Human experimental pain models for assessing the therapeutic efficacy of analgesic drugs.

Pain models in animals have shown low predictivity for analgesic efficacy in humans, and clinical studies are often very confounded, blurring the evaluation. Human experimental pain models may therefore help to evaluate mechanisms and effect of analgesics and bridge findings from basic studies to the clinic. The present review outlines the concept and limitations of human experimental pain models and addresses analgesic efficacy in healthy volunteers and patients. Experimental models to evoke pain and hyperalgesia are available for most tissues. In healthy volunteers, the effect of acetaminophen is difficult to detect unless neurophysiological methods are used, whereas the effect of nonsteroidal anti-inflammatory drugs could be detected in most models. Anticonvulsants and antidepressants are sensitive in several models, particularly in models inducing hyperalgesia. For opioids, tonic pain with high intensity is attenuated more than short-lasting pain and nonpainful sensations. Fewer studies were performed in patients. In general, the sensitivity to analgesics is better in patients than in healthy volunteers, but the lower number of studies may bias the results. Experimental models have variable reliability, and validity shall be interpreted with caution. Models including deep, tonic pain and hyperalgesia are better to predict the effects of analgesics. Assessment with neurophysiologic methods and imaging is valuable as a supplement to psychophysical methods and can increase sensitivity. The models need to be designed with careful consideration of pharmacological mechanisms and pharmacokinetics of analgesics. Knowledge obtained from this review can help design experimental pain studies for new compounds entering phase I and II clinical trials.