Documente Academic
Documente Profesional
Documente Cultură
Type 1 Diabetes
Simona Ghetti, PhD, Joshua K. Lee, BA, Clare E. Sims, BA, Dana M. DeMaster, BA, and Nicole S. Glaser, MD
Objective We tested the hypothesis that diabetic ketoacidosis (DKA) results in memory deficits typical of hypoxic/
ischemic injury because recent studies suggest that cerebral metabolic changes similar to those observed in hypoxic/ischemic cerebral injury are observed in children with DKA, even without symptoms suggesting cerebral injury.
Study design Thirty-three children with type 1 diabetes mellitus (T1DM) and a history of DKA and 29 children with
T1DM without a history of DKA were enrolled from an academic hospital pediatric endocrinology clinic. These
groups were comparable on demographic and disease-related variables. These groups ability to recall events in
association with specific details, the memory function most directly affected by mild hypoxia/ischemia, was compared on 2 tasks (ie, event-color associations and eventspatial position associations).
Results In multivariate analyses controlling for other critical variables, children with DKA history had significantly
lower rates of accurate memory on both tasks (mean, 0.34 0.13 on the color task and 0.57 0.15 on the spatial
task) than did children without DKA history (mean, 0.44 0.11 and 0.65 0.18, P < .01).
Conclusions DKA disrupts memory function, underscoring the importance of DKA prevention when T1DM is
known and prompt diagnosis of children with new onset of T1DM. (J Pediatr 2010;156:109-14).
hildren with type 1 diabetes mellitus (T1DM) are typically average in overall intellective abilities1,2 but exhibit deficits
on neuropsychological measures of memory.3-7 The origin of these deficits is not fully understood, in part because specific complications of T1DM, rather than T1DM per se, may place subsets of patients at increased risk. The effect of
complications of T1DM (eg, hypoglycemia) on memory has been evaluated,3-7 but the effect of diabetic ketoacidosis (DKA)
has not been reported.
DKA is a state of metabolic decompensation with high blood glucose concentrations and hepatic ketone production resulting in
acidosis.8 Twenty-five to forty percent of children with new onset of T1DM may present with DKA, and in established patients,
DKA may be the consequence of illness, poor compliance, or malfunction of diabetes care equipment (eg, insulin pumps). Newer
data suggest that DKA may cause subtle brain injury.9 Cerebral edema associated with severe disturbances in neurological function
and high mortality rates has long been recognized as a rare but serious complication of DKA in children.10,11 Overt manifestations
of cerebral edema may represent only the most severe cerebral injury caused by DKA. Subtle cerebral injury may occur in the
majority of children with DKA, even among those with minimal or no alterations in mental status.9,12-15 Further, studies using
magnetic resonance (MR) imaging in children and in an animal model demonstrate changes in cerebral water distribution, blood
flow, and metabolism during DKA similar to those often observed in hypoxic/ischemic cerebral injury.9,15,16
Cerebral hypoxia or ischemia associated with other conditions (eg, high-altitude exposure, cardiac arrest, and sleep-related
breathing disorders in childhood) has been linked to memory deficits.17-20 Research with animal models confirms a causal relation between mild hypoxia or ischemia and specific damage to the hippocampus,21 a brain structure that is responsible for
formation and retrieval of memories of the association between an event and the context in which it originally occurred.22,23 We
studied whether DKA is associated with such specific memory deficits in children with T1DM. We examined the effects of DKA
on memory for item-context associations while accounting for other variables that are related with memory impairment in
children with T1DM.3-7 We hypothesized that if DKA results in mild hypoxia or ischemia, then specific deficits in memory
functions mediated by the hippocampus should be expected.
Methods
Patients with T1DM between the ages of 7 and 16 years were recruited from the
pediatric endocrinology clinic at the University of California, Davis, Medical
ANCOVA
DKA
GCS
MR
T1DM
Analysis of covariance
Diabetic ketoacidosis
Glasgow Coma Scale
Magnetic resonance
Type 1 diabetes mellitus
109
www.jpeds.com
Center. The study was approved by the University of California Davis Internal Review Board, and written consent was obtained from the parents or guardians. Patients were eligible
for participation if they did not have a history of attention
and learning disorders (eg, disorders such as attention
deficit hyperactivity disorder and dyslexia, which are not associated with hypoxic/ischemic injury), psychiatric conditions, history of head trauma, premature delivery (<36
weeks of gestation), low birth weight (<5.5 lb), color blindness, or were not English speaking. The patients blood glucose concentration was required to be within 70 to 300 mg/
dL just before participating. Participants were excluded
from analysis if their glucose concentration fell outside of
the desired range or if IQ scores were 2 standard deviations
below the mean.
After obtaining consent and blood glucose testing the
experimental procedure began. The experimenter was blinded
to DKA history at the time of testing. We included the 2 memory tasks described below to investigate memory deficits across
different types of contextual features. After the memory tasks,
a 5- to-10-minute break preceded the administration of the
Wechsler Abbreviated Scale of Intelligence, described below.
While participants were tested, their parents or guardians
provided demographic and disease-related information on
a questionnaire, which was integrated with information
from the medical records. Early onset of T1DM was defined
as diagnosis before age 7 years because children diagnosed
with diabetes within the first 6 years of life have been found
to have more extensive cognitive deficits than do children
who have diabetes later in life.3,7 A history of severe hypoglycemia (defined as hypoglycemia associated with loss of consciousness, seizure, or other severe alteration in mental
status) was obtained from the parent or guardian as well as
by review of the childs medical record. A history of DKA
was obtained by review of the medical record and by questioning the parent or guardian about any hospitalizations or
emergency department visits for DKA occurring at other facilities. Medical records were reviewed to determine whether
any child had clinically apparent DKA-related cerebral edema
(defined as deterioration in mental status during DKA with
radiographic evidence of cerebral edema and/or requiring
treatment with mannitol or hypertonic saline).
Hemoglobin A1c was measured on the day of participation
using a Bayer DCA 2000 HbA1c analyzer (Block Scientific,
Nutley, New Jersey). As an additional comparison of overall
blood glucose control among patients in each group, we
determined the percentage of glucose readings from the
patients meter either below 70 mg/dL or above 300 mg/dL
during the month before participation.
Memory Measures
Versions of published tasks were used.24 In the color task,
80 black-ink drawings on a white, square background were
presented on a computer screen with either a green, red, yellow, or blue border (20 of each color in a random order), and
participants were instructed to try to remember both the item
110
ORIGINAL ARTICLES
January 2010
Characteristics
Female participants, No. (%)
Age at testing, mean (SD), y
Family income, mean (SD), $1000
Early onset (before age 7), No. (%)
Severe hypoglycemia, No. (%)
HbA1c, mean % (SD)
Frequency of high and low glucose, % (SD)*
Frequency of ketosis resolved at home, No (%)
Glucose before testing, mg/dL(SD)
Total IQ, mean score (SD)
Verbal IQ, mean score (SD)
Performance IQ, mean score (SD)
Color task
d 0 , mean (SD)z
Spatial task
d 0 , mean (SD)z
Children with
DKA history
(n = 33)
Total
(n = 62)
P
value
20 (55%)
12.4 (3.3)
65.9 (27.4)
12 (36%)
14 (42%)
8.1 (1.4)
27 (13%)
5 (17%)
199 (75)
106.0 (10.6)
104.5 (11.8)
106.6 (12.4)
16 (61%)
12.5 (2.4)
61.3 (29.2)
13 (45%)
9 (31%)
8.1 (1.7)
28 (16%)
7 (21%)
181 (79)
109.1 (11.0)
109.4 (13.2)
106.4 (9.4)
36 (58%)
12.4 (2.9)
63.53 (28.2)
25 (40%)
23 (37%)
8.1 (1.6)
.28 (15%)
12 (19%)
190 (77)
107.5 (10.8)
106.8 (12.6)
106.5 (11.0)
.80
.83
.53
.61
.43
.90
.86
.69
.37
.26
.13
.94
1.97 (.67)
1.96 (.79)
1.97 (.72)
.97
2.06 (.78)
2.34 (.79)
2.19 (.79)
.16
*This index of glucose variability was calculated as the percentage of all measured glucose values below 70 or above 300 mg/dL during the month participation.
This index refers to the number of patients with episodes of ketosis treated at home in consultation with the childs physician.
zd 0 is the normalized difference between hit rates and false-alarm rates; it may range from 0 to 4; higher scores indicate better performance.
Results
Sixty-six children with T1DM enrolled in the study; 3 were
excluded from the analyses because their pretest blood glucose concentrations were above the specified range (n = 2)
or because the IQ score was more then 2 standard deviations below the mean. One additional patient was excluded
because his parents elected to end participation before
completion of the memory tasks. Descriptive data for the
remaining 62 children who comprised the study sample
are presented in the Table. None of the children had clinically apparent cerebral edema. None of the participants exhibited clinically relevant mental status deterioration as
measured by the Glasgow Coma Scale (GCS) scores obtained roughly hourly during DKA treatment. A few patients had minimal abnormalities in GCS scores. One
patient had a decline in GCS score to 13, and 5 patients
had GCS scores decline to 14. GCS information was missing for 3 patients, but there was no indication in their clinical records that would suggest concern about clinically
apparent cerebral edema or brain injury.
Thirty-seven additional eligible patients declined to participate in the study. These patients did not significantly differ
from the study sample in mean age at first contact (12.2
2.4 years; P = .61), sex (62% females, P = .58), frequency of
early T1DM onset (35% early onset, P = .52), HbA1c (8.4
1.9, P = .30), frequency of DKA (36% had a history of
DKA, P = .14), and severe hypoglycemic episodes recorded
in the medical record (14% versus 22% in patients who participated, P = .30; this comparison is based on the information included in the medical records alone, given that
parental reports were not available for patients who declined
to participate).
Children with and without a history of DKA did not significantly differ in any demographic or T1DM-related variables
111
www.jpeds.com
Figure 2. Significant interaction effect on rates of item-context associations between DKA history, early T1DM onset,
and sex [F (1, 52) = 6.07, P < .05]: Males who had DKA and an
early T1DM onset had disproportionately lower memory
scores.
with late onset, those who at the time of testing were 12 years
of age or older seemed resilient to the effects of DKA. Sex was
a significant covariate [F (1, 52) = 13.99, P < .001].
In the third ANCOVA analysis examining main and interactive effects of DKA, early T1DM onset, and hypoglycemia
adjusting for age at testing and sex, the effect of DKA was
confirmed to be significant [F (1, 52) = 8.74, P < .01]. Although a main effect of severe hypoglycemia was not found,
there was a significant interaction with age at T1DM onset,
such that lower performance was observed in participants
who had a history of hypoglycemia and early T1DM onset
[F (1, 52) = 4.15, P < .05]. Age and sex were significant covariates [Fs (1, 52) $ 7.76, Ps # .01].
All of the main and interactive effects of DKA reported
above retained significance even after controlling for IQ
[Fs (1, 51) $ 5.56, Ps < .05], HbA1c [Fs (1, 51) $ 5.93,
Ps < .05], and glucose levels at time of testing [Fs (1, 51) $
4.69, Ps < .05]. These effects were also retained when the
DKA group was modified to include only patients who
had 1 DKA episode that occurred at the time of diagnosis
[Fs (1, 42) $ 4.29, Ps < .05].
Finally, we examined the correlation between time in years
since DKA (average 2.54 years and varied from 0.11 to 14.54
years) and current memory for item-context associations in
patients who only had 1 DKA episode at diagnosis. After controlling for current age, we found a significant correlation
(r = .37, P < .05), such that memory performance was
worse in children whose DKA was in the more distant past.
Discussion
Figure 1. Rates of item-context associations as a function of
DKA history and type of task: Children with DKA history had
significantly lower rates of accurate memory on both tasks
than did children without DKA history [F (1, 52) = 11.30,
P < .001].
112
The ability to remember specific events along with information about the context in which these events occurred is a fundamental capacity, supporting mundane yet crucial abilities
as well as more complex senses such as continuity of self.
Our study demonstrated that this capacity is significantly
Ghetti et al
ORIGINAL ARTICLES
January 2010
Figure 3. Significant interaction effect on rates of item-context associations between DKA history, early T1DM, and age
at testing [F (1, 52) = 7.00, P = .01]. Performance was most
strongly affected by DKA in children with early diabetes onset;
children with late onset and who at the time of testing were 12
years of age or older appeared protected from the effects of
DKA.
the most frequent abnormality, again suggesting hippocampal involvement. Although the authors did not find an association between mesial temporal sclerosis and DKA episodes,
their study included relatively few participants with DKA,
and the investigators recorded only moderate or severe
DKA episodes rather than all episodes of DKA. The lack of
association between mesial temporal sclerosis and DKA
may therefore reflect these methodological and sample size
differences from our report.
Mild cerebral edema is present in at least 50% of children
during DKA treatment.12 These findings are observed consistently despite use of different imaging techniques including MR imaging, CT, or MR diffusion weighted imaging.14,15
Alterations in levels of cerebral metabolites typically observed during hypoxic/ischemic cerebral injury have also
been demonstrated in children with DKA evaluated with
MR spectroscopy during treatment.9 Measurements of apparent diffusion coefficient values using MR diffusion
weighted imaging suggest that untreated DKA is associated
with cytotoxic cerebral edema (cell swelling) but that vasogenic cerebral edema, with increased extracellular fluid accumulation, develops during DKA treatment.15,29 These data
imply not only that untreated DKA is associated with cerebral injury but that further cerebral injury may result during
DKA treatment, as cerebral tissues are reperfused. The current data are consistent with these studies and with the occurrence of subtle but permanent cerebral injury in
children with DKA. Future research should examine directly
how brain injury mediates the relationship between DKA
and memory deficits.
In the current study, several variables other than DKA were
associated with differences in memory function. We found
that male sex, young age at onset of diabetes, and severe hypoglycemia were associated with reduced memory performance either directly or in interaction. Previous studies
have similarly reported that males are more likely than females to show detrimental effects of T1DM on intellectual
development and memory specifically,6 suggesting sex-specific vulnerabilities with regard to the consequences of diabetes for cognitive development. Earlier studies also suggest
that children who are diagnosed with T1DM earlier in life exhibit more extensive cognitive deficits (in memory, attention,
and executive functioning) than do children who had diabetes later in childhood.3,7 Our results further suggest that DKA
in early age may be particularly detrimental for memory
function. Finally, severe hypoglycemia has also been related
to damage to the hippocampus through an excitotoxic process.3,5 Although we did not find an overall effect of hypoglycemia on memory function, our data suggest that a subgroup
of children diagnosed with T1DM at an early age may be
more vulnerable to hypoglycemic injury.
The current study has some limitations. First, we relied on
medical record reviews and parents or guardians recollections of DKA episodes and hypoglycemic events to classify patients according to these variables. It is possible that some
episodes of DKA or severe hypoglycemia may have been
missed because the child received treatment at an outside
113
www.jpeds.com
References
1. Kovacs M, Goldston D, Lyengar S. Intellectual development and academic performance of children with insulin-dependent diabetes mellitus: a longitudinal study. Dev Psychol 1992;28:676-84.
2. Hagen J, Barclay CR, Anderson BJ, Feeman DJ, Segal SS, Bacon G, et al.
Intellective functioning and strategy use in children with insulin-dependent diabetes mellitus. Child Dev 1990;61:1714-27.
3. Hershey T, Perantie DC, Warren SL, Zimmerman EC, Sadler M,
White NH. Frequency and timing of severe hypoglycemia affects spatial
memory in children with type 1 diabetes. Diabetes Care 2005;28:
2372-7.
4. Perantie DC, Lim A, Wu J, Weaver P, Warren SL, Sadler M, et al. Effects
of prior hypoglycemia and hyperglycemia on cognition in children with
type 1 diabetes mellitus. Pediatr Diabetes 2008;9:87-95.
5. Hershey T, Lillie R, Sadler M, White NH. A prospective study of severe hypoglycemia and long-term spatial memory in children with
type 1 diabetes. Pediatr Diabetes 2004;5:63-71.
6. Fox MA, Chen RS, Holmes CS. Gender differences in memory and learning in children with insulin-dependent diabetes mellitus (IDDM) over
a 4-year follow-up interval. J Pediatr Psychol 2003;28:569-78.
114
Ghetti et al