Background: The ideal postoperative analgesia management of elderly surgical patients in intensive care units (ICUs) is continually being investigated. Objective: The purpose of this study was to assess the effectiveness and tolerability of IV administration of dexmedetomidine as an adjunct to a low-dose epidural bupivacaine infusion for postoperative analgesia after abdominal surgery in elderly patients in the ICU. Methods: ICU patients aged >70 years undergoing abdominal surgery were eligible for the study. A lumbar epidural catheter was inserted at the beginning of the surgery with no medication. On arrival at the ICU, the catheter was loaded with 0.25% bupivacaine 25 mg at the T8 to T10 sensory level, and a continuous infusion of 0.125% bupivacaine was started at 4 to 6 mL/h in combination with patient-controlled epidural analgesia (PCEA) of fentanyl (4 μg/bolus) for pain treatment. Patients in the treatment group received dexmedetomidine as an IV loading dose of 0.6 pg/kg for 30 minutes followed by continuous infusion at 0.2 μg/kg · h -1. Patients in the control group were not administered dexmedetomidine. The effectiveness of the pain relief was determined using a visual analog scale (VAS) (0 = no pain to 10 = worst pain imaginable) at rest. VAS score, heart rate (HR), systolic blood pressure (SBP), diastolic blood pressure, and arterial blood gases were monitored periodically for 24 hours after surgery. If required, tenoxicam (20-mg IV bolus) was used to ensure a VAS score of ≤3. The number of times PCEA and tenoxicam were administered and the occurrence of adverse events (AEs) were also recorded. Results: Sixty patients (34 men, 26 women; mean [SD] age, 75.96 [4.25] years; mean [SD] weight, 74.13 [10.62] kg) were included in the study. VAS scores were significantly lower in the dexmedetomidine group compared with the control group at hours 1, 2, and 12 (VAS [hour 1]: 2.8 [0.4], P < 0.001; VAS [hour 2]: 2.7 [0.5], P < 0.001; and VAS [hour 12]: 0.9 [0.7], P 0.044). The mean number of administrations of fentanyl via PCEA was significantly greater in the control group compared with the dexmedetomidine group (2.20 vs 6.63 times; P < 0.001). The mean number of administrations of tenoxicam was significantly lower in the treatment group than the control group (0.27 vs 1.07 times; P < 0.001). In the control group, the decreases in sedation at 0, 8, 12, 16, and 20 hours were significant compared with baseline ( P = 0.024, P = 0.001, P = 0.020, P < 0.001, and P = 0.005, respectively). Mean HR, SBR and AEs (eg, bradycardia [HR <60 beats/min], respiratory depression [respiratory rate <8 breaths/min], hypotension \\SBP <90 mm Hg], oversedation, hypoxia, and hypercapnia) decreased significantly in the dexmedetomidine group (all, P < 0.05). Significantly more patients in the dexmedetomidine group rated their satisfaction with postoperative pain control as excellent compared with the control group (12 vs 6 patients; P = 0.014). Conclusion: Intravenous dexmedetomidine was effective and generally well tolerated as an analgesic adjunct to epidural low-dose bupivacaine infusion for pain treatment, with lower need for opioids after abdominal surgery in these elderly intensive care patients than in the control group.