Errors of Fact and Reasoning in Consideration of Shaken Baby Syndrome To the Editor: M iller and Miller 1 reanalyze data from others to suggest that male sex in- creases the risk of subdural hemorrhage after birth or head injury in infants and children. As pediatric intensivists involved daily in the care of severely head-injured children, we are aware of fundamental errors in the analysis and interpretation that undermine this finding. Their putative evidence is a recorded excess of males with subdural bleeding from a variety of causes. Any comment about a male preponderance in subdural bleeding has to consider a male prepon- derance in the incidence of injury. Only if subdural bleeding occurs in a higher proportion of male children after severe head injury (rather than simply numeri- cally more children) can their hypothesis be sustained. It is regrettable that despite 117 references, neither Miller and Miller 1 nor the reviewers thought it appropriate to consider the unambiguous data that there is a very significant male prepon- derance in severe head injury at all ages. The seminal article of Tilford et al, 2 ana- lyzing 98,023 critically head-injured chil- dren, describes 69,150 males (70.5%). Similar proportions are evident in the French data set (67% of 585 cases) 3 and the UK Paediatric Intensive Care Audit data (69% of 620 cases). 4 These data com- bined give us 69,957/99,223 = 70.5% male preponderance. This difference remains in the youngest age groups. 4 Miller and Miller provide tables de- scribing the absolute number of cases by sex of subdural bleeding after accidental (Table 2; 62.2%, males 163/262) and inflicted trauma (Table 3; 62.6% males, 1007/1609). If we assume that their fig- ures are correctly extracted from the lit- erature then it is a simple matter of estimation of the difference between pro- portions in comparison to the data previ- ously mentioned. We observe an absolute risk reduction for subdural bleeding with male sex of around 7% in both scenarios (95% confidence interval, 2%Y13% for accidental [P = 0.003] and 5%Y9% for inflicted injury [P G 0.0002]; significance of difference between 2 independent pro- portions). Put simply, there is no evidence that a higher proportion of male children develop subdural hemorrhages than would be expected from the baseline trauma epi- demiology. Miller and Miller’s hypothesis therefore can be refuted. The situation should not be confused by considering birth-related subdural bleed- ing where head size may indeed contribute to a differential level of trauma during childbirth (including subdural bleeding) in male neonates. Unfortunately, the article is further weakened by misquotations of the litera- ture in describing subdural hemorrhage frequency. For example, of the 4 largest series quoted in Table 3, figures extracted from 3 of these series are incorrect. The figures presented from Wells et al 5 are those for all types of intracranial hemor- rhage (including 80 cases with epidural hemorrhage, 36 with subarachnoid, and 13 with intraventricular bleeding) rather than subdural bleeding as described. The Canadian data are similarly misquoted in that 364 cases are presented by Miller and Miller, whereas the number with subdural bleeding is presented as 313 in the Canadian article. 6 The article of Jenny et al 7 suffers the same fate as that of Miller and Miller, incorrectly analyzing 173 cases (the number of head injuries) not of cases with subdural hematomas (n = 150). Perhaps the most extraordinary ele- ment of this article is the interpretation applied to these flawed data. The authors suggest (Fig. 1) that slowly enlarging sub- dural hemorrhages in at-risk cases under- lie what is misdiagnosed as shaken baby syndrome. This is inconsistent with the typical history of acute collapse, the im- aging of thin-film subdural hemorrhages (not space-occupying hematoma), 8 and the pathology of the triad of predomi- nantly diffuse hypoxic-ischemic encepha- lopathy. 9 The authors themselves restate these features. 8,9 Articles such as this that lack preci- sion in data collection and analysis and then build dramatic interpretations on these flawed analyses do not advance knowledge. Worse, they may contribute to misdiagnosis of inflicted head injury, thereby exposing vulnerable children to risk. AUTHOR DISCLOSURE INFORMATION The authors declare no conflicts of interest. Joe Brierley, MBChB, MRCP FRCPCH, MA Mark J. Peters, MBChB, MRCP FRCPCH, PhD Paediatric Intensive Care Unit Great Ormond St Hospital London, UK brierj@gosh.nhs.uk REFERENCES 1. Miller R, Miller M. Overrepresentation of males in traumatic brain injury of infancy and in infants with macrocephaly: further evidence that questions the existence of shaken baby syndrome. Am J Forensic Med Pathol. 2010;31(2):165Y173. 2. Tilford JM, Aitken ME, Anand KJ, et al. Hospitalizations for critically ill children with traumatic brain injuries: a longitudinal analysis. Crit Care Med. 2005;33(9):2074Y2081. 3. Ducrocq SC, Meyer PG, Orliaguet GA, et al. Epidemiology and early predictive factors of mortality and outcome in children with traumatic severe brain injury: experience of a French pediatric trauma center. Pediatr Crit Care Med. 2006;7(5):461Y467. 4. Parslow RC, Morris KP, Tasker RC, et al. Epidemiology of traumatic brain injury in children receiving intensive care in the UK. Arch Dis Child. 2005;90(11):1182Y1187. 5. Wells RG, Vetter C, Laud P.Intracranial hemorrhage in children younger than 3 years: prediction of intent. Arch Pediatr Adolesc Med. 2002;156(3):252Y257. 6. King WJ, MacKay M, Sirnick A. Shaken baby syndrome in Canada: clinical characteristics and outcomes of hospital cases. CMAJ. 2003;168(2):155Y159. 7. Jenny C, Hymel KP, Ritzen A, et al. Analysis of missed cases of abusive head trauma. JAMA. 1999;281(7):621Y626. 8. Geddes JF, Vowles GH, Hackshaw AK,et al. Neuropathology of inflicted head injury in children. II. Microscopic brain injury in infants. Brain. 2001;124(Pt 7):1299Y1306. 9. Geddes JF, Hackshaw AK, Vowles GH,et al. Neuropathology of inflicted head injury in children. I. Patterns of brain damage. Brain. 2001;124(Pt 7):1290Y1298. RESPONSE Reply: I n response to the letter by Peters and Brierley regarding our study, we offer the following comments 1 : Their main criti- cism is the control group that we used. We compared our results of male/female LETTER TO THE EDITOR e12 www.amjforensicmedicine.com Am J Forensic Med Pathol & Volume 33, Number 3, September 2012 Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.