Acknowledgements: Preparation of this chapter was supported in part by National Institute of Neurological Disorders and Stroke grants R01 NS 46308, R01 NS 29462, and Department of Education grant H133G040279.
Introduction
Since Caffey and Kempe et al’s groundbreaking reports linking battering and shaking of infants to cerebrovascular lesions and mental retardation, numerous studies have examined mechanisms of injury and predictors of outcome in this most vulnerable population of children. Assessment of specific outcomes across different clinical populations varies depending on the level of analysis and time frame. The six major areas of outcome research include death, disease, disability, discomfort, dissatisfaction, and destitution. , In acute care studies, death and disease are used as outcome markers, although post-acute studies typically examine disability, impairment, and health-related quality of life. This chapter examines the literature regarding post-acute outcomes in children with abusive head trauma (AHT). To facilitate evaluation of the strength of the empirical foundation for determining outcomes, findings are examined in relation to study methodology. Outcomes from studies examining prospectively recruited samples of children with AHT evaluated using rating scales and/or standardized measures of cognition and behavior are compared with outcomes of children with noninflicted head trauma (NHT) or to sociodemographically similar community comparison children. Findings from descriptive studies examining rates of disability and impairment are also examined. Predictors of post-acute outcomes, including coma rating scales, biomarkers, neuroimaging findings, and family factors, are reviewed. Finally, directions for future research on outcomes following early brain injury are discussed.
Neurobehavioral and Neuropsychological Outcomes
Studies examining outcomes after inflicted childhood neurotrauma vary along several key dimensions, including determination of the external cause of injury, recruitment and follow-up methodology, and the specific outcome domains evaluated. Determination of the presence of AHT is difficult since the history is often not provided or is incompatible with the clinical presentation. Different approaches to identifying AHT have attendant strengths and weaknesses and can introduce different biases. Several investigators applied an algorithm similar to Duhaime et al to categorize injuries as presumptive or suspicious for AHT based on the presenting injury in relation to the history and associated findings. This approach is affected by the circular reasoning inherent in using specific clinical findings often associated with abusive injury to support diagnosis of abusive injury. Hymel et al used a priori criteria, including admitted abusive acts, infants with acute cardiorespiratory compromise linked to traumatic injuries, developmentally inconsistent histories, changing explanations of the trauma, and presence of at least two noncranial injuries considered moderately or highly associated with abuse. Application of these criteria in a multicenter study resulted in classification of the external cause of injury as “undetermined” in 24%. Developmental outcomes of the undetermined group were significantly more favorable than in children classified as either AHT or NHT. Because of the difficulty diagnosing putative inflicted injury in milder cases, a bias exists toward inclusion of children with more severe traumatic brain injury (TBI) in the inflicted injury group. This bias is most likely to affect retrospective studies.
Post-acute outcome studies are unavoidably biased by the willingness of parents and guardians to volunteer to participate. High rates of attrition approaching 50% within 6 months to 1 year after injury are common. Long-term outcome studies are likely to retain children from intact families who have resources to facilitate participation, in situations in which the perpetrator is not a close relative, and in children with very poor outcomes who continue to require medical care.
Outcomes in children with AHT who have been assessed from subacute to chronic stages of recovery include neurological findings, global outcome ratings, or specific neurodevelopmental outcomes assessed using standardized instruments. Table 48-1 provides findings from studies examining outcome using rating scales and standardized mental, motor, and/or adaptive behavior outcomes in children with AHT. The Glasgow Outcome Scale (GOS) and pediatric variants (Pediatric Outcome Performance Category) are rating scales that assess outcome on a five- or six-point scale ranging from good outcome to death. Standard scores have a mean of 100 and a standard deviation of 15 or 16.
| Author | Sample Size | Months of Age at Injury (Range) Mean | Study Design | Follow-Up Interval | Glasgow Outcome Scale or Pediatric Outcome Performance Category (%) | Standardized Assessment (M) | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Comparison of Abusive Head Trauma to Noninflicted TBI and/or Control Groups | ||||||||||||
| Good/mild | Moderate | Severe | PVS | Mental | Motor | Adaptive | ||||||
| Beers (2007) | a = 15 n = 15 | a = 5.8* n = 17.2 | Prospective cross-sectional cohort | 6 mo | a = 69.0* n = 97.3 | — — | 95.9* 115.8 | |||||
| Ewing-Cobbs (1998) | a = 20 n = 20 | (0-59) a = 10.6* n = 35.6 | Prospective longitudinal cohort | 1.3 mo | a: n: | 20 55 | 65 25 | 15 20 | 0* 0 | a = 78.2* n = 87.7 | 80.3* 84.3 | — — |
| Ewing-Cobbs (1999) | a = 28 c = 28 | (0-42) a = 9.3 c = 9.4 | Prospective longitudinal cohort | 4.6 mo | a: | 25 | 61 | 14 | 0 | a = 82.1* c = 97.7 | 81.9* 100.5 | — — |
| Hymel (2007) | a = 11 n = 30 ? = 13 | (0-35) a = 10.5 n = 9.0 ? = 11.5 | Prospective longitudinal cohort and convenience sample | 6 mo follow-up in: 4 iTBI, 16 nTBI, 6 ? TBI | a = 60.0* n = 94.4 ? = 107.3 | 59.8* 101.8 102.2* | — — — | |||||
| Keenan (2004) | a = 62 n = 50 | (0-23) a = 4.0* n = 7.5 | Prospective population-based longitudinal cohort | Discharge | a: n: | 55 82 | 45* 18 | |||||
| (Dichotomized into good/mild vs moderate/severe) | ||||||||||||
| Keenan (2006a) | a = 41 n = 31 | Follow-up of prospective population-based longitudinal cohort | 1 y | a: n: | 53 77 | 20 16 | 27 7 | 0* 0 | — — | — — | 96.4* mdn 100.0 mdn | |
| Keenan (2007) | a = 25 n = 23 c = 31 | Follow-up of prospective population-based longitudinal cohort | 1-3 y | a = 68* n = 84 | 55* 92 | 94 100 | ||||||
| Descriptive Studies of Inflicted TBI | ||||||||||||
| Barlow (2005) | a = 25 | (0-34) 2.3 mdn | Cross-sectional and prospective longitudinal | 59 m | 48 | 16 | 36 | — | ||||
| Bonnier (2003) | a = 25 | (0-13) 4.2 | Retrospective cohort | 6 y (2.5-13 y) | 4 | 35 | 48 | 12 | a = 63.3 | |||
| Duhaime (1996) | a = 14 | (1-24) 6.4 | Follow-up of prospective cohort | 9 y | 14 | 36 | 43 | 7 | ||||
| Ghahreman (2005) | a = 56 | (0.5-46) 8.2 | Retrospective record review | Mdn 20 mos | 39 | 19 | 26 | 10 | ||||
| Johnson (1995) | a = 28 | 5 | Retrospective record review | — | 32 | 25 | 42 | 4 | ||||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
