• Olubukola O Nafiu, Aleda Thompson, S Devi Chiravuri, Benjamin Cloyd, and Paul I Reynolds.
• From the Department of Anesthesiology, Section of Pediatric Anesthesiology, University of Michigan, Ann Arbor, Michigan.
• Anesth. Analg. 2019 Jun 1; 128 (6): 122512331225-1233.

BackgroundMany children recovering from anesthesia experience pain that is severe enough to warrant intravenous (IV) opioid treatment within moments of admission to the postanesthesia care unit (PACU). Postoperative pain has several negative consequences; therefore, preventing significant PACU pain in children is both a major clinical goal and a moral/ethical imperative. This requires identifying patient-level and perioperative factors that may be used to predict PACU IV opioid requirement. This should allow for the development of personalized care protocols to prevent clinically significant PACU pain in children. Our objective was to develop prediction models enabling practitioners to identify children at risk for PACU IV opioid requirement after various painful ambulatory surgical procedures.MethodsAfter Institutional Review Board approval, clinical, demographic, and anthropometric data were prospectively collected on 1256 children 4-17 years of age scheduled for painful ambulatory surgery (defined as intraoperative administration of analgesia or local anesthetic infiltration). Three multivariable logistic regression models to determine possible predictors of PACU IV opioid requirement were constructed based on (1) preoperative history; (2) history + intraoperative variables; and (3) history + intraoperative variables + PACU variables. Candidate predictors were chosen from readily obtainable parameters routinely collected during the surgical visit. Predictive performance of each model was assessed by calculating the area under the respective receiver operating characteristic curves.ResultsOverall, 29.5% of patients required a PACU IV opioid, while total PACU analgesia requirement (oral or IV) was 41.1%. Independent predictors using history alone were female sex, decreasing age, surgical history, and non-Caucasian ethnicity (model area under the receiver operating characteristic curve [AUROC], 0.59 [95% confidence interval {CI}, 0.55-0.63]). Adding a few intraoperative variables improved the discriminant ability of the model (AUROC for the history + intraoperative variables model, 0.71 [95% CI, 0.67-0.74]). Addition of first-documented PACU pain score produced a substantially improved model (AUROC, 0.85 [95% CI, 0.82-0.87]).ConclusionsPostoperative pain requiring PACU IV opioid in children may be determined using a small set of easily obtainable perioperative variables. Our models require validation in other settings to determine their clinical usefulness.

31 keywords...

hide…