There is currently no consensus concerning the optimal therapeutic strategy for neuropathic pain, despite an increasing number of clinical trials demonstrating successful pain relief with several drugs. Treatments have generally been selected on the basis of evidence for efficacy in randomized placebo-controlled trials conducted in disease-based groups of patients, notably in postherpetic neuralgia and diabetic polyneuropathy. These studies plead in favour of the overall efficacy of tricyclic antidepressants, standard and newer antiepileptics, opioids, tramadol, systemic and topical local anaesthetics, and some NMDA receptor antagonists; whereas evidence for efficacy is less for selective serotonin reuptake inhibitors, antiarrhythmics (mexiletine), and capsaicin. Pharmacological tests, notably therapeutic infusions, have been proposed for predicting the effectiveness of long-term treatments, but are not routinely performed. An analysis of the various neuropathic symptoms, aimed at selecting treatments targeted at mechanisms, may ultimately help the choice of different pharmacologic agents.