The reliability of preemptive analgesia is controversial. Its effectiveness may vary among anatomical areas or surgical types. We evaluated preemptive analgesia by epidural morphine in six surgery types in a randomized, double-blind manner. Pain intensity was rated using a visual analog scale, a verbal report, and a measurement of postsurgical morphine consumption. Preemptive analgesia was effective in limb surgery and mastectomy, but ineffective for gastrectomy, hysterectomy, hemiorrhaphy, and appendectomy. Relief of postsurgical pain in hemiorrhaphy was more rapid than that in the other surgery types. Preemptive analgesia was effective in Limb surgery and mastectomy, but not in surgeries involving laparotomy, regardless of whether the surgery was major (gastrectomy and hysterectomy) or minor (herniorrhaphy and appendectomy). These results suggest that viscero-peritoneal nociception is involved in postsurgical pain. The abdominal viscera and peritoneum are innervated both heterosegmentally tin duplicate or triplicate by the vagus and/or phrenic nerves) and segmentally Coy the spinal nerves). Therefore, supraspinal and/or cervical spinal neurons might be sensitized, despite the blockade of the segmental nerves with epidural morphine. The rapid retreat of the pain after hemiorrhaphy suggests that central sensitization remits soon after minor surgery, but that in appendicitis, it may be protracted by additional noxious stimuli, such as infection. Implications: Epidural preemptive analgesia was reliably effective in limb and breast surgeries but ineffective in abdominal surgery, suggesting involvement of the brainstem and cervical spinal cord via the vagous and phlenic nerves.