Complex regional pain syndrome.

Since the early musings in the mid-1800s of Claude Bernard and his French neurological colleagues on the association of pain with the sympathetic nervous system, complex regional pain syndrome (CRPS) has both fascinated and perplexed practitioners. Some of the clearest and most interesting descriptions of `causalgia' come from the American Civil War by one of Bernard's students, Silas Weir-Mitchell. The low-velocity, high-mass missiles used in this confrontation (the `Minnie ball') seemed to be particularly effective in inducing neuropathic pain associated with intense autonomic dysregulation. Weir-Mitchell's depictions are clear and elegant, and as good as any clinical description that can be found in this century. Many great minds have struggled with the pathophysiology of what came to be called `re ̄ex sympathetic dystrophy' in the later part of the 1900s and what has, since the Orlando consensus-based workshop of 1999, come to be called complex regional pain syndrome (CRPS) (Table 1). 63 85 From Leriche and his vicious circles we have progressed through Livingston and Sunderland with the turbulence theory, and ®nally to the solid physiological information generated by the various animal models of causalgia, especially the chronic constriction injury model of Bennett and Xie. Recently, the effort to understand the syndrome has turned towards consensus symposia. The ®rst of these concerned taxonomy, as above. 85 A second Dahlem-type conference was conducted in regard to the guidelines for therapy, and recently the International Association for the Study of Pain (IASP) sponsored a symposium in Cardiff, Wales in 2000 to discuss issues of pathophysiology and to amend the diagnostic considerations. The epidemiology of the syndrome is very unclear. Although the syndrome has traditionally been considered rare, its `discovery' by personal injury lawyers in the United States has caused a radical increase in the reporting of the syndrome (at least in the USA). The current diagnostic criteria, as set forth by the Committee of Classi®cation of Chronic Pain of the IASP, have contributed to the liberalization of the diagnosis (Table 1). This effort was extremely important in providing standardized diagnostic criteria, and caused a vast improvement in clinical communication and research homogeneity. It provided the hope that results could be generalized across studies, and in fact widespread use of these standardized criteria has helped all these things considerably. These criteria, while being very sensitive, greatly lack speci®city. 21 31 The intent of the Orlando conference in 1994 was that these criteria should evolve on the basis of experience and empirical testing, and that they should be subject to systematic validation research over time. 62 85 This has been accomplished to some extent, and through a process of internal and external validation the opportunity to improve the speci®city of the bedside diagnostic criteria is available. 21 31 Although the original IASP criteria required only subjective and potentially only historical signs and symptoms, the suggestions for improving these criteria are that some objecti®cation and observed evidence be included. It is recommended that the diagnostic criteria be modi®ed to include at least one symptom in each of the four diagnostic categories derived by factor analysis: sensory (hyperaesthesia), vasomotor (temperature and/or skin colour asymmetry), sudomotor/oedema (reports of asymmetrical oedema in the affected limb and/or a sweating asymmetry), and a new symptom set which was identi®ed by factor analysis: the motor/trophic set (reports of motor dysfunction or trophic changes). 21 31 Our data also suggest that the patient should display at least one quasi-objective sign (observed by the physician) in two or more of these categories (Table 2). 31 This approach maintained a sensitivity of 0.70 while increasing speci®city to 0.94. This level of speci®city is important in research. However, the original IASP criteria could be used if maximal clinical sensitivity was desired. A variety of schemes in which the clinician/researcher can virtually set the sensitivity and speci®city, depending on the diagnostic scheme used, have been discussed. 31 British Journal of Anaesthesia 87 (1): 99±106 (2001)