meaningful decisions based on ancillary test data is illustrated by the following:
In a study specifically designed to assess the response of patients with painful distal sensory neuropathies to the 5% lidocaine patch, no relation- ship between treatment response and distal leg skin biopsy, QST, or sensory nerve conduction study results could be established . From a mechanistic perspective, the hypothesis that the lidocaine patch would be most effective in patients with relatively intact epidermal innervation, whose neuropathic pain is presumed attributable to ‘‘irritable nociceptors,’’ and least effective in patients with few surviving epidermal nociceptors, presumably with ‘‘deafferentation pain,’’ was unproven . The possible explanations are multiple and outside the scope of this review. However, these findings, cou- pled with the disparity in C-fiber subtype involvement in diabetic small-fiber neuropathy , and the recently reported inability of enzyme replacement therapy in Fabry disease to influence intraepidermal innervation density, while having mixed effects on cold and warm QST thresholds, and beneficial effects on sudomotor findings [26,27], when therapeutic benefit was demon- strated , lead one to conclude that the specificity of ancillary testing in neuropathic pain is inadequate at present, and reinforce the aforementioned caveats about inferential conclusions from indirect data. The diagnosis of neuropathic pain mechanisms is in its nascent stages and ancillary testing remains ‘‘subordinate,’’ ‘‘subsidiary,’’ and ‘‘auxiliary’’ as defined in Web- ster’s Third New International Dictionary.
As a consequence of these difficulties, the recent approach by Bennett and his colleagues  may have merit. They have hypothesized (and provide data in support) that chronic pain can be more or less neuropathic on a spec- trum between ‘‘likely,’’ ‘‘possible,’’ and ‘‘unlikely,’’ based on patient re- sponses on validated neuropathic pain symptom scales, when compared with specialist pain physician certainty of the presence of neuropathic pain on a 100-mm visual analog scale. The symptoms most associated with neuropathic pain were dysesthesias, evoked pain, paroxysmal pain, thermal pain, autonomic complaints, and descriptions of the pain as being sharp, hot, or cold, with high sensitivity. Higher scores for these symptoms correlated with greater clinician certainty of the presence of neuropathic pain mechanisms. Considering each individual patient’s chronic pain as be- ing somewhere on a continuum between ‘‘purely nociceptive’’ and ‘‘purely neuropathic’’ may have diagnostic and therapeutic relevance by enhancing specificity, but this requires clinical confirmation. Thus, symptom assess- ment remains indispensable in the evaluation of neuropathic pain, ancillary testing notwithstanding .
 Woolf CJ. Pain: moving from symptom control toward mechanism-specific pharmacologic management. Ann Intern Med 2004;140:441–51.