Objective     To observe the sedation of midazolam（MIZ）injected into subarachnoid cavity for lumbar anesthesia and its dose-response relationship

and compare with that of MIZ injected via muscles and veins.      Methods     Two hundred patients (82 males and 118 females) at the age of 16-50 (mean 34.5±16.3) admitted to our hospital for lower abdominal and extremity operation under lumbar anesthesia from December 2008 to June 2009 were enrolled in this study. The patients were randomly divided groups A to D (50 in each group). Patients in group A received lumber anesthesia with 2 mg MIZ through the vein and those in group B received lumber anesthesia with 2 mg MIZ through muscles

while patients in group C received lumber anesthesia with 1.6 ml 0.75% bupivacaine +0.4 ml 50% glucose solution +0.2 ml 5 mg/ml MIZ through muscles and those in group D received lumber anesthesia with 1.6 ml 0.75% bupivacaine +0.4 ml 50% glucose solution +0.4 ml 5 mg/mL MIZ through vein. From 5 to 60 min after MIZ injection

sedation

blood pressure

heart rate

pulse oxygen saturation

and glossoptosis in 4 groups were compared.      Results     The sedation score was lower in group B than in group A and higher in groups D and C than in group A（P0.01 or P0.05）. No glossocoma occurred in groups A to C except in group D（P0.05）.      Conclusion     Lumber anesthetics plus a small dose of MIZ can produce a rather good sedation. The sedation of MIZ is more significant than that of other anesthetics injected through muscles or veins

which may be related with the inhibition of central nervous system due to the binding of MIZ to its corresponding receptors in subarachnoid cavity.