Sedation can increase patient comfort during spinal anaesthesia. Understanding the relationship between the propofol effect-site concentration (Ce) and patient sedation level could help clinicians achieve the desired sedation level with minimal side effects. We aimed to model the relationship between the propofol Ce and adequate and deep sedation and also incorporate covariates. Thirty patients scheduled for orthopaedic surgery received spinal anaesthesia with 0.5% bupivacaine. Propofol was administered via an effect-site target-controlled infusion device using the Schnider pharmacokinetic model. The pharmacodynamic models for both adequate sedation [Observer's Assessment of Alertness/Sedation (OAA/S) scores of 3-4] and deep sedation (OAA/S scores of 1-2) were developed using nonlinear mixed-effects modelling. Increments in the propofol Ce were associated with increased depths of sedation. In the basic model, the estimated population Ce50 values for adequate and deep sedation were 0.94 and 1.52 mug/ml, respectively. The inclusion of the patient's age and sensory block level for adequate sedation and of age for deep sedation as covariates significantly improved the basic model by decreasing the objective function's minimum value from 10696.72 to 10677.92 (p = 0.0003). The simulated Ce50 values for adequate sedation in 20-year-old patients with a T12 sensory level and in 80-year-old patients with a T4 level were 1.63 and 0.53 mug/ml, respectively. Both age and sensory block level should be considered for adequate sedation, and the propofol concentration should be reduced for elderly patients with a high spinal block to avoid unnecessarily deep levels of sedation.