We conducted a blinded randomized clinical trial (registered at https://clinicaltrials.gov with the identifier NCT04867824, under the title “The Use of Lavender vs Vanilla Essential Oil as Complementary Analgesia for Frenotomy in Healthy Newborns”). Our hospital Ethics Committee (CEIm-PSMAR) reviewed and approved this study on May 4, 2021 (approval number: 2021/9731/I). This study was conducted according to the ethics code of the Barcelona Medical Association and the principles of the Helsinki-Fortaleza Declaration 2013.
This study was conducted at the neonatal unit of a tertiary care hospital in Barcelona, Spain, which experiences approximately 1,400 births per year from a multiethnic population with Spanish, Pakistani and Bangladeshi being the most frequent nationalities of our patients . We have high breastfeeding rates: around 85% at discharge from the maternity ward (86.8% in 2018) , 82% at the age of three months, and 54% at the age of six months. We assess for the presence of ankyloglossia as part of the routine neonatal evaluation and classify it based on Coryllos’s criteria  and the Hazelbaker tool . A lingual frenulum is symptomatic if it scores eight points or less in appearance and/or 11 points or less in function according to Hazelbaker. Advice and help with positioning and attachment for breastfeeding is provided to all the mothers by lactation support providers.
During the study period, if we identified a symptomatic patient with a type 3 tongue-tie which affected breastfeeding, we offered the patient's parents to participate. We chose type 3 tongue-ties because they are the most common in our population  and, due to their anatomical features (thick and submucosal), they seem to make breastfeeding more difficult. Patients were enrolled if their parents agreed to and signed a written informed consent prior to the procedure, then they were allocated into the experimental or the control group by simple random sampling using the program OxMAR (Online Minimization and Randomization for Clinical Trials) . Prior to recruitment, we generated a list of 142 numbers, where each number was randomized to either the “VEO” or “control (LEO)” group. Patients were enrolled in numerical order and assigned into the pre-determined group. The group into which a patient had been enrolled was not known by the attending personnel until the moment of performing the frenotomy.
To perform the frenotomy, the neonate was taken to the neonatal unit and monitored with a pulse-oximeter (COVIDIEN Nellcor Portable SpO2 Patient Monitoring System PM10N, Covidien Ireland Limited, IDA Business & Technology Park, Tullamore, Ireland) before, during and after the procedure. For both groups we swaddled, administered 1 mL of oral sucrose, and let the newborn suck for two minutes prior to the procedure. The control group had a 7 x 7 cm gauze pad with one drop of 100% pure LEO (Pranarôm España S.L.) placed 2 cm under their nose for two minutes prior to starting the frenotomy and for the duration of the procedure; whereas in the experimental group the drop on the gauze pad was of 100% pure VEO (Pranarôm España S.L.) instead. The bottles of both LEO and VEO have a dropper that always dispenses the same amount of oil per drop. We did not start the procedure until the patients were calm and had a NIPS score of 0. Frenotomy was performed by one of the three staff neonatologists by placing a sterile groove director under the tongue holding the frenulum in place with visualization of tongue base and frenulum, then snipping the frenulum with a blunt tip scissor along the underside of the tongue to its base just proximal to the genioglossus muscle until a full release was achieved . We assessed pain by means of crying time and the highest Neonatal Infant Pain Scale (NIPS) score in the five minutes post procedure, and whether there was an increase in heart rate (HR) and decrease in oxygen saturation (SatO2). NIPS evaluates facial expression, crying, breathing pattern, arm and leg position, and state of arousal on a scale from 0 to 7, where 0-2 means no pain to mild pain, 3-4 mild to moderate pain, and >4 severe pain . A blinded neonatologist who did not perform the frenotomy evaluated vital signs through the screen of the pulse-oximeter, NIPS score and crying time from a neighboring room through a glass, for which he/she could not smell or see which oil was being used. Vital signs, whether the baby cried or not, the seconds crying lasted, and the post procedure NIPS score were recorded on a data collection sheet. If a neonate cried, the attending staff who performed the frenotomy provided calming techniques such as holding, swaddling, and sucking regardless of which essential oil was being used. These persons were not blinded. Once the procedure was completed, we removed the gauze pad and returned the neonate to the mother for breastfeeding.
Calculation of sample size: In an exploratory preliminary study prior to the intervention, we observed a mean (SD) crying time of 19.80 (21.14) seconds. We used this data as our baseline. In order to detect a difference of 10 seconds in crying time, we calculated that we needed a sample size of 71 patients per group to be able to draw conclusions with a confidence interval (CI) 95% and a power of 80%. We used crying time to calculate sample size because it is an objective way to measure pain, whereas NIPS could be more person-specific.
The observer recorded demographic (sex, gestational age, birth weight, age in hours at the time of frenotomy) and clinical data (HR and SatO2 before, during, and after the procedure, whether the patient cried or not during the procedure, length of crying time in seconds, presence of side effects during the procedure (apnea, desaturation, others) and highest NIPS score within the first five minutes after the procedure) on a data collection sheet. The primary outcome was difference in crying time between the experimental and the control group, and secondary outcomes were: difference of NIPS score, HR and SatO2 pre and post-procedure between the experimental and the control group. Participants’ confidentiality was maintained because neither the name nor the medical record number were recorded on the data collection sheet.
Statistical analysis: Quantitative variables are described using the mean, standard deviation, and 95% CI; the experimental and the control groups were compared with a Student's t test. Qualitative variables are presented in percentages and compared using Fisher’s exact test. We compared NIPS scores between both groups using the Wilcoxon rank-sum (Mann-Whitney) test. Statistical significance was set for a p <.05. To perform statistical analyses we used STATA version 16.1 (StataCorp, College Station, TX, USA).