The identification, in advance, of patients at risk for chronic post-operative pain is of obvious importance. Diffuse noxious inhibitory control (DNIC) is an endogenous mechanism of pain suppression, which can be explored experimentally by a 'pain inhibits pain' testing paradigm, activating the brainstem-spinal descending modulation system. We aimed to assess whether DNIC as tested in pre-operative pain-free patients, can predict the risk for developing chronic post-thoracotomy pain. DNIC was assessed before thoracotomy in 33 patients by (i) administration of painful contact heat stimulus to the dominant forearm for 30", (ii) 1' immersion of the non-dominant hand in 46.5oC water, and (iii) repeated administration of stimulus i during the second half of stimulus ii. The extent of pain reduction between initial and subsequent contact stimuli was calculated as DNIC, representing the individual's ability to modulate pain. Patients were followed up for acute pain scores during hospitalization and for chronic pain levels at clinic visits 6 weeks or more (mean follow-up time 10.5±3.6) after surgery. The extent of pre-surgery DNIC was -16.6±17.4, range -63 (most effective) to +10 (least effective). Patients examined at 6 weeks or more reported mean pain scores of 41.3±28.8. Linear regression model revealed that chronic post-operative pain intensity was strongly predicted by DNIC (p=0.0005), and was not related to time elapsed since surgery. Acute post-operative pain, was also associated with greater chronic pain, but, was not predicted by DNIC. After using partial correlation to control for effects of acute post-operative pain, DNIC was found as the stronger predictor of chronic post-operative pain. DNIC can predict development of chronic post thoracotomy pain, suggesting this pain test as a useful one in determining the likelihood to develop chronic pain in a wide range of potentially painful medical and surgical interventions.