This study hypothesized that different types of anaesthesia management would result in similar postoperative sleep quality. In this prospective single-blind investigation, 219 patients undergoing elective thoracic surgery were randomized into three arms: general anaesthesia, as the control group (group C); general anaesthesia combined with thoracic epidural anaesthesia (TEA) (group E); and general anaesthesia combined with infusion of 1μg/kg dexmedetomidine (group D). Plasma samples were obtained to measure the amine and inflammatory cytokine concentrations. All patients underwent assessment with the bispectral index (BIS) for sleep quality and the visual analogue scale (VAS) for pain. The primary outcomes were inflammatory cytokine [interleukin-6 (IL-6) and tumor necrosis factor alpha (TNF-α)] secretion and postoperative sleep quality on the first and second postoperative nights. The secondary outcomes were amine (adrenaline and noradrenaline) secretion during the surgical period and haemodynamic stability. The postoperative BIS area under the curve was significantly lower in group E (75.7%) than in group C (87.3%) or group D (86.5%). Patients in group E had the highest BIS of sleep efficiency index (29.2%,P< 0.05) and the lowest VAS scores (3.5,P< 0.05). Group E had lower IL-6 levels than the other two groups 24 h after surgery (P< 0.05). Patients given TEA may show reduced sleep disturbances on the first night after surgery, perhaps due to better pain management and inhibition of IL-6 release.