David O. Warner1
1 Professor of Anesthesiology, Department of Anesthesiology, Mayo Clinic, Rochester; USA
Randall P. Flick2
2 Associate Professor of Anesthesiology and Pediatrics, Departments of Anesthesiology and Pediatrics and Adolescent Medicine, Mayo Clinic, Rochester; USA
Corresponding author: Dr. D.O. Warner ([email protected]); Department of Anesthesiology, Mayo Clinic, 200 First Street SW, Rochester, MN, USA, 55905.
We read with interest the recent review by Dr. Hansen of anesthetic neurotoxicity (1). Another paper from his group (2) describes the analysis of educational outcomes of a cohort of children who underwent a neurosurgical procedure prior to their first birthday compared to a sample of the national population not so exposed. Perhaps not surprisingly, those children who survived scored more poorly on group tests of achievement. Referring this paper in his review, Dr. Hansen makes the remarkable claim that “…[this] new cohort study have (sic) unambiguously demonstrated that pooling of data from multiple surgeries in studies on anesthesia-related neurotoxicity and the developing brain is not a legitimate approach.” As authors of several of these “illegitimate” studies (3), we wish to respond.
The logic of this conclusion seems to be that because one set of underlying conditions that necessitate surgery (neurological diseases) can affect neurodevelopment, all can, presumably in varying degrees. Thus, any analysis that includes a mix of procedures is hopelessly confounded, as any observed associations can be explained by the underlying condition or the consequences of surgery. To be clear, all published studies examining the question of anesthesia neurotoxicity in children are observational, all have the potential for this and other confounds, and all extensively discuss this potential as a limitation (as first pointed out by Wilder et al (3)). However, in each of these studies, major procedures performed in children with underlying pathology that may contribute to the risk of adverse developmental outcomes (such as neurosurgery) are only a minor fraction of the total number of procedures. For example, in the Mayo Clinic cohort (3), of the total of 875 surgeries performed prior to age 4 years, 9 (1.1%) were neurosurgical and 15 (1.7%) were cardiac procedures. Of interest, 4 children receiving neurosurgery and 6 children receiving cardiac surgery had multiple procedures; none of these children developed learning disabilities, the primary outcome of these studies. It also would seem that if the underlying condition or consequences of certain surgical procedures were determinative of adverse outcomes, this would bias against finding associations between anesthesia/surgery and these outcomes in studies that include heterogeneous procedures.
Conversely, the choice of a single procedure is not certain to eliminate or even reduce the opportunity for confounding. For example, inguinal hernia repair, the procedure used as a surrogate for anesthetic exposure in a previous study from Dr. Hansen’s group (4), is unambiguously associated with prematurity and the potential for reduced achievement. Other weaknesses of this study include that group achievement tests may not be a sufficiently sensitive indicator of injury, its inability to differentiate between singly and multiply exposued children, and that not all exposed children were tested for the primary outcome. We mention these weaknesses not to unduly criticize this work, but to emphasize that every observational study has strengths and weakness.
We would argue, as in the past has Dr. Hansen (5), that examining neurodevelopmental outcomes across wide range of procedures and anesthetic types is in fact both necessary and a strength when seeking the associations between exposures and outcomes. One of the ten criteria proposed by Sir Bradford Hill to establish a causal relationship between the two (6) is consistency, with similar effects observed across a range of conditions. The fact that significant effect sizes (hazard ratios of approximately 2) are consistently observed for the association between multiple exposures and outcomes when a wide range of procedures are analyzed together speaks to this criterion.
This is but one example of the considerable complexities involved in generating (and interpreting) data regarding this issue. We were thus somewhat surprised that the Dr. Hansen was ready to conclude based on current limited evidence that “anesthesia-related neurotoxicity and the developing animal brain is not a significant problem in children”. This may be meant as a rhetorical device to provoke discussion, and there is certainly much that is not yet certain, but many of us are not quite yet ready to reach such a sanguine conclusion, especially since the potential public health impact of any such effects are so important. We agree that all current studies have limitations (as has been discussed and acknowledged on multiple occasions), much remains to be learned, and that children should not be denied essential surgery. However, the scientific process is ill-served by the blanket dismissal of most published human studies. We fervently hope that in fact anesthesia-related neurotoxicity is not a significant (or perhaps minor?) problem, but believe that such a conclusion is not warranted at this time – we simply do not yet know.
Conflicts of interest: No conflicts of interest declared.
1 Hansen TG. Anesthesia-related neurotoxicity and the developing animal brain is not a significant problem in children. Paediatr Anaesth 2014 Sep 30. doi: 10.1111/pan.12548. [Epub ahead of print]
2 Hansen TG, Pedersen JK, Henneberg SW, et al. Neurosurgical conditions and procedures in infancy are associated with mortality and academic performances in adolescence: a nationwide cohort study. Paediatr Anaesth 2014 Sep 24. doi: 10.1111/pan.12533. [Epub ahead of print]
3 Wilder RT, Flick RP, Sprung J, et al. Early exposure to anesthesia and learning disabilities in a population-based birth cohort. Anesthesiology 2009; 110: 796-804.
4 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: 1076-1085.
5 Hansen TG, Flick R. Anesthetic effects on the developing brain: insights from epidemiology. Anesthesiology 2009; 110: 1-3.
6 Flick RP, Warner DO. A Users’ Guide to Interpreting Observational Studies of Pediatric Anesthetic Neurotoxicity: The Lessons of Sir Bradford Hill. Anesthesiology 2012. 117: 459-462.