Three types of diabetes-related biomedical variables have been linked to the appearance of neurocognitive anomalies in children with diabetes: moderately severe episodes of hypoglycemia, ketoacidosis, and chronic hyperglycemia. our understanding of these associations remains imperfect, unfortunately, because so few studies have adequately ascertained those biomedical variables in pediatric samples. Rather than capturing most, or even a representative number, of those events over the course of the diabetic child's disease, the best investigators have been able to do is to estimate metabolic control from one or a handful of glycosylated hemoglobin values, count severe hypoglycemic episodes retrospectively (missing virtually all episodes of nocturnal hypoglycemia), and rely on often incomplete medical records or parents' delayed recall to quantify the number, duration, and severity of episodes of ketoacidosis.
Animal research (86) and a series of dramatic case reports (77, 87-92) have provided compelling support for the argument that under certain circumstances, hypoglycemia is capable of producing irreversible structural and functional damage to the cNs, although the exact pathophysiological processes remain incompletely understood (93). Yet in virtually all of those instances, the degree of hypoglycemia was "profound," i.e., blood glucose levels typically fell below 1.5mmol/l (27mg/dl), persisted for an extended period of time, and were accompanied by coma or seizure, or an essentially flat (isoelectric) EEG. Moreover, most of the diabetic patients described in these case reports were older adults who may have also had some other conditions that could affect brain integrity (e.g., history of chronic alcohol abuse) or who may have manifested wildly fluctuating blood glucose levels (e.g., history of "brittle" diabetes). Although this form of profound hypo-glycemia can lead to extraordinary brain morbidity, it seems to be atypical, occurring remarkably infrequently in diabetic patients as a group, with even fewer published reports of such events occurring in children and adolescents.
On the other hand, "severe" hypoglycemic events, defined as a blood glucose value below 3.8mmol/l (70mg/dl) with loss of consciousness or seizure, and "moderately severe" hypoglycemic events, defined as a blood glucose value below 3.0mmol/l (55mg/dl) without loss of consciousness or seizure (43), are far more common in diabetic patients of any age (see Chapter 6). Rates of severe hypoglycemia in children and adolescents are approximately 20 per 100 patient years, although the exact value varies somewhat, depending on type of diabetes management regimen and level of glycemic control (54).
Poorer neurocognitive outcomes have been associated with recurrent episodes of severe hypoglycemia in several small, cross-sectional pediatric studies (51, 64, 70, 94, 95), but as noted above, these effects were often moderated by age at onset. That is, children with an earlier age of diabetes onset and a history of severe hypoglycemia were most likely to perform poorly on a limited number of cognitive tests or show more EEG abnormalities, whereas those with a later age at onset were less likely to show deficits, regardless of the presence or absence of severe hypoglycemia. Other studies of children, adolescents, and young adults have either found no robust relationship between severe hypoglycemic events and multiple measures of brain function (14, 96-99) or have been unable to accurately attribute neu-rocognitive dysfunction to severe hypoglycemia in subjects who may have had elevated (or highly variable) blood glucose values over an extended time period (13, 40, 60).
Null results from a subgroup analysis of the Diabetes Control and Complication Trial (DCCT) study cohort provide the most compelling evidence to date that severe hypoglycemia does not lead to readily detectable cognitive dysfunction in adolescents or young adults (16). Subjects who enrolled in the DCCT as adolescents (13-19 years of age) were followed over a period of approximately 18 years and were reassessed repeatedly with a comprehensive battery of neuropsychological tests as well as with multiple detailed biomedical measures, including careful prospective ascertainment of severe hypoglycemia (defined as seizure or coma). Of the 249 adolescents who entered the DCCT, 175 (76% of surviving eligible subjects) completed the follow-up cognitive test battery, with half the subjects (N = 88) never having an episode of severe hypoglycemia while the remainder experiencing one or more events. Despite a high incidence of severe hypoglycemia (N = 249 episodes), there was no relationship between cumulative number of hypoglycemic events and change in cognitive functioning over time within any of the eight cognitive domains assessed. The major limitation of this study is its focus on adolescence - a period when most brain development has already occurred (100). Whether the brain of the younger diabetic child would be more vulnerable to the effects of severe hypoglycemia is an issue that has not yet been settled, although similar null results have been reported from two smaller clinical studies that carefully ascertained hypoglycemic events and measured cognitive function in diabetic children who were 6-15 years of age (98, 99).
Diabetic ketoacidosis (DKA) results from an absolute or relative deficiency in circulating insulin, combined with increased levels of counter-regulatory hormones. The resulting acceleration in catabolism is accompanied by increased glucose production in the kidney and liver and reduced peripheral glucose utilization which leads to hyperglycemia, hyperosmolar-ity, ketogenesis, and metabolic acidosis (101) (see Chapter 7). DKA occurs frequently during the onset of diabetes, with rates ranging from 37 (for those diagnosed before 4 years of age) to 15% in older adolescents (102). For a very small proportion of children (<1%), DKA eventuates into clinically significant cerebral edema, although perhaps as many as half of the non-symptomatic children may manifest evidence of subclinical brain swelling, characterized by a narrowing of the lateral ventricles (103).
The effects of DKA on neurocognitive function have been studied only infrequently, with the focus generally being on acute changes in brain electrical activity, as indexed by EEG recording (69), and brain tissue integrity, as measured by diffusion-weighted MRI (104). Most of the studies previously described in this chapter have either not collected information on DKA or have not incorporated that information into their analyses (13-15, 37, 49). The few reports linking cognitive dysfunction to DKA have tended to be secondary findings; to the best of our knowledge, no neurocognitive study has been explicitly designed to explicitly compare outcomes in individuals with and without a clearly documented history of DKA. Nevertheless, several casual observations suggest possible links between DKA and changes in the CNS measured several years after the event. For example,
26 of the adolescents participating in the DCCT experienced one or more episodes of DKA and subsequently showed declines over the 18-year follow-up period on measures of learning; in contrast, those with no history of DKA showed improvements over time (16). Smaller studies have also reported that children with DKA were more likely to manifest reductions in cerebral blood flow (70) and earn somewhat lower scores on certain cognitive tests, with the magnitude and pattern of results varying depending on the age when DKA occurred (18). Those latter observations led the authors to speculate that DKA-associated cognitive dysfunction may require an extended period of time to develop, with specific manifestations being dependent on the stage of brain maturation and the sensitivity of different brain substrates to metabolic insults. These very thoughtful speculations clearly require additional study.
Chronic Hyperglycemia A very large body of research on adults with diabetes now demonstrates that the risk of developing a wide range of neurocognitive changes - poorer cognitive function, slower neural functioning, abnormalities in cerebral blood flow and brain metabolites, and reductions or alterations in gray-and white-brain matter - is associated with chronically elevated blood glucose values and the occurrence of clinically significant microvascular and macrovascular diabetic complications [for reviews see (105,106) and Chapter 11]. In contrast, no pediatric neurocognitive studies have yet evaluated the impact of comorbid medical conditions like elevated blood pressure or early microvascular changes like background retinopathy or subclinical peripheral neuropathy, despite the fact that these are evident in children within 5 years of diagnosis (107) and have been found to predict cognitive dysfunction in adults with and without diabetes (108-110). On the other hand, multiple small studies have examined the relationship between HbA1c values and neurocognitive measures in children. In some instances chronic hyperglycemia has been operationalized by averaging several HbA1c values collected over an extended period of time (60); in other studies, composite estimates of chronic hyperglycemia have been generated by calculating the percentage of time from diagnosis that the child exceeded a "poor control" threshold [HbA1c > 9.5%] (13) or by adding the z-score of median lifetime HbA1c values to the z-score of diabetes duration (14). This approach has often [but not invariably - see (37, 95)] revealed statistically reliable relationships between exposure to hyperglycemia and neural slowing (59, 60), structural changes in brain gray and white matter (75), and poorer performance on at least a subset of neuropsychological tests (13-15, 111), although the magnitude of these effects may be moderated by age at onset, i.e., the relationships are strongest in those with an early onset of diabetes (15).
Animal research also supports the position that chronic hyperglycemia, as induced by streptozotocin (STZ), can adversely affect the integrity of the CNS (see Chapters 16 and 17). Compared to healthy controls, young adult diabetic rats showed slowed neural processing (longer latencies of auditory- and visual-evoked potentials) after 3-4 months of diabetes. As duration increased further, there was a corresponding increase in evoked potential latencies which was subsequently reversed following the initiation of insulin therapy (112). Spatial learning skills, assessed in a water maze, were also impaired, with the magnitude of those learning deficits linked to reductions in hippocampal plasticity (113).
Very young rats that experienced STZ-induced hyperglycemia between 4 and 8 weeks of age showed especially notable structural cNS abnormalities. Within the cortex of the hyperglycemic rats, neurons were more closely packed together, were smaller than normal, and had less myelin, as indicated by reductions in the amount of protein, fatty acids, and cholesterol. Similar results were seen in the hippocampus, where the density of both astrocytes and neurons was greater, with many more smaller neurons. In contrast, recurrent bouts of moderately severe hypoglycemia had minimal effects on brain structure (114). More recently, this group has identified changes in the neuronal structure that were characterized by reductions in dendritic branching and spine density (115). These structural abnormalities were accompanied by increased levels of sorbitol (indicative of alterations in the activity of the cerebral polyol pathway) and decreases in taurine (thought to serve as a trophic factor for normal neuronal development). While the hyperglycemic rats showed normal learning performance on a water maze task, their delayed recall on that task was significantly impaired. Again, these changes were limited to animals made diabetic with STZ at 4 weeks of age; recurrent bouts of hypoglycemia had no impact on cNS structure or function.
Taken together, the limited animal research on this topic complements the extant pediatric work and provides quite compelling support for the view that even relatively brief bouts of chronically elevated blood glucose values can induce structural and functional changes to the brain. Rodent models have been used previously to study the impact of diabetes on the cNS in adults (116). The very elegant work by Malone and his associates (114, 115) is some of the first to explore the effects of diabetes and glucose fluctuations in the developing organism and demonstrates the value of animal models in searching for the pathophysiological basis for brain dysfunction in children with diabetes.
In addition to the biomedical risk factors described above, one needs to entertain the possibility that some of the neurocognitive problems noted in children may be exacerbated (or obscured) by the occurrence of anxiety and mood disorders - most typically depression and anxiety - that are often (33, 35,117) although not invariably (118) seen in children and adolescents with diabetes. A growing literature on nondiabetic children has demonstrated that depression is associated with reductions in whole brain volumes, with changes particularly apparent in the frontal lobes (119), the amygdala (120), and the basal ganglia (121); whether the hippocampus is affected remains unresolved (120, 122). Generalized anxiety disorders are also associated with structural changes, with one study reporting increases in white- and gray-matter volumes in the superior temporal gyrus (123). It is noteworthy that young adults with a childhood onset of diabetes manifest changes in this same brain region, although they are more likely to show decreases in gray-matter density, rather than increases (76). Why this region appears to be sensitive to anxiety and diabetes remains unknown. Magnetic resonance spectroscopy studies have identified changes in brain metabolites (124), particularly brain choline levels, suggestive of diminished cell growth or myeli-nation in the left dorsolateral frontal cortex (125), although there is no complete agreement in this regard (126). Although research has begun on the possible synergistic effects of depression and diabetes on CNS integrity in adults with type 2 diabetes (127), similar studies have not yet been initiated in diabetic children.
Was this article helpful?