Association between surgical procedures under general anesthesia in infancy and developmental outcomes at 1 year: the Japan Environment and Children’s Study

Recently, the influence of anesthesia on children’s neurodevelopment has attracted attention. Numerous animal studies have indicated that exposing the developing brain to anesthetics may lead to widespread apoptotic neurodegeneration, interference with synaptogenesis during brain development, and long-term neurocognitive impairment [1‐6]. However, the detailed mechanism has not been clarified. Results in human children, however, remain controversial. Three large-scale prospective studies, the General Anesthesia compared to Spinal anesthesia (GAS) trial [7, 8], the Pediatric Anesthesia Neurodevelopment Assessment (PANDA) study [9], and the Mayo Anesthesia Safety in Kids (MASK) study [10], demonstrated that a single, relatively short exposure to general anesthesia in infants and toddlers is unlikely to cause clinically detectable neurodevelopmental deficits or serious behavior disorders at age 5 years, 8–15 years, and 8–12 or 15–20 years, respectively. However, the effects of repeated and prolonged anesthetic exposure in children are unknown. The US Food and Drug Administration (FDA) warned that repeated or lengthy use of general anesthetics in children younger than 3 years of age may affect brain development [11, 12].

Although numerous studies have been conducted, there is no large-scale Japanese study investigating the neurotoxicity of anesthesia. The Japan Environment and Children’s Study (JECS) is an ongoing prospective nationwide birth cohort study in Japan, and the participating children will be followed up until they reach 13 years of age. In this study, we investigated the association of surgery under general anesthesia in infancy with developmental outcomes at age 1 year using data from the JECS.

Subject characteristics are shown in Table 1. Among the 64,141 infants included, 746 (1.2%) had surgery under general anesthesia once, 90 (0.1%) had twice, and 71 (0.1%) had three times or more.

Table 2 shows the cutoff scores at 12 months for each domain of the J-ASQ-3 and the numbers and percentages of normal and delayed infants. The percentage of infants with communication delays was the lowest (7.0%), while personal–social delays were the highest (16.8%).

Table 3 shows the numbers and percentages of infants with delays for each domain by the number of surgical procedures under general anesthesia. Among infants who had surgery under general anesthesia more than once, the percentages of delays were high in all domains. The tendency toward developmental delay in all five domains with increasing number of surgical procedures was recognized by the Cochran–Armitage test (p < 0.001 in all domains).

Figure 2 depicts the adjusted odds ratios (aORs) obtained from logistic analyses for all five domains. The aORs of gross motor delay in infants who had surgical procedures under general anesthesia compared with those who did not have surgery were 4.69 [95% confidence interval (CI) 2.82–7.81] for three or more procedures, 3.22 [95% CI 2.04–5.09] for two procedures, and 1.45 [95% CI 1.19–1.78] for one procedure. For the fine motor, problem solving, and personal–social domains, the aORs of developmental delays were significantly increased in infants who had surgery under general anesthesia three times or more (2.99 [95% CI 1.70–5.28], 2.47 [95% CI 1.45–4.20], and 2.55 [95% CI 1.51–4.31], respectively). In these three domains, the aORs of developmental delay were significantly higher among infants exposed to surgery twice than among unexposed infants, but this was not the case with infants exposed once. For the communication domain, infants exposed to surgery three times or more had a significantly higher risk than unexposed infants did (aOR 3.32 [95% CI 1.78–6.20]). In all five domains, the risk of developmental delays was increased with three or more surgical procedures under general anesthesia.

The above analyses were conducted using maternal age as a continuous variable. When it was used as a categorical variable, the results were much the same (Additional file 1). When paternal age, parents’ educational level, and household income were additionally adjusted, the risks in the infants exposed to surgery were elevated, although the number of subjects decreased (from 64,141 to 30,448) owing to the limited number of registered fathers (Additional file 2). In the additional analyses after excluding infants who had congenital heart diseases, the aORs of developmental delays were also increased among infants who had surgery under general anesthesia twice or more (Additional file 3). In the analyses using the cutoff scores of the J-ASQ-3 reported for Japanese children, the results were approximately similar, and the aORs were generally higher (Additional files 4, 5, 6).

Recently, the neurotoxicity of general anesthesia to the developing brains has been a topic of concern. In the JECS study, the participating children will be followed up until they reach 13 years of age. In the present study, we evaluated whether surgery under general anesthesia in infancy was associated with developmental delay, using data from the JECS. A large percentage of infants who had surgery under anesthetic exposure displayed developmental delays in all five J-ASQ-3 domains at age 1 year, and this percentage increased significantly with the number of surgical procedures under general anesthesia. In multivariate analysis, the risk of developmental delays in all five J-ASQ-3 domains increased when the surgical procedures under general anesthesia numbered three or more. Gross motor development was the most affected, and the risk of the gross motor delay was increased by exposure to surgery even once.

Many studies have investigated the association between general anesthetic exposure and adverse neurodevelopmental outcomes in human children. Of the prospective studies, the GAS trial [7, 8] was an international, multicenter, randomized, controlled, equivalence trial. A total of 722 infants under 60 weeks’ postmenstrual age, who were undergoing inguinal hernia surgery, were randomly assigned to one of two anesthetic techniques: general anesthesia with inhaled sevoflurane or awake regional anesthesia (with no general anesthesia exposure). The intelligence quotient (IQ) at age 5 years [7] and neurodevelopmental outcomes at age 2 years [8] were assessed using the Bayley Scales of Infant and Toddler Development, Third Edition (BSID-III) [22]; no significant differences in the primary outcomes were found between the two groups. The PANDA study [9], a multicenter, ambidirectional observational study, examined a sibling cohort discordant for exposure to general anesthesia for inguinal hernia repair. The sibling pairs were under 36 months of age. There was no statistically significant difference in the IQ scores between sibling pairs at age 8–15 years. Despite their contributions to the literature, GAS and PANDA did not clearly elucidate the risks of repeated and prolonged anesthetic exposure. In the MASK study [10], an ambidirectional observational study, children who were unexposed, singly exposed, or multiply exposed to anesthesia before age 3 years underwent neuropsychological testing at age 8–12 years, or age 15–20 years. Although the IQ did not differ by exposure status, it was suggested that multiple, but not single, exposures were associated with a pattern of changes in specific neuropsychological domains associated with behavioral and learning difficulties.

Various retrospective studies have also investigated the association between general anesthesia in childhood and adverse neurodevelopmental outcomes. While some have concluded that there is no association [23‐31], others have suggested otherwise [21, 32‐42]. Among the studies reporting that anesthesia does not affect development, many have either set the timing of developmental evaluation at adolescence [23‐26], early after surgery [27, 28], or have considered the period before the age of 3–6 years as the time of exposure to anesthesia [23, 24, 27‐30]. However, there are also studies considering the effects of anesthesia received in infancy that are similar to our study [25, 26]. Hansen et al. [25] reported that single, relatively brief anesthetic exposure in connection with hernia repair in infancy did not reduce academic performance at age 15 or 16 years, after adjusting for known confounders. In studies of anesthesia and development, a variety of exposure periods and outcomes has been reported, including neuropsychological assessment [21, 32‐35], degree of learning disability [36‐39], and academic performance [40, 41]. Some described the effects of anesthesia exposure in infancy [21, 34, 40, 42], while some had multiple anesthesia exposures [38, 39]. Wilder et al. [39] reported that a single exposure to anesthesia by age 4 years was not associated with an increased risk of learning disability, but those with greater exposure were at an increased risk of a learning disability (hazard ratios = 1.59 for two anesthetics and 2.60 for three or more anesthetics). The MASK study [10] reevaluated the same topic using an ambidirectional design. Despite controversial results and a lack of consensus on the effects of multiple and long-term anesthesia exposures, the results of the GAS trial, MASK study, and PANDA study suggest that a single and brief exposure to anesthesia may not be associated with neurodevelopmental disorders and behavioral problems [7‐12, 43].

Our study demonstrates that surgery under general anesthesia in infancy is associated with developmental delay, as assessed by the J-ASQ-3 method at age 1 year. The timing of the outcome evaluation, however, is important due to the plasticity of neural development in early life: most studies reporting that anesthesia has no influence tend to evaluate development several months after surgery or even in adolescence. We evaluated development at age 1 year, which is relatively early. The JECS conducts follow-ups until age 13 years, and so, there is an opportunity to follow changes in this cohort over a long period. When evaluating neurotoxicity due to anesthesia, it is important to select an outcome measure that can comprehensively evaluate development [43‐45]. Reviewing studies according to outcome measure, Davidson et al. [43] found that academic performance or school readiness tests are not greatly affected by anesthesia exposure in early life. Furthermore, results were inconsistent when abnormalities in neurocognitive function or behavior (based on validated neuropsychological or behavior assessment tools) were the outcome. The J-ASQ-3 is a developmental assessment tool and may lead to inconsistent outcomes.

Adverse developmental outcomes increased with the number of anesthetic exposures in previous studies, which our study confirmed [38, 39]. Also, two studies have used the BSID-III [22] to evaluate development at age 1 year among children exposed to anesthesia in infancy [21, 34]. Walker et al. [21] reported that infants who underwent cardiac surgery within 90 days of birth had significantly lower mean scores in all five domains of the BSID-III, and non-cardiac major surgery mean scores were lower in four domains compared with controls. Although the outcomes used differed from our study, the results were similar. Furthermore, it is interesting to note that gross motor delay was most affected in both the cardiac and non-cardiac surgery groups in Walker et al.’s study [21] and our study. On the contrary, Ing et al. [32] reported that motor function was not impaired among children exposed to anesthesia before age 3 years. Furthermore, in the MASK study [10], processing speed and fine motor abilities were decreased in children after multiple anesthetic exposures. Thus, there is no consensus regarding which domain is most susceptible to anesthesia, and further studies are needed. In our study, we used the cutoff scores of the original ASQ-3, but the cutoff scores of the J-ASQ-3 for Japanese children were published only recently. The cutoff scores reported for Japanese children were considerably lower than original scores, particularly for the communication and personal–social domains. When we also conducted a sensitivity analysis using the Japanese cutoff score (Additional files 4, 5, 6), the results were approximately similar, and the aORs were generally higher.

The main strength of our study was that we could examine children who had surgery under general anesthesia in a large-scale birth cohort study, including the 64,141 infants. However, our study had some limitations. First, information on surgery under general anesthesia was obtained from the mothers’ questionnaire responses, not medical records, and we did not confirm their records with the medical institution. In addition, details such as time and type of surgery and anesthesia were unknown. Second, when the subjects were selected, chromosomal anomalies and anomalies of the head or brain were the only congenital diseases excluded. Therefore, diseases that required multiple operations and long-term and multiple hospitalizations, such as congenital heart disease and digestive system disease, might have been inadvertently included. Considering the developmental effects of these diseases themselves and the increases in hospitalization they lead to, it is difficult to say that our findings are purely related to general anesthesia. Third, developmental outcomes were evaluated only at age 1 year. Individual differences in infancy are large, and it is unclear to what degree developmental delays at age 1 year can predict future prognosis. In addition, some researchers have questioned the accuracy of the ASQ-3 in children less than 2 years of age [46, 47]. Fourth, the associations of developmental delays with surgery under general anesthesia were analyzed in a cross-sectional manner because information on both of the surgical procedures and the responses to the J-ASQ were obtained from the 1-year questionnaire. However, our study is based on a birth cohort study, and the children will be followed up until age 13 years. Therefore, the association between surgery under general anesthesia in infancy and developmental outcomes can be further elucidated.

Here, we would like to emphasize that general anesthetic and sedation drugs are essential for young children who require surgery or other painful and stressful procedures. Some clinical studies have demonstrated that inadequate sedation, anesthesia, and analgesia may actually increase the risk of adverse postoperative outcomes in infants and children [48‐52]. Therefore, as recommended by the FDA [11], concerns about anesthesia affecting the developing brain should not stand in the way of necessary surgeries or procedures. At this stage, it is necessary to balance the benefits of appropriate anesthesia in young children with the potential risks, especially for procedures that may last longer than 3 h or in the case of children who require multiple procedures before age 3 years [11, 12]. If surgical procedures under general anesthesia in infancy are required, we think that the treatment should be performed with the short and minimum times of anesthesia. In addition, it is desirable that the neurodevelopment with the growth of the child will be followed up after surgery.

We investigated whether surgery under general anesthesia in infancy was associated with developmental delays at age 1 year using the JECS data. A significantly high percentage of infants who had surgery under general anesthesia displayed delays in all five J-ASQ-3 domains, and there was a tendency toward greater development delays as the number of surgical procedures increased. In all the five domains, the risk of developmental delays may increase among infants who had surgery under general anesthesia three times or more, especially in the gross motor domain. However, follow-up studies are essential to confirm the results and to elucidate whether the observed delays persist, worsen, or improve as the children grow. At this stage, the balance of the benefits of appropriate anesthesia in young children with the potential risks should be considered.