Long-term Differences in Language and Cognitive Function After Childhood Exposure to Anesthesia

Dr. Gregory and colleagues are right that it is difficult to draw causal inferences from observational epidemiology, due mainly to biases from many possible sources. Despite the limitations inherent in observational study designs, epidemiological data are important for understanding the long-term effects of anesthetic agents on neurodevelopmental outcomes in children. Our study involved prospective assessments in four different cognitive domains, which are likely more sensitive than those used in previously published studies. Our findings of differences in some cognitive domains and not others may also be useful in directing neuropsychological testing in future studies.

In response to the question about the effect of influential observations, regression diagnostics were performed and the outliers did not influence our results. As for including children who received anesthesia for CT or MRI scans, we included these patients in the cohort, but did not evaluate them independently. As a clarification, parental questionnaires regarding patient related data were filled out during follow up visits at ages 1, 2, and 3.

We agree that our results present an additional challenge to anesthesiologists, pediatricians, and surgeons who need to speak to parents of children requiring surgical procedures. We do acknowledge the limitations of our results, and emphasize that our findings alone are insufficient to be used as the basis for clinical decision-making. We hope the signal we identified in our study however will stimulate further research in this area.

We agree with Ueda regarding the importance of the window of vulnerability in children. We chose our age cutoff based on data from animal studies which suggest that vulnerability to anesthetic neurotoxicity occurs during peak synaptogenesis. In humans, peak synaptogenesis of the different brain regions occurs at up to 3 years of age.[1] However, future studies will be needed to establish the specific age of vulnerability. Only with such information, will we be able to advise parents and providers regarding the appropriate timing of elective surgery.

We appreciate the comments by Smith, and have therefore performed a sensitivity analysis with additional data we have since obtained. When we excluded all patients with congenital conditions in the same manner as Hansen et al., aRR between exposed and unexposed children for CELF-R, CELF -E, CELF-T, and CPM remained significant.[2] We plan to further examine the role of comorbidity in the association of neurocognitive deficit with early childhood anesthesia exposure in our future study.

References:

1. Huttenlocher PR, Dabholkar AS. Regional differences in synaptogenesis in human cerebral cortex. J Comp Neurol 1997;387:167-78.

2. Hansen TG, Pedersen JK, Henneberg SW, et al. Academic performance in adolescence after inguinal hernia repair in infancy: a nationwide cohort study. Anesthesiology 2011;114:1076-85.

Conflict of Interest:

None declared

To: Editor of Pediatrics

We read with interest the paper by Ing and associates "Long-term Differences in Language and Cognitive Function after Exposure to Anesthesia" in the September, 2012 edition of Pediatrics. The authors attempted to answer a very important clinical question, "does being anesthetized during the first three years of life cause central nervous injury"? There are, however, many questions that remain unanswered. When read uncritically or by a layperson not familiar with either statistics or the tests used, the article implies that anesthesia in the early years of life causes neuropsychological disability in a 1-to-1 cause-effect relationship. This is not true; the overwhelming majority of patients in both groups was normal.

The major shortcoming of this paper is that the investigators had no access to the patients or to their records, including the records of those who underwent anesthesia and surgery. Most of the data presented were gathered for a different purpose. No data were available about problems that might have occurred during surgery, including hypoxia, ischemia, volume depletion, cardiac arrest, or other problems. It appears that much of the patient related data came from questionnaires filled out by the parents when the children were 10 years of age. Furthermore, underlying diagnoses of these children remain unknown and obviously can impact greatly on cognitive and developmental outcomes.

It would have been very helpful to have data about whether either of the groups contained patients who had other problems, e.g., Down Syndrome (1/445 births in Western Australia [1]) or cerebral palsy (2/1,000 in the same area [2]) or other congenital, genetic, or perinatal problems. If there were patients with these problems, were they evenly distributed between the groups? A few of such patients in either group could greatly affect the conclusion and not be related to anesthesia.

Had the data been presented as a continuum, the reader could have judged whether the differences were due to a small number of so-called "outlier" patients in one group or the other. The larger standard deviations of the Exposed group compared to the Unexposed group hint that there may have been a few individuals who scored very low. Given the small sample size of the Exposed group compared to that of the Unexposed group makes it possible for a few very poorly functioning children to have made a dent in lowering the mean of this group. By choosing to present and highlight the outcome data (the test scores) as either OK or as "disability", rather than as a wide range of neuropsychological outcomes among both the Exposed and the Unexposed groups and by focusing on the small percentage of children in both groups who were in the "disability" range rather than on the very large percentages in both groups that did NOT show "disability" at the age of 10 years, the message comes across as a single stark conclusion: anesthesia during the first three years of life produces neurocognitive "disability" later. By introducing the term "disability" (in language and cognition) in the Abstract, the main source of most science news, and doing so without defining it, the investigators biased the conclusion about the danger from early anesthesia on later language and cognition. Use of "disability" in data presentation in Tables 5 and 6, combined with the words "risk" and "exposure to anesthesia" in the title of Table 6, contributed to amplifying that danger.

As to the outcome measures, the use of the term "neurocognitive" for describing the area of the outcome that was tested is somewhat misleading. The choice of tests is quite limited in number and quality and covers only a small portion of the wide range of skills that must be sampled during testing to arrive at any conclusions about the level of neurocognitive functioning, not to mention arriving at a qualitative judgment of a "neurocognitive disability".

More than 60 percent of the surgeries were "minor", making it likely that the length of anesthesia was <20-30 minutes (Table 3 in the paper). Animal data indicate a clear dose-time relationship between exposure to anesthetics (multiple hours) and CNS injury [3]. To get a similar length of anesthesia in humans would require days to weeks, not minutes of anesthesia. Thus it seems difficult to ascribe this short duration of anesthesia (<20 - 30 minutes) as the causative agent responsible for the "problems" reported in the children in this paper?

Lack of sufficient information makes it exceedingly difficult for parents, pediatricians, and anesthesia providers to determine what they should do based on the data presented in this paper. Should these children undergo anesthesia or not? If so under what circumstances should they do so? (It is unfortunate that children undergoing anesthesia for CT and MRI scans were excluded from the study. This might have given some insight into whether the changes presented were due to the surgery or the anesthesia, although this would not help us discern the effects of the underlying problem initiating the investigations.) Without further information, we are left in a quandary as to what to do. Thus, how are we to use these well-publicized data? Should we be counseling our patients against anesthesia and surgery for procedures that should be corrected early in life? Unfortunately, the conclusions drawn from the way the findings of this study have been presented have caused a media stir that will significantly impact our ability to care for our children in the first few years of life. This is unfortunate because what their data show, at most, is that among children exposed to early anesthesia a few are more likely to show low performance on mental tasks in later years compared to unexposed children, some of whom will also perform poorly on such tasks. However, because of the lack of information about the children themselves or the anesthetics used, or of intraoperative problems, it is difficult to attribute their poorer language and cognition outcomes at 10 years of age to having been anesthetized in the early years.

References: 1. Bower C, Ryan A, Rudy E, and Cosgrove P. Report of the Births Defects Registry of Western Australia 1980- 2002. King Edward Memorial Hospital, No. 10, 2003 (1 in 445 prevalence is derived from1995-2002 data). 2. Watson L, Blair E, Stanley F. Report of the western Australian cerebral palsy register to birth year 1999. 3. Hayashi H, Dikkes P, Soriano SG. Repeated administration of ketamine may lead to neuronal degeneration in the developing rat brain. Paediatr Anaesth. 2002 Nov;12:770-4

George A. Gregory, MD Professor Emeritus Anesthesia and Pediatrics University of California, San Francisco

Donna M. Ferriero, MD MS W.H. And Marie Wattis Distinguished Professor & Chair University of California, San Francisco Department of Pediatrics Physician-in-Chief UCSF Benioff Children's Hospital

Rita J. Jeremy, PhD Developmental Psychologist University of California, San Francisco Clinical & Translational Science Institute-Pediatric Clinical Research Center

Conflict of Interest:

None declared

Dear Editor,

I read the paper published by Dr. Caleb Ing et al., entitled "Long- term Differences in Language and Cognitive Function After Childhood Exposure to Anesthesia." This in-depth birth cohort analysis reveals a risk for disability in language and cognitive function in children exposed to general anesthesia. While this could be the most recent and most detailed data to discuss the relationship between general anesthesia and pediatric development, I would like to add a few cautionary comments about the results.

Infantile development is one of the most important concerns for pediatricians, and as such pediatricians must judge whether any treatment is truly all right for the patient's growth and development in the long run. Any analysis of results of developmental research should therefore be done carefully, in particular when determining the analysis methods to adjust the confounder variables.

The authors discussed several demographic features such as single/multiple anesthesia and parental education. However I believe the most important factor is the timing of the anesthesia, in other words the age at which anesthesia is administered.

Other articles have already discussed the effects of anesthesia on the immature brain, showing that the more immature the brain, the more vulnerable it is likely to any insults, which may affect the consequential neurological development(1-2).Sprung JF et al., discussed the association between anesthesia for cesarean delivery and learning disabilities(3). Discussing the timing of surgery would be important for a decision making because physicians are allowed to discuss the timing of the operation in the case of elective surgeries for the balance of the risk of anesthesia on the developing brain and the benefit of surgical repairs.

Given the effects of anesthesia on the immature brain, the most important research question would be "when is the appropriate timing of elective surgery under general anesthesia?". While Dr. Ing's analysis of the Raine Study data is certainly worthy of further discussion, it has provided no conclusive answers about this issue.

References (1)Jevtovic-Todorovic V.et.al., Developmental synaptogenesis and general anesthesia: a kiss of death? Curr Pharm Des. 2012 Jul 2[Epub ahead of print]. (2)Patel P et.al.,Update on neonatal anesthetic neurotoxicity: insight into molecular mechanisms and relevance to humans.Anesthesiology 2009 110(4)703-8. (3)Sprung JF et.al.,Anesthesia for cesarean delivery and learning disabilities in a population-based birth cohort. Anesthesiology 2009 111(2) 302-310

Conflict of Interest:

None declared

The authors believe that the fact that the vast majority of patients underwent minor procedures reduces the potential for confounding due to co -morbidity.[1]

However, the study by Hansen et al, showed that, even among Danish children undergoing inguinal hernia repair (a minor procedure), there was a significant risk of test score non-attainment at age 15-16 (which they regarded as "unusual" as 87% of the population attained scores). Although they could not indentify the reason for this non attainment, after excluding children with other cogenital malformations, the risk of non- attainment was attenuated.[2] This is an indicator that even minor procedures could be a marker for significant co-morbidity.

References

[1]Long-term Differences in Language and Cognitive Function After Childhood Exposure to Anesthesia Caleb Ing, Charles DiMaggio, Andrew Whitehouse, Mary K. Hegarty, Joanne Brady, Britta S. von Ungern-Sternberg, Andrew Davidson, Alastair J.J. Wood, Guohua Li, and Lena S. Sun Pediatrics peds.2011-3822;

[2]: Academic Performance in Adolescence after Inguinal Hernia Repair in Infancy: A Nationwide Cohort Study TG Hansen, JK Pedersen, SW Henneberg... - Anesthesiology, 2011

Conflict of Interest:

None declared