Physiopathologie de la douleur neuropathique : revue des modèles expérimentaux et des mécanismes proposés

Abstract

Neuropathic pain can be conceptualized as the result of an "aberrant learning" process, associated with maladaptive plasticity of the nervous system. A number of modifications of the peripheral nervous system have been described in animal models of neuropathic pain, but their relation with different symptoms in humans is far from fully understood. We note in particular ectopic discharges in damaged myelinated fibers, abnormal activity in undamaged fibers, overexpression of calcium channels increasing the release of excitatory neurotransmitters, and sympathetic sprouting towards the spinal ganglia. Spinal mechanisms involve central sensitization, kindling and potentiation phenomena. Underlying these phenomena may be connectivity changes--still controversial--of non-nociceptive terminals and variations in the sensitivity of postsynaptic receptors. Also contributing to these pathophysiologic modifications are attenuation of spinal inhibition by selective neuronal loss and the development of inflammatory phenomena, including cytokine secretion by macrophages and glial cells. Changes in the dorsal horn modify the activity of projections towards the brainstem and increase spinal hyperactivity still further by feedback loops. These effects are delayed, suggesting that maintenance of spinal sensitization requires the involvement of mechanisms of descending facilitation involving the brainstem. These phenomena induce changes in the activity of thalamocortical networks, which develop autonomous processes that maintain the pain. The cortical representation of body areas change after nervous lesions, and these changes may correlate with the emergence of pain. Neuropathic allodynia and hyperalgesia are supported by cortical modifications that experimental models reproduce very incompletely. Experimental allodynia and neuropathic allodynia share the activation of the cortical pain matrix as well as the bilateralization of insular activity. However, although experimental allodynia tends to increase the activity of limbic and affective networks of the perigenual and orbitofrontal cortex, in neuropathic allodynia, analgesic procedures lead to increased activity in these structures. This suggests that their role in experimental allodynia would likely be reactive and protective, and that inability to generate their activation may contribute to the clinical expression of neuropathic pain.