Caroline Helwick

October 25, 2012

NEW ORLEANS — Although much of the data come from preclinical and observational studies, evidence is mounting that exposure to anesthesia in early childhood may increase the risk for learning disabilities and other forms of neurodevelopmental impairment in adolescence, according to speakers at the American Academy of Pediatrics (AAP) 2012 National Conference and Exhibition.

"Concerns regarding the safety of anesthetics in young children have been present within anesthesiology for most of the past decade, and the concern is spreading," said Randall Flick, MD, MPH, associate professor of anesthesiology and pediatrics and chair of pediatric anesthesiology at the Mayo Clinic, Rochester, Minnesota.

But, in spite of noticeable trends, he urged pediatricians not to jump to conclusions and change their practices yet.

Preclinical Evidence of Neurotoxicity

Sulpicio G. Soriano, MD, professor of anesthesia at Harvard Medical School, Boston, Massachusetts, said the scientific evidence is "irrefutable" that exposure to anesthetic drugs induces neurotoxicity and neurodevelopmental disorders in laboratory animals.

Dr. Soriano has led some of the seminal work on the effects of ketamine in infant laboratory animals.

"Recent studies question the safety of the anesthetic drugs we use. Are we causing actual harm to our patients?" he asked during the session. He believes preclinical studies are very informative on this question.

Although a single dose of ketamine in young lab animals does not appear to increase neurodegeneration, continuous infusion (> 9 hours) induces apoptosis of neurons. Likewise, in primates, large and continuous doses of ketamine (24-hour exposure) increase neuronal apoptosis and necrosis in the fetal period, and in 5-day-old animals, it is associated with long-lasting cognitive declines.

Similarly, a single exposure to sevoflurane does not impact tests of neurocognitive function in mice, but 2 hours of exposure over 3 days is associated with neurocognitive decline.

Isoflurane also affects neurogenesis and long-term neurocognitive function, Dr. Soriano noted. Rats exposed to isoflurane in utero during early gestation are behaviorally abnormal as adults, showing deficits in locomotor activity, hippocampal-dependent learning and memory, and increased anxiety.

"There is a time-dependent and dose-dependent increase in neuroapoptosis, and there are functional consequences," he reported.

But Dr. Soriano cautioned against overextrapolation to human neonates. "Human babies are not large rat pups!" he emphasized. "And in laboratory experiments, we use doses 7 times higher than we use in humans, and use them for longer durations. There are problems with animal experiments, but [this does demonstrate] causation."

The Human Evidence

"The available human studies cannot exclude the possibility that the anesthesia-induced neurotoxicity observed in many animal studies may also occur in children," Dr. Flick said in his presentation.

Although there have been about a dozen studies in young children, all are retrospective and observational, and all rely on existing databases, and therefore, "all are flawed," he cautioned.

Nevertheless, he added, "virtually every drug we use in the care of children is implicated. We cannot say with certainty that any are safe."

These include gamma-aminobutyric acid (GABA) agonists, N-methyl-D-aspartic acid (NMDA) antagonists, benzodiazepines, propofol/barbiturates, nitrous oxide, volatile agents, ketamine, and etomidate.

Mayo Clinic Studies

The Mayo Clinic has produced much of the work in this area, involving some of the largest cohorts, and Dr. Frick described their findings.

The Rochester Learning Disabilities Birth Cohort includes 5718 children born from 1976 to 1982 and for whom records are available from schools, medical providers, and reading/dyslexia centers. Participants were individually administered intelligence quotient (IQ) and achievement tests for diagnosing learning disabilities before age 19. Of this large cohort, 593 children received general anesthesia before age 4, allowing investigators to look for associations between anesthetic exposure and the development of a learning disability.

In the oft-quoted Wilder study, which Dr. Frick coauthored, duration of anesthesia was shown to be important (Wilder RT et al, Anesthesiology 2009;110:796-804). A single exposure did not increase risk, but the hazard ratio for children receiving 2 anesthetics was 1.59, and the hazard ratio for those receiving 3 or more was 2.60.

Although 20% of children with 1 exposure had a learning disability — a rate no different from that of the unexposed children — this rose to 35% among children with multiple exposures, Dr. Frick reported.

The risk for a learning disability increased with longer cumulative duration of exposure, with the hazard ratio reaching 1.56 with exposure ≥ 120 minutes.

"These differences were not subtle," he noted. Differences persisted even after controlling for comorbidities related to learning disabilities. Differences were also shown with respect to other endpoints, including speech and language deficits and attention-deficit/hyperactivity disorder (ADHD).

Other Studies, Mixed Results

Not all studies have been as clear, though taken as a group, the trend is for multiple exposures to be associated with neurodevelopmental disorders later in life.

In a 2009 discordance study of 1143 monozygotic twin pairs, early exposure to anesthesia (< age 3) was not associated with a difference in school performance compared with the unexposed twin, though this study relied on parental reporting of anesthesia exposure (Bartels M et al, Twin Res Hum Gen 2009;12:246-253).

But in another retrospective cohort study that year, the 383 children who underwent inguinal hernia repair before age 3 had a 2.3 increased risk for a developmental or behavioral disorder, compared with control participants (DiMaggio C et al, J Neurosurg Anesthesiol 2009;21:286-91). In a second study by the same authors, a single exposure did not increase these risks in comparison with siblings, though multiple exposures did (DiMaggio C et al, Anesth Analg 2001;113:1143-51).

In a very large national cohort study from Denmark involving 2689 children who underwent inguinal hernia repair in infancy, no evidence was found that a single, relatively brief anesthetic exposure reduced academic performance in adolescence (Hansen TG et al, Anesthesiology 2011;114:1076-85).

In the most recent study, from the University of Iowa, duration of surgery and anesthesia correlated negatively with scores on academic achievement tests; however, the authors felt that "other explanations," such as referral bias, were possible (Block RI et al, Anesthesiology 2012;117:494-503).

What the Studies Mean: "Enormous Public Health Problem"

"We have to be careful about interpreting observational studies," Dr. Frick emphasized, "but if there is an increased likelihood of a learning disability after anesthesia exposure, studies indicate that for every 6 children exposed to multiple anesthetics, 1 additional child will develop a learning disability who would not otherwise. This suggests we have an enormous public health problem that must be addressed."

Summing up "what is known," he made the following points:

  • Single anesthetics do not seem to have an effect.

  • Repeated anesthetic exposures consistently show an effect.

  • The effect persists despite adjustment for comorbidity.

  • Learning (reasoning), speech, and language but not behavior appear to be affected.

  • ADHD follows a similar pattern.

  • There are many opportunities for bias, confounding, and so forth in observational studies.

Three large prospective trials are being initiated or are already in progress, and these may provide a better quality of evidence on which to base decisions, Dr. Flick explained.

Meanwhile, what should pediatricians tell parents? "That the associations are not necessarily causal," Dr. Flick advised, "and that brief, single-anesthetic exposures do not seem to cause problems."

As for providers, he maintained, "We need more information before we change our practice, which could have unintended consequences."

Dr. Soriano agreed. "Have I changed my practice? I flat out say 'no.' "

Subspecialists Express Concern

Several subspecialists who participated in a panel discussion after the presentation said the findings are alarming.

Diego Preciado, MD, a pediatric laryngologist at Children's National Medical Center in Washington, DC, said, "These talks are very salient to our practice and very concerning."

"We treat many children with airway disorders, and they are repeat offenders. Some have had surgery more than 30 times, and many are under prolonged sedation — for example, when we do airway reconstruction. This is a group that is ripe to study, and we clearly need more information on these risks."

Brian Shaw, MD, a pediatric orthopedic surgeon in Roanoke, Virginia, commented, "This is big news to me. I haven't given much thought to anesthesia risk."

He said his specialty is in an ideal position to make changes. "We can do so much now under regional anesthesia, and we can give nerve blocks to reduce the amount of general anesthesia we need," he said. "The good news is that general anesthesia is usually a single exposure in orthopedic surgery, and also, many procedures — such as simple deformities — can be delayed or deferred until the child gets a little older. These findings may change some of our practices."

Dr. Soriano, Dr. Flick, Dr. Preciado, and Dr. Shaw have disclosed no relevant financial relationships.

American Academy of Pediatrics (AAP) 2012 National Conference and Exhibition. Presented October 20, 2012.