Objective. Preclinical evaluation is an essential step in the assessment of new antiinflammatory or analgesic drugs. This study was undertaken to develop a new mode of evaluation of drug effectiveness based on behavior indicating well-being in a rat model of chronic inflammatory pain. We chose to examine the circadian pattern of spontaneous behavior.Methods. The work was performed with a model of chronic monarthritis induced by Freund's complete adjuvant. Variations in behavioral patterns during the time course of arthritis were analyzed. In a second phase, the impact of acetaminophen and 2 nonsteroidal antiinflammatory drugs (aspirin and celecoxib), which are currently used in clinical practice to treat chronic inflammation, was studied after 7 days of treatment.Results. The nocturnal pattern of activity of healthy rats comprised 3 main bursts. Chronic painful monarthritis altered this spontaneous pattern of nocturnal behavior (normal period of activity). Monarthritic rats showed a decrease in the total time spent in activity during the night, and lost their pattern of activity. These behavioral disturbances were reversed after long-term treatment with acetaminophen or celecoxib, with celecoxib appearing to be more effective. Aspirin was ineffective.Conclusion. These results enabled us to test this new procedure as a means of assessing well-being or illbeing during stages of chronic inflammatory pain in rats, and the effectiveness of repeated pharmacologic treatments.Arthritis and other rheumatic conditions are among the most common chronic diseases: they affected 70 million US adults in 2001, and are the major cause of disability among US adults (60% of US population older than 65 years) (1). These conditions are characterized by persistent inflammatory pain, which is treated with nonsteroidal antiinflammatory drugs or corticosteroids (2). However, long-term use of these drugs induces numerous gastrointestinal side effects (3,4), and their real impact on patient quality of life remains uncertain.The search for new pharmacologic treatments for these conditions continues (5-9), but preclinical evaluation of their effectiveness has lagged behind. The predictive ability of current techniques used to assess drug effectiveness remains approximate because they are based on reflex (10-12) or pseudo-affective (13,14) reactions to acute or subacute painful stimuli in animals, which have limited relevance to predictability in humans. Investigators have attempted to identify specific behaviors as markers of discomfort generated by pain (paw position [15], weight load on each leg [16], gait analysis [17], or locomotion [18]). However, there is only putative evidence that they may be of use in pharmacologic studies since they adopt a functional approach to disease rather than constituting a real evaluation of spontaneous pain.Many clinical studies on pain are based on genSupported by a grant from the Association pour la Recherche sur la Poly-arthrite.