Effect of midazolam in autism spectrum disorder: A retrospective observational analysis

While midazolam is commonly used as premedication for uncooperative patients, its effects are difficult to predict in patients with autism spectrum disorder for whom abnormalities in gamma-aminobutyric acid have been reported. This study aimed to investigate the influence of autism spectrum disorder on the effect of midazolam when used as premedication.

This retrospective observational study was performed between April 2017 and August 2018. Before inducing general anesthesia with sevoflurane for dental treatment, 390 uncooperative patients received premedication with midazolam. Ordinal logistic regression analysis was performed with the Observer's Assessment of Alertness/Sedation score 30 min after premedication as the objective variable. Age, sex, American Society of Anesthesiologists physical status class, premedication route, dose per body weight, presence of specific disorders (autism spectrum disorder, intellectual disability, epilepsy, cerebral palsy, and other psychiatric disorders), and regular benzodiazepine or non-benzodiazepine psychotropic administration were included as explanatory variables. Kendall's rank correlation coefficient was used to assess the correlation between the Observer's Assessment of Alertness/Sedation score and cooperation level (1, obvious negative response; 2, negative response; 3, positive reaction; 4, obvious positive reaction) during admission and inhalation induction. All data were extracted from anesthesia and medical records.

Age (odds ratio 1.437 [95% confidence interval (CI) 1.213-1.708], P < .001), autism spectrum disorder (1.318 [1.079-1.612], P = .007), benzodiazepine medication (0.574 [0.396-0.827], P = .002), and intramuscular route (1.478 [1.137-1.924], P = .004) were significantly associated with the Observer's Assessment of Alertness/Sedation score, while the score was negatively associated with cooperation levels during admission (τ = -0.714, P < .001) and inhalation induction (τ = -0.606, P < .001).

Patients with autism spectrum disorder may be susceptible to premedication with midazolam; however, regular benzodiazepine administration may reduce the effect.