Long-term impact of sevoflurane exposure on behavioural and neurocognitive outcomes in preschool children: protocol for a prospective cohort study
Long-term impact of sevoflurane exposure on behavioural and neurocognitive outcomes in preschool children: protocol for a prospective cohort study
The use of sevoflurane, a commonly used paediatric anaesthetic, raises concerns about potential long-term neurotoxic effects on behavioural and neurocognitive development, particularly during critical neurodevelopmental stages in preschool children. Endogenous hydrogen sulfide (H2S), a neuroprotective gasotransmitter, may be affected by anaesthetic exposure, but its role in sevoflurane-induced neurotoxicity remains unclear.
This prospective cohort study aims to evaluate behavioural and neurocognitive outcomes in 200 preschool children aged 4–6 years (1:1 allocation), with exposure to sevoflurane general anaesthesia (GA) as the primary predictor. A family-centred, professionally guided questionnaire-based assessment approach will be employed. Data accuracy and reliability will be ensured through the integration of real-time medical records and standardised instruments. Moreover, by investigating changes in serum H2S levels among children in the exposed group, this study offers a novel perspective on the potential neurotoxic mechanisms of GA and may inform the development of targeted neuroprotective interventions.
Ethical approval was obtained from Shengjing Hospital of China Medical University Ethics Examination Committee (2024PS1204K). We will present the results of the study at national and/or international conferences and in peer-reviewed journals. The study began in October 2024 and is expected to be completed in December 2025.
ChiCTR2400090174.
http://creativecommons.org/licenses/by-nc/4.0/
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
As medical advancements continuously evolve, an increasing number of infants, young children and pregnant women require general anaesthesia (GA) for essential diagnostic procedures, surgeries and other medical interventions. Approximately 10% of children in the USA undergo anaesthesia annually, with 95% receiving GA.1 Similarly, the incidence of non-obstetric surgeries during pregnancy ranges from 0.75% to 2.0%, with most requiring GA as well.2 Historically, GA was considered a reversible alteration of consciousness. However, given that the brains of infants and young children are in critical stages of development, there is ongoing debate about whether short-term exposure to general anaesthetics could induce neurotoxicity. This potential neurotoxicity might lead to irreversible brain functional and morphological changes, which could affect cognitive development and result in long-term intellectual consequences.3–6 As a result, research into the neurotoxic effects of general anaesthetics during developmental periods has garnered significant attention and has become a critical global public health issue.7 Over the past two decades, an increasing number of studies have demonstrated that commonly used general anaesthetics can cause pathological brain damage and long-term cognitive and behavioural abnormalities, particularly when administered during sensitive developmental periods.8 9 Despite the large number of animal experiments confirming anaesthesia-induced neurotoxicity, the evidence in humans remains highly controversial.10
Previous animal models have demonstrated the association between GA and abnormal central nervous system development, even triggering neurobehavioural disorders.11 12 This evidence amplifies the potential clinical significance of anaesthesia-related neurotoxicity. Notably, the agents used to induce altered states of consciousness under GA meet the criteria for neurotoxins. According to the US National Toxicology Program, neurotoxicity is defined as any adverse effect on the structure or function of the central and/or peripheral nervous systems caused by biological, chemical or physical agents, leading to baseline deviations that impair an organism’s survival, reproduction or adaptability.8 In 2017, the US Food and Drug Administration (FDA) issued a warning that prolonged or repeated exposure to anaesthesia in children under 3 years of age or in pregnant women during the third trimester may affect brain development. However, clinical evidence of the neurotoxic effects of general anaesthetics on the infant brain remains limited. At present, three major large-scale clinical studies have investigated the neurological effects of anaesthesia in infants and young children: PANDA (Paediatric Anaesthesia Neurodevelopment Assessment), MASK (Mayo Anaesthesia Safety in Kids), and GAS (General Anaesthesia compared with Spinal anaesthesia). The PANDA study reported that children who underwent a single hernia repair under GA did not differ significantly in IQ or on a range of neurocognitive and behavioural tests at ages 8–15 when compared with siblings who had not been exposed to anaesthesia.13 The MASK study found that children who received multiple (≥2) GA before the ages of 2–4 years had a significantly increased risk of developmental learning disabilities, memory impairments, language deficits and attention deficit hyperactivity disorder compared with an unexposed group.14 Although the GAS study showed that GA of less than 1 hour in early infancy does not alter neurodevelopmental outcomes at 5 years of age compared with conscious regional anaesthesia,15 a recent secondary analysis of GAS data revealed that multiple GA exposures were associated with decreased neurocognitive ability at 5 years of age.16 Moreover, a recent large-scale retrospective study found that long-term survivors of childhood acute lymphoblastic leukaemia who were exposed to GA exhibited higher rates of cognitive impairment and neuroimaging abnormalities, independent of known chemotherapy-related neurotoxicity. Furthermore, these findings were positively correlated with the cumulative dose and duration of anaesthesia exposure.17 Given these findings, the FDA’s cautionary stance remains justified. In the absence of conclusive evidence, prolonged or repeated anaesthesia should be avoided in infants, and the necessity of early elective surgeries should be carefully reconsidered.
Sevoflurane is a volatile general anaesthetic widely used in paediatric anaesthesia worldwide due to its smooth induction, minimal irritation and pleasant odour.10 18 However, the developing neonatal brain is highly sensitive to environmental stimuli, and early exposure to sevoflurane may adversely affect brain development.10 19–21 Preclinical studies have demonstrated that repeated or prolonged sevoflurane exposure leads to long-term cognitive deficits in neonatal rodents.22–24 Nevertheless, such preclinical evidence cannot be directly extrapolated to clinical practice, underscoring the need for continued exploration of the mechanisms and clinical implications of sevoflurane-induced developmental neurotoxicity.
Hydrogen sulfide (H2S) is recognised as the third endogenous gaseous signalling molecule in the human body, following nitric oxide (NO) and carbon monoxide (CO). It serves as a crucial neuroprotective and regulatory agent in the central nervous system, contributing to the maintenance of homeostasis through its anti-inflammatory, antioxidant, anti-apoptotic and mitochondrial protective mechanisms.25–28 In the body, H2S is primarily synthesised from cysteine by cystathionine β-synthase (CBS), cystathionine γ-lyase (CSE) and 3-mercaptopyruvate sulfurtransferase (3-MST), with tissue-specific distribution. CBS and 3-MST are predominantly found in the central nervous system, while CSE is mainly located in peripheral organs such as the liver and cardiovascular system.29–33 Researchers have observed a reduction in endogenous H2S levels in various experimental models of cognitive dysfunction.34 However, the mechanisms underlying the beneficial effects of H2S across different models are complex (figure 1), and there is a lack of preclinical and clinical studies elucidating the precise relationship between H2S and central nervous system protection.
Figure 1
Summary of the neuroprotective properties of H2S. AD, Alzheimer’s disease; AMPK, adenosine 5’-monophosphate-activated protein kinase; BDNF, brain-derived neurotrophic factor; IRI, ischaemic reperfusion injury; KATP, ATP-sensitive potassium channel; MMP-9, matrix metalloproteinase; mTOR, mammalian target of rapamycin; mtDNA, mitochondrial DNA; NO, nitric oxide; PD, Parkinson’s disease; PKC, protein kinase C; PND, perioperative neurocognitive disorders; RNS, reactive nitrogen species; ROS, ractive oxygen species; TBI, traumatic brain injury.
Interestingly, clinical evidence suggests that reduced levels of H2S in the brain and plasma are directly associated with the development of Alzheimer’s disease (AD). Specifically, AD patients exhibit lower H2S and S-adenosyl-L-methionine (a CBS activator) levels in the brains, alongside homocysteine accumulation (an upstream synthetic substrate of H2S).35 Another study observed that plasma H2S levels were significantly lower in patients with cognitive dysfunction phenotypes associated with AD, vascular dementia, and cerebrovascular disease.36 Preclinical data further indicate that reduced endogenous H2S or disruption of its synthesis underlies various animal models of cognitive dysfunction, while supplementation with H2S donors or precursors can rescue these deficits.34 37–40 Collectively, these findings highlight the critical role of H2S and its endogenous metabolic pathways in cognitive function, suggesting their potential preventive and therapeutic significance in the study of GA-induced developmental neurotoxicity and neurocognitive impairment. Excitingly, a study demonstrated that H2S can mitigate isoflurane-induced neuronal apoptosis and cognitive deficits in developing rat brains,40 further underscoring the urgency of investigating the role of H2S in the context of sevoflurane–induced developmental neurotoxicity in clinical settings. In summary, although animal models provide robust evidence of sevoflurane-induced neurotoxicity in developing brains, clinical confirmation remains limited. This study addresses that gap by evaluating the long-term behavioural and neurocognitive outcomes of sevoflurane GA in preschool children. It also investigates serum H2S levels and their endogenous metabolic pathways as a potentially neurocognitive function-associated mechanism, offering new insights into preventing and treating sevoflurane-related neurodevelopmental impairments.
This prospective cohort study investigates the long-term effects of sevoflurane GA on behavioural and neurocognitive function in preschool children (aged 4–6 years, n=200, 1:1 allocation), and its potential association with alterations in endogenous H2S metabolic pathways. Participants are divided into two groups: an exposed group undergoing sevoflurane GA and an age-matched unexposed group from kindergarten. The exposed group comprised patients who received a standardised balanced anaesthetic regimen, involving intravenous induction with propofol (2.5–3.5 mg/kg administered slowly over 20–40 seconds), followed by maintenance with sevoflurane inhalation.Baseline assessments include demographic data and standardised behavioural and neurocognitive evaluations using the Child Behaviour Checklist (CBCL/4–16) and the Behaviour Rating Inventory of Executive Function–Preschool Version (BRIEF-P). In the exposed group, serum samples are collected preoperatively and postoperatively to measure H2S levels and key metabolic enzymes (CBS and 3-MST). At 6 months post-enrolment, all participants undergo follow-up assessments with the CBCL/4–16, BRIEF-P and a 50-Item Preschool Intelligence Scale (50-item). A summary of the study protocol is shown in figure 2.
Between September 2024 and June 2025, preschool children aged 4–6 years scheduled to undergo surgery with sevoflurane GA at Shengjing Hospital of China Medical University (Shenyang, China) will be enrolled as the exposed group (n=100). We selected preschool children aged 4–6 years because, during this period, synaptic remodelling and myelination proceed rapidly, and core cognitive and socioemotional functions undergo initial consolidation without interference from formal schooling, rendering this cohort particularly vulnerable to anaesthesia-related neurotoxicity.41–44 Moreover, existing clinical studies have predominantly focused on children under 3 years of age or encompassed broader age ranges,8 leaving a lack of evidence specific to the population aged 4–6 years; thus, our study addresses this critical gap in the literature. Serum samples from clinical residual biospecimens will be collected preoperatively (the day before surgery) and postoperatively (day 0) to measure serum levels of H2S, CBS and 3-MST at the Benxi Training Base laboratory. An age-matched (1:1) unexposed group will comprise healthy preschool children from urban kindergartens in Shenyang with no prior GA or surgical history (n=100). The inclusion and exclusion criteria for research subjects are as follows:
In addition to the above exclusion criteria, children in the exposure group will also be excluded if they have previously undergone GA but the specific anaesthetic technique is unknown, or if they have clearly received anaesthesia maintenance with agents other than sevoflurane (e.g., Total Intravenous Anaesthesia). There was no patient or public involvement in the design and conduct of this study.
Baseline demographic data will be collected through structured telephone and electronic questionnaires (Basic Information Questionnaire, BIQ) following written informed consent, see online supplemental supplement 1). All participants will complete baseline cognitive-behavioural assessments (CBCL/4–16 and BRIEF-P) immediately after enrolment, with follow-up evaluations and a single 50-item assessment conducted 6 months later. During the follow-up, any significant events potentially influencing study outcomes, such as injuries, illnesses, hospitalisations or anaesthesia exposure, will be documented. Perioperative-related clinical data will be obtained through preoperative assessments and intraoperative anaesthesia records (figure 3).
Figure 3
Study timeline of data collection and milestones. The incident record includes significant events that may affect the results of the study, such as injury, illness, hospitalisation or exposure to anaesthesia. BIQ, Basic Information Questionnaire; BRIEF-P, Behavioural Rating Inventory of Executive Function for Preschool; CBCL/4-16, Child Behavioural Checklist for Ages 4–16; CBS, cystathionine β-synthase; H2S, hydrogen sulfide; 50-item, 50-Item Preschool Intelligence Scale; 3-MST, 3-mercaptopyruvate sulfurtransferase.
GA was administered by anaesthetists participating in the study using a standardised balanced anaesthetic regimen. Anaesthesia was induced with intravenous propofol (2.5–3.5 mg/kg), followed by maintenance with sevoflurane (1.5%–2.5%) in combination with nitrous oxide (50%–70%) and oxygen (30%–50%). The exact doses and concentrations, as well as the use of opioid agents, were determined at the discretion of the attending anaesthetist based on the clinical context, in order to maximise the applicability of the study findings to routine clinical practice. Demographic and perioperative characteristics of the study population are summarised in table 1. The surgical procedures included in this study were primarily derived from paediatric orthopaedics, paediatric urology, and paediatric general surgery, all of which are associated with minimal anatomical disruption and physiological impact on the child.
Table 1
Demographic characteristics and perioperative information
Sample size
The primary outcome of this study is CBCL/4–16 total problem scores in preschool children. As an observational and exploratory study, formal sample size calculation is not mandatory. However, in order to ensure the reliability of the research results and the robustness of the model, we have designed a precise sample size calculation plan. Previous studies report a 2–3 points difference in CBCL total problem scores between GA-exposed and unexposed children, with SDs of 10–13 points.45 This range aligns with the normative population used to develop the CBCL scale (SD=10).46 Although previous findings have demonstrated statistically significant between-group differences, clinicians often require a more substantial difference to qualify as clinically meaningful.13 14 Therefore, a 0.5 SD threshold is widely recognised as a minimal clinically important difference (MCID) for health-related scales, particularly in paediatric psychological and related measurement contexts.47–50 Based on these considerations, we conducted a sample size calculation for a multiple regression (random model) model using PASS software (NCSS) in ‘effect size’ mode. We prespecified an MCID of 5 points (0.5 SD, with SD=10) for the CBCL total problem score, a two-sided α of 0.05, and a power of 80%, while including 5–10 covariates and designating the group factor (exposed vs unexposed) as the primary predictor. Under these assumptions, the estimated effect size is f²=0.06, indicating a minimum of 146 participants. Allowing for an anticipated 20% attrition rate, a total of 200 children (100 per group) will be recruited.
All surgeries will involve mild tissue dissection, resulting in limited surgical trauma and minimal potential impact on neurocognitive function. The subjects’ recruitment methods will include:
Data measurement and collection
During the hospital visits of children in the exposed group, an in-person interview will be conducted, during which the study’s objectives and significance will be thoroughly explained to the guardians. The parents’ willingness to participate in the study will be clarified, and written informed consent will be obtained. The participant’s ID will then be confirmed, and the baseline assessment using the CBCL/4–16 scale will be conducted. Parents will also receive on-site structured training from professionals on how to complete the scale assessments.
To ensure privacy protection and data reliability, the collection of BIQ will be conducted through follow-up phone interviews combined with an electronic questionnaire. This information will primarily cover the participant’s medical history (all previous hospital diagnoses and treatments, including birth circumstances), history of anaesthesia and surgery (exposure to GA, sedation or surgery), and family and school environments. Additionally, details of the mother’s pregnancy and delivery history will be gathered, and the absence of any exclusion criteria in the subject will be confirmed during this information collection.
Information regarding the family environment will include the parents’ educational level, socioeconomic status (occupation, income, housing, insurance type, etc), marital status (emotional well-being, separation status), and the child’s birth order. Among them, the family socioeconomic status will be included as a main covariate in the final model for analysis (table 2). Data on the school environment will cover whether the child enjoys school, how well they get along with peers and whether the child has any formal diagnoses or parental concerns about learning difficulties.
Table 2
Socioeconomic and family characteristics of parents
Furthermore, where information is available, objective medical records and parental memory feedback will provide information about the children’s previous anaesthetic procedures. For children in the exposed group who have previously undergone GA and surgery at Shengjing Hospital, a thorough review of their past anaesthesia audit records will be conducted. This review will collect detailed information on the child’s age at the time of the procedure, the specific nature of the surgery, the duration of GA, the anaesthetic drugs and techniques used and any notable perioperative anaesthetic or surgical complications. During this GA procedure, according to the standards of the ASA, all children will be monitored intraoperatively using pulse oximetry, blood oxygen saturation, non-invasive arterial blood pressure, ECG and end-tidal CO₂ (PETCO₂).
●Sample collection
Preoperative and postoperative blood samples from children in the exposed group will be obtained from routine clinical residual biological samples (1–1.5 mL of serum) at the Clinical Laboratory of Shengjing Hospital, China Medical University. A total of 200 blood samples will be collected. After the acquisition, the samples will be stored in liquid nitrogen tanks and subsequently transported to the Laboratory for biomarker analysis. Each sample will be retained for 3 days after testing to allow for potential retesting. On completion of the retention period, all samples will be uniformly deactivated and destroyed. The samples and associated data will be used exclusively for this study. No blood samples will be collected or analysed from participants in the unexposed group.
Both groups of children will undergo a BIQ and a baseline assessment of the CBCL/4–16 and BRIEF-P immediately after confirmation of enrolment. The 50-item questionnaire will only be surveyed 6 months after enrolment.
Our team, which includes professionals specialising in paediatric rehabilitation and psychotherapy, will provide prior training to the children’s guardians, thoroughly explaining the content of the scales and the key points to consider during the assessment. Any issues encountered during testing will be promptly communicated with professionals, and online guidance will be encouraged if necessary. All individuals involved in administering the scales and analysing the results will be qualified professionals who have undergone standardised test training to ensure the reliability of the outcomes. The CBCL/4–16 is a behavioural scale that is widely used and comprehensive in content (online supplemental 2). It was compiled by American psychologist Achenbach in 1976 and revised and introduced by the Shanghai Mental Health Centre in 1991, establishing the norm for Chinese children aged 4–16.51 The CBCL/4–16 is a revised version of the key questions in the CBCL/4–18 and is suitable for children aged 4–16.52 The scale consists of three parts, with the third part focusing on behavioural problems, comprising 113 items. This section is the primary focus of the scale and the core subject of this research. Each behavioural problem is assigned a raw score (0, 1 or 2). The total raw score is obtained by summing all 113 items, where higher scores indicate more severe behavioural problems and lower scores suggest milder issues. Based on statistical analysis of a large sample, the upper limit of normal was determined, with cut-off values as follows: for boys aged 4–5, 6–11 and 12–16, the values are 42, 40–42 and 38, respectively; for girls of the same age groups, the values are 42–45, 37–41 and 37, respectively. Scores exceeding these cut-offs are considered abnormal. Furthermore, multivariate statistical analysis reveals that behavioural problems can be grouped into 8–9 factors. Each factor comprises a subset of the 113 items, varying slightly by gender and age group. Some items may appear in multiple factors. The crude scores for items within each factor are summed to calculate the factor score. These scores are often converted for statistical purposes into standardised T-scores (
), which map raw scores to a distribution with a mean of 50 and an SD of 10. The normal range for factor scores is defined as the 69th to 98th percentile (T-scores of 55–70). Scores above the 98th percentile (T>70) are considered potentially abnormal. The entire assessment process involves parents filling in each item on the scale, and it takes about 15–20 min to complete an assessment. The Chinese version of the CBCL/4–16 is widely used in China and has demonstrated good reliability and validity.53 54 For background information and detailed content of BRIEF-P and 50-item, see online supplemental 3 and 4).
Primary outcome
CBCL/4–16 total problem scores for behavioural problems 6 months after enrolment.
Secondary outcome
Incidence of abnormalities in each factor of CBCL/4–16 after 6 months of enrolment.
BRIEF-P scores on five factors (Inhibition, Shifting, Emotional Control, Initiation, Working Memory) and three dimensions (Inhibitory Self-Control Index/Flexibility Index/Emergent Metacognition Index) after 6 months of enrolment.
50-item raw total score after 6 months of enrolment.
50-item scores in each ability area (Self-awareness, Motor skills, Memory skills, Observation skills, Thinking skills, General knowledge) 6 months after enrolment.
Differences in preoperative and postoperative serum H2S, CBS and 3-MST levels in exposed groups (any day before surgery and day 0 postsurgery).
This prospective cohort study investigates behavioural and neurocognitive outcomes in preschool children (aged 4–6) 6 months after exposure to sevoflurane GA. A parallel unexposed group of healthy, age-matched kindergarten children without GA exposure was established for comparison.
The primary outcome—the total problem score on the CBCL/4–16—will be assessed using multivariable linear regression, adjusting for baseline CBCL scores, age, sex, parental education, socioeconomic status and perioperative variables. To enhance the robustness of the findings, generalised linear mixed models (GLMMs) with random intercepts will also be employed, accounting for within-subject correlations between baseline and 6 month follow-up assessments and allowing flexible handling of partially missing data. Comparisons between these modelling approaches will assess the consistency of the exposure–outcome relationship.
Subgroup analyses within the exposed group will explore associations between the number of sevoflurane exposures (categorised as 1, 2 or ≥3) and outcomes. To mitigate baseline imbalances due to the non-randomised design, propensity score matching (PSM) will be conducted using a multivariable logistic regression model that incorporates key clinical and socioeconomic variables (eg, surgical indication, perioperative factors and detailed family background). Nearest-neighbour matching with a calliper of 0.2 SD of the logit of the propensity score will be used to optimise balance and retain sample size.
A multiplatform follow-up system (via telephone, WeChat, email and surgeons) was established to maximise retention. Reasons for attrition and missing data will be documented, and multiple imputation (MI) and maximum likelihood estimation methods will be used to address missingness. Baseline characteristics of completers versus non-completers will be compared with assess potential selection bias. For repeated measures, GLMMs will be applied to handle within-subject correlation and incomplete data.
To develop a predictive model of neurodevelopmental risk, multivariable linear regression will be performed with independent variables including cumulative sevoflurane and opioid doses, surgical duration, gestational age, birth weight, parental education and perinatal/perinatal adverse events (see table 3). For significant predictors, restricted cubic splines will be used to explore potential nonlinear dose–response relationships—particularly between sevoflurane exposure and CBCL total scores.
Table 3
Multivariate linear regression model analysis of CBCL/4-16 scale and sevoflurane exposure
In parallel, preanaesthesia and postanaesthesia serum levels of H2S/CBS/3-MST will be measured in the exposed group. Correlations between changes in these biomarkers and behavioural outcomes will be examined to investigate potential mechanistic links.
All statistical analyses will be performed using SPSS V.25.0 (IBM). Continuous variables will be tested for normality using the Shapiro-Wilk test. Normally distributed data will be presented as means (±SD), and non-normally distributed data as medians (IQR). Between-group comparisons will be conducted using independent t-tests, Mann-Whitney U tests, χ2 tests, continuity-corrected χ2 tests or Fisher’s exact tests, as appropriate.
The primary analysis will follow a per-protocol set (PPS), including participants who fully adhered to the protocol and completed all assessments. Given the expected subtle effect sizes, this strategy enhances data precision. To validate findings, sensitivity analyses will be performed using the full analysis set (FAS), which includes all eligible participants contributing data, ensuring robustness and generalisability of results.
The study protocol and informed consent procedures have been reviewed and approved by Shengjing Hospital of China Medical University Ethics Examination Committee (2024PS1204K) and strictly adhere to the Declaration of Helsinki as well as relevant ethical guidelines and regulations. Any significant protocol modifications or adverse events occurring during the study will be promptly reported to the ethics committee, with updates to informed consent forms and relevant documents as necessary. Written informed consent will be obtained from the legal guardians (parents or other authorised representatives) of all participating children. Prior to enrolment, research staff will clearly explain to guardians, in understandable language, the study’s objectives, potential risks and anticipated benefits, specific procedures, confidentiality safeguards and their right to withdraw at any time. Participation in the study will commence only after the guardian fully understands and voluntarily signs the informed consent form.
This study exclusively uses residual serum samples from routine clinical testing, ensuring no additional invasive procedures are performed on the children. Blood samples collected before and after surgery are processed by the clinical laboratory as part of standard care, and only surplus serum will be used for the current research. After obtaining consent from guardians, personal identifiers will be removed, samples will be anonymised and coded, and then stored in liquid nitrogen tanks for subsequent analysis of serum H2S and related enzyme indicators. Following analysis and validation, all remaining specimens will be appropriately destroyed or securely stored for a defined period in strict accordance with ethical guidelines.
To protect participants’ privacy, all research data will be deidentified using a unique coding system, with personal identifiers such as names replaced by anonymised codes. Access to data or biological samples is restricted exclusively to authorised research personnel. Results published or presented in conferences will not contain personally identifiable information. Electronic data will be securely stored on encrypted, password-protected servers, while paper documents and samples will be secured in facilities accessible only to authorised individuals.
Findings will be communicated mainly within scientific circles. We will submit full manuscripts to peer-reviewed journals and present interim data at relevant conferences and symposia. A concise project webpage and short updates on the host institution’s social-media channels will make key results accessible to interested clinicians, researchers and decision-makers. This streamlined approach keeps the emphasis on rigorous academic exchange while providing a modest online presence for broader visibility.
This study aims to investigate the long-term impact of sevoflurane GA exposure on behavioural problems and neurocognitive function in preschool children (4–6 years). Additionally, it explores changes in H2S levels, a known endogenous gaseous signaling molecule with established neuroprotective properties, during the perioperative period. By examining the interplay between sevoflurane GA, H2S dynamics and paediatric behavioural and neurocognitive outcomes, this research seeks to uncover potential mechanistic links and novel insights into the implications of anaesthesia in early childhood.
Currently, no direct association is established between GA and neurotoxicity linked to cognitive or behavioural sequelae in children. Moreover, studies suggest that for most healthy children requiring surgery, the neurotoxic effects of general anaesthetics are unlikely to be the primary contributors to adverse neurodevelopmental outcomes. Instead, biological, environmental and social factors are expected to exert a more substantial influence on neurodevelopment.55 Considering this, we plan to adopt a PPS analysis akin to that used in randomised controlled trials. Only data from participants who complete all questionnaires and follow-ups will be included in the final analysis. Using an intention-to-treat or FAS in this context is anticipated to yield more conservative results, potentially diluting the observed impact of sevoflurane GA and obscuring differences between groups. To minimise such bias and ensure the reliability of the findings, we will primarily employ PPS, thereby enhancing the robustness and interpretability of the study outcomes.
This study has several strengths (1) Focus on single sevoflurane: Previous studies have rarely focused on the impact of specific anaesthetic agents on behavioural and cognitive outcomes in infants and young children. This study, however, examines sevoflurane—the most used anaesthetic for paediatric anaesthesia worldwide—with particular emphasis on exploring its independent effects, making this research both timely and of significant clinical value. (2) Family-Centred Approach: The concept of ‘family-centred’ care originates from the Association for the Care of Children’s Health in the USA, emphasising family strengths and choices.56 57 This study prioritises family involvement by enabling parents to conduct cognitive and behavioural assessments of their children under the supervision and guidance of healthcare professionals. Unlike previous studies that relied solely on assessments conducted in healthcare institutions, this approach facilitates follow-up, supports the development of family-centred neurodevelopmental monitoring and early intervention systems and encourages active parental participation. This enhances patient–provider interaction and promotes healthy child development. (3) Real-time data collection: Previous clinical research on anaesthesia-induced neurotoxicity in developing brains has heavily relied on retrospective epidemiological data.8 In contrast, this study integrates real-time medical records and cross-sectional data collection with structured telephone or electronic surveys, ensuring the authenticity of the collected information and strengthening the reliability of the study outcomes.58 (4) Exploration of H2S in Developing Brains: Existing evidence supports the neuroprotective effects of H2S, including its anti-inflammatory, antioxidative and anti-apoptotic properties.59 However, most studies have focused on ischaemic-hypoxic brain injuries and neurodegenerative diseases,60 primarily in elderly populations and disease-specific contexts. This study addresses the gap in research on H2S’s role in the developing brain. (5) Innovative Focus on H2S and anaesthesia: Prior studies have linked plasma H2S levels and endogenous synthase activity to the progression of neurodegenerative diseases, as well as their therapeutic significance in adverse cerebrovascular events like brain ischaemia-hypoxia.29 However, these findings are predominantly derived from animal models, with limited clinical evidence available. This study leverages the H2S/CBS/3-MST endogenous sulphur metabolism system to investigate the effects of sevoflurane anaesthesia exposure on serum H2S levels, offering both theoretical feasibility and high innovation potential.
Nevertheless, this study has several limitations. First, the narrow age range of the participants restricts the generalisability and societal value of the findings, even though the preschool stage represents a critical period for neurocognitive development in children. Second, while the design and selection of the questionnaires were rigorous and comprehensive, with reliability and validity tested in the Chinese population, the absence of globally recognised assessment scales limits the precision of neurodevelopmental and cognitive function measurements. Standardised scales such as the Wechsler and Gesell tests require follow-ups conducted by professionals in hospital settings, significantly increasing time costs and implementation difficulty. As an exploratory study, our focus is on ensuring efficient and stable execution, as well as screening for behavioural and cognitive issues, rather than establishing definitive diagnoses. This approach represents a pragmatic compromise to support a family-centred framework for child health monitoring and improve participant adherence. Third, this study evaluates changes in H2S levels only within the exposed group. We could not fully address baseline differences between children undergoing GA and healthy children, nor entirely resolve potential confounding factors. Due to ethical concerns, obtaining blood samples from healthy children outside hospital settings poses challenges. However, reusing clinically discarded blood samples from non-surgical or outpatient paediatric patients warrants further consideration. Finally, we fully acknowledge that the primary methodological limitation of this study will lie in confounding by indication. Children undergoing surgery and GA will inherently differ from healthy controls due to underlying medical conditions and surgical needs, which may independently influence neurodevelopmental outcomes. This will complicate the isolation of sevoflurane GA-specific effects. To address this concern, we will apply stringent inclusion criteria—restricting enrolment to children undergoing common, minimally invasive procedures and classified as ASA physical status I–II—to minimise clinical heterogeneity. Key covariates will be adjusted for multivariable regression models, and sensitivity analyses using GLMMs will be conducted to validate the robustness of our findings. To further mitigate baseline imbalances, PSM will be implemented using detailed clinical and socioeconomic variables. Subgroup analyses stratified by surgical indication will also be performed to evaluate and control for residual confounding. Despite these extensive mitigation efforts, we recognise that the observational nature of this study will preclude complete elimination of unmeasured or unknown confounders. Therefore, our findings will need to be interpreted with caution, particularly regarding their generalisability to broader populations. Potential selection bias, arising from differential recruitment or parental participation, will be minimised through multichannel recruitment strategies and systematic documentation of refusal or withdrawal reasons. Baseline characteristics of participants and non-participants will be compared with assess any participation-related bias. Information bias, particularly from parent-reported outcome measures such as the CBCL/4–16, will be reduced through standardised administration protocols, structured training sessions, ongoing follow-up support and rigorous quality control using validated tools.
In conclusion, this study will provide valuable insights into the timing and clinical strategies for sevoflurane-based GA in preschool children. Furthermore, it offers a preliminary exploration of the potential role and significance of the endogenous gasotransmitter H₂S in the context of anaesthesia-induced neurotoxicity in the developing brain.
Patient recruitment started in October 2024. The project was originally planned to be fully completed by June 2025, but due to the slow progress of patient enrolment, we expect that recruitment will be completed by July 2025, with final follow-up completion, data analysis and publication of relevant research results by December 2025. The current protocol version is 1.4, dated 23 December 2024.
Not applicable.
This study is particularly grateful to the National Regional Medical Center for Children of China–Shengjing Hospital of China Medical University for providing a research platform and abundant patient enrolment resources. Figdraw assisted in completing the schematic diagram of the graphical abstraction in figure 1.
