Decompression theory is the study and modelling of the transfer of the inert gas component of breathing gases from the gas in the lungs to the tissues and back during exposure to variations in ambient pressure. In the case of underwater diving and compressed air work, this mostly involves ambient pressures greater than the local surface pressure, but astronauts, high altitude mountaineers, and travellers in aircraft which are not pressurised to sea level pressure, are generally exposed to ambient pressures less than standard sea level atmospheric pressure. In all cases, the symptoms caused by decompression occur during or within a relatively short period of hours, or occasionally days, after a significant pressure reduction.
Decompression in the context of diving derives from the reduction in ambient pressure experienced by the diver during the ascent at the end of a dive or hyperbaric exposure and refers to both the reduction in pressure and the process of allowing dissolved inert gases to be eliminated from the tissues during this reduction in pressure.
When a diver descends in the water column the ambient pressure rises. Breathing gas is supplied at the same pressure as the surrounding water, and some of this gas dissolves into the diver's blood and other fluids. Inert gas continues to be taken up until the gas dissolved in the diver is in a state of equilibrium with the breathing gas in the diver's lungs, (see: "Saturation diving"), or the diver moves up in the water column and reduces the ambient pressure of the breathing gas until the inert gases dissolved in the tissues are at a higher concentration than the equilibrium state, and start diffusing out again.
The absorption of gases in liquids depends on the solubility of the specific gas in the specific liquid, the concentration of gas, customarily measured by partial pressure, and temperature. The main variable in the study of decompression theory is pressure.
Once dissolved, distribution of the dissolved gas may be by diffusion, where there is no bulk flow of the solvent, or by perfusion where the solvent (blood) is circulated around the diver's body, where gas can diffuse to local regions of lower concentration. Given sufficient time at a specific partial pressure in the breathing gas, the concentration in the tissues will stabilise, or saturate, at a rate depending on the solubility, diffusion rate and perfusion.
If the concentration of the inert gas in the breathing gas is reduced below that of any of the tissues, there will be a tendency for gas to return from the tissues to the breathing gas. This is known as outgassing, and occurs during decompression, when the reduction in ambient pressure or a change of breathing gas reduces the partial pressure of the inert gas in the lungs.
The combined concentrations of gases in any given tissue will depend on the history of pressure and gas composition. Under equilibrium conditions, the total concentration of dissolved gases will be less than the ambient pressure, as oxygen is metabolised in the tissues, and the carbon dioxide produced is much more soluble. However, during a reduction in ambient pressure, the rate of pressure reduction may exceed the rate at which gas can be eliminated by diffusion and perfusion, and if the concentration gets too high, it may reach a stage where bubble formation can occur in the supersaturated tissues. When the pressure of gases in a bubble exceed the combined external pressures of ambient pressure and the surface tension from the bubble - liquid interface, the bubbles will grow, and this growth can cause damage to tissues. Symptoms caused by this damage are known as Decompression sickness.
The actual rates of diffusion and perfusion, and the solubility of gases in specific tissues is not generally known, and it varies considerably. However mathematical models have been proposed which approximate the real situation to a greater or lesser extent, and these models are used to predict whether symptomatic bubble formation is likely to occur for a given dive profile.
Two rather different concepts have been used for decompression modelling. The first assumes that dissolved gas is eliminated while in the dissolved phase, and that bubbles are not formed during asymptomatic decompression. The second, which is supported by experimental observation, assumes that bubbles are formed during most asymptomatic decompressions, and that gas elimination must consider both dissolved and bubble phases.
Early decompression models tended to use the dissolved phase models, and adjusted them by more or less arbitrary factors to reduce the risk of symptomatic bubble formation. Dissolved phase models are of two main groups: Parallel compartment models, where several compartments with varying rates of gas absorption (half time), are considered to exist independently of each other, and the limiting condition is controlled by the compartment which shows the worst case for a specific exposure profile. These compartments represent conceptual tissues and are not intended to represent specific organic tissues, merely to represent the range of possibilities for the organic tissues. The second group uses serial compartments, where gas is assumed to diffuse through one compartment before it reaches the next.
A recent variation on the serial compartment model is the Goldman interconnected compartment model (ICM).
More recent models attempt to model bubble dynamics, also by simplified models, to facilitate the computation of tables, and later to allow real time predictions during a dive. The models used to approximate bubble dynamics are varied, and range from those which are not much more complex that the dissolved phase models, to those which require considerably greater computational power.
None of the decompression models can be shown to be an accurate representation of the physiological processes, although interpretations of the mathematical models have been proposed which correspond with various hypotheses. They are all approximations which predict reality to a greater or lesser extent, and are acceptably reliable only within the bounds of calibration against collected data.
- 1 Physics and physiology of decompression
- 1.1 Solubility
- 1.2 Diffusion
- 1.3 Inert gas uptake (Ingassing)
- 1.4 Inert gas elimination (Outgassing)
- 1.5 Bubble formation, growth and elimination
- 1.6 Isobaric counterdiffusion (ICD)
- 1.7 Doppler ultrasonic bubble detection
- 1.8 Decompression sickness and injuries
- 1.9 Factors influencing uptake and elimination of dissolved gases and decompression risk
- 2 Decompression models
- 2.1 Validation of models
- 2.2 Residual inert gas
- 2.3 Deterministic models
- 2.3.1 Perfusion limited tissues and parallel tissue models
- 2.3.2 Diffusion limited tissues and the "Tissue slab", and series models
- 2.3.3 Bubble models
- 2.4 Probabilistic models
- 3 See also
- 4 References
- 5 Further reading
- 6 External links
Physics and physiology of decompression
Decompression involves a complex interaction of gas solubility, partial pressures and concentration gradients, diffusion, bulk transport and bubble mechanics in living tissues.
This section provides an introductory discussion of some of the factors influencing inert gas uptake and elimination in living tissues.
Solubility is the property of a gas, liquid or solid substance (the solute) to be held homogeneously dispersed as molecules or ions in a liquid or solid medium (the solvent).
In decompression theory the solubility of gases in liquids is of primary importance.
Solubility of gases in liquids is influenced by three main factors:
- The nature of the solvent liquid and the solute gas
- Temperature (gases are less soluble in water but may be more soluble in organic solvents, at higher temperatures.)
- Pressure (solubility of a gas in a liquid is proportional to the partial pressure of the gas on the liquid – Henry's Law)
The presence of other solutes in the solvent can also influence solubility.
|Gas||Molecular weight||Water solubility||Lipid solubility||Water/lipid solubility ratio|
Diffusion is the movement of molecules or ions in a medium when there is no gross mass flow of the medium, and can occur in gases, liquids or solids, or any combination.
Diffusion is driven by the kinetic energy of the diffusing molecules – it is faster in gases and slower in solids when compared with liquids due to the variation in distance between collisions, and diffusion is faster when the temperature is higher as the average energy of the molecules is greater. Diffusion is also faster in smaller, lighter molecules of which helium is the extreme example. Diffusivity of helium is 2.65 times faster than nitrogen.
In decompression theory the diffusion of gases, particularly when dissolved in liquids, is of primary importance.
Partial pressure gradient
Also known as concentration gradient, this can be used as a model for the driving mechanism of diffusion. The partial pressure gradient is the variation of partial pressure (or more accurately, the concentration) of the solute (dissolved gas) from one point to another in the solvent. The solute molecules will randomly collide with the other molecules present, and tend over time to spread out until the distribution is statistically uniform. This has the effect that molecules will diffuse from regions of higher concentration (partial pressure) to regions of lower concentration, and the rate of diffusion is proportional to the rate of change of the concentration.
Molecules of solute will also tend to aggregate in areas of greater solubility in a non-homogeneous solvent medium.
Inert gas uptake (Ingassing)
In this context, inert gas refers to a gas which is not metabolically active. Atmospheric nitrogen (N2) is the most common example, and helium (He) is the other inert gas commonly used in breathing mixtures for divers.
Atmospheric nitrogen has a partial pressure of approximately 0.78bar at sea level. Air in the alveoli of the lungs is diluted by saturated water vapour (H2O) and carbon dioxide (CO2), a metabolic product given off by the blood, and contains less oxygen (O2) than atmospheric air as some of it is taken up by the blood for metabolic use. The resulting partial pressure of nitrogen is about 0,758bar.
At atmospheric pressure the body tissues are therefore normally saturated with nitrogen at 0.758bar (569mmHg). At increased ambient pressures due to depth or habitat pressurisation, a diver's lungs are filled with breathing gas at the increased pressure, and the partial pressures of the constituent gases will be increased proportionately.
- For example: At 10 meters sea water (msw) the partial pressure of nitrogen in air will be 1.58 bar.
The inert gases from the breathing gas in the lungs diffuse into blood in the alveolar capillaries ("move down the pressure gradient") and are distributed around the body by the systemic circulation in the process known as perfusion.
Perfusion is the mass flow of blood through the tissues. Dissolved materials are transported in the blood much faster than they would be distributed by diffusion alone (order of minutes compared to hours).
The dissolved gas in the alveolar blood is transported to the body tissues by the blood circulation. There it diffuses through the cell membranes and into the tissues, where it may eventually reach equilibrium. The greater the blood supply to a tissue, the faster it will reach equilibrium with gas at the new partial pressure.
Saturation and supersaturation
If the supply of gas to a solvent is unlimited, the gas will diffuse into the solvent until there is so much dissolved that equilibrium is reached and the amount diffusing back out is equal to the amount diffusing in. This is called saturation.
If the external partial pressure of the gas (in the lungs) is then reduced, more gas will diffuse out than in. This is a condition known as supersaturation. The gas will not necessarily form bubbles in the solvent at this stage.
Most decompression models work with slow and fast tissue compartments. These are imaginary tissues which are designated as fast and slow to describe the rate of saturation. Real tissues will also take more or less time to saturate, but the models do not need to use actual tissue values to produce a useful result. Models with from one to 16 tissue compartments have been used to generate decompression tables.
- For example: Tissues with a high lipid content can take up a larger amount of nitrogen, but often have a poor blood supply. These will take longer to reach equilibrium, and are described as slow, than tissues with a good blood supply and less capacity for dissolved gas, which are described as fast.
Tissue half times
Half time of a tissue is the time it takes for the tissue to take up or release 50% of the difference in dissolved gas capacity at a changed partial pressure. For each consecutive half time the tissue will take up or release half again of the cumulative difference in the sequence ½, ¾, 7/8, 15/16, 31/32, 63/64 etc. The number of half times chosen to assume full saturation depends on the decompression model, and typically ranges from 4 (93.75%) to 6 (98.44%).
- For example: A 5 minute tissue will be 50% saturated in 5 minutes, 75% in 10 minutes, 87.5% in 15 minutes and for practical purposes, saturated in about 30 minutes (98.44% saturated at 6 half times)
Tissue compartment half times range from 1 minute to 720 minutes or more in current decompression models.
A specific tissue compartment will have different half times for gases with different solubilities and diffusion rates. This model may not adequately describe the dynamics of outgassing if it includes gas phase bubbles.
Outgassing of tissues
Gas remains in the tissues until the partial pressure of that gas in the lungs is reduced sufficiently to cause a concentration gradient with the blood at a lower concentration then the relevant tissues. A lowered partial pressure in the lungs will result in more gas diffusing out of the blood into the lung gas and less from the lung gas into the blood. A similar situation occurs between the blood and each tissue. As the concentration in the blood drops below the concentration in the adjacent tissue, the gas will diffuse out of the tissue into the blood, and will then be transported back to the lungs where it will diffuse into the lung gas and then eliminated by exhalation.
If the ambient pressure reduction is limited, this desaturation will take place in the dissolved phase, but if the ambient pressure is lowered sufficiently, bubbles may form and grow, both in blood and other supersaturated tissues.
When the gas in a tissue is at a concentration where more diffuses out than in it is called supersaturated, though some authorities define supersaturation in this context as when the partial pressure of inert gas dissolved in a tissue exceeds the total ambient pressure on the tissue, and there is a theoretical possibility of bubble formation.
There is a metabolic reduction of total gas pressure in the tissues.
The sum of partial pressures of the gas that the diver breathes must necessarily balance with the sum of partial pressures in the lung gas. In the alveoli the gas has been humidified by a partial pressure of approximately 63 mbar (47 mmHg) and has gained about 55 mbar (41 mmHg) carbon dioxide from the venous blood. Oxygen has also diffused into the arterial blood, reducing the partial pressure of oxygen in the alveoli by about 67 mbar(50 mmHg) As the total pressure in the alveoli must balance with the ambient pressure, this dilution results in an effective partial pressure of nitrogen of about 758 mb (569 mmHg) in air at normal atmospheric pressure.
At a steady state, when the tissues have been saturated by the inert gases of the breathing mixture, metabolic processes reduce the partial pressure of the less soluble oxygen and replace it with carbon dioxide, which is considerably more soluble in water. In the cells of a typical tissue, the partial pressure of oxygen will drop to around 13 mbar (10 mmHg), while the partial pressure of carbon dioxide will be about 65 mbar (49 mmHg). The sum of these partial pressures (water, oxygen, carbon dioxide and nitrogen) comes to roughly 900 mbar (675 mmHg), which is some 113 mbar (85 mmHg) less than the total pressure of the respiratory gas. This is a significant saturation deficit, and it provides a buffer against supersaturation and a driving force for dissolving bubbles.
Experiments suggest that the degree of unsaturation increases linearly with pressure for a breathing mixture of fixed composition, and decreases linearly with fraction of inert gas in the breathing mixture. As a consequence, the conditions for maximising the degree of unsaturation are a breathing gas with the lowest possible fraction of inert gas – i.e. pure oxygen, at the maximum permissible partial pressure.
Inert gas elimination (Outgassing)
For optimised decompression the driving force for tissue desaturation should be kept at a maximum, provided that this does not cause symptomatic tissue injury due to bubble formation and growth (symptomatic decompression sickness), or produce a condition where diffusion is retarded for any reason.
There are two fundamentally different ways this has been approached. The first is based on an assumption that there is a level of supersaturation which does not produce symptomatic bubble formation and is based on empirical observations of the maximum decompression rate which does not result in an unacceptable rate of symptoms. This approach seeks to maximise the concentration gradient providing there are no symptoms. The second assumes that bubbles will form at any level of supersaturation where the total gas tension in the tissue is greater than the ambient pressure and that gas in bubbles is eliminated more slowly than dissolved gas. These philosophies result in differing characteristics of the decompression profiles derived for the two models: The critical supersaturation approach gives relatively rapid initial ascents, which maximise the concentration gradient, and long shallow stops, while the bubble models require slower ascents, with deeper first stops, but may have shorter shallow stops.
The critical supersaturation approach
Critical ratio model
J.S. Haldane originally used a pressure ratio of 2 to 1 for decompression on the principle that the saturation of the body should at no time be allowed to exceed about double the air pressure. This principle was applied as a pressure ratio of total ambient pressure and did not take into account the partial pressures of the component gases of the breathing air. His experimental work on goats and observations of human divers appeared to support this assumption. However, in time, this was found to be inconsistent with incidence of decompression sickness and changes were made to the initial assumptions. This was later changed to a 1.58:1 ratio of nitrogen partial pressures.
Critical difference models
Further research by people such as Robert Workman suggested that the criterion was not the ratio of pressures, but the actual pressure differentials. Applied to Haldane's work, this would suggest that the limit is not determined by the 1.58:1 ratio but rather by the difference of 0.58 atmospheres between tissue pressure and ambient pressure. Most tables today, including the Bühlmann tables, are based on the critical difference model.
At a given ambient pressure, the M-value is the maximum value of absolute inert gas pressure that a tissue compartment can take without presenting symptoms of decompression sickness. M-values are limits for the tolerated gradient between inert gas pressure and ambient pressure in each compartment. Alternative terminology for M-values include "supersaturation limits", "limits for tolerated overpressure", and "critical tensions".
Gradient factors are a way of modifying the M-value to a more conservative value for use in a decompression algorithm. The gradient factor is a percentage of the M-value chosen by the algorithm designer, and varies linearly between the maximum depth and the surface. They are expressed as a two number designation, where the first number is the percentage of the deep M-value, and the second is a percentage of the shallow M-value.
For example: A 30/85 gradient factor would limit the allowed supersaturation at depth to 30% of the designer's maximum, and to 85% at the surface.
In effect the user is selecting a lower maximum supersaturation than the designer considered appropriate. Use of gradient factors will increase decompression time, particularly in the depth zone where the M-value is reduced the most. Gradient factors may be used to force deeper stops in a model which would otherwise tend to produce relatively shallow stops, by using a gradient factor with a small first number.
Gradient factors produce an M-value which is linearly variable in proportion to ambient pressure.
The critical volume approach
The critical-volume criterion assumes that whenever the total volume of gas phase accumulated in the tissues exceeds a critical value, signs or symptoms of DCS will appear. This assumption is supported by doppler bubble detection surveys. The consequences of this approach depend strongly on the bubble formation and growth model used, primarily whether bubble formation is practicably avoidable during decompression.
This approach is used in decompression models that assume that during practical decompression profiles, there will be growth of stable microscopic bubble nuclei which always exist in aqueous media, including living tissues.
Efficient decompression will minimise the total ascent time while limiting the total accumulation of bubbles to an acceptable non-symptomatic critical value. The physics and physiology of bubble growth and elimination indicate that it is more efficient to eliminate bubbles while they are very small. Models which include bubble phase have produced decompression profiles with slower ascents and deeper initial decompression stops as a way of curtailing bubble growth and facilitating early elimination, in comparison with the models which consider only dissolved phase gas.
The no-supersaturation approach
According to the thermodynamic model of LeMessurier and Hills, this condition of optimum driving force for outgassing is satisfied when the ambient pressure is just sufficient to prevent phase separation (bubble formation).
The fundamental difference of this approach is equating absolute ambient pressure with the total of the partial gas tensions in the tissue for each gas after decompression as the limiting point beyond which bubble formation is expected.
The model assumes that the natural unsaturation in the tissues due to metabolic reduction in oxygen partial pressure provides the buffer against bubble formation, and that the tissue may be safely decompressed provided that the reduction in ambient pressure does not exceed this unsaturation value. Clearly any method which increases the unsaturation would allow faster decompression, as the concentration gradient would be greater without risk of bubble formation.
The natural unsaturation increases with depth, so a larger ambient pressure differential is possible at greater depth, and reduces as the diver surfaces. This model leads to slower ascent rates and deeper first stops, but shorter shallow stops, as there is less bubble phase gas to be eliminated.
Bubble formation, growth and elimination
The location of micronuclei or where bubbles initially form is not known. Heterogeneous nucleation and tribonucleation are considered the most likely mechanism for bubble formation. Homogeneous nucleation requires much greater pressure differences than experienced in decompression. The spontaneous formation of nanobubbles on hydrophobic surfaces is a possible source of micronuclei, but it is not yet clear if these can grow to symptomatic dimensions as they are very stable.
The incorporation of bubble formation and growth mechanisms in decompression models may make the models more biophysical and allow better extrapolation.
Flow conditions and perfusion rates are dominant parameters in competition between tissue and circulation bubbles, and between multiple bubbles, for dissolved gas for bubble growth.
Equilibrium of forces on the surface is required for a bubble to exist. These are:
- Ambient pressure, exerted on the outside of the surface, acting inwards
- Pressure due to tissue distortion, also on the outside and acting inwards
- Surface tension of the liquid at the interface between the bubble and the surroundings. This is along the surface of the bubble, so the resultant acts towards the centre of curvature. This will tend to squeeze the bubble, and is more severe for small bubbles as it is an inverse function of the radius.
- The resulting forces must be balanced by the pressure on the inside of the bubble. This is the sum of the partial pressures of the gases inside due to the net diffusion of gas to and from the bubble.
- The force balance in the bubble may be modified by a layer of surface active molecules which can stabilise a microbubble at a size where surface tension on a clean bubble would cause it to collapse rapidly.
- This surface layer may vary in permeability, so that if the bubble is compressed it may become impermeable to diffusion at sufficient compression.
If the solvent outside the bubble is saturated or unsaturated, the partial pressure will be less than in the bubble, and the surface tension will be increasing the internal pressure in direct proportion to surface curvature, providing a pressure gradient to increase diffusion out of the bubble, effectively "squeezing the gas out of the bubble", and the smaller the bubble the faster it will get squeezed out. A gas bubble can only grow at constant pressure if the surrounding solvent is sufficiently supersaturated to overcome the surface tension or if the surface layer provides sufficient reaction to overcome surface tension.
Clean bubbles that are sufficiently small will collapse due to surface tension if the supersaturation is low. Bubbles with semipermeable surfaces will either stabilise at a specific radius depending on the pressure, the composition of the surface layer, and the supersaturation, or continue to grow indefinitely, if larger than the critical radius.
Bubble formation occurs in the blood or other tissues, possibly in crevices in macromolecules.
A solvent can carry a supersaturated load of gas in solution. Whether it will come out of solution in the bulk of the solvent to form bubbles will depend on a number of factors. Something which reduces surface tension, or adsorbs gas molecules, or locally reduces solubility of the gas, or causes a local reduction in static pressure in a fluid may result in a bubble nucleation or growth. This may include velocity changes and turbulence in fluids and local tensile loads in solids and semi-solids. Lipids and other hydrophobic surfaces may reduce surface tension (blood vessel walls may have this effect). Dehydration may reduce gas solubility in a tissue due to higher concentration of other solutes, and less solvent to hold the gas.
Another theory presumes that microscopic bubble nuclei always exist in aqueous media, including living tissues. These bubble nuclei are spherical gas phases that are small enough to remain in suspension yet strong enough to resist collapse, their stability being provided by an elastic surface layer consisting of surface-active molecules which resists the effect of surface tension.
Once a micro-bubble forms it may continue to grow if the tissues are still supersaturated. As the bubble grows it may distort the surrounding tissue and cause damage to cells and pressure on nerves resulting in pain, or may block a blood vessel, cutting off blood flow and causing hypoxia in the tissues normally perfused by the vessel.
If a bubble or an object exists which collects gas molecules this may reach a size where the internal pressure exceeds the combined surface tension and external pressure and the bubble will grow.
If the solvent is sufficiently supersaturated, the diffusion of gas into the bubble will exceed the rate at which it diffuses back into solution. If this excess pressure is greater than the pressure due to surface tension the bubble will grow. When a bubble grows, the surface tension decreases, and the interior pressure drops, allowing gas to diffuse in faster, and diffuse out slower, so the bubble grows or shrinks in a positive feedback situation. The growth rate is reduced as the bubble grows by the fact that the surface area increases as the square of the radius, while the volume increases as the cube of the radius. If the external pressure is reduced (due to reduced hydrostatic pressure during ascent, for example) the bubble will also grow.
The Variable Permeability Model ordering hypothesis states that nuclei are neither created nor totally eliminated during the pressure cycle, and the initial ordering according to size is preserved. therefore each bubble count is determined by the properties and behaviour of a nominal "critical" nucleus which is at the threshold of bubble-formation – all larger nuclei will form bubbles, and all smaller nuclei will not.
Decompression bubbles appear to form mostly in the systemic capillaries where the gas concentration is highest, often those feeding the veins draining the active limbs. They do not generally form in the arteries, as arterial blood has recently had the opportunity to release excess gas into the lungs. The bubbles carried back to the heart in the veins may be transferred to the systemic circulation via a patent foramen ovale in divers with this septal defect, after which there is a risk of occlusion of capillaries in whichever part of the body they end up in.
Bubbles are also known to form within other tissues, where they may cause damage leading to symptoms of decompression sickness. This damage is likely to be caused by mechanical deformation and stresses on the cells rather than local hypoxia, which is the assumed mechanism in the case of gas embolism of the capillaries.
Bubbles which are carried back to the heart in the veins will normally find their way to the right side of the heart, and from there they will normally enter the pulmonary circulation and eventually pass through or be trapped in the capillaries of the lungs, which are around the alveoli and very near to the respiratory gas, where the gas will diffuse from the bubbles though the capillary and alveolar walls into the gas in the lung. If the number of lung capillaries blocked by these bubbles is relatively small, the diver will not display symptoms, and no tissue will be damaged (lung tissues are adequately oxygenated by diffusion).
The bubbles which are small enough to pass through the lung capillaries may be small enough to be dissolved due to a combination of surface tension and diffusion to a lowered concentration in the surrounding blood, though the Varying Permeability Model nucleation theory implies that most bubbles passing through the pulmonary circulation will lose enough gas to pass through the capillaries and return to the systemic circulation as recycled but stable nuclei.
Bubbles which form within the tissues must be eliminated in situ by diffusion, which implies a suitable concentration gradient.
Isobaric counterdiffusion (ICD)
Isobaric counterdiffusion is the diffusion of gases in opposite directions caused by a change in the composition of the external ambient gas or breathing gas without change in the ambient pressure. During decompression after a dive this can occur when a change is made to the breathing gas, or when the diver moves into a gas filled environment which differs from the breathing gas.
While not strictly speaking a phenomenon of decompression, it is a complication that can occur during decompression, and that can result in the formation or growth of bubbles without changes in the environmental pressure. Two forms of this phenomenon have been described by Lambertsen:
Superficial ICD (also known as Steady State Isobaric Counterdiffusion) occurs when the inert gas breathed by the diver diffuses more slowly into the body than the inert gas surrounding the body.
An example of this would be breathing air in an heliox environment. The helium in the heliox diffuses into the skin quickly, while the nitrogen diffuses more slowly from the capillaries to the skin and out of the body. The resulting effect generates supersaturation in certain sites of the superficial tissues and the formation of inert gas bubbles.
Deep Tissue ICD
Deep Tissue ICD (also known as Transient Isobaric Counterdiffusion) occurs when different inert gases are breathed by the diver in sequence. The rapidly diffusing gas is transported into the tissue faster than the slower diffusing gas is transported out of the tissue.
This can occur as divers switch from a nitrogen mixture to a helium mixture (diffusivity of helium is 2.65 times faster than nitrogen), or when saturation divers breathing hydreliox switch to a heliox mixture.
There is another effect which can manifest as a result of the disparity in solubility between inert breathing gas diluents, which occurs in isobaric gas switches near the decompression ceiling between a low solubility gas (typically helium, and a higher solubility gas, typically nitrogen)
An inner ear decompression model by Doolette and Mitchell suggests that a transient increase in gas tension after a switch from helium to nitrogen in breathing gas may result from the difference in gas transfer between compartments. If the transport of nitrogen into the vascular compartment by perfusion exceeds removal of helium by perfusion, while transfer of helium into the vascular compartment by diffusion from the perilymph and endolymph exceeds the counterdiffusion of nitrogen, this may result in a temporary increase in total gas tension, as the input of nitrogen exceeds the removal of helium, which can result in bubble formation and growth. This model suggests that diffusion of gases from the middle ear across the round window is negligible. The model is not necessarily applicable to all tissue types.
- If the diver is surrounded by or saturated with nitrogen, they should not breathe helium rich gases.
- Gas switches that involve going from helium rich mixtures to nitrogen rich mixtures would be acceptable, but changes from nitrogen to helium should include recompression.
However Doolette and Mitchell's more recent study of Inner Ear Decompression Sickness (IEDCS) shows that the inner ear may not be well-modelled by common (e.g. Bühlmann) algorithms. Doolette and Mitchell propose that a switch from a helium-rich mix to a nitrogen-rich mix, as is common in technical diving when switching from trimix to nitrox on ascent, may cause a transient supersaturation of inert gas within the inner ear and result in IEDCS. They suggest that:
- Breathing-gas switches from helium-rich to nitrogen-rich mixtures should be carefully scheduled either deep (with due consideration to nitrogen narcosis) or shallow to avoid the period of maximum supersaturation resulting from the decompression. Switches should also be made during breathing of the largest inspired oxygen partial pressure that can be safely tolerated with due consideration to oxygen toxicity.
A similar hypothesis to explain the incidence of IEDCS when switching from trimix to nitrox was proposed by Steve Burton, who considered the effect of the much greater solubility of nitrogen than helium in producing transient increases in total inert gas pressure, which could lead to DCS under isobaric conditions.
Burton argues that effect of switching to Nitrox from Trimix with a large increase of nitrogen fraction at constant pressure has the effect of increasing the overall gas loading within particularly the faster tissues, since the loss of helium is more than compensated by the increase in nitrogen. This could cause immediate bubble formation and growth in the fast tissues. A simple rule for avoidance of ICD when gas switching at a decompression ceiling is suggested:
- Any increase in gas fraction of nitrogen in the decompression gas should be limited to 1/5 of the decrease in gas fraction of helium.
This rule has been found to successfully avoid ICD on hundreds of deep trimix dives.
Doppler ultrasonic bubble detection
Doppler bubble detection equipment uses ultrasonic signals reflected from bubble surfaces to identify and quantify gas bubbles present in venous blood. This method was used by Dr Merrill Spencer of the Institute of Applied Physiology and Medicine in Seattle, who published a report in 1976 recommending that the then current no-decompression limits be reduced on the basis that large counts of venous gas bubbles were detected in divers exposed to the US Navy no-decompression limits. These non-symptomatic bubbles have become known as "silent bubbles", and are thought to be nitrogen bubbles released from solution during ascent.
Decompression sickness and injuries
Problems due to vascular decompression bubbles
Bubbles may be trapped in the lung capillaries, temporarily blocking them. If this is severe, the symptom called "chokes" may occur.
If the diver has a patent foramen ovale (or a shunt in the pulmonary circulation), bubbles may pass through it and bypass the pulmonary circulation to enter the arterial blood. If these bubbles are not absorbed in the arterial plasma and lodge in systemic capillaries they will block the flow of oxygenated blood to the tissues supplied by those capillaries, and those tissues will be starved of oxygen. Moon and Kisslo concluded that "the evidence suggests that the risk of serious neurological DCI or early onset DCI is increased in divers with a resting right-to-left shunt through a PFO. There is, at present, no evidence that PFO is related to mild or late onset bends."
Bubbles form within other tissues as well as the blood vessels. Inert gas can diffuse into bubble nuclei between tissues. In this case, the bubbles can distort and permanently damage the tissue. As they grow, the bubbles may also compress nerves as they grow causing pain.
Extravascular bubbles usually form in slow tissues such as joints, tendons and muscle sheaths. Direct expansion causes tissue damage, with the release of histamines and their associated affects. Biochemical damage may be as important as, or more important than mechanical effects.
Factors influencing uptake and elimination of dissolved gases and decompression risk
The exchange of dissolved gases between the blood and tissues is controlled by perfusion and to a lesser extent by diffusion, particularly in heterogeneous tissues. The distribution of blood flow to the tissues is variable and subject to a variety of influences. When the flow is locally high, that area is dominated by perfusion, and by diffusion when the flow is low. The distribution of flow is controlled by the mean arterial pressure and the local vascular resistance, and the arterial pressure depends on cardiac output and the total vascular resistance. Basic vascular resistance is controlled by the sympathetic nervous system, and metabolites, temperature, and local and systemic hormones have secondary and often localised effects, which can vary considerably with circumstances. Peripheral vasoconstriction in cold water decreases heat loss without increasing oxygen consumption until shivering begins, at which point oxygen consumption will rise, though the vasoconstriction can persist.
Breathing gas composition
The composition of the breathing gas during pressure exposure and decompression is the most significant factor in inert gas uptake and elimination for a given pressure exposure profile, for two main reasons:
Gas fraction and partial pressure of the component inert gas
Breathing gas mixtures for diving will typically have a different gas fraction of nitrogen to that of air. The partial pressure of each component gas will differ to that of nitrogen in air at any given depth, and uptake and elimination of each inert gas component is proportional to the actual partial pressure over time. The two foremost reasons for use of mixed breathing gases are the reduction of nitrogen partial pressure by dilution with oxygen, to make Nitrox mixtures, primarily to reduce the rate of nitrogen uptake during pressure exposure, and the substitution of helium (and occasionally other gases) for the nitrogen to reduce the narcotic effects under high partial pressure exposure. Depending on the proportions of helium and nitrogen, these gases are called Heliox, if there is no nitrogen, or Trimix, if there is nitrogen and helium along with the essential oxygen.
Solubility characteristics of the inert gases in the mixture
The inert gases used as substitutes for nitrogen have different solubility and diffusion characteristics in living tissues to the nitrogen they replace. For example, the most common inert gas diluent substitute for nitrogen is helium, which is significantly less soluble in living tissue, but also diffuses faster due to the relatively small size and mass of the He atom in comparison with the N2 molecule.
Body temperature and exercise
Blood flow to skin and fat are affected by skin and core temperature, and resting muscle perfusion is controlled by the temperature of the muscle itself. During exercise increased flow to the working muscles is often balanced by reduced flow to other tissues, such as kidneys spleen and liver.
Blood flow to the muscles is lower in cold water, but exercise keeps the muscle warm and flow elevated even when the skin is chilled. Blood flow to fat normally increases during exercise, but this is inhibited by immersion in cold water. Adaptation to cold reduces the extreme vasoconstriction which usually occurs with cold water immersion.
Variations in perfusion distribution do not necessarily affect respiratory inert gas exchange, though some gas may be locally trapped by changes in perfusion. Rest in a cold environment will reduce inert gas exchange from skin, fat and muscle, whereas exercise will increase gas exchange. Exercise during decompression can reduce decompression time and risk, providing bubbles are not present, but can increase risk if bubbles are present.
Inert gas exchange is least favourable for the diver who is warm and exercises at depth during the ingassing phase, and rests and is cold during decompression.
Other factors which can affect decompression risk include oxygen concentration, carbon dioxide levels, body position, vasodilators and constrictors, positive or negative pressure breathing. and dehydration (blood volume).
Individual susceptibility to decompression sickness has components which can be attributed to a specific cause, and components which appear to be random. The random component makes successive decompressions a poor test of susceptibility. Obesity and high serum lipid levels have been implicated as risk factors, and risk seems to increase with age. Other factors, such as gender and previous injury provide inconsistent results.
A more recent study has shown that older subjects tended to bubble more than younger subjects for reasons not yet known. No trends between weight, body fat, or gender and bubbles were identified, and the question of why some people are more likely to form bubbles than others remains unclear.
A fundamental problem in the design of decompression tables is that the rules that govern a single dive and ascent do not apply when some tissue bubbles already exist, as these will delay inert gas elimination and equivalent decompression may result in decompression sickness.
One attempt at a solution was the development of multi-tissue models, which assumed that different parts of the body absorbed gas at different rates. Each tissue, or compartment, has a different half-life. Fast tissues absorb gas relatively quickly, but will release it quickly during ascent. A fast tissue may become saturated in the course of a normal sports dive, while a slow tissue may hardly have absorbed any gas. By calculating the levels in each compartment separately, researchers are able to construct better algorithms. In addition, each compartment may be able to tolerate more or less supersaturation than others. The final form is a complicated model, but one that allows for the construction of algorithms and tables suited to a wide variety of diving. A typical dive computer has an 8–12 tissue model, with half times varying from 5 minutes to 400 minutes. The Bühlmann tables have 16 tissues, with half times varying from 4 minutes to 640 minutes.
The ideal decompression profile creates the greatest possible gradient for inert gas elimination from a tissue without causing bubbles to form, but it is not certain whether this is practically possible: some of the decompression models assume that stable bubble micronuclei always exist. However, the dissolved phase decompression models are based on the assumption that bubble formation can be avoided. The bubble models make the assumption that there will be bubbles, but there is a tolerable total gas phase volume or a tolerable gas bubble size, and limit the maximum gradient to take these tolerances into account. A number of empirical modifications to dissolved phase models have been made since the identification of venous bubbles by doppler measurement in asymptomatic divers soon after surfacing.
Repetitive diving, multiple ascents within a single dive, and surface decompression procedures are significant risk factors for DCS.
The function of decompression models has changed with the availability of Doppler ultrasonic bubble detectors, and is no longer merely to limit symptomatic occurrence of decompression sickness, but also to limit asymptomatic post-dive venous gas bubbles.
Validation of models
It is important that any theory be validated by carefully controlled testing procedures. As testing procedures and equipment become more sophisticated, researchers learn more about the effects of decompression on the body. Initial research focused on producing dives that were free of recognizable symptoms decompression sickness (DCS). With the later use of Doppler ultrasound testing, it was realized that bubbles were forming within the body even on dives where no DCI signs or symptoms were encountered. This phenomenon has become known as "silent bubbles". The US Navy 1956 tables were based on limits determined by external DCS signs and symptoms. Later researchers were able to improve on this work by adjusting the limitations based on Doppler testing. However the US Navy CCR tables based on the Thalmann algorithm also used only recognisable DCS symptoms as the test criteria.
Since the testing procedures are lengthy and costly, it is common practice for researchers to make initial validations of new models based on experimental results from earlier trials. This has some implications when comparing models.
Residual inert gas
A considerable amount of inert gas will remain in the tissues after a diver has surfaced. This residual gas may be dissolved or in sub-clinical bubble form, and will continue to outgas while the diver remains at the surface. If a repetitive dive is made, the tissues are preloaded with this residual gas which will make them saturate faster.
In repetitive diving, the slower tissues can accumulate gas day after day. This can be a problem for multi-day multi-dive situations. Multiple decompressions per day over multiple days can increase the risk of decompression sickness because of the build up of asymptomatic bubbles, which reduce the rate of off-gassing and are not accounted for in most decompression algorithms. Consequently, some diver training organisations make extra recommendations such as taking "the seventh day off".
Deterministic decompression models are a rule based approach to calculating decompression. These models work from the idea that "excessive" supersaturation in various tissues is "unsafe" (resulting in decompression sickness). The models usually contain multiple depth and tissue dependent rules based on mathematical models of idealised tissue compartments. There is no objective mathematical way of evaluating the rules or overall risk other than comparison with empirical test results. The models are compared with experimental results and reports from the field, and rules are revised by qualitative judgment and curve fitting so that the revised model more closely predicts observed reality, and then further observations are made to assess the reliability of the model in extrapolations into previously untested ranges. The usefulness of the model is judged on its accuracy and reliability in predicting the onset of symptomatic decompression sickness and asymptomatic venous bubbles during ascent.
It may be reasonably assumed that in reality, both perfusion transport by blood circulation, and diffusion transport in tissues where there is little or no blood flow occur. The problem with attempts to simultaneously model perfusion and diffusion is that there are large numbers of variables due to interactions between all of the tissue compartments and the problem becomes intractable.
A way of simplifying the modelling of gas transfer into and out of tissues is to make assumptions about the limiting mechanism of dissolved gas transport to the tissues which control decompression. Assuming that either perfusion or diffusion has a dominant influence, and the other can be disregarded, can greatly reduce the number of variables.
Perfusion limited tissues and parallel tissue models
The assumption that perfusion is the limiting mechanism leads to a model comprising a group of tissues with varied rates of perfusion, but supplied by blood of approximately equivalent gas concentration. It is also assumed that there is no gas transfer between tissue compartments by diffusion. This results in a parallel set of independent tissues, each with its own rate of ingassing and outgassing dependent on the rate of blood flowing through the tissue. Gas uptake for each tissue is generally modelled as an exponential function, with a fixed compartment half-time, and gas elimination may also be modelled by an exponential function, with the same or a longer half time, or as a more complex function, as in the exponential-linear elimination model.
Critical ratio hypothesis
This hypothesis predicts that the development of bubbles will occur in a tissue when the ratio of dissolved gas partial pressure to ambient pressure exceeds a particular ratio for a given tissue. The ratio may be the same for all tissue compartments or it may vary, and each compartment is allocated a specific critical supersaturation ratio, based on experimental observations.
John Scott Haldane
Haldane introduced the concept of half times to model the uptake and release of nitrogen into the blood. He suggested 5 tissue compartments with half times of 5, 10, 20, 40 and 75 minutes.
In this early hypothesis (Haldane 1908) it was predicted that if the ascent rate does not allow the inert gas partial pressure in each of the hypothetical tissues to exceed the environmental pressure by more than 2:1 bubbles will not form.
Basically this meant that one could ascend from 30 m (4 bar) to 10 m (2 bar), or from 10 m (2 bar) to the surface when saturated, without a decompression problem.
To ensure this a number of decompression stops were incorporated into the ascent schedules.
The ascent rate and the fastest tissue in the model determine the time and depth of the first stop. Thereafter the slower tissues determine when it is safe to ascend further.
This 2:1 ratio was found to be too conservative for fast tissues (short dives) and not conservative enough for slow tissues (long dives). The ratio also seemed to vary with depth.
The ascent rates used on older tables were 18 m/min, but newer tables use 9 m/min.
Critical difference hypothesis
Robert D. Workman
Haldane's approach to decompression modeling was used from 1908 to the 1960s with minor modifications, primarily changes to the number of compartments and half times used. The 1937 US Navy tables were based on research by O. D. Yarborough and used 3 compartments. The 5 and 10 min compartments were dropped. In the 1950s the tables were revised and the 5 and 10 minute compartments restored, and a 120 minute compartment added.
In the 1960s Robert D. Workman of the U.S. Navy Experimental Diving Unit (NEDU) undertook a review of the basis of the model and subsequent research performed by the US Navy. Tables based on Haldane's work and subsequent refinements were observed to still be inadequate for longer and deeper dives.
Workman revised Haldane's model to allow each tissue compartment to tolerate a different amount of supersaturation which varies with depth. He introduced the term "M-value" to indicate the maximum amount of supersaturation each compartment could tolerate at a given depth and added three additional compartments with 160, 200 and 240 minute half times.
Workman presented his findings as an equation which could be used to calculate the results for any depth and stated that a linear projection of M-values would be useful for computer programming.
Albert A. Bühlmann
A large part of Bühlmann's research was to determine the longest half time compartments for Nitrogen and Helium, and he increased the number of compartments to 16. He investigated the implications of decompression after diving at altitude and published decompression tables that could be used at a range of altitudes. Bühlmann used a method for decompression calculation similar to that proposed by Workman, which included M-values expressing a linear relationship between maximum inert gas pressure in the tissue compartments and ambient pressure, but based on absolute pressure, which made them more easily adapted for altitude diving.
Bühlmann's algorithm was used to generate the standard decompression tables for a number of sports diving associations, and are used in several personal decompression computers, sometimes in a modified form.
Thermodynamic model and deep stops
Torres Strait pearl divers
B.A. Hills and D.H. LeMessurier studied the empirical decompression practices of Okinawan pearl divers in the Torres Strait and observed that they made deeper stops but reduced the total decompression time compared with the generally used tables of the time. Their analysis strongly suggested that bubble presence limits gas elimination rates, and emphasised the importance of inherent unsaturation of tissues due to metabolic processing of oxygen.
A "Pyle stop" is an additional brief deep-water stop, which is increasingly used in deep diving (named after Richard Pyle, an early advocate of deep stops). Typically, a Pyle stop is 2 minutes long and at the depth where the pressure change halves on an ascent between the bottom and the first conventional decompression stop.
For example, a diver ascends from a maximum depth of 60 metres (200 ft), where the ambient pressure is 7 bars (100 psi), to a decompression stop at 20 metres (66 ft), where the pressure is 3 bars (40 psi). The Pyle stop would take place at the halfway pressure, which is 5 bars (70 psi) corresponding to a depth of 40 metres (130 ft).
Pyle found that on dives where he stopped periodically to vent the swim-bladders of his fish specimens, he felt better after the dive, and based the deep stop procedure on the depths and duration of these pauses. The hypothesis is that these stops provide an opportunity to eliminate gas while still dissolved, or at least while the bubbles are still small enough to be easily eliminated, and the result is that there will be considerably fewer or smaller venous bubbles to eliminate at the shallower stops as predicted by the thermodynamic model of Hills.
Diffusion limited tissues and the "Tissue slab", and series models
The assumption that diffusion is the limiting mechanism of dissolved gas transport in the tissues results in a rather different tissue compartment model. In this case a series of compartments has been postulated, with perfusion transport into one compartment, and diffusion between the compartments, which for simplicity are arranged in series, so that for the generalised compartment, diffusion is to and from only the two adjacent compartments on opposite sides, and the limit cases are the first compartment where the gas is supplied and removed via perfusion, and the end of the line, where there is only one neighbouring compartment.
The simplest series model is a single compartment, and this can be further reduced to a one-dimensional "tissue slab" model.
Bubble decompression models are a rule based approach to calculating decompression based on the idea that microscopic bubble nuclei always exist in water and tissues that contain water and that by predicting and controlling the bubble growth, one can avoid decompression sickness. Most of the bubble models assume that bubbles will form during decompression, and that mixed phase gas elimination occurs.
Decompression models that assume mixed phase gas elimination include:
- The arterial bubble decompression model of the French Tables du Ministère du Travail 1992
- The U.S.Navy Exponential-Linear (Thalmann) algorithm used for the 2008 US Navy air decompression tables (among others)
- Hennessy's combined perfusion/diffusion model of the BSAC'88 tables
- The Varying Permeability Model (VPM) developed by D.E. Yount and others at the University of Hawaii
- The Reduced Gradient Bubble Model (RGBM) developed by Bruce Wienke at Los Alamos National Laboratory
Probabilistic decompression models are designed to calculate the risk (or probability) of decompression sickness (DCS) occurring on a given decompression profile. These models can vary the decompression stop depths and times to arrive at a final decompression schedule that assumes a specified probability of DCS occurring. The model does this while minimizing the total decompression time. This process can also work in reverse allowing one to calculate the probability of DCS for any decompression schedule.
Goldman Interconnected Compartment Model
In contrast to the independent parallel compartments of the Haldanean models, in which all compartments are considered risk bearing, the Goldman model posits a relatively well perfused "active" or "risk-bearing" compartment in series with adjacent relatively poorly perfused "reservoir" or "buffer" compartments, which are not considered potential sites for bubble formation, but affect the probability of bubble formation in the active compartment by diffusive inert gas exchange with the active compartment.
During compression, gas diffuses into the active compartment and through it into the buffer compartments, increasing the total amount of dissolved gas passing through the active compartment. During decompression, this buffered gas must pass through the active compartment again before it can be eliminated. If the gas loading of the buffer compartments is small, the added gas diffusion through the active compartment is slow.
The interconnected models predict a reduction in gas washout rate with time during decompression compared with the rate predicted for the independent parallel compartment model used for comparison.
The Goldman model differs from the Kidd-Stubbs series decompression model in that the Goldman model assumes linear kinetics, where the K-S model includes a quadratic component, and the Goldman model considers only the central well-perfused compartment to contribute explicitly to risk, while the K-S model assumes all compartments to carry potential risk. The DCIEM 1983 model associates risk with the two outermost compartments of a four compartment series.
The mathematical model based on this concept is claimed by Goldman to fit not only the Navy square profile data used for calibration, but also predicts risk relatively accurately for saturation profiles. A bubble version of the ICM model was not significantly different in predictions, and was discarded as more complex with no significant advantages. The ICM also predicted decompression sickness incidence more accurately at the low-risk recreational diving exposures recorded in DAN's Project Dive Exploration data set. The alternative models used in this study were the LE1 (Linear-Exponential) and straight Haldanean models.
- Decompression (diving)
- Decompression practice
- Decompression sickness
- History of decompression research and development
- Chris W Dueker, MD, Scuba Diving in Safety & Health, ISBN 0-9614638-0-5
- Burton, Steve (December 2004). "Isobaric Counter Diffusion". ScubaEngineer. Retrieved 3 February 2011.
- Hills, Brian A (1978). "Effect of decompression per se on nitrogen elimination". J Appl Physiol 45 (6): 916–921. PMID 730597. Retrieved 2011-10-31.
- Bühlmann Albert A. (1984). Decompression–Decompression Sickness. Berlin New York: Springer-Verlag. ISBN 0-387-13308-9.
- Yount 1991, p. 137.
- Huggins 1992, chpt. 1 page 7
- Hills, Brian A (1978). "A fundamental approach to the prevention of decompression sickness". South Pacific Underwater Medicine Society Journal 8 (2): 20–47. ISSN 0813-1988. OCLC 16986801. Retrieved 2011-10-31.
- Wienke 2002, p. 10
- Behnke, Albert R (1967). "The isobaric (oxygen window) principle of decompression". Trans. Third Marine Technology Society Conference, San Diego. The New Thrust Seaward. Washington DC: Marine Technology Society. Retrieved 19 June 2010.
- Van Liew, Hugh D; Conkin, J; Burkard, ME (1993). "The oxygen window and decompression bubbles: estimates and significance". Aviation, Space, and Environmental Medicine 64 (9): 859–65. ISSN 0095-6562. PMID 8216150.
- Boycott, AE; Damant, GCC; Haldane, John Scott (1908). "Prevention of compressed air illness". Journal of Hygiene 8 (3): 342–443. doi:10.1017/S0022172400003399. PMC 2167126. PMID 20474365. Retrieved 30 May 2010.
- Beresford, M.: CMAS-ISA Normoxic Trimix Manual
- Workman, Robert D (1957). "Calculation of air saturation decompression tables". Navy Experimental Diving Unit Technical Report. NEDU-RR-11-57. Retrieved 2011-10-31.
- Baker, Erik (1998). "Understanding M-values". Immersed 3 (3): 23–27.
- Matti Anttila, Gradient Factors, http://www.diverite.com/education/rebreather/tips/gradient%20factors/ accessed May 2, 2012
- LeMessurier and Hills. (1965) Decompression Sickness. A thermodynamic approach arising from a study on Torres Strait diving techniques. Hvalradets Skrifter, Nr. 48, 54–84.
- Papadopoulou, Virginie; Robert J. Eckersley, Costantino Balestra, Thodoris D. Karapantsios, Meng-Xing Tang (2013). "A critical review of physiological bubble formation in hyperbaric decompression". Advances in Colloid and Interface Science (Elsevier) (191–192): 22–30.
- Yount 1991, p. 131.
- Yount 1991, p. 132.
- Hills BA (March 1992). "A hydrophobic oligolamellar lining to the vascular lumen in some organs". Undersea Biomed Res 19 (2): 107–20. PMID 1561717. Retrieved 2011-10-31.
- Yount 1991.
- Yount 1991, pp. 131,136.
- Hamilton & Thalmann 2003, pp. 477–478.
- Lambertson, Christian J (1989). Relations of isobaric gas counterdiffusion and decompression gas lesion diseases. In Vann, RD. "The Physiological Basis of Decompression". 38th Undersea and Hyperbaric Medical Society Workshop UHMS Publication Number 75(Phys)6-1-89. http://archive.rubicon-foundation.org/6853. Retrieved 10 January 2010.
- D'Aoust, BG; White, R; Swanson, H; Dunford, RG; Mahoney, J (1982). "Differences in Transient and Steady State Isobaric Counterdiffusion". Report to the Office of Naval Research. http://archive.rubicon-foundation.org/4629. Retrieved 10 January 2010.
- Doolette, David J; Mitchell, Simon J (June 2003). "Biophysical basis for inner ear decompression sickness". Journal of Applied Physiology 94 (6): 2145–50. doi:10.1152/japplphysiol.01090.2002 (inactive 2012-01-10). PMID 12562679. Retrieved 10 January 2010.
- Burton, Steve (December 2004). "Isobaric Counter Diffusion". ScubaEngineer. http://www.scubaengineer.com/isobaric_counter_diffusion.htm. Retrieved 10 January 2010.
- Huggins 1992, chpt. 4 page 6
- Vann, R.D.(ed) (1989), The Physiological basis of decompression: an overview. pp1-10, Proceedings of the thirty-eighth undersea and hyperbaric medical society workshop, Undersea and Hyperbaric Medical Society, Bethesda, Maryland. http://archive.rubicon-foundation.org/6853
- Moon, Richard E; Kisslo, Joseph (1998). "PFO and decompression illness: An update". South Pacific Underwater Medicine Society Journal 28 (3). ISSN 0813-1988. OCLC 16986801. Retrieved 2011-10-31.
- Scharlin, P.; Battino, R. Silla, E.; Tuñón, I.; Pascual-Ahuir, J. L. (1998). "Solubility of gases in water: Correlation between solubility and the number of water molecules in the first solvation shell". Pure & Appl. Chem. 70 (10): 1895–1904. doi:10.1351/pac199870101895
- Clifford A. Hampel (1968). The Encyclopedia of the Chemical Elements. New York: Van Nostrand Reinhold. pp. 256–268. ISBN 0-442-15598-0.
- Williams, ST; Prior, F; Bryson, PJ (2005), Haematocrit change in recreational Scuba divers following single dive exposure. http://archive.rubicon-foundation.org/1691
- Bookspan, J (May 2003). "Detection of endogenous gas phase formation in humans at altitude". Medicine & Science in Sports & Exercise Suppl. 35 (5,): S164. doi:10.1097/00005768-200305001-00901. Retrieved 2012-05-07.
- Gorman, Des F (1989). "Decompression tables: their use and problems". South Pacific Underwater Medicine Society Journal 19 (3): 111–113. Retrieved 2011-10-31.
- Gorman, Desmond F; Pearce, A; Webb, RK (1988). "Dysbaric illness treated at the Royal Adelaide Hospital 1987, a factorial analysis". South Pacific Underwater Medicine Society Journal 18 (3): 95–101.
- JP Imbert, D Paris, J Hugon, Divetech, France. 2004; The Arterial Bubble Model for Decompression Tables Calculations, EUBS 2004, http://gtuem.praesentiert-ihnen.de/tools/literaturdb/project2/pdf/Imbert%20JP.%20-%20EUBS%202004.pdf
- Thalmann 1984, p. 24
- Thalmann 1985, p. 5
- Kindwall, Eric P; Baz, A; Lightfoot, EN; Lanphier, Edward H; Seireg, A (1975). "Nitrogen elimination in man during decompression". Undersea Biomedical Research 2 (4): 285–297. ISSN 0093-5387. OCLC 2068005. PMID 1226586. Retrieved 2011-10-31.
- Lang, Michael A; Vann, Richard D (1991). Proceedings of the AAUS Repetitive Diving Workshop. Duke University, Durham, NC: American Academy of Underwater Sciences. p. 339. Retrieved 2011-10-31.
- Cole, Bob (2008). "Diver Behaviour – Micro-bubble Control". SAA Buhlmann Deep Stop System Handbook. Sub-Aqua Association. pp. 4–2. ISBN 0-9532904-8-4. "The SAA recommends that you to [sic] take at least the seventh day off to allow your body to off-gas and return to some level of normality"
- Doolette David J (2005). "Development and testing of deterministic and probabilistic decompression models". South Pacific Underwater Medicine Society Journal 35 (1). Retrieved 2012-01-10.
- "Decoweenie Manual" (PDF). decoweenie.com. Retrieved 2008-09-26.
- Pyle, Richard L (2007-09-27). "Deep Decompression Stops". Bishop Museum. Retrieved 2009-09-09.
- Pyle, Richard L (1997). "The importance of deep safety stops: Rethinking ascent patterns from decompression dives". South Pacific Underwater Medicine Society Journal (reprinted from: Deep Tech) 27 (2). Retrieved 2011-10-31.
- Goldman, Saul; Goldman, Ethel (2010). "Coming soon to a Dive Computer near you". Alert Diver (European edition) (Roseto degli Abruzzi, Italy: DAN Europe) (4th quarter, 2010): 4–8.
- Goldman, Saul (April 19, 2007). "A new class of biophysical models for predicting the probability of decompression sickness in scuba diving". Journal of Applied Physiology 103 (2): 484–493. doi:10.1152/japplphysiol.00315.2006.
- Goldman, Saul; Goldman, Ethel (2014). "To stop or not to stop and why?". Alert Diver (DAN South Africa) 6 (2): 34–37. ISSN 2071-7628. Retrieved 10 September 2014.
- Goldman, Saul (23 September 2013). "How SAUL relates to the PADI dive tables". Modern decompression. Retrieved 10 September 2014.
- Ball, R; Himm, J; Homer, LD; Thalmann, ED (1995). "Does the time course of bubble evolution explain decompression sickness risk?". Undersea and Hyperbaric Medicine 22 (3): 263–280. ISSN 1066-2936. PMID 7580767.
- Brubakk, A. O.; Neuman, T. S. (2003). Bennett and Elliott's physiology and medicine of diving (5th Revised ed.). United States: Saunders. ISBN 0-7020-2571-2.
- Gerth, Wayne A; Doolette, David J. (2007). "VVal-18 and VVal-18M Thalmann Algorithm – Air Decompression Tables and Procedures". Navy Experimental Diving Unit, TA 01-07, NEDU TR 07-09. Retrieved 2012-01-27.
- Hamilton, Robert W; Thalmann, Edward D (2003). "10.2: Decompression Practice". In Brubakk, Alf O; Neuman, Tom S. Bennett and Elliott's physiology and medicine of diving (5th Revised ed.). United States: Saunders. pp. 455–500. ISBN 0-7020-2571-2. OCLC 51607923.
- Huggins, Karl E. (1992). "Dynamics of decompression workshop". Course taught at the University of Michigan. Retrieved 2012-01-10.
- Lippmann, John (1990). Deeper into Diving (1st ed.). Melbourne, Australia: J L Publications. ISBN 0-9590306-3-8.
- Parker, E. C.; S.S. Survanshi, P.K. Weathersby, and E.D. Thalmann. (1992). "Statistically Based Decompression Tables VIII: Linear Exponential Kinetics". Naval Medical Research Institute Report. 92-73. Retrieved 2008-03-16.
- Powell, Mark (2008). Deco for Divers. Southend-on-Sea: Aquapress. ISBN 1-905492-07-3.
- Thalmann, E. D. (1984). "Phase II testing of decompression algorithms for use in the U.S. Navy underwater decompression computer". Navy Exp. Diving Unit Res. Report 1–84. Retrieved 2008-03-16.
- Thalmann, E. D. (1985). "Development of a Decompression Algorithm for Constant Oxygen Partial Pressure in Helium Diving". Navy Exp. Diving Unit Res. Report 1–85. Retrieved 2008-03-16.
- Wienke, Bruce R; O'Leary, Timothy R (13 February 2002). "Reduced gradient bubble model: Diving algorithm, basis and comparisons". Tampa, Florida: NAUI Technical Diving Operations. Retrieved 25 January 2012.
- Yount, DE (1991). "Gelatin, bubbles, and the bends". International Pacifica Scientific Diving... Hans-Jurgen, K; Harper Jr, DE (eds.), (Proceedings of the American Academy of Underwater Sciences Eleventh Annual Scientific Diving Symposium held 25–30 September 1991. University of Hawaii, Honolulu, Hawaii). Retrieved 2012-01-25.
- Powell, Mark (2008). Deco for Divers. Southend-on-Sea: Aquapress. ISBN 1-905492-07-3.
- Hills. B. (1966); A thermodynamic and kinetic approach to decompression sickness. Thesis
- Gribble, M. de G. (1960); A comparison of the High-Altitude and High-Pressure syndromes of decompression sickness, Brit. J. industr. Med., 1960, 17, 181.
- Lippmann, John; Mitchell, Simon (2005). Deeper into Diving (2nd ed.). Melbourne, Australia: J L Publications. ISBN 0-9752290-1-X. Section 2 chapters 13–24 pages 181–350