Four etiological theories have evolved to reconcile the varying observations of onset time, severity, response to therapy, and histopathology. They are (1) gas embolism; (2) venous infarction; (3) autochthonous ("in situ") bubbles; and (4) hemorrhage or inflammation.
The first theory for DI of the spinal cord was developed by Boycott and Damant. Lesions in the spinal cord of goats were found to consist, almost entirely, of white matter lesions 9. Indeed human pathology, although rarely observed in this mostly non-fatal condition, has also shown similar punctate, white matter lesions and hemorrhage. However, embolic injury to the spinal cord is, generally speaking, very rare. This is believed to be due to the relative difference in blood flow favoring embolization of the brain. Experimental spinal cord embolism has also been shown to produce ischemic grey matter pathology rather than white matter lesions 73. To confuse the matter further a type of DI was identified that began as rapid onset cerebral arterial gas embolism but then evolved into a particularly resistant form of spinal injury. This has been called "combined", "concurrent" or "Type III decompression sickness" 30,74-76. Although the exact mechanism is still unknown, the predominant theory is related to growth of arterial gas emboli in tissues saturated with inert gas.
In 1975, based on Batson's experiments on tumor embolization via epidural veins 77, Hallenbeck et al postulated that DI of the spinal cord was due to bubble accumulation in the epidural venous plexus with subsequent venous infarction of the spinal cord. Although confirmed in extreme decompression 78-80, loss of function only occurred after several minutes and therefore did not offer an explanation for ultra-short-latency disease. In addition, the pattern of dysbaric injury was different to that observed in other causes of venous infarction of the spinal cord that typically affects the central grey matter 81.
Francis proposed that rapid-onset spinal cord damage may be related to spontaneous bubble formation in the spinal cord white matter. He felt that this was the only mechanism that could explain both the rapid onset and the distribution of lesions observed in the spinal cord. In his classic experiment, the spinal cord of decompressed dogs was rapidly perfusion-fixed at the moment of maximal disruption of somatosensory evoked potentials. He consistently found extravascular, white matter lesions - called autochthonous bubbles 82,83. The puzzling piece is how these small, scattered and isolated space occupying lesions (making up no more than 0.5% of the spinal cord and being no more than 20-200um in diameter) are able to produce such catastrophic clinical effects 84. It has been postulated that autochthonous bubbles could account for loss of function if more than 30% of the axons became dysfunctional due to direct injury, stretching or compression, inflammation, biochemical injury or hemorrhage. In support of this, Hills et al has also shown that small lesions (able to increase the spinal cord volume by 14-31%) can cause an increase in tissue pressures with a resulting spinal compartment syndrome 85.
Hemorrhage and Inflammation:
In his studies on autochthonous bubbles, Francis made three important additional observations 86: (1) animals that only developed abnormalities after 30 minutes had no demonstrable space occupying lesions suggesting another mechanisms for the dysfunction; (2) animals sacrificed some time after the development of rapid-onset dysfunction no longer had bubbles suggesting that they are temporary; and (3) the histological appearance of spinal cords from dogs with late-onset spinal symptoms was similar to that of spinal embolism or ischemic lesions 73. Interestingly in the animals harvested some time after rapid onset illness, hemorrhage and inflammation were observed in the same areas where autochthonous bubble injuries were seen in those harvested early. This could explain why some cases of rapid onset spinal cord DI appear resistant to recompression and would suggest caution in the use of anti-coagulants in DI of the spinal cord.
Intriguingly, all of the mechanisms appear to converge within a particular area: a c-shaped area around the spinal cord grey matter. This area represents a watershed zone between the anterior and posterior spinal cord circulation and would therefore be susceptible to both inert gas accumulation as well as subsequent bubble-related ischemia. The cervical and lumbar enlargements are particularly vulnerable and also correspond to the areas of greatest clinical importance in DI.
It is unlikely that one single mechanism can account for the wide variety of latencies and presentations of DI of the spinal cord and the decompression schedules leading to them. Rather it is probably the result of several interacting, compressive-ischemic mechanisms. DI of the spinal cord should be thought of as a spectrum of cause-and-effect over a 48-hour time-continuum. It is an interplay between a variety of distinct, yet synergistic pathophysiological processes - some of which are amenable to recompression and adjunctive medical therapy and some which, unfortunately, are not.
Finally, in spite of the disturbing vulnerability of the spinal cord, it has a remarkable capacity of recovery. Many divers with residual deficits after recompression therapy continue to improve for years afterwards. However, this does not indicate that the injury has been reversed, only that the body has compensated for it 87,88.
Was this article helpful?