of the procedure is to wash N2 out of the lungs by ventilating the patient with 100% O2. This prevents relatively insoluble and non-metabolizable N2 from creating an “air lock” in the lavaged lung. The lung is then flooded with normal saline at 37°C (98.6°F) at a rate based on one-half the patient's O2 uptake (1 to 2 mL O2/kg body weight per minute) through a tube connected to a reservoir of saline 45 to 50 cm above the mid-axillary line. Prior lung volume measurements are used to estimate the volume of fluid that the lung will hold. If HBO is required, compression begins during flooding as the hemoglobin saturation decreases. Compression of the chamber continues until hemoglobin saturation can be maintained at 90%.

After filling the lung, the hemithorax is percussed vigorously as 400 to 500 mL of fluid is drained by gravity into a graduated receiver 40 to 45 cm below the patient. Some investigators have advocated nearly complete emptying of the lung with each aliquot, but smaller tidal aliquots minimize the hypoxemia associated with the emptying phases of the lavage.77,78 After emptying, repeated tidal aliquots of saline are instilled and emptied until the effluent is nearly clear to inspection. This usually requires 10 to 25 L of fluid per lung. The efficiency of this method with respect to yield of wet debris is comparable to that achieved with nearly complete emptying of the lung (Fig. 112-4). Afterward, the lung is drained and suctioned before reinflating with 100% O2 and administration of positive end-expiration pressure. The availability of HBO allows the patient after one lung lavage to remain well oxygenated during the contralateral procedure. Hence, both lungs can be lavaged in sequence on the same day. This allows a single period of general anesthesia and a shorter hospitalization for the patient. Some institutions require a 2- to 4-day recovery period after one lavage before performing the procedure on the second lung. After the procedure, the patient is weaned rapidly from 10 cm H2O positive end-expiration pressure and a high inspired O2 fraction in the ICU and usually can be extubated within 12 hours.