Retinal Hemorrhages in Children
Retinal Hemorrhages in Children
Agrawal S, Peters MJ, Adams GG, Pierce CM
Pediatrics. 2012;129:e1388-e1396
Retinal hemorrhages, subdural hemorrhages, and acute severe encephalopathy are considered highly suggestive, rather than diagnostic, of abusive head trauma. Evidence shows, for example, that retinal hemorrhages occur as a result of multiple events such as cardiopulmonary resuscitation and bleeding disorders. In previous research, retinal hemorrhages caused by abusive head trauma often had a different shape (such as a flame shape), involved multiple layers of the retina, and were more likely to be bilateral.
Agrawal and colleagues used a prospective cohort design to examine the prevalence and types of retinal hemorrhages in critically ill children aged 6 weeks to 16 years. Children with known abusive head trauma were excluded. The children were admitted to a single pediatric intensive care unit from 2008 to 2009 and were evaluated for coagulopathy and other critical clinical conditions. The clinical presentation of each child was classified as 1 of the following:
Data on clinical parameters, such as the results of laboratory testing, imaging, medications used, and medical comorbid conditions, were also collected. All patients underwent dilated funduscopic examination, and photographs were taken of the retina to assess for the outcome of interest. Most of the 159 children (median age: 18 months, slightly more boys than girls) were evaluated within 48 hours of admission to the intensive care unit. Retinal hemorrhages were found in 15% (n=24) of the patients. Two thirds of the hemorrhages were mild, 11 were unilateral, and most were single-layered within the retina. Two patients had moderate retinal hemorrhage: one after enteroviral-induced sepsis and the other after witnessed accidental trauma. Finally, 6 patients (3.8%) had severe retinal hemorrhages, all of which were multilayered; 5 were bilateral. Comorbid conditions in the children with severe hemorrhage included leukemia in 50%, accidental traumatic brain injury in 2 patients, and a severe coagulopathy due to late-onset hemorrhagic disease of the newborn in 1.
In multivariable analysis, age and gender were not associated with the prevalence of retinal hemorrhage. However, a finding of coagulopathy and an admission diagnosis of sepsis were both associated with risk for retinal hemorrhage, and both odds ratios were >2.5. In contrast, respiratory failure as the admission diagnosis was associated with a reduced risk for retinal hemorrhage (odds ratio, 0.23; 95% confidence interval, 0.06-0.8).
Agrawal and colleagues concluded that 15% of children with critical illness have retinal hemorrhages, but most were mild. Only one fourth of children had severe, multilayered retinal hemorrhages typically seen in abusive head trauma, and all of these hemorrhages were associated with severe or fatal witnessed accidental trauma, severe coagulopathy, or sepsis.
Evaluation of children for physical abuse can be one of the most challenging scenarios faced by healthcare providers. Few communities have specialists or specialty teams to evaluate abuse and neglect, so many pediatric providers are truly at the front line for these conditions. Take-home points from this study include the fact that evaluation for coagulopathy should be completed in any child in whom retinal hemorrhage is identified. In addition, I am struck that the history of the injury as well as the clinical picture matter greatly in deciding whether a retinal hemorrhage is caused by a clinical condition or possible abusive head trauma. I don't think this study provides any absolutes, but it does help clarify the fact that severe, bilateral, multilayered retinal hemorrhages occur primarily in the most critically ill children, suggesting that their presence in children without such illness is highly suggestive of abuse.
Abstract
Prevalence of Retinal Hemorrhages in Critically Ill Children
Agrawal S, Peters MJ, Adams GG, Pierce CM
Pediatrics. 2012;129:e1388-e1396
Retinal Hemorrhages
Retinal hemorrhages, subdural hemorrhages, and acute severe encephalopathy are considered highly suggestive, rather than diagnostic, of abusive head trauma. Evidence shows, for example, that retinal hemorrhages occur as a result of multiple events such as cardiopulmonary resuscitation and bleeding disorders. In previous research, retinal hemorrhages caused by abusive head trauma often had a different shape (such as a flame shape), involved multiple layers of the retina, and were more likely to be bilateral.
Study Summary
Agrawal and colleagues used a prospective cohort design to examine the prevalence and types of retinal hemorrhages in critically ill children aged 6 weeks to 16 years. Children with known abusive head trauma were excluded. The children were admitted to a single pediatric intensive care unit from 2008 to 2009 and were evaluated for coagulopathy and other critical clinical conditions. The clinical presentation of each child was classified as 1 of the following:
Traumatic brain injury;
Sepsis;
Respiratory pathology; or
Nontraumatic encephalopathy.
Data on clinical parameters, such as the results of laboratory testing, imaging, medications used, and medical comorbid conditions, were also collected. All patients underwent dilated funduscopic examination, and photographs were taken of the retina to assess for the outcome of interest. Most of the 159 children (median age: 18 months, slightly more boys than girls) were evaluated within 48 hours of admission to the intensive care unit. Retinal hemorrhages were found in 15% (n=24) of the patients. Two thirds of the hemorrhages were mild, 11 were unilateral, and most were single-layered within the retina. Two patients had moderate retinal hemorrhage: one after enteroviral-induced sepsis and the other after witnessed accidental trauma. Finally, 6 patients (3.8%) had severe retinal hemorrhages, all of which were multilayered; 5 were bilateral. Comorbid conditions in the children with severe hemorrhage included leukemia in 50%, accidental traumatic brain injury in 2 patients, and a severe coagulopathy due to late-onset hemorrhagic disease of the newborn in 1.
In multivariable analysis, age and gender were not associated with the prevalence of retinal hemorrhage. However, a finding of coagulopathy and an admission diagnosis of sepsis were both associated with risk for retinal hemorrhage, and both odds ratios were >2.5. In contrast, respiratory failure as the admission diagnosis was associated with a reduced risk for retinal hemorrhage (odds ratio, 0.23; 95% confidence interval, 0.06-0.8).
Agrawal and colleagues concluded that 15% of children with critical illness have retinal hemorrhages, but most were mild. Only one fourth of children had severe, multilayered retinal hemorrhages typically seen in abusive head trauma, and all of these hemorrhages were associated with severe or fatal witnessed accidental trauma, severe coagulopathy, or sepsis.
Viewpoint
Evaluation of children for physical abuse can be one of the most challenging scenarios faced by healthcare providers. Few communities have specialists or specialty teams to evaluate abuse and neglect, so many pediatric providers are truly at the front line for these conditions. Take-home points from this study include the fact that evaluation for coagulopathy should be completed in any child in whom retinal hemorrhage is identified. In addition, I am struck that the history of the injury as well as the clinical picture matter greatly in deciding whether a retinal hemorrhage is caused by a clinical condition or possible abusive head trauma. I don't think this study provides any absolutes, but it does help clarify the fact that severe, bilateral, multilayered retinal hemorrhages occur primarily in the most critically ill children, suggesting that their presence in children without such illness is highly suggestive of abuse.
Abstract
