Neurodevelopmental outcome at 5 years of age after general anaesthesia or awake-regional anaesthesia in infancy (GAS): an international, multicentre, randomised, controlled equivalence trial

August 01, 2019


McCann ME, Graaff JC, Dorris L, Disma N, Withington D, Bell G, Grobler A, Stargatt R, Hunt RW, Sheppard SJ, Marmor J, Giribaldi G, Bellinger DC, Hartmann PL, Hardy P, Frawley G, Izzo F, Sternberg BS, Lynn A, Wilton N, Mueller M, Polaner DM, Absalom AR, Szmuk P, Morton N, Berde C, Soriano S, Davidson AJ, for the GAS Consortium. Lancet 2019; 393 (10172):664-677. doi: 10.1016/S0140-6736(18)32485-1. PMID: 30782342


Ankita Shukla, MD
Assistant Professor of Pediatrics,
Arkansas Children’s Hospital
University of Arkansas for Medical Sciences

Vikas Chowdhary, MD
Assistant Professor of Pediatrics,
Arkansas Children’s Hospital
University of Arkansas for Medical Sciences


Treatment, Prognosis


(P) For infants less than 60 weeks postmenstrual age, that were born at more than 26 weeks gestational age and undergoing inguinal herniorrhaphy, (I) effect of general anesthesia was compared with (C) awake regional anesthesia (O) on neurodevelopmental outcomes (T) over a period of 5 years.


  • Design: multicenter, international, parallel-group, randomized, assessor-masked, controlled, equivalence trial
  • Allocation:   Concealed- Anesthetists were aware of group allocation, but individuals administering the neurodevelopmental assessments were not. Parents were informed of their infants group allocation upon request, but were told to mask this information from assessors
  • Blinding: Blinded
  • Follow-up period: Infants were followed until 5 years of age
  • Setting: 28 hospitals in seven countries- Australia, Italy, the USA, the UK, Canada, the Netherlands, and New Zealand
  • Patients:
    • Inclusion Criteria:  infants of less than 60 weeks’ postmenstrual age who were born at more than 26 weeks’ gestation and were scheduled for inguinal herniorrhaphy
    • Exclusion Criteria:
      • Any contraindication for either anesthetic technique used in the study
      • History of congenital heart disease requiring surgery or pharmacotherapy
      • Mechanical ventilation immediately before surgery
      • Known chromosomal abnormalities or other known acquired or congenital abnormalities that might affect neurodevelopment
      • Previous exposure to volatile general anesthesia or benzodiazepines as a neonate or in the third trimester in utero
      • Any known neurological injury such as cystic periventricular leukomalacia or grade three or four intraventricular hemorrhage
      • Any social or geographical factor that might make follow-up difficult
      • Use of a primary language at home in a region where neurodevelopmental tests were not available in that language
  • Intervention: Awake-regional (AWR) group received a spinal, caudal, or combined caudal and spinal anesthetic, according to institutional preferences. Bupivacaine or levobupivacaine at a dose of 0·75–1 mg/kg was administered for spinal anesthesia. Caudal anesthesia was with 0·25% bupivacaine or levobupivacaine up to a total dose of 2·5 mg/kg. General anesthesia (GA) group received sevoflurane for induction and maintenance in a mix of air and oxygen. The concentration of sevoflurane, choice of airway device, ventilation technique, and use of neuromuscular blocking agents were left to the preference of the anesthetist.
  • Outcomes: Neuropsychological assessments were to be done within 4 months of the child turning 5 years of age
    • Primary outcome:  Wechsler Preschool and Primary Scale of Intelligence, third edition (WPPSI-III) full-scale intelligence quotient (FSIQ) score
    • Secondary outcomes:  Selected NEuroPSYchological (NEPSY-II) subtests to assess attention and executive function; the Wechsler Individual Achievement Test, second edition (WIAT-II),orthe BVN (the Italian equivalent of the WIAT- II); selected subtests of the Children’s Memory Scale (CMS); the global executive composite of the Behaviour Rating Inventory of Executive Function, Preschool version (BRIEF-P); the Adaptive Behaviour Assessment System, second edition (ABAS-II); and the Child Behaviour Checklist caregiver questionnaire (CBCL)
  • Analysis and sample size: The authors hypothesized that the WPPSI-III FSIQ score at age 5 years would be equivalent between study groups.
    • Outcomes were analyzed on a per-protocol basis to ensure a conservative estimate of the treatment effect in the direction of non-equivalence. Equivalence was defined a priori as the 95% CI of the difference in means of the FSIQ lying within –5 and +5 IQ points.  Sample size assumed a 90% chance that the 95% CI would exclude a difference of more than 5 points.  Accounting for a 10% loss to follow-up and a 10% rate of major protocol violations, about 720 participants were needed.
    • Multiple imputations were done under a multivariate normal distribution to impute missing outcome data in the primary analysis of all outcomes, with a sensitivity analysis done on only complete cases.
    • This study screened 4023 infants, enrolling 722 participants. 5-year assessment was done for 447 patients, 205 of which received awake-regional anesthesia and 242 received general anesthesia. The variables used in the multiple imputation models included baseline, post-randomization, 2-year cognitive variables and 5-year outcome variables.
  • Patient follow-up: Of the enrolled infants, 447 infants were analyzed for difference in neurodevelopmental outcomes based on type of anesthesia at the end of 5 years.


There was no difference in baseline demographics, pregnancy and birth details, familial demographics as well as anesthesia variables (i.e. need for fluid bolus for hypotension, duration of surgery and sevoflurane exposure, significant postoperative apnoea, use of rescue IV glucose or vasoactive drugs). The study population had a male preponderence  (84% participants were males), given the procedure performed. The chronological age at surgery was 68.9 (mean 30.8) days vs 71.1 (mean 31.7) days and weight of child at surgery was 4.2 kg (1.1) vs 4.3 (1.1) kg in AWR and GA groups respectively

The mean duration of general anesthesia was 54 mins (IQR 41-70)

Infants that received sevoflurane required more fluid boluses (17% vs 5%) and vasoactive medications (5% vs 1%).

19%  ( 69 of 361) infants in AWR group had some exposure to the genreal anesthetics due to required deviation during surgery when optimal sedation or pain control could not be acheived


The mean FSIQ score was 99·08 (SD 18·35) in the AWR group and 98·97 (19·66) in the GA group. The adjusted mean difference in FSIQ for AWR minus GA was 0·16 (95% CI −2·45 to 2·78) in the intention-to-treat analysis and 0·23 (95% CI −2·59 to 3·06) in the per-protocol analysis. This difference was not significant.


There were no significant differences in a range of secondary neurocognitive and behavioral outcomes


The authors conclude that less than 1 h of general anesthesia in early infancy does not alter neurodevelopmental outcome at age 5 years compared with awake-regional anesthesia in a predominantly male study population.


Neurotoxicity risks with general anesthesia administration have been a much-debated topic in the recent decade. Preclinical studies in rodents, rats and monkeys have repeatedly shown accelerated apoptosis and neuronal cell death in rapidly growing preterm and term brains on exposure to general anesthetic agents. (1,2,3) In 2017, FDA issued an advisory statement cautioning that usage of general anesthesia “for lengthy periods of time or over multiple surgeries or procedures may negatively affect brain development in children younger than 3 years”, leading to mediaction label warnings.(4) Clinical evidence though remains limited and divided. In some smaller observational cohort studies and case reports, general anesthesia use has been shown to have no negative effect on behavour and cognition (5), yet others reported association of general anesthetic agents with learning disabilities (6,7,8). Most cohort studies are hampered with small numbers and varied population as well as variable outcome measures, making their generalizability restricted.

This is the first multi center randomized controlled equivalence trial comparing the effects of general vs regional anesthesia on neurodevelopmental outcomes for newborns greater than  26 weeks gestation and younger than 60 weeks post menstrual age, that underwent inguinal hernia repair. The trial showed strong evidence for equivalence in both primary and secondary outcomes measured at 2 years (interim analysis) and then at the 5 year follow up.

A major strength of this study is its size, possible only with multi center collaboration lending strong credibility to the results. Other strengths would be the use of comprehensive analytical techniques such as multiple imputations to account for missing data and loss to follow up. The analysis was done on a per-protocol basis adjusting for gestational age at birth and country, an intention to treat analysis was also performed on both subsets. The conduct of this trial is elaborate and commendable .

A major limitation of the study is that it only reviews a single exposure of an anesthesic agent for less than 1 hour in predominantly male infants, whereas all previous preclinical and clinical evidence alludes to a dose-dependant adverse effect of anesthetic agents. Thus the scientific community needs to exercise prudence in extrapolating these reults to infants who are older, require multiple doses and/or multiple neurotropic drugs in the same surgery, multiple surgeries requiring repeated anesthetic exposure as well as prologed exposure to these agents.  It should also be noted than despite the best efforts of the authors, the sample size of 598 infants was not met for the primary outcome (447 complete assessments at 5 years).  A sensitivity analysis might have been reassuring, if full sample size had been attained and results showed minimal change in outcome.

The study establishes that single, short duration exposure (<60 min) to sevoflurane general anesthesia does not lead to different neurodevelopmental outcomes as assessed by the WPPSI-III when compared to awake regional anesthesia for infants undergoing inguinal hernia repair. Determining the dose-dependent effect of anesthesia on neuronal development would require the designing of a thoughtfully constructed trial delineating the complex surgical neonates from those that require uncomplicated surgical procedures.


  1. Jevtovic-Todorovic V. Exposure of developing brain to general anesthesia: what is the animal evidence? 2018;128:832–39.
  2. Coleman K, Robertson ND, Dissen GA, Neuringer MD, Martin LD et al. Isoflurane AnesthesiaHas Long-term Consequences on Motor and Behavioral Development in Infant Rhesus Macaques.  2017;126:74-84.
  3. Lunardi N, Sica R, Atluri N, Salvati KA, Keller C, Beenhakker MP et al. Disruption of Rapid Eye Movement Sleep Homeostasis in Adolescent Rats after Neonatal Anesthesia. 2019;130:981-994.
  4. US Food and Drug Administration. Drug safety communication: FDA approves label changes for use of general anesthetic and sedation drugs in young children. 2017.
  5. Warner DO, Chelonis JJ, Paule MG, Frank RD, Lee M, Zaccariello MJ, Katusic SK. Performance on the Operant Test Battery in young children exposed to procedures requiring general anaesthesia: the MASK study. Br J Anaesth.2019;122:470-479.
  6. Sun LS, Li G, DiMaggio CJ, Byrne MW, Ing C, Miller TL Feasibility and pilot studyof the Pediatric Anesthesia NeuroDevelopment Assessment (PANDA) project. J Neurosurg Anesthesiol. 2012;24:382-8.
  7. O’Leary JD, Warner DO. What do recent human studies tell us about the association between anaesthesiain young children and neurodevelopmental outcomes? Br J Anaesth. 2017;119:458-464.
  8. Zaccariello MJ, Frank RD, Lee M, Kirsch AC, Schroeder DR, Hanson AC. Patterns of neuropsychological changes after general anaesthesiain young children: secondary analysis of the Mayo Anesthesia Safety in Kids study (MASK study). Br J Anaesth. 2019;122:671-681.