Pain management of first-trimester surgical abortion: effects of selection of local anesthesia with and without lorazepam or intravenous sedation

ObjectiveThis prospective observational study compared the effectiveness and acceptability of three commonly used regimens to reduce pain during first-trimester surgical abortion with local anesthesia.MethodWomen undergoing suction curettage at less than 14 weeks gestation received preoperative ibuprofen and an intraoperative paracervical block. Participants then selected (1) no additional medication (local arm), (2) sublingual lorazepam (0.5 to 1 mg, approximately 20 min preoperatively) or (3) intravenous sedation with fentanyl and midazolam (doses from 50 to 125 μg of fentanyl and 1–2 mg of midazolam). Our primary outcome was intraoperative pain as measured on an 11-point verbal rating scale (0=no pain and 10=worst pain ever).ResultsThree-hundred thirty women were enrolled, 105 chose local, 106 opt for lorazepam and 119 picked intravenous sedation. Mean pain scores were 6.2 for local, 6.8 for lorazepam and 5.7 for intravenous sedation. Increased preoperative anxiety, depression and expected pain and a lower volume of local anesthesia used (10 mL vs. 20 mL) were positive predictors of intraoperative pain. Using multivariable analyses to control for these confounders, plus gestational age, alcohol use and body mass index, intravenous sedation was associated with a 0.86 point lower pain score (p<.005), and the local and lorazepam groups were not significantly different. Compared with the referent group (local), participants who received lorazepam were significantly less satisfied with pain control (23.8% vs. 6% unsatisfied; OR=1.93, 95% CI=1.13 to 3.26).ConclusionIn women who self-selected their type of anesthesia, intravenous sedation using fentanyl and midazolam decreased pain with suction curettage under local anesthesia. Sublingual lorazepam as studied did not decrease pain compared with local anesthesia alone and was associated with more dissatisfaction with pain control.