Inert gas bubble formation from improper ascent — Haldanian tissue compartment theory
As a diver descends, ambient pressure increases by 1 atmosphere (ATA) for every 10 meters of seawater. Because the diver breathes compressed gas at ambient pressure, the partial pressure of inert nitrogen delivered to the lungs rises in lockstep — and by Henry’s Law, more gas dissolves into blood and tissue the higher that partial pressure climbs. This is the physical starting point of all decompression physiology.
Two gas laws govern everything that follows in this simulation. Boyle’s Law (PV = constant, at fixed temperature) describes how a fixed quantity of gas compresses as ambient pressure rises — a balloon taken from the surface to 10 m (2 ATA) shrinks to half its volume. Henry’s Law describes the companion process for gas already dissolved in a liquid: the amount of gas that dissolves into a liquid is proportional to the partial pressure of that gas in contact with it.
When a diver breathes compressed air at depth, the nitrogen partial pressure delivered to the alveoli rises with ambient pressure (at 30 m / 4 ATA, inspired N₂ partial pressure reaches 3.16 ATA, versus 0.79 ATA at the surface). Blood equilibrates with alveolar gas within a single pulmonary transit, then carries the now higher-pressure dissolved nitrogen out to every perfused tissue in the body, where it continues to diffuse down its own local pressure gradient until each tissue’s dissolved N₂ pressure matches the new arterial value.
Oxygen is metabolized continuously and never accumulates, but nitrogen is metabolically inert — it does nothing except dissolve, and it must later be carried back out exactly the way it came in. This inertness is the entire reason decompression sickness exists as a phenomenon.
Not all tissues load nitrogen at the same rate. Well-perfused, blood-rich tissue (blood itself, brain, kidney) equilibrates with a new ambient pressure within minutes — these are called "fast" compartments. Poorly perfused tissue (fat, tendon, cartilage, joint capsules) may take hours to reach the same equilibrium — "slow" compartments.
This difference is captured mathematically as an exponential half-time: the time required for a compartment to close half the gap between its current dissolved gas pressure and the new ambient partial pressure. A theoretical fast compartment might have a 5-minute half-time; a slow compartment used in modern algorithms can have a half-time of 635 minutes — over two orders of magnitude slower.
Because fat tissue holds roughly five times more dissolved nitrogen at saturation than blood (nitrogen is about five times more soluble in lipid than in water), joints and fatty tissue are simultaneously the slowest to load and the slowest — and most dangerous — to unload during ascent.
Long before scuba diving existed, 19th-century caisson workers who dug tunnel and bridge foundations under compressed air suffered mysterious joint pain and paralysis after emerging — including many of the workers who built the Brooklyn Bridge (1870–83), whose chief engineer Washington Roebling was himself severely disabled by what was then called "caisson disease."
In 1908, physiologist John Scott Haldane, commissioned by the Royal Navy, published the first mathematical model of nitrogen uptake and elimination in the body. He proposed that tissues could be modeled as a set of parallel compartments, each loading and unloading gas exponentially, and that each compartment could tolerate a certain degree of supersaturation on ascent — roughly a 2:1 ratio of tissue to ambient pressure — without forming symptomatic bubbles. Haldane’s staged-ascent decompression tables, and the compartment concept itself, remain the conceptual foundation of every dive table and dive computer algorithm used today.
While a diver remains at a constant depth ("bottom time"), every tissue compartment in the Haldanian model asymptotically approaches equilibrium with the ambient nitrogen partial pressure. How close each compartment gets depends entirely on how long the diver stays and how fast that particular compartment equilibrates — this is exactly what dive tables and dive computers track in real time.
Haldane’s 1908 model represented the body as five parallel, independent tissue compartments with half-times of 5, 10, 20, 40, and 75 minutes. Each compartment loads (and later unloads) dissolved gas according to the same exponential curve:
P(t) = P0 + (Pamb − P0) × (1 − 2^(−t / half-time))
Where P0 is the compartment’s starting dissolved gas pressure, Pamb is the new ambient partial pressure, and t is elapsed time. The elegance of this formula is that it needs only one parameter — the half-time — to describe an entire tissue’s uptake behavior, yet it reproduces the real physiological pattern remarkably well: a rapid initial uptake that decelerates smoothly as equilibrium is approached.
A tissue compartment reaches 50% saturation after one half-time, 75% after two, 87.5% after three, 93.75% after four, 96.9% after five, and 98.4% after six half-times. By convention, six half-times is treated as "fully saturated" for practical purposes.
Most dive computers today run a descendant of Haldane’s model developed by Swiss physician Albert A. Bühlmann: ZH-L16, using sixteen parallel tissue compartments with half-times spanning roughly 4 minutes (fastest) to 635 minutes (slowest, representing the most poorly perfused fat and connective tissue).
Each compartment is assigned an "M-value" — the maximum inert gas pressure that compartment can tolerate at a given ambient pressure before bubble formation becomes likely. M-values are expressed as a linear function of depth: M = M0 + ΔM × depth, where M0 is the surface M-value and ΔM describes how tolerance changes with depth. Fast compartments have higher M0 but tolerate less absolute overpressure at depth; slow compartments have lower M0 but a gentler slope, which is why slow-tissue DCS often appears after a dive is already over, once the diver has fully surfaced.
Printed dive tables (the classic US Navy tables, or recreational tables like the PADI Recreational Dive Planner) pre-calculate safe bottom-time limits for a given depth by simulating the slowest few compartments and ensuring none exceeds its M-value on direct ascent to the surface.
Modern dive computers do the same calculation continuously and in real time for all sixteen compartments simultaneously, using actual depth and time data from an onboard pressure sensor rather than a fixed dive profile. This lets computers credit multi-level dives (spending time shallower before ascending) far more generously than table-based planning, but the underlying physiology — exponential compartment loading toward whatever the current ambient nitrogen pressure happens to be — is identical to what Haldane described in 1908.
The instant a diver begins ascending, ambient pressure starts falling — but the nitrogen already dissolved in tissue cannot leave instantly. For a window of time, tissue holds more dissolved nitrogen than the new, lower ambient pressure would allow at equilibrium. This state, supersaturation, is not itself dangerous — within limits it is the normal, expected condition of every ascent — but exceeding those limits is what turns an ordinary ascent into decompression sickness.
Haldane observed that tissue could tolerate a certain degree of supersaturation — dissolved gas pressure exceeding ambient pressure — without bubbles forming, and estimated this tolerance at roughly a 2:1 ratio for the fastest tissues (later refined toward ~1.58:1 in reanalysis of his data). Below that ratio, dissolved nitrogen simply diffuses out through the blood, is carried to the lungs, and is exhaled — no different in principle from gas loading, just running in reverse.
Modern algorithms express this same idea more precisely with M-values: the maximum tissue nitrogen pressure tolerated at a given ambient pressure, varying by compartment. Ascending too far or too fast pushes one or more compartments past their M-value line, and it is precisely at that boundary that dissolved gas can no longer stay in solution.
A slow, controlled ascent gives dissolved nitrogen time to diffuse out of tissue into blood and be carried to the lungs at roughly the same rate that ambient pressure is falling — keeping every compartment inside its M-value envelope throughout the ascent. A rapid ascent drops ambient pressure faster than any diffusion process can keep pace with, driving one or more compartments past their tolerance line.
Recreational diving standards today specify an ascent rate of 9–10 meters per minute — slower than the 18 m/min once specified by older US Navy tables, which produced an unacceptable rate of "undeserved" DCS hits even in divers who stayed within their computed no-decompression limits. The single most effective behavioral change a diver can make to reduce DCS risk, more than any other variable, is simply ascending more slowly.
Doppler ultrasound studies monitoring divers after ascent show that adding a 3-minute stop at 5 m (15 ft) — a "safety stop" — can reduce detectable venous gas bubble counts by roughly 40–50% compared to a direct ascent, even when both ascents stay within computed no-decompression limits.
A safety stop works because the pressure reduction from 5 m to the surface (0.5 ATA) is the single largest proportional pressure drop in the entire ascent — going from 1.5 ATA to 1.0 ATA is a 33% relative pressure reduction, larger proportionally than any equivalent depth change deeper in the water column. Pausing before that final, steepest pressure drop gives fast and medium compartments extra time to off-gas at low absolute risk, shrinking the supersaturation gap that the final ascent to the surface will create.
Deep stops (brief pauses well below the last required decompression stop) were proposed on similar logic but have since been found in controlled trials to increase, not decrease, bubble formation in some cases, likely because they slow off-gassing in fast compartments without meaningfully protecting slow ones — current guidance favors shallow safety stops over deep stops for no-decompression dives.
When tissue supersaturation exceeds a compartment’s tolerance, dissolved nitrogen molecules aggregate into gas-phase bubbles — the same physics as the fizz that erupts when a shaken carbonated drink is opened. This process, called nucleation, is the mechanistic pivot point between an ordinary ascent and decompression sickness.
Pure liquids can hold substantial supersaturation without spontaneously forming bubbles because nucleating a new gas phase from nothing requires overcoming significant surface tension — this is why a perfectly still, particle-free liquid can be supersaturated far beyond what blood or tissue can tolerate. In the body, however, microscopic "gas nuclei" — tiny, stable pockets of gas trapped in crevices on vessel walls, on connective tissue surfaces, or stabilized by surfactant-like molecules — provide ready-made nucleation sites.
Once ambient pressure drops enough that dissolved gas pressure exceeds local tissue pressure by more than the nucleus’s surface tension can resist, dissolved nitrogen diffuses into these micronuclei and they begin to grow into visible bubbles, drawing in more dissolved gas from the surrounding supersaturated tissue as they expand — a self-reinforcing process once started.
Bubbles form both intravascularly (venous gas emboli, VGE) and extravascularly within tissue itself — joint capsules, tendons, spinal cord white matter, and inner ear tissue are common sites given their relatively slow off-gassing and vulnerability to local distension. Venous bubbles are usually filtered harmlessly by the pulmonary capillary bed and exhaled as the lungs act as an effective bubble trap for moderate bubble loads.
Precordial Doppler ultrasound, placed over the heart after a dive, can detect these venous bubbles as characteristic "chirping" sounds even when a diver has no symptoms at all — bubble presence alone does not equal decompression sickness. Bubble burden is graded on scales such as Spencer’s (Grade 0 = no bubbles, Grade IV = bubbles present continuously, obscuring heart sounds), and clinical DCS risk rises steeply once grade III–IV bubble loads are reached.
Studies of divers who stayed within their computer’s no-decompression limits still show Doppler-detectable "silent" venous bubbles on up to 40% of dives — asymptomatic microbubble formation is a routine, expected byproduct of ascent, not proof that something went wrong.
A patent foramen ovale (PFO) — a small flap-like opening between the heart’s right and left atria that fails to fully seal after birth — is present in roughly 25% of the general population and usually causes no problems whatsoever in everyday life. But under the right conditions (a Valsalva maneuver, coughing, or simply elevated right-heart pressure after a dive), it can allow venous bubbles to bypass the pulmonary filter entirely and cross directly into the arterial circulation.
Once in the arterial system, bubbles can travel to the brain, spinal cord, or coronary arteries and cause an arterial gas embolism — a much more acute and dangerous event than venous bubbles, since arterial bubbles are not filtered by any downstream organ before reaching end arteries. Divers with a known large PFO are advised to adopt more conservative profiles, and unexplained neurological DCS is one of the clinical clues that prompts investigation for an undiagnosed PFO.
Bubbles injure the body through two overlapping mechanisms: direct mechanical effects (distending tissue, compressing nerves, and physically obstructing blood vessels) and biochemical effects (activating complement, damaging vascular endothelium, and triggering platelet aggregation and inflammation). The clinical picture ranges from nagging joint pain to immediately life-threatening neurological and cardiopulmonary events — and definitive treatment is recompression.
Clinicians traditionally split DCS into two severity classes. Type I ("mild") DCS includes musculoskeletal pain — "the bends" proper — a deep, dull ache typically in a large joint (shoulder, elbow, knee, hip) caused by bubbles distending joint capsules and periarticular tissue, and cutaneous DCS, presenting as itching or a mottled, marbled skin rash (cutis marmorata) from bubbles affecting cutaneous blood flow and lymphatics.
Type II ("serious") DCS involves the nervous system or lungs. Spinal cord DCS is the most common serious presentation, classically affecting the thoracolumbar cord — a watershed region with a particularly vulnerable venous drainage pattern — producing back pain followed by paresthesia, weakness, or paralysis in the legs, and bladder or bowel dysfunction. Pulmonary DCS ("the chokes") results from a massive shower of venous bubbles overwhelming the lungs’ filtering capacity, causing substernal chest pain, cough, and shortness of breath, and can progress to cardiovascular collapse if untreated.
Arterial gas embolism (AGE) occurs when bubbles reach the arterial circulation, either via a right-to-left cardiac shunt (such as a PFO) or, more rarely, via pulmonary barotrauma from breath-holding on ascent that forces alveolar gas directly into pulmonary veins. Because arterial bubbles travel directly to end-organs with no filtering step, AGE can produce stroke-like focal neurological deficits, sudden loss of consciousness, or cardiac arrhythmia within minutes of surfacing — dramatically faster onset than typical joint-pain DCS, which may take hours to become apparent.
AGE and severe Type II DCS are both diving medical emergencies: 100% surface oxygen should be started immediately, the patient kept supine, and emergency transport to the nearest operational hyperbaric chamber arranged without delay, since neurological outcome correlates strongly with how quickly recompression begins.
Recompression is the only treatment that directly addresses the bubbles themselves, using the same two gas laws that created the problem. By Boyle’s Law, increasing ambient pressure physically compresses existing bubbles back down in volume. By Henry’s Law, breathing 100% oxygen at that elevated pressure creates the steepest possible pressure gradient for dissolved and bubble-phase nitrogen to diffuse out of tissue and out of the bubble itself, back into the blood and out through the lungs.
The standard protocol, US Navy Treatment Table 6, recompresses the patient to 18 meters (60 feet of seawater) inside a hyperbaric chamber, cycling periods of 100% oxygen breathing with brief air breaks (to limit the risk of oxygen toxicity seizures during prolonged high-partial-pressure O2 exposure), for a minimum total treatment time of 4 hours 45 minutes, with standard extensions available for incomplete symptom resolution before the slow, controlled final ascent back to surface pressure.
US Navy Treatment Table 6 holds the patient at 18 m (60 fsw) breathing 100% oxygen in alternating cycles with air breaks, for a minimum of 4 hours 45 minutes — deliberately mirroring, in reverse and under medical supervision, the same slow off-gassing physics that a safe ascent is meant to achieve in the first place.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| Type I — Musculoskeletal ("the bends") | Deep, dull ache in a large joint (shoulder, elbow, knee, hip); worsens with movement | Bubbles distend joint capsule / periarticular connective tissue | Urgent — 100% O2 + recompression, good prognosis if treated |
| Type I — Cutaneous | Itching, mottled "marbled" rash (cutis marmorata), sometimes with edema | Bubbles affect dermal blood flow and lymphatic drainage | Urgent — evaluate for underlying Type II; recompress if marbling present |
| Type II — Neurological (spinal cord) | Back pain, leg paresthesia/weakness, paralysis, bladder/bowel dysfunction | Bubbles + venous congestion in thoracolumbar spinal cord watershed | Emergency — immediate 100% O2, urgent recompression, time-critical |
| Type II — Pulmonary ("the chokes") | Substernal chest pain, dry cough, dyspnea, possible cardiovascular collapse | Massive venous gas embolism overwhelms pulmonary capillary filter | Emergency — 100% O2, IV fluids, immediate recompression |
| Arterial Gas Embolism (AGE) | Sudden stroke-like deficits, loss of consciousness, seizure, arrhythmia — onset within minutes | Bubbles cross PFO/shunt or from pulmonary barotrauma into arterial circulation | Critical emergency — 100% O2 supine, fastest possible recompression |