Gait abnormalities differentially indicate pain or structural joint damage in monoarticular antigen-induced arthritis Michael K. Boettger a, * , Konstanze Weber a , Manuela Schmidt b , Mieczyslaw Gajda c , Rolf Bräuer c , Hans-Georg Schaible a a Institute of Physiology I, University Hospital, Jena, Germany b Institute of Systematic Zoology and Evolutionary Biology, Friedrich-Schiller-University, Jena, Germany c Institute of Pathology, University Hospital, Jena, Germany article info Article history: Received 22 April 2009 Accepted 1 June 2009 Keywords: Antigen-induced arthritis Joint pain Gait analysis Pain-related behavior Locomotor behavior Etanercept abstract Gait abnormalities have been suggested to provide an objective measure for joint pain in animal models. Here, we aimed to assess whether parameters of gait analysis correlate with measures of pain-related behavior in experimental monoarthritis. For this purpose, antigen-induced arthritis was induced in the left knee joints of 68 female Lewis rats, of which 30 were treated with tumor necrosis factor- alpha(TNF)-neutralizing compounds. During the course of arthritis, paw print analysis parameters and measures for mechanical and thermal hyperalgesia were obtained. Knee joints harvested on either day 3 or day 21 were scored histologically for signs of inflammation and cartilage and bone destructions. Data were compared to those obtained from 33 immunized control rats and correlated for days 3 and 21. Arthritic rats showed distinct asymmetric gait abnormalities. In the acute stage of antigen-induced arthritis, but not in the chronic phase, there was a significant correlation between the gait parameter ‘left–right distance’ and measures of primary and secondary hyperalgesia. Both in the acute and chronic phases, however, the gait parameter ‘angle between paws’ indicating outward rotation of paws mainly correlated with joint destruction as assessed using histology. Etanercept treatment exhibited pronounced anti-nociceptive and pro-locomotional effects, but the described correlations remained. In conclusion, some parameters of gait analysis may represent a good measure for arthritis pain, mainly in acute inflam- mation, while others are increasingly influenced by mechanical joint deformation as indicated by carti- lage and bone destructions. Thus, gait abnormalities may not unequivocally be suitable for objective pain assessment in all stages of experimental arthritis. Ó 2009 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. 1. Introduction Animal models of experimental arthritis are being used to un- cover putative pathological mechanisms underlying human joint disease and to develop or improve therapeutic strategies. In con- trast to end-point analyses, in which mainly histological or inflam- matory changes are being assessed at one defined time point or after one specific treatment, functional studies have become more and more important in recent years [22]. This owes to the fact that pain and loss of normal joint function are common findings in both experimental and human arthritis [1,2,5,9]. For both phenomena, several behavioral tests have been developed that can be used in animal models [22]. One major problem in animal pain research is, however, to identify changes in behavior that reliably indicate a painful sensation and that, in addition, mimic the complaints of respective patients. When testing for pain-related behavior, stimulus-evoked changes in behavior are most widely used [3,4,22]. In these tests, limb withdrawal or vocalization is regarded to indicate discomfort, and thresholds to evoke these reactions can be interpreted as objective parameters for pain-related behavior. Hence, phenomena like hyperalgesia (or allodynia), resulting in decreased thresholds, can be quantified. However, patients suffering from arthritis do not necessarily complain about stimulus-evoked pain, but rather about pain at rest or movement-induced pain [17,36]. The assess- ment of such pain states in animal models proves unequally more difficult. Models of inflammation or neuropathy in the fore- or hindlimbs of animals have repeatedly employed gait analysis, since changes in locomotion are very likely to be associated with experi- enced pain and respective guarding behavior [8,24,37]. In some studies, gait analysis has even been discussed to reflect an objec- tive measure for pain in animal models of arthritis [7,34,35]. This appears reasonable, since the pain described by patients suffering from deep somatic pain like in rheumatoid arthritis or osteoarthri- tis is often worsened upon utilization of the joint, e.g. when 0304-3959/$36.00 Ó 2009 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.pain.2009.06.006 * Corresponding author. Address: Institute of Physiology I/Neurophysiology, University Hospital Jena, Teichgraben 8, 07743 Jena, Germany. Tel.: +49 3641 938864; fax: +49 3641 938812. E-mail address: Michael.Boettger@mti.uni-jena.de (M.K. Boettger). www.elsevier.com/locate/pain PAIN Ò 145 (2009) 142–150