Introduction

Since survival rates in prematurity have improved in the past decades, increasing numbers of former preterm infants have become eligible for surgery, with one of the most common operations being inguinal hernia repair. The preterm infants presenting for anesthesia in their early infancy are at increased risk of perioperative complications related to the immaturity involved with multiple organ systems1,2,3. Among the overall intraoperative adverse events in pediatric anesthesia, more than half were related to respiratory complications, especially being higher in infants below 1 year of age4,5.

Endotracheal tube (ETT) with mechanical ventilation is traditionally performed in neonatal anesthesia to secure the airway, especially in preterm infants, those with lower body weight and those with more co-morbidity 3,6. However, the risk of cardiovascular and respiratory disturbance, including bradycardia, fluctuation in blood pressure, and severe desaturation, is more commonly associated with tracheal intubation in neonates, with the risk inversely correlated to age 7,8. In a multicenter observational study, the incidence of difficult tracheal intubation was reported to be 5.8% in neonates and in infants, among whom 40% experienced desaturated and 8% developed bradycardia 9.

Supraglottic airway device (SAD) is a less invasive airway device compared to ETT and is an alternative airway to achieve effective ventilation in neonatal resuscitation when tracheal intubation is unsuccessful10,11. In pediatric anesthesia, it is related to less perioperative respiratory adverse events compared to ETT in children and in older infants12,13. In preterm infants receiving surgeries for retinopathy, SAD has been reported as a feasible airway device for general anesthesia14,15, with a retrospective survey finding that SAD is associated with less postoperative apnea, postoperative desaturation, and delayed extubation in former preterm infants receiving general anesthesia6,16. Aside from these surveys, prospective studies between SAD and ETT in this preterm population appear scarce. We conducted a randomized controlled trial to compare the perioperative respiratory adverse events between using supraglottic airway device and endotracheal tube in former preterm infants receiving general anesthesia for hernia surgery.

Materials and methods

Ethical approval and consent to participate

This randomized controlled trial was conducted from September 2019 to May 2021 at Kaohsiung Medical University Hospital. It was approved by the hospital’s Research Ethics Committee (Kaohsiung Medical University Hospital Institutional Review Board, KMUHIRB-F(II)-20190017), and registered at ClinicalTrials.gov (NCT03931902, first registered on 30/04/2019). Parental written informed consents were obtained from all families before surgeries. All the procedures were conducted in accordance with the Helsinki Declaration-2013.

Eligibility criteria and randomization

We enrolled former preterm neonates (gestational age < 37 weeks) in our neonatal ward (complete neonatal care unit and neonatal intensive care unit) scheduled to receive inguinal herniorrhaphy. Infants who were already intubated right before surgery, required long-term oxygen or mechanical ventilation dependence, having severe congenital cardiopulmonary disease, lacking birth history, and with postmenstrual age > 52 weeks or body weight < 2000 g at surgery were all excluded, as were those requiring emergent surgery as that in combination with other major organ surgeries. In our institute, elective surgeries in infants with respiratory tract infection within two weeks are routinely postponed. All participants were enrolled by two anesthesiologist (PY Hu and MP Su) who performed the anesthesia during this study.

The participants were randomly allocated into two equal-sized groups: the supraglottic airway device group (SAD group) or the endotracheal tube group (ETT group) using computer-generated random numbers which generated by one staff not involved in this study. The patients and their family were blinded to the allocation. The airway device used in the SAD group was a size 1 first-generation disposable SAD (Vital-Seal Laryngeal Mask, C-Bona Medical Corporation, Taiwan), and in the ETT group was a size 3 uncuffed ETT (Henan Tuoren Endotracheal Tube, Henan Tuoren Medical Device Co., Ltd., China). All patients received vital sign monitors after coming to the operation room including electrocardiography, noninvasive blood pressure monitor and peripheral oxygen oximetry.

The protocol of anesthesia

Anesthesia was induced with inhaled sevoflurane. Routine premedication was not used for our participants. Neuromuscular blockade was not applied in either group because the half-life of non-depolarizing neuromuscular blocking agents (cisatracurium and rocuronium at our hospital) was longer than the average surgical time for hernia repair at our institution6. Additionally, there are controversial debates regarding the use of succinylcholine in neonatal intubation. The allocated airway device was inserted by an experienced anesthesiologist after adequate anesthetic depth had been achieved. If desaturation (SpO2 < 90%) occurred before the airway device was successfully placed, the procedure would be paused immediately and the patient would be manually ventilated via a face mask. If successful ventilation could not be established through the assigned device within two insertion attempts, the other airway device, which meant ETT for the SAD group and SAD for the ETT group, would be applied in the third manipulation, and the patient would be excluded from the final analysis.

Anesthesia was subsequently maintained with sevoflurane. If spontaneous breathing could not achieve adequate ventilation, pressure support or pressure control ventilation would be used. Local analgesia was given by the surgeon before incision. Opioids and muscle relaxants were avoided in both groups during the whole study period. The initial setting of the fraction of inspired oxygen (FiO2) was 65% during induction and emergence and 45% during maintenance of anesthesia, although the anesthesiologist in charge of each patient could adjust the FiO2 any time for the patient’s safety at their expertise and discretion.

After the surgery was completed and the patient had awoken, the SAD or ETT was removed according to the clinical criteria for extubation; then, the patient was transferred to the post-anesthesia care unit (PACU) accompanied by a pediatric anesthesiologist and an anesthetic nurse. After they meeting the PACU discharge criteria, the patient would be further transported to a neonatal ward accompanied by a pediatric doctor and a ward nurse with a portable oxygen monitor being used during all transportation. In the neonatal ward, infants are routinely monitored by electrocardiography and peripheral pulse oximeter.

Data collection

Patient demographics were collected from medical records. Respiratory-related medical conditions since birth and within two weeks prior to surgery were also documented including apnea, oxygen therapy, and methylxanthines (aminophylline and theophylline) treatment. Regarding intraoperative respiratory adverse events, coughing lasting more than 5 s and stridor during emergent period were recorded. Laryngospasm and bronchospasm indicated airway obstruction with increased respiratory effort, wheezing or vocal cord tightly closed as observed from direct laryngoscopy. Desaturation was defined as the readings of a peripheral pulse oximetry below 90%. Delayed extubation meant the patient could not met the clinical extubation criteria immediate after surgery and the airway device remained in situ when the patient left the operation room. Any postoperative 24-h respiratory adverse events were recorded from PACU and the neonatal ward. Bradycardia was defined as heart rate below 100 beats per min, while apnea was defined as loss of spontaneous breathing for more than 20 s or if associated with desaturation/bradycardia. The need for supplemental oxygen was up to the clinical decision from the pediatric physicians in charge of the neonatal ward.

Sample size and statistical analysis

Based on the retrospective data in our hospital6, the overall incidence of perioperative respiratory adverse events in former preterm infants undergoing general anesthesia for herniorrhaphy was 47% in whom receiving ETT as airway and 0% in those using SAD. To reduce the perioperative respiratory adverse events from 45 to 5% and considering the drop-off rate of 20%, a size of 20 patients in each group was required at a power of 80% and type I error of 0.05. Statistical analysis was performed using the SPSS software (version 28.0, IBM Corp., Armonk, NY, USA). Differences among each group were calculated by the Pearson Chi-square test or Fisher’s exact test for categorical data including laryngospasm, bronchospasm, severe cough, delayed extubation or prolonged oxygen dependence, apnea and bradycardia as appropriate, the Mann–Whitney U test was applied for consecutive data, while multivariable logistic regression was performed to detect potential risk factors for intraoperative and postoperative adverse event. P values less than 0.05 were considered statistically significant.

Results

Among the 43 eligible infants during the study period, two were declined to participate by parents and one was excluded before surgery because of the shortage of studied airway device due to the COVID pandemic (Fig. 1). Of the 40 former preterm infants enrolled into randomization, the assigned airway devices failed to successfully ventilate within two insertion attempts in two infants: one in the SAD group and the other in the GA group. They both experienced desaturation during the induction of anesthesia, but were successfully ventilated by the other airway device (SAD for the one in the ETT group, and vice versa) without other adverse events and were excluded in the final analysis.

Figure 1
figure 1

Flow diagram.

Full size image

Patient characteristics are shown in Table 1. Two out of 40 were former extremely preterm (gestational age below 28 weeks), 11 were very preterm (between 28 and 32 weeks), and the other 27 infants were moderate to late preterm (between 32 and 27 weeks). Very low birth weight infants (birth weight below 1500 g) accounted for 40%. Twenty-one infants were diagnosed as apnea of prematurity while 32 had history of using supplemental oxygen or invasive/non-invasive respiratory support. The postmenstrual age at surgery was between 34 to 45 weeks for all studied infants.

Table 1 Patient characteristics.
Full size table

The mean surgical time was 20 min (IQR 15–30 min in the SAD group, 15–25 min in the ETT group, p = 0.968), and the mean anesthesia time was 35 min (IQR 25–48 min) in the SAD group and 39 min (IQR 26.25–50 min) in the ETT group without significant difference (p = 0.445). All airway devices were removed in the operation room immediately after surgery, and no patient was re-intubated within postoperative 24 h.

Thirty-eight infants, 19 in each group, were enrolled in the analysis for intraoperative and postoperative respiratory adverse events. The desaturation rate was significantly lower in the SAD group then in the ETT group (risk ratio = 0.286) (Table 2). In the SAD group, desaturations occurred during induction in 4/19 (21.1%) of infants, during emergence in 1/19 (5.3%), and none in emergence. For the ETT group, the infants who ever recorded SpO2 below 90% were 9/19 (47.4%) during induction, 1/19 (5.3%) during maintenance, and 9/19 (47.4%) in emergence (Table 3). All the desaturation events during the induction of anesthesia occurred after the start of SAD or ETT manipulation. No participants developed bradycardia accompanied with these desaturation events.

Table 2 Perioperative respiratory adverse events.
Full size table
Table 3 Details of the timing distribution in desaturated infants.
Full size table

The second attempt for airway device insertion was required in three patients: one in the SAD group and two in the ETT group. The potential risk factors for intraoperative desaturation were surveyed by multivariable logistic regression including gestational age, postmenstrual age, weight at birth and at surgery, latest hematocrit level, time of surgery and anesthesia, preoperative apnea, aminophylline or theophylline treatment and supplementation of oxygen within two weeks prior to surgery. No independent risk factor was identified except for the airway device.

Laryngospasm and bronchospasm occurred in three infants in each group during induction of anesthesia. One of whom in the ETT group developed stridor after extubation but recovered soon before transferring to PACU; the other infants' respiratory events were relieved by inhaled anesthetics. One patient in the ETT group experienced a coughing incident for more than 5 s during intubation.

During postoperative 24 h, the prevalence of postoperative apnea, bradycardia and need of supplemental oxygen in the total 40 former preterm infants receiving general anesthesia for herniorrhaphy was 5%, 7.5% and 7.5% respectively, without significant deference between airway devices. Two patients receiving theophylline treatment for apnea of prematurity during the perioperative period experienced postoperative apnea but the symptom was not identified within two weeks prior to anesthesia. In the three patients experiencing postoperative bradycardia, one was currently under theophylline treatment, one was free from theophylline for four days, and the other one was free from caffeine therapy for 12 days. In the three infants requiring supplemental oxygen, one had accompanying postoperative apnea, one had bradycardia, and the other was currently using theophylline since before surgery.

Discussion

To our knowledge, this is the first randomized controlled trial comparing SAD and ETT in neonatal anesthesia for former preterm neonates receiving short-duration surgery.

The higher desaturation rate in the ETT group, both during induction and emergence, might be related to several reasons. Firstly, the time for airway device placement is usually shorter by using SAD then using ETT. In one report, the SAD insertion time is averagely less then 10 s, and the ETT intubation time is suggested to be limited in 30 s in neonatal resuscitation; in another study compared the use of SAD and ETT for surfactant administration, the time for SAD placement is significantly shorter than for ETT intubation17,18. However, the median time from ceasing positive ventilation to SpO2 < 90% is 22 s in preterm infants8. The longer procedure duration for airway establishment with ETT might contribute to the higher desaturation rate during induction of anesthesia. Secondly, we observed that a significant number of infants in the ETT group had transient breath-holding or periodic breathing immediately after extubation, which may relate to the desaturation during the emergence of anesthesia, although in a meta-analysis comparing the airway complications of tracheal extubation and SAD removal in children19, no difference was found in incidence of breath-holding between the two airway devices. Nonetheless, in the population of former preterm infants, the physiological change between tracheal extubation and SAD removal might require further evaluation. Thirdly, the relatively low FiO2 in our study design might have increased the risk of perioperative desaturation, as we avoided higher FiO2 in the perioperative period owing to the potential for hyperoxia-induced injury in prematurity20. In our study, most of the transient desaturation during the induction and emergence of anesthesia ranged in SpO2 from 85 to 89%. In a large prospective meta-analysis21, the lower and higher SpO2 targets (85–89% vs. 91–95%) showed no significant difference in the composite of death or major disability in extremely preterm infants, although the lower SpO2 target was related to a higher risk of death and necrotizing enterocolitis but a lower risk of premature retinopathy. Adequate FiO2 in neonatal anesthesia, especially in preterm infants, requires further study.

Our report represents no delayed extubation in both SAD and ETT groups. In two previous prospective studies comparing general anesthesia and regional anesthesia in former preterm or high-risk infants undergoing herniorrhaphy22,23, the incidence of prolonged ventilator use after general anesthesia was 20% and 35.7% respectively, and all the infants with general anesthesia received neuromuscular blockade to facilitate tracheal intubation. In a retrospective study comparing SAD and ETT for general anesthesia in a similar population, 40% patients in the ETT group had delayed extubation, with all receiving neuromuscular blockade6. We suppose that the muscle relaxant agent might be the mainly potential cause of postoperative prolonged ventilatory support in these high-risk infants receiving short-duration surgery.

Typically, desaturation of surgical neonates may accompany by breath-holding, bradycardia, wheezing or tachypnea. Immediate treatment to restore patent airway and adequate oxygenation is mandatory such as close monitoring, oxygen supplement via face mask, airway management including clean secretion and gentle positive pressure ventilation. In our study overall 18 of 40 neonates developed desaturation (4 in SAD group, 14 in ETT group) and none of them had major adverse events after timely airway intervention.

Incidences of postoperative apnea in PACU and neonatal ward in our study showed no statistically significant differences between groups, although two cases both occurred in the ETT group. The relatively small number of patients in the study design might limit this result. In our experience, more breath-holding immediately after tracheal extubation in the operation room was also observed after SAD removal, although patients recovered soon before transferring to PACU. In a retrospective study in preterm infants receiving surgery for retinopathy, SAD was related to less apnea after extubation than ETT16. In a multicenter randomized prospective trial, the incidence of apnea in infants below 60 weeks of postmenstrual age was similar following general anesthesia and regional anesthesia in postoperative 12 h, but was slightly higher in infants receiving general anesthesia within postoperative 30 min24. Further large studies focusing on the timing of life-threatening apnea in this former preterm population are recommended.

There are several limitations in this prospective study: firstly, the small sample size might mask the difference in postoperative respiratory adverse events between groups, especially for the incidence of postoperative apnea. Secondly, the time-to-time intraoperative anesthetic depth and the ventilatory setting could not be retrieved from our electronic anesthetic equipment, which might be related to the breath-holding phenomenon in the emergence period. Thirdly, the airway insertion time was not recorded according to the initial study design, because both the anesthesiologist involved in this study being experienced and the practice are assumed to be consistent with clinical guidelines. Fourthly, concerning the logistical regression model, too many risk factors are added for the outcome. This limitation suggests that the associations identified could be less reliable and will be cautiously considered in future analyses to enhance the validity of our findings.

In conclusion, our results provide strong evidence that SAD use is associated with fewer complication in former preterm infants receiving general anesthesia for herniorrhaphy as it much decreases the incidence of intraoperative desaturation compared to ETT. The desaturation related to ETT could occur in both induction and emergence periods; however, no significant difference in other perioperative respiratory adverse events between these two airway devices including the incidence of apnea in postoperative 24 h was found, although the small sample size of this study might limit this result. Infants receiving both airway devices were all successfully extubated in the operation room and this might be related to the avoidance of muscle relaxant agent use in our study design.