Prophylactic analgesic administration reduces pain behavior after pediatric bilateral myringotomy and pressure equalization tube placement (BMT). We hypothesized that postoperative pain in children treated with intraoperative ketorolac would, among several exposures of interest, be strongly associated with ear condition. We conducted a retrospective cohort study of healthy children (9 months to 7 years) who underwent BMT at the Children's Hospital of Philadelphia or its ambulatory surgery centers from 2013 to 2016. Anesthetic care included preoperative oral midazolam, sevoflurane/nitrous oxide (N2O)/air/oxygen (O2) by mask, and intramuscular ketorolac. Demographic and procedural information included left and right tympanic membrane (normal, retracted, or bulging) and middle ear (normal/no, serous, mucoid, or purulent effusion) conditions. Because tympanic membrane and middle ear conditions were highly concordant and mean maximum Face, Legs, Activity, Cry and Consolability (FLACC) scores (0-10) were not different across the array of abnormal findings, we categorized each ear as normal or abnormal based on middle ear effusion alone. We then defined the ear condition of each child (primary exposure) using bilateral findings: normal/normal, normal/abnormal, and abnormal/abnormal. Secondary exposures included age, BMT history, procedure duration, facility location, and attending surgeon/anesthesiologist pair. The primary outcome was maximum postanesthesia care unit FLACC score: 4-10 (moderate-to-severe pain) versus 0-3 (no-to-low pain). Rescue oxycodone, acetaminophen administration, and emergence agitation were secondary outcomes. Statistical analysis incorporated generalized linear mixed models with random intercepts accounting for the clustering effect of provider pairs. Adjusting for multiple comparisons, significance level was set at P = .004. Excluding recurrent cases, 1922 unique evaluable subjects remained. The probability of moderate-to-severe pain behavior (FLACC, 4-10) was 52.4% (95% confidence interval [CI], 50.2-54.6) overall. In a confounder-adjusted model, ear condition was significantly associated with moderate-to-severe pain: compared to bilateral abnormal (effusions), odds ratio (OR) (95% CI) for bilateral normal was 2.2 (1.6-2.9), P < .0001. Younger age (OR, 1.1 [1.1-1.2] per year; P = .001) and longer procedure duration (OR, 1.1 [1.0-1.2] per minute; P = .0008) were likewise related to higher pain. With surgeon added to the model, variance explained by provider pairs decreased from 9.60% to 1.05%. Two secondary outcome associations also emerged: comparing bilateral normal to abnormal ears, ORs were 1.7 (1.3-2.2), P = .0001, for rescue oxycodone and 2.0 (1.2-3.3), P = .008, for emergence agitation. Pain behavior after BMT varies by surgeon and is strongly associated with ear condition. Ketorolac as a single prophylactic analgesic appears less effective in younger children with normal middle ear findings.
Read full abstract