Change in blood– brain barrier permeability during pediatric diabetic ketoacidosis treatment

Diabetic ketoacidosis (DKA) is a frequent presentation of type I diabetes in children, occurring with an incidence of 15% to 70% worldwide and approximately 25% in the United States (1, 2). DKA also occurs in children with known diabetes with acute illness or poor adherence to the recommended diabetes management. Cerebral edema (CE) is a rare but devastating complication of pediatric DKA, occurring in <1% of cases. With a suggested mortality rate of 20% to 50%, CE represents the leading cause of death in pediatric type 1 diabetes and leaves significant neurologic morbidity in 35% to 40% of survivors (3–7). However, the etiology of CE and the strategies to prevent this complication remain unknown. Investigations into the pathophysiology of CE in pediatric DKA suggest both cytotoxic and vasogenic pathways (6 – 8). A large retrospective case– control study comparing clinical features of children with DKA with and without CE suggested hyperventilation, elevated blood urea nitrogen, and treatment with bicarbonate were associated with a higher incidence of CE with increases in relative risk as high as 4.2 for the individual risk factors (6). The authors concluded that dehydration and cerebral ischemia cause cytotoxic injury and thus the observed CE. Yet, despite the epidemiologic data associating hypocapnia with clinical CE, markers of cerebral ischemia are not elevated during DKA treatment (6). Alternatively, proponents of a vasogenic mechanism of CE formation note that normal cerebral blood flow velocities have been demonstrated despite hypocapnia in DKA, demonstrating relative cerebral hyperemia and that cerebral autoregulation is abnormal (7). Magnetic resonance imaging (MRI) examinations of children acutely ill with DKA show a decrease in apparent diffusion coefficient from initial to after recovery in most brain areas and shorter mean transit times, both suggesting a vasogenic process, as the predominant mechanism of edema formation rather than osmotic cellular swelling (8). The etiology of DKA-related CE is likely multifactorial and not simply the result of either cerebral oligemia (low cerebral blood flow) or cerebral hyperemia (high cerebral blood flow) alone, but rather the result of an interplay of complex pathophysiological processes involving the brain (9 –13). The formation of vasogenic edema may have several additive components such that only a small proportion of patients with DKA progresses to clinical CE. We reasoned that vasogenic edema would be worsened when blood– brain barrier (BBB) integrity was compromised. Although microvascular complications are well described in diabetes and BBB dysfunction and increased BBB permeability lead to CE in animal studies (9), no human studies characterize the BBB in DKA. We hypothesized that BBB permeability is high early in the course of DKA. The aim of this study, therefore, was to examine BBB permeability, during illness and recovery, as an indicator of BBB integrity in DKA.