Plasma concentration based response surface model predict better than effect-site concentration based model for wake-up time during gastrointestinal endoscopy sedation.

BACKGROUND: Sedation for esophagogastroduodenoscopy (EGD) and colonoscopy is characterized by rapid patient induction and emergence. The drugs midazolam and alfentanil have long been used for procedural sedation; however, the relationship between plasma or effect-site concentrations (Cp or Ce, respectively) and emergence remains unclear. The aim of this study is to develop patient wake-up prediction models for both Cp and Ce using response surface modeling, a pharmacodynamics tool for assessing patients' responses.

METHODS: The Observer's Alertness/Sedation (OAA/S) score was used to monitor sedation depth during the examinations. Concentration pairs of midazolam and alfentanil were calculated for each of Cp and Ce using pharmacokinetic simulation software. Response surface models were developed using the Greco construct. Temporal analysis was done by comparing model-predicted wake-up time with true patient wake-up time.

RESULTS: Thirty-three patients with an average body mass index of 21.85 ± 2.3 kg/m2 were pooled for analysis. The average duration of examination were 2.9 ± 1.4 min for EGD and 6.6 ± 2.7 min for colonoscopy. Seventy-five concentration pairs of midazolam and alfentanil were obtained for each Cp and Ce. The Cp-based Greco response surface model showed significant synergy between midazolam and alfentanil and was a better predictor of patient wake-up time, with an average deviation of 1.0 ± 3.9 min, while the Ce model show time deviation greater than 20 min.

CONCLUSION: The early phases of drug distribution are unique and complicated by nonsteady-state concentrations, and our study revealed that Ce-based wake-up time prediction is more difficult under these circumstances.