Abstract
AIMS: Dexmedetomidine is highly specific α2-adrenoceptor agonist. A single bolus of dexmedetomidine can achieve clinical therapeutic effect. Therefore, it is essential to know the safety margin between the clinical effectiveness dosages of dexmedetomidine and its side effect. METHODS: A total of 42 patients who underwent elective thyroidectomy were enrolled in this study. Dexmedetomidine was given as a single bolus injection 30 min towards the end of surgery. The up-and-down sequential schedule was used in this study. The starting dose of dexmedetomidine was set at 0.1 μg/kg in the first patient and the next patient would then receive a dose of dexmedetomidine decremented by 0.05 μg/kg if the prior patient's baseline heart rate (HR) had a decrease of ≥20% and/or mean arterial blood pressure (MAP) increase or decrease of ≥20%, otherwise, the following patient would receive an incremental 0.05 μg/kg dose of dexmedetomidine. The analytic techniques of linear, linear-logarithmic, exponential regressions and centred isotonic regression were used to determine the ED50 of dexmedetomidine and the residual standard errors were calculated for the comparison of goodness of fit among the different models. RESULTS: The median (interquartile range [range]) lowest HR was 57 beats/min (53-63.3[46-76]) with an average HR decrease of 8.0 beats/min (5-13 [4 to 23]). The median (interquartile range [range]) highest MAP was 98 mmHg (91.8-105 [83-126]) with a MAP increase of 10.0 mmHg (6.8-18.0 [2-24]). The ED50 (95% confidence interval) from 4 different statistical approaches (linear, linear-logarithmic, exponential regressions and centred isotonic regression) were 0.262 μg/kg (0.243, 0.306), 0.252 μg/kg (0.238, 0.307), 0.283 μg/kg (0.238, 0.307), and 0.278 μg/kg, respectively. Among the 4 models, the exponential regression had the least residual standard error (0.03618). CONCLUSION: The ED50 derived from 4 statistical models for an intravenous bolus of dexmedetomidine without significant haemodynamic effects was distributed in a narrow range of 0.252-0.283 μg/kg, and the exponential regression was the model to best match the study data.