The Impact of Acute Hypoglycemia on Neuropsychological and Neurometabolite Profiles in Children With Type 1 Diabetes

Case 1 was a 7-year-old girl with a 2-year history of type 1 diabetes who presented to the emergency department, Royal Children’s Hospital, following a severe hypoglycemic event with seizure. She awoke with a blood glucose level of 3.6 mmol/l. This was not treated, and she was given her usual morning dose of short- and intermediate-acting insulin. She had a seizure within 20 min of insulin administration. The seizure resolved within 5 min after treatment with glucagon and glucose gel. Her most recent HbA_1c (A1C) was 7.5%.

Case 2 was a 9-year-old girl with a 6-year history of type 1 diabetes who experienced a hypoglycemic seizure immediately before breakfast after a nocturnal supplemental dose of rapid-acting insulin. She was treated with glucagon, with seizure resolution within 5 min. Her blood glucose level immediately after glucagon administration was 2.4 mmol/l. Her most recent A1C was 8.7%.

Case 3 was a 10-year-old boy with a 3-year history of type 1 diabetes who felt unwell for several days and experienced a hypoglycemic seizure during an afternoon nap. His blood glucose level was 2.9 mmol/l. Oral hypoglycemic agents were unsuccessful, and he was consequently taken to the Royal Children’s Hospital and treated with intravenous dextrose. The duration of his seizure was unknown. His most recent A1C was 7.8%.

Families of participants were informed about the study, and consent was obtained before study commencement. Ethical approval was obtained from the human ethics research committee at the Royal Children’s Hospital. All children were euglycemic before neuroimaging and neuropsychological assessment.

Tissue metabolite profiles were assessed with single-voxel magnetic resonance spectroscopy in 2-cm isotropic regions of interest (ROIs). The ROIs were chosen based on their known sensitivity to diabetes-related variables, including hypoglycemia (2). Bilateral temporal lobe ROIs were selected in the coronal plane with maximal inclusion of the high-intensity area in both the superior temporal gyrus seen in the structural scout image and the medial temporal lobe (centered ∼45 mm from the temporal pole). Basal ganglia and frontal lobe spectra were prescribed from axial images. Spectra were recorded at 3T using short-echo point-resolved spectroscopy using 128 transients, 30 ms echo time, 3,000 ms relaxation delay with a standard birdcage head coil. Neurometabolites measured included N-acetyl aspartate (NAA), associated with neuronal integrity; myoinositol, associated with gliosis and osmotic regulation; trimethylamines (TMAs), markers of membrane turnover; and glutamate plus glutamine, involved in neurotransmitter cycling and osmoregulation.

The neuropsychological assessments at baseline and follow-up were designed to measure verbal abilities (3,4), visuo-spatial reasoning (3,4), short- and long-term memory (5,6), attention (7), and mental flexibility (8,9). These constructs were examined, as they are thought to be mediated by frontal and mesio-temporal regions, areas known to be sensitive to diabetes-related variables, and have been assessed in previous work (10,11) investigating the impact of diabetes on brain function.

The most consistent finding on neuropsychological measures was the performance improvement on selective attention over time (Table 1). All case subjects showed below age-expected performance at baseline, and all showed significant improvement at follow-up. However, there was variability between case subjects on other neuropsychological measures, indicating that severe hypoglycemia impacts on different skills in each patient. For example, in addition to improvements in attentional processes, case 1 improved on tasks of long-term memory and case 2 improved on visuospatial reasoning and mental flexibility tasks over time. Interestingly, compared with baseline, case 3 performed more poorly at the follow-up assessment on measures of vocabulary, visuospatial reasoning, and short-term memory.

Neurometabolite profiles also showed variability between case subjects and brain regions, suggesting that hypoglycemia differentially affected neuronal integrity, gliosis, and membrane turnover in each ROI for each case subject (Table 2). Total NAA in frontal lobe showed a baseline concentration that was somewhat reduced compared with control data and returned toward normal levels at follow-up for cases 1 and 2. However, case 3 showed a reduction in NAA concentration over time. Similarly, in the temporal lobe, compared with control data, cases 1 and 2 had elevated baseline TMA concentrations with some recovery at follow-up, whereas levels for case 3 were unchanged over time. Myoinositol concentrations did not differ from expected levels in any ROI for cases 1 and 2, but case 3 showed a higher baseline concentration in the frontal lobe. Glutamate plus glutamine concentrations were comparable with control data at baseline and follow-up in the frontal lobe and temporal lobe for all case subjects. However, in the basal ganglia, case 3 had elevated concentrations at baseline and reduced concentration at follow-up compared with control data.

This is the first study to show that acute hypoglycemia has transient affects on neurometabolite and neuropsychological profiles in diabetic children. In the frontal lobe, NAA levels were reduced following the hypoglycemic seizure but returned to more normal levels 6 months later. This “recovery” accords with Perros et al. (12) who reported that NAA concentrations did not differ between euglycemic diabetic patients with or without a history of severe hypoglycemia. Unlike the frontal lobe, temporal lobe and basal ganglia NAA changes were minimal, suggesting that neuronal integrity in anterior brain regions is particularly susceptible to acute hypoglycemia. Similar findings have been reported (13,14) for event-related potentials and regional cerebral blood flow studies that indicate that frontal regions are activated during acute hypoglycemia. Frontal brain regions are thought to be involved in higher-order cognitive processing, including attention. Given the improvement in attentional processes over time in all patients, our data support the notion that frontally mediated processes are highly susceptible to very-low blood glucose levels.

Results from the current study also suggest that temporal lobes are affected by hypoglycemia. TMA levels were somewhat elevated immediately after the seizure but returned to more normal levels by 6 months. Neuropsychological findings from case 1 also support this, with an improvement in memory skills 6 months postseizure. Pathological studies (15) in humans and animals indicate that hypoglycemia-induced cell death preferentially occurs in temporal lobe regions. Thus, the increased levels of TMA recorded immediately following the seizures may reflect neurodegeneration.

The findings from this pilot study indicate that the frontal lobe, temporal lobe, and basal ganglia are somewhat affected by severe hypoglycemia, albeit transiently. It is important to note that the effect on these regions differed between case subjects, highlighting the variable impact of hypoglycemia on brain. It should also be noted that the changes in neurometabolites and neuropsychological performance over time were generally subtle. It is unlikely that neurometabolite changes reflect developmental processes, as there is very little change in metabolite profiles in healthy children from 12 months of age (16). The improvement in cognitive ability over time exceeds that expected by practice effects, with the test-retest reliabilities of the measures ranging from 0.71 to 0.85 for a test-retest interval much shorter than used in the present study (maximum of 12 weeks). Therefore, these subtle changes should be considered clinically significant, as even mild decrements in children still acquiring new skills and knowledge are important. The heterogeneity of findings also raise the possibility that protective factors may minimize the impact of hypoglycemia on brain. For example, there is some evidence that various nonglucose metabolic fuels are able to support cognitive function during hypoglycemia (17). In addition, avoidance of even mild hypoglycemia can restore/improve subjective hypoglycemia awareness, and, more importantly, the hierarchy of hormonal and symptomatic responses that occur before cognitive dysfunction is apparent (17). These factors may contribute to individual variation in response to hypoglycemia.

In summary, it is difficult to ascertain the exact impact of a hypoglycemic seizure on brain given the limited sample and lack of premorbid functioning. Nevertheless, preliminary findings suggest that acute hypoglycemia has transient effects on neurometabolites and function of the frontal lobe, temporal lobe, and basal ganglia. Larger samples and longer-duration follow-up may elucidate the long-term implications of these findings.