Introduction Sepsis, a systemic inflammatory response to infective etiologies, has a high mortality rate that is linked both to excess cytokine activity and apoptosis of critical immune cells. Dexmedetomidine has recently been shown to improve outcome in a septic cohort of patients when compared to patients randomized to a benzodiazepine-based sedative regimen. We sought to compare the effects of dexmedetomidine and midazolam, at equi-sedative doses, on inflammation and apoptosis in an animal model of severe sepsis.Methods After central venous access, Sprague Dawley rats underwent cecal ligation and intestinal puncture (CLIP) with an 18 G needle without antibiotic cover and received either saline, or an infusion of comparable volume of saline containing midazolam (0.6 mg.kg-1.h-1) or dexmedetomidine (5 ug.kg-1.h-1) for 8 hours. Following baseline measurements and CLIP, blood was sampled for cytokine measurement (tumour necrosis factor (TNF)-alpha and interleukin (IL)-6; n = 4-6 per group) at 2, 4 and 5 hours, and animal mortality rate (MR) was monitored (n = 10 per group) every 2 hours until 2 hours had elapsed. In addition, spleens were harvested and apoptosis was assess* by immunoblotting (n = 4 per group).Results The 24 hour MR in CLIP animals (90％) w< significantly reduced by sedative doses of eith dexmedetomidine (MR = 20%) or midazoiam (MR = 30% While both sedatives reduced systemic levels of tt inflammatory cytokine TNF-alpha (P < 0.05); or dexmedetomidine reduced the IL-6 response to CLIP, thouc this narrowly missed achieving significance (P = 0.05 Dexmedetomidine reduced splenic caspase-3 expression (P 0.05), a marker of apoptosis, when compared to eith midazolam or saline.Conclusions Sedation with midazolam and dexmedetomidii both improve outcome in polymicrobial severely septic ra1 Possible benefits conveyed by one sedative regimen ov another may become evident over a more prolonged tim course as both IL-6 and apoptosis were reduced I dexmedetomidine but not midazolam. Further studies a required to evaluate this hypothesis.