The Unspoken Question!

See Articles, p 723 and 734

Why do so many otherwise healthy children acquire neurological impairments after treatment for a serious illness? It is rare for us to see a chronically ill child who has had many surgical procedures who does not develop a neurological impairment. This observation has been corroborated in the literature.¹ Therefore, the hypothesis that general anesthesia is neurotoxic in infants and children does not seem farfetched, and has fueled an exponential rise of both basic and clinical investigations and publications.² The evolving nature of preclinical investigations in anesthetic neurotoxicity has not only identified subtle behavioral impairments in nonhuman primates, but also uncovered new insights into mechanisms of general anesthesia.

In his article in this issue of the Journal, Barnes³ declares that it is “time to stop” pediatric anesthesia neurotoxicity research based on clinical findings of studies supporting the lack of neurocognitive deficits associated with general anesthesia. He points to no causation between anesthesia exposure and later neurocognitive difficulties in twin studies, sibling studies, and the only randomized prospective trial to date, the General Anesthesia Compared to Spinal Anesthesia Study (GAS) trial, which compared the outcomes of infants randomized to either general anesthesia or regional anesthesia for inguinal hernia repair.^4–6 Taken together, these reports provide compelling evidence that short-to-moderate length anesthesia exposure at a young age does not affect neurocognitive outcomes in terms of overall intelligence and school readiness. However, these rigorous investigations tell us nothing about the long-term effects of prolonged or repeated anesthetic exposures in young children. Recently, a cohort of 212 survivors of childhood acute lymphoblastic leukemia from a single tertiary-care pediatric institution underwent neurocognitive testing and neuroimaging.⁷ After adjusting for chemotherapeutic doses and age of diagnosis, higher propofol cumulative dose, volatile anesthetic exposure, and longer anesthesia duration were individually associated with neurocognitive impairment. On average, these children underwent almost 16 hours of anesthesia and 27 anesthetic exposures during their treatment. However, as Barnes³ points out, “Children are given anesthetics in order to undergo surgery or other procedures, making it extraordinarily difficult to eliminate the possibility that any findings are indicative of association but not causation.” Because these patients were from a single institution subjected to the same clinical protocols, the children with more frequent anesthetic exposures probably required more procedures due to complications of treatment or severity of initial illness. Both lead to prolonged hospitalizations, social isolation, and concurrent medications and therapies. These are known confounders that directly impact neurocognitive development.

An alternative explanation for the neurological impairments in this group of young oncology patients is that stressed and frail patients may be vulnerable to the ostensible toxic effects of anesthesia drugs. This might fit into the stress diathesis model in which patients who have an underlying vulnerability when stressed are likely to develop long-term adverse sequelae, whereas resilient patients given the same stressors will likely not develop adverse sequelae.⁸ The stressors occurring during anesthetic exposure might include anesthetic neurotoxicity, hypotension, hypocapnia, or hypoxia.⁹ Other stressors occurring as a result of the surgical procedures include systemic inflammation, which has been shown to affect neurocognition in adults. It is impossible to determine which explanation is correct, or if both explanations are correct, from this retrospective study design. So, essentially, we are left in the same place we were when the first retrospective large cohort studies were published many years ago that showed an association between general anesthesia exposure and later neurocognitive disabilities. Another take-home message from this study was the sheer number of general anesthetics that children are exposed to if they are unlucky enough to develop a serious illness. The factors that lead to poor neurocognitive outcomes in patients with major illnesses, including the effects of frequent and prolonged anesthetic exposures, need to be studied.

In this same issue, Ing et al¹⁰ reported increased utilization of medications for attention deficit hyperactivity disorder (ADHD) in children who were exposed to general anesthesia at a young age. This finding supports a previous report linking anesthesia and surgery to ADHD. For instance, the data analyzed by Ing et al¹⁰ may be deeply confounded by the choice of procedures. Figures 2D and 3E reveal a relatively high hazard ratio for persistent utilization of ADHD medication in patients exposed after 2 years of age and undergoing tonsillectomy and adenoidectomy, respectively. Because the most common indication for tonsillectomy and adenoidectomy in the United States is obstructive sleep apnea, it stands to reason that there would be a high percentage of these patients treated for ADHD as well as other behavioral problems, such as enuresis, because these conditions are associated with obstructive sleep apnea.¹¹

From a public health standpoint, research and funding should also be directed toward examining the factors and confounders that impact pediatric anesthesia care to improve neurocognitive outcomes. Ultimately, clinical answers are what our colleagues demand, not experimental observations in the laboratory. Most of the retrospective studies have relied on databases that are already populated with information about outcome measures such as school readiness tests, referrals for educational help or neurobehavioral medical diagnoses. When compared to randomized controlled trials or prospective comparisons of cohort, retrospective database analyses are less expensive and time consuming to perform. The effort, time, and money to explore these databases are minimal. So the argument that a great deal of research money is funding studies in which the research question is already settled does not hold water. The argument can be and should be made that these type of retrospective database analyses are so confounded that any associations found are specious. Furthermore, retrospective observations do not confirm causation.

The fact that parents, families, and caregivers being unduly concerned about the neurotoxic effects of anesthesia is a real issue. In our experience, the best response to this concern is for anesthesiologists to be aware and conversant about the available literature. Generally, parents in our practice have their concerns allayed about both potential neurotoxicity and overall immediate risks of general anesthesia when they are presented with the data. In a perverse way, undue parental concern may be a marker of those parents who need extra information and time to feel comfortable with their child undergoing anesthesia and surgery. It is the responsibility of the anesthesiologists to make sure that proceduralists, be they surgeons, endoscopists, or imaging specialists, are up to date on the real risks of general anesthesia for young children. If the potential of repeated anesthetic or prolonged duration of anesthetic exposure makes some proceduralist pause in their deliberations about whether a certain examination is absolutely necessary, that is probably a good thing.

Finally, we would like to make the argument that research on pediatric anesthetic neurotoxicity has broader implications for all those children who require prolonged sedation in intensive care units. The mean length of time for sedation and neuromuscular paralysis for patients undergoing procedures for long gap esophageal atresia repair by traction sutures is reported to be 14 days for primary repairs and 35 days for secondary repairs.¹² Although we agree with Barnes³ that the sole concern should not be on developing anesthetic regimens that are less neurotoxic, we do feel that research into developing alternative regimens for sedation for prolonged intensive care may benefit patients by decreasing anesthetic neurotoxicity as well as improving hemodynamic stability. Prolonged sedation may also provoke withdrawal syndrome in these vulnerable patients. The real costs of prolonged sedation for our young patients are just being explored, and the potential for neurotoxicity of these agents is just one of the many concerns regarding these children.

Barnes,³ appropriately, is trying to shift the attention away from “potential toxins” to the real concerns about how we need to improve the conduct of pediatric anesthesia. We wholeheartedly agree and, in fact, would like to broaden the mandate of pediatric anesthesiologists to examine all aspects of pediatric care and determine ways that it can be improved. As the specialty that cares for children in the operating room, intensive care units, and pain clinics, we are uniquely positioned to take on the challenge of determining which factors are most important in assuring optimal care for our patients. This approach will have to be collaborative, innovative, and multifactorial. Now that the question has been spoken, it behooves us to diligently develop strategies to keep neurologically intact children normal during serious illness and to limit any further adverse neurological sequelae in children who are already compromised.

DISCLOSURES

Name: Mary Ellen McCann, MD, MPH.

Contribution: This author helped conceive and design the editorial, analyze and interpret the data, draft and revise the manuscript, and approve the final manuscript.

Name: Sulpicio G. Soriano, MD.

Contribution: This author helped conceive and design the editorial, analyze and interpret the data, draft and revise the manuscript, and approve the final manuscript.

This manuscript was handled by: James A. DiNardo, MD, FAAP.