Distinct attentional profiles associated with diabetic ketoacidosis and hypoglycemia in pediatric type 1 diabetes: MOXO-CPT assessment

Abstract

Introduction

Type 1 diabetes (T1D) in children and adolescents has been associated with alterations in attentional and executive functioning. Acute metabolic complications, particularly diabetic ketoacidosis (DKA) and clinically significant hypoglycemia (CSH), may adversely affect neurocognitive development. This study evaluated the associations between acute metabolic events and attentional functioning in pediatric patients with T1D.

Materials and methods

The study included 117 children and adolescents with T1D (55 females, 62 males; mean age 12.86 ± 3.16 years). Participants were stratified according to the presence of DKA and CSH. Cognitive functioning was assessed using the MOXO Continuous Performance Test (MOXO-CPT), which evaluates sustained attention, timeliness, impulsivity, and hyperactivity. Associations between cognitive performance and clinical variables were analyzed using ANOVA, ANCOVA, and multivariable regression models.

Results

Patients with a history of CSH demonstrated significantly lower sustained attention (p < 0.001) and timeliness scores (p < 0.001) compared with controls. DKA was associated with poorer impulsivity scores (p = 0.008). After adjustment for age, sex, HbA1c, diabetes duration, and treatment modality, the differences in sustained attention and timeliness, whereas the association with impulsivity was attenuated.

Conclusions

Acute metabolic complications in pediatric T1D were associated with domain-specific alterations in attentional functioning. CSH was primarily related to impaired sustained attention and timeliness, whereas DKA was associated mainly with impulsivity. Early identification of attentional difficulties may support individualized diabetes care and improve long-term metabolic safety.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s12020-026-04800-7.

Introduction

Type 1 diabetes (T1D) is one of the most common chronic diseases in childhood and adolescence, with a rising incidence globally [1, 2]. Previous research, including our recently published analysis, has demonstrated that children and adolescents with T1D tend to show lower cognitive performance compared to their healthy peers [3, 4]. Building on these findings, the present study represents a secondary analysis focusing specifically on the associations between acute metabolic complications and domain-specific attentional functioning in children and adolescents living with T1D.

Effective self-management of T1D requires intact cognitive, executive, and adaptive functioning, with attention as a key component of daily disease management.

Attention deficits in youth with T1D have been associated with poor metabolic regulation, recurrent episodes of severe hypoglycemia, and diabetic ketoacidosis (DKA) [5, 6]. During development, cognitive functions mature in parallel with structural brain growth and environmental interaction. However, the developing central nervous system (CNS) is particularly vulnerable to metabolic insults [7, 8].

Neuroimaging and longitudinal studies have shown that repeated severe hypoglycemia can negatively affect attention, working memory, and processing speed [9, 10]. Similarly, even a single episode of DKA may disrupt CNS development, with evidence of altered brain growth trajectories and lower IQ observed in children with early DKA exposure [5, 11].

This study investigates selected domains of cognitive functioning (sustained attention, timeliness, hyperactivity, impulsivity) in children and adolescents with T1D, focusing on the presence of CSH and DKA throughout the course of the disease. However, the specific impact of acute metabolic events on distinct attentional domains remains insufficiently characterized.

Materials and methods

Study design

This study employed an observational, cross-sectional design based on clinical data and standardized neuropsychological assessments. It represents a secondary analysis within a larger research project on neurodevelopmental outcomes in youth with T1D. Written informed consent was obtained from all participants aged ≥ 16 years and their parents or legal guardians. For participants younger than 16 years, written informed consent was obtained from their parents or legal guardians. The study adhered to the principles outlined in the Declaration of Helsinki [12] and was approved by the Bioethics Committee at Poznan University of Medical Sciences (No. 1140/19).

Setting

The study was conducted at the Department of Pediatric Diabetology, Auxology and Obesity at Poznan University of Medical Sciences (Poland). All cognitive assessments were performed in hospital-based testing rooms under standardized conditions during patients’ clinical stays.

Participants’ demographic characteristics and group stratification

A total of 117 children and adolescents with T1D, aged 6–18 years, were recruited consecutively during routine clinical visits. Inclusion criteria were confirmed diagnosis of T1D, age between 6 and 18 years, and clinical stability at the time of assessment. Exclusion criteria comprised hypoglycemia during testing (BGL < 70 mg/dL [< 3.9 mmol/L]), as well as neurological disorders, psychiatric conditions, severe comorbidities requiring intensive treatment, newly diagnosed T1D, and lack of consent to participate in the study.

The final sample included 55 females (47.0%) and 62 males (53.0%). The mean age was 12.86 years (SD = 3.16), and the mean duration of diabetes was 5.64 years (SD = 4.01). The mean HbA1c level was 7.49% (SD = 1.53). The mean blood glucose level during the MOXO test was 151.3 ± 56.6 mg/dL (8.41 ± 3.14 mmol/L). Treatment methods included multiple daily injections (MDI; 43.8%) and continuous subcutaneous insulin infusion (CSII; 46.2%). None of the participants used a hybrid closed-loop (HCL) insulin delivery system. Overall, 39 participants (33.3%) used continuous or flash glucose monitoring (CGM/FGM), whereas 78 (66.7%) performed self-monitoring of blood glucose (SMBG). The participants were recruited from similar urban areas with comparable access to education and healthcare, ensuring relative homogeneity of the study population.

The primary exposure variables were a history of DKA, defined as at least one episode since diagnosis, with the last episode occurring at least six months before study enrollment. CSH, defined as the occurrence of at least one episode per month during disease. DKA was defined according to the criteria established by the International Society for Pediatric and Adolescent Diabetes (ISPAD), as a blood glucose level > 200 mg/dL (11.0 mmol/L), venous pH < 7.30 or serum bicarbonate (HCO₃⁻) < 18 mmol/L, and the presence of ketones in urine or elevated serum ketone concentrations. CSH was defined as a blood glucose level below 54 mg/dL (3.0 mmol/L) [13] corresponding to level 2 hypoglycemia according to the American Diabetes Association (ADA) criteria [14]. Based on these criteria, participants were stratified into four exposure groups, where 25 individuals (21.4%) experienced DKA without CSH (group I), 32 (27.4%) had CSH without DKA (group II), 22 (18.8%) had both conditions (group III), and 38 (32.5%) had neither (group IV).

An exploratory analysis was performed by stratifying participants in Group I according to the number of previous DKA episodes (one, two, or more than four episodes).

To examine whether the frequency of mild hypoglycemia (< 70 mg/ dl, ≥ 54 mg/dl; < 3.9 mmol/l, ≥ 3.0 mmol/L), corresponding to Level 1 (L1) hypoglycemia according to the ADA classification [14], was associated with cognitive performance. No participants experienced ADA Level 3 hypoglycemia, defined as a severe hypoglycemic event requiring assistance from another person.

Group IV, participants with no history of DKA or CSH, served as the reference group for comparisons with Groups I–III. Group IV was further stratified according to the reported frequency of hypoglycemic episodes (daily, 2–3 times per week, 1 per week, sporadically). Figure 1 presents the group distribution of the study population, while Table 1 summarizes the detailed characteristics of the study groups.

Fig. 1.

Participants’ group stratification

Table 1.

Detailed characteristics of the study groups

	Group I (n = 25)	Group II (n = 32)	Group III (n = 22)	Group IV (n = 38)	Total	p
Age, years(median [IQR])	14 [12–16]	12.5 [9.25–15.75]	14 [11–16]	12 [9–16]	13 [10–16]	0.348
Sex
Female, n (%)	12 (48%)	10 (31.3%)	14 (63.6%)	19 (50%)	55 (47%)	0.123
Male, n (%)	13 (52%)	22 (68.8%)	8 (36.4%)	19 (50%)	62 (53%)
Duration of disease
≤ 5 years, n (%)	15 (60%)	18 (56.3%)	10 (45.5%)	19 (50%)	62 (53%)	0.737
> 5 years, n (%)	10 (40%)	14 (43.8%)	12 (54.5%)	19 (50%)	55 (47%)
Treatment method
MDI, n (%)	17 (68%)	18 (56.3%)	13 (59.1%)	15 (39.5%)	63 (53.8%)	0.139
CSII, n (%)	8 (32%)	14 (43.8%)	9 (40.9%)	23 (60.5%)	54 (46.2%)
Glycemic monitoring method
CGM/FGM, n (%)	8 (32%)	11 (34.4%)	5 (22.7%)	15 (39.5%)	39 (33.3%)	0.619
SMBG, n (%)	17 (68%)	21 (65.6%)	17 (77.3%)	23 (60.5%)	78 (66.7%)
L1 hypoglycemia episodes
Daily, n (%)	4 (16%)	7 (21.9%)	11 (50%)	10 (26.3%)	32 (27.4%)	0.109
2-3x week, n (%)	5 (20%)	10 (31.3%)	2 (9.1%)	13 (34.2%)	30 (25.6%)
1x week, n (%)	7 (28%)	8 (25%)	2 (9.1%)	8 (21.1%)	25 (21.4%)
Occasionally, n (%)	9 (36%)	7 (21.9%)	7 (31.8%)	7 (18.4%)	30 (25.6%)
HbA1c %, median [IQR]	7.4 [6.8–8.85]ab	7 [6.3–7.2]ac	7.4 [7.05–9.58]cd	6.8 [6.2–7.58]bd	7.1 [6.6–7.9]	0.005

Note: IQR - interquartile range; SD - standard deviation; MDI - multiple daily injections; CSII - continuous subcutaneous insulin injection, HbA1c – glycated hemoglobin; n-sample size; L1 hypoglicemia episodes, level 1 hypoglycemia episodes according ADA criteria; CGM/FGM, continuous/flash glucose monitoring; SMBG, self-monitoring of blood glucose; P, p-value; a, b, c, d - post hoc differences

Data sources and outcome variables

Data for this study were obtained from multiple sources, including hospital medical records, glucose monitoring devices, structured clinical interviews, and standardized neuropsychological assessments. Medical records provided information on T1D duration, HbA1c levels, treatment method, and documented episodes of DKA. Information on hypoglycemic episodes (frequency and severity) was obtained from glucometer records, continuous glucose monitoring (CGM/FGM) systems when available, and structured patient interviews. In participants using CGM/FGM, hypoglycemic episodes were included only if confirmed by a capillary blood glucose measurement.

Cognitive outcomes were assessed using the MOXO Continuous Performance Test (MOXO-CPT), which yields standardized measures of attentional and executive functioning. The primary outcome variables were the four MOXO-CPT domains: Sustained Attention, Timeliness, Impulsivity, and Hyperactivity, expressed as age- and sex-adjusted Z-scores.

Cognitive assessment

To assess cognitive functioning, with a particular focus on attentional parameters, the MOXO-CPT was selected. This computerized test was developed to identify attentional and executive difficulties while providing an objective and standardized measurement of neuropsychological functions under conditions that closely simulate real-life environments [15]. All tests were conducted under standardized conditions, ensuring procedural consistency and reliability across participants. Cognitive assessments were performed during scheduled outpatient follow-up visits or planned hospital admissions for routine diabetes evaluation and re-education. Immediately before testing, capillary blood glucose was measured using a glucometer, and participants with blood glucose levels < 70 mg/dL were not assessed. In addition, routine venous blood biochemical tests confirmed a stable metabolic state. Thus, cognitive testing was performed only in participants without evidence of acute metabolic decompensation.

Test’s structure

The MOXO-CPT is available in two age-specific versions: for children aged 6–12 years, with a duration of 15 min, and for adolescents and adults aged 13 years and above, with a duration of 18.5 min. The Polish-language version of the MOXO-CPT was used. In the present study, 68 participants (58.1%) completed the version intended for children aged 6–12 years, whereas 49 participants (41.9%) completed the version for participants aged 13 years and older. Before the test, each participant received standardized instructions and completed a brief practice trial to ensure comprehension of task demands and response criteria. The full test consists of eight sequential phases, each varying in the type, intensity, and modality of distractors (visual, auditory, or combined audiovisual), designed to simulate real-world cognitive challenges [16]. The primary task requires participants to respond as quickly and accurately as possible to a predefined visual target stimulus by pressing the spacebar, while ignoring concurrently presented non-target stimuli. These distractors simulate real-life environmental stimuli and allow the assessment of attentional control under interference [15].

Cognitive indicators

The MOXO-CPT provides results across four key domains of executive functioning. Sustained attention measures the ability to consistently detect and respond to relevant stimuli over time. Timeliness indicates temporal accuracy of responses within the appropriate response window. Impulsivity reflects premature or inappropriate responses to non-target stimuli. Hyperactivity evaluates excessive or purposeless motor and/or verbal activity during task performance [15, 16].

Data analysis and interpretation

Test results were interpreted using age- and sex-adjusted normative data, presented as Z-scores. Each cognitive indicator was categorized according to the following scale: Z ≥ 0 is high normative performance, − 0.825 ≤ Z < 0 is standard performance, − 1.65 ≤ Z < − 0.825 is low performance (within the low normative range), Z < − 1.65 indicates functional difficulties. For scores falling below the normative range (Z < − 1.65), the severity level of impairment was further categorized as low, moderate, high, or extreme. Higher Z-scores indicate better cognitive performance across assessed domains [15]. Figure 2 presents an example of the MOXO-CPT attentional profile obtained from a study participant.

Fig. 2.

Comparison MOXO-CPT results to norms (in Z-score). Note. Participant’s profile. Source: Own work based on materials from Neurotech Solutions.

Statistical analysis

Statistical analyses were performed using IBM SPSS Statistics version 30. Continuous variables are presented as mean ± standard deviation (SD), and categorical variables as frequencies and percentages. Normality of distribution was assessed using the Shapiro–Wilk test (Table 2). Differences between groups were evaluated using one-way ANOVA, followed by Tukey post-hoc tests. Analysis of covariance (ANCOVA) was performed separately for each MOXO-CPT domain (Sustained Attention, Timeliness, Hyperactivity, and Impulsivity), with the MOXO-CPT score as the dependent variable, clinical group as the fixed factor, and age, sex, HbA1c level, diabetes duration, and treatment modality as covariates.

Table 2.

Shapiro–Wilk test results for MOXO variable distributions

MOXO Parameters	Min	Max	M	SD	Me	W	P
Sustained attention	-14.98	3.14	-0.72	2.92	0.27	0.69	< 0.001
Timeliness	-5.46	2.52	-0.95	1.76	-0.62	0.96	0.002
Hyperactivity	-15.27	2.14	0.17	2.47	0.88	0.57	< 0.001
Impulsivity	-21.28	2.66	-1.02	3.08	-0.19	0.78	< 0.001

Note: p-p value; M – mean, SD - standard deviation, Min - minimum, Max - maximum, Me-mean; W - Shapiro–Wilk test statistic

Multivariable linear regression models with backward elimination were used to identify independent predictors of cognitive performance. Where appropriate, effect size measures were calculated. For non-parametric comparisons using the Mann–Whitney U test, the effect size was expressed as r, calculated from the Z statistic, and interpreted as follows: r ≈ 0.10 (small effect), r ≈ 0.30 (medium effect), and r ≥ 0.50 (large effect). Statistical significance was set at p < 0.05. Post hoc power analyses were conducted for the four primary MOXO-CPT outcomes, based on the effect sizes observed in the primary one-way ANOVA. The achieved statistical power exceeded 99% for sustained attention (η² = 0.231; Cohen’s f = 0.548) and timeliness (η² = 0.200; Cohen’s f = 0.500), exceeded 90% for impulsivity (η² = 0.099; Cohen’s f = 0.331), and was lower for hyperactivity (η² = 0.048; Cohen’s f = 0.225), consistent with the smaller effect size.

Result

The distribution of MOXO parameters was evaluated using the Shapiro–Wilk test (Table 2). All variables deviated significantly from normal distribution (p < 0.05). The strongest deviations were observed for sustained attention (W = 0.69), hyperactivity (W = 0.57), and impulsivity (W = 0.78). Although the assumptions were not fully met, ANOVA was applied due to its robustness to moderate deviations from normality, particularly in samples of comparable size. Additionally, non-parametric analyses (Kruskal–Wallis and Mann–Whitney U tests) were performed to confirm the robustness of the findings. No significant differences between groups were observed in terms of sex distribution, age, duration of diabetes, or treatment modality (all p > 0.05). HbA1c levels differed significantly among the study groups (F = 4.79, p = 0.0035). Post hoc pairwise analyses demonstrated significantly higher HbA1c levels in Groups I and III than in Groups II and IV (Table 1). The remaining clinical and demographic characteristics did not differ significantly among the four study groups. Therefore, HbA1c was included as a covariate in the subsequent ANCOVA analyses and as an independent variable in the multivariable linear regression analyses to account for its potential confounding effect.

Differences in cognitive performance among the four clinical groups were analyzed using one-way analysis of variance (ANOVA). Non-parametric analyses yielded results consistent with the ANOVA findings. Specifically, significant group differences were found in sustained attention (F(3,113) = 11.29, p < 0.001, η² = 0.23), timeliness (F(3,113) = 9.42, p < 0.001, η² = 0.20), and impulsivity (F(3,113) = 4.12, p = 0.008, η² = 0.10). In contrast, hyperactivity scores did not significantly differ between the groups (F(3,113) = 1.92, p = 0.13). Table 3 presents the comparison of MOXO-CPT parameters between the study groups, while Fig. 3 illustrates these differences using boxplots.

Table 3.

Comparison of MOXO-CPT parameters (Z-score, mean ± SD) between study groups (ANOVA)

MOXO-CPT Parameters	Group I (n = 25)	Group II (n = 32)	Group III (n = 22)	Group IV (n = 38)	F(3,113)	P	η2
	M ± SD
Sustained attention	0.47ab ± 1.27	-2.33ac ± 3.61	-2.05bd ± 3.82	0.64cd ± 0.53	11.29	< 0.001	0.23
Timeliness	-0.3ab ± 1.31	-1.84ac ± 1.87	-1.72bd ± 1.73	-0.18cd ± 1.41	9.42	< 0.001	0.20
Hyperactivity	0.1 ± 2.79	-0.25 ± 3.22	-0.42 ± 2.11	0.91 ± 1.36	1.92	0.130	0.05
Impulsivity	-2.55a ± 4.88	-0.92 ± 2.6	-1.34 ± 2.43	0.09a ± 1.51	4.12	0.008	0.10

Annotation: M – mean, SD – standard deviation, F – ANOVA statistic, η² – size effect, p – p value, a,b, c,d - post hoc differences, Group I – DKA+/CSH−; Group II – DKA−/CSH+; Group III – DKA+/CSH+; Group IV – DKA−/CSH− (reference group).

Fig. 3.

Comparison of MOXO-CPT parameters between study groups. Note: Values are presented as age- and sex-adjusted standardized Z-scores. Boxes represent the interquartile range (IQR), the horizontal line indicates the median, and whiskers extend to the most extreme values within 1.5 × IQR, and circles represent outliers. Group I – DKA+/CSH−; Group II – DKA−/CSH+; Group III – DKA+/CSH+; Group IV – DKA−/CSH− (reference group).

Post-hoc comparisons using Tukey’s test showed that patients with CSH (Group II) had significantly lower sustained attention and timeliness scores compared with the control group (both p < 0.001).

Patients experiencing both clinical conditions (Group III) also demonstrated significantly poorer performance in sustained attention and timeliness compared with the control group (both p < 0.001). No significant differences were observed between Group I (DKA only) and the control group (Group IV) in these domains. For impulsivity, patients in Group I showed significantly lower scores compared with the control group (p = 0.003), indicating poorer performance. No statistically significant differences were observed between groups in the domain of hyperactivity (p > 0.05).

To evaluate DKA and CSH as independent predictors of MOXO-CPT performance, a two-way ANOVA was performed (Fig. 4). DKA was significantly associated with lower Impulsivity scores (F(1,113) = 7.53, p = 0.007, η² = 0.06), whereas CSH was associated with lower Sustained Attention (F(1,113) = 31.37, p < 0.001, η² = 0.22) and Timeliness score (F(1,113) = 26.20, p < 0.001, η² = 0.19). No significant effects of DKA or CSH were observed for hyperactivity. No statistically significant interactions between DKA and CSH were observed, although the interaction effect for Impulsivity yielded a p-value of 0.050 (F(1,113) = 3.92, η² = 0.03).

Fig. 4.

Interaction effects of DKA and CSH on MOXO-CPT parameters (Z-score). Note: Values are presented as standardized z-scores. DKA, diabetic ketoacidosis; CSH, clinically significant hypoglycemia

To control for potential confounding variables, analysis of covariance (ANCOVA) was performed, including age, sex, duration of diabetes, HbA1c level, and treatment modality as covariates. After adjustment, differences between clinical groups remained statistically significant for sustained attention (F = 10.35, p < 0.001) and timeliness (F = 9.05, p < 0.001). In contrast, no significant differences were observed for hyperactivity (p = 0.21), and impulsivity was no longer significant after adjustment (p = 0.079). The results are presented in Table 4.

Table 4.

Summary of ANOVA and ANCOVA results for MOXO-CPT parameters

MOXO-CPT parameters	ANOVA F (3,113)	P	Effect size (η²)	ANCOVA F (3,107)	P
Sustained attention	11.29	< 0.001	0.231	10.35	< 0.001
Timeliness	9.42	< 0.001	0.200	9.05	< 0.001
Hyperactivity	1.92	0.130	0.048	1.55	0.207
Impulsivity	4.12	0.008	0.099	2.33	0.079

Note: ANOVA, one-way analysis of variance; ANCOVA, analysis of covariance; F, F-statistic; p, p-value; η², eta-squared effect size

To identify independent predictors of cognitive performance, multivariable linear regression analyses were performed, including age, sex, HbA1c, duration of diabetes, and treatment modality, and clinical group as independent variables, with the control group (Group IV) as the reference. For sustained attention, the overall model was significant (F(2,114) = 13.88, p < 0.001, R²adj = 0.18), with lower performance independently associated with belonging to Group II (β = − 0.40, p < 0.001) and Group III (β = − 0.29, p = 0.001). Similarly, for timeliness, the model was significant (F(2,114) = 12.45, p < 0.001, R²adj = 0.17), with lower scores observed in Group II (β = − 0.37, p < 0.001) and Group III (β = − 0.29, p = 0.001). For hyperactivity, a weaker but significant model was observed (F(1,115) = 4.48, p = 0.037, R² = 0.04), with longer duration of diabetes associated with higher scores (β = 0.19, p = 0.037). For impulsivity, the model was also significant (F(4,112) = 8.73, p < 0.001, R²adj = 0.21). Higher impulsivity scores were independently associated with lower HbA1c (β = − 0.30, p < 0.001), longer duration of diabetes (β = 0.24, p = 0.005), and not belonging to Group I (β = − 0.20, p = 0.020). Sex and treatment modality were not significant predictors of any MOXO-CPT parameter. The results of the multivariable regression analyses are presented in Table 5.

Table 5.

Multiple regression predicting MOXO-CPT parameters

MOXO-CPT parameters	Predictor	Β	P	Model (F, R²adj)
Sustained attention	Group II	–0.40	< 0.001	F(2,114) = 13.88; R²adj = 0.18
	Group III	–0.29	0.001
Timeliness	Group II	–0.37	< 0.001	F(2,114) = 12.45; R²adj = 0.17
	Group III	–0.29	0.001
Hyperactivity	Duration of diabetes	0.19	0.037	F(1,115) = 4.48; R²=0.04
Impulsivity	HbA1c	–0.30	< 0.001	F(4,112) = 8.73; R²adj = 0.21
	Duration of diabetes	0.24	0.005
	Group I	–0.20	0.020

Note: β, standardized coefficient; t, t test statistic; F, overall model test; R²_adj/ R², coefficient of determination; p, p-value. Multivariable regression models were developed using backward elimination

Finally, exploratory analyses were conducted to further investigate associations between selected clinical variables and MOXO-CPT outcomes.

Specifically, Spearman correlation analysis was performed to assess whether the cumulative number of DKA episodes was associated with MOXO-CPT outcomes. Among the 25 participants with a history of DKA, the mean number of DKA episodes was 1.76 ± 1.74 (median 1, range 1–8). Most participants (72%) experienced a single DKA episode, while only three participants (12%) had four or more episodes. No significant correlations were observed for Sustained Attention, Timeliness, or Hyperactivity. A weak inverse correlation was found between the number of DKA episodes and Impulsivity (Spearman’s ρ = −0.243, p = 0.008). Exploratory correlation analyses between clinical and demographic variables and MOXO-CPT outcomes are presented in Table S1.

Additionally, exploratory correlation analyses using Pearson’s r showed that higher impulsivity scores were associated with lower HbA1c (r = − 0.30, p < 0.01). In subgroup analyses, higher HbA1c levels were significantly associated with lower sustained attention in Group III (r = − 0.76, p < 0.001) and Group IV (r = − 0.60, p < 0.01).

Additionally, Group IV was further analyzed by stratifying patients according to the frequency of L1 hypoglycemic episodes (daily, 2–3 times per week, once per week, and occasional episodes. Comparisons were performed using one-way analysis of variance (ANOVA), and the results are presented in Table 6. No statistically significant differences were observed between subgroups in terms of sustained attention (F(3,34) = 0.57, p = 0.639, η² = 0.05), timeliness (F(3,34) = 1.16, p = 0.338, η² = 0.09), hyperactivity (F(3,34) = 1.34, p = 0.276, η² = 0.11), or impulsivity (F(3,34) = 1.63, p = 0.200, η² = 0.13). All patients in Groups I-III reported episodes of L1 hypoglycemia (Table 1); its independent effect could not be evaluated due to overlap with DKA and CSH.

Table 6.

Comparison of MOXO-CPT parameters according to the frequency of hypoglycemic episodes in the control group

MOXO-CPT parameters	Group IV (n = 38)				F(3,34)	P	η2
	L1 hypoglycemia episodes
	Daily(n = 10)	2–3 x week(n = 13)	1x week(n = 8)	Occasional (n = 7)
	M ± SD
Sustained attention	0.65 ± 0.39	0.75 ± 0.6	0.64 ± 0.71	0.42 ± 0.3	0.57	0.639	0.05
Timeliness	0.24 ± 0.86	0.0 ± 1.4	-0.94 ± 2.04	-0.24 ± 1.08	1.16	0.338	0.09
Hyperactivity	0.55 ± 2.17	1.27 ± 0.52	0.34 ± 1.47	1.4 ± 0.26	1.34	0.276	0.11
Impulsivity	-0.44 ± 1.61	-0.25 ± 1.71	0.7 ± 1.16	0.77 ± 0.96	1.63	0.200	0.13

Note: M, mean; SD, standard deviation; F, F statistic; p, p value; η², effect size (eta-squared)

Discussion

The present study suggests that CSH is the primary determinant of attentional dysfunction in children and adolescents with T1D, whereas DKA is mainly associated with alterations in impulsivity, highlighting domain-specific differences in neurocognitive functioning of acute metabolic disturbances as assessed by the MOXO-CPT.

Differential effects of acute metabolic events on attentional functioning

In our study, CSH was associated with marked impairments in sustained attention and timeliness and remained an independent predictor of these parameters. These cognitive domains are essential for learning, information processing, and everyday functioning. Hypoglycemia, particularly when severe or recurrent, may be associated with neuroglycopenia and affect brain regions such as the prefrontal cortex and hippocampus, which are crucial for attentional regulation and executive control [17, 18, 19] . Recurrent hypoglycemia may also induce adaptive changes in cerebral glucose transport and neuronal metabolism, potentially affecting cognitive processing speed and attentional regulation [20, 21].

Previous studies have primarily focused on severe hypoglycemia accompanied by neurological complications such as seizures or loss of consciousness [6, 22]. The present study extends these findings by demonstrating that clinically significant hypoglycemia, even in the absence of such events, may be associated with alterations in attentional functioning.

In contrast, DKA in our study was primarily associated with poorer performance in the domain of impulsivity, suggesting potential difficulties in behavioral inhibition and regulation of motor responses, while sustained attention and timeliness appeared relatively preserved. However, in multivariable regression analyses, DKA was not identified as an independent predictor of attentional performance. Instead, higher HbA1c levels and longer diabetes duration were significantly associated with poorer outcomes in selected cognitive domains. This observation suggests that the influence of DKA on cognitive functioning may be partly mediated by overall metabolic control and cumulative disease burden rather than by the acute metabolic episode itself. The increased impulsivity observed in patients with a history of DKA may reflect altered or delayed maturation of neural circuits involved in behavioral inhibition and self-monitoring. These findings are consistent with previous studies suggesting that episodes of DKA, particularly when occurring in early childhood, may affect frontostriatal networks involved in executive and behavioral regulation [5, 11].

Interestingly, higher glucose levels during testing were also associated with higher impulsivity scores. However, this finding does not necessarily indicate improved cognitive functioning. Rather, it may reflect increased arousal or response activation under conditions of acute hyperglycemia, potentially leading to faster or more frequent responses without improved inhibitory control. Similar observations have been reported in previous studies examining the effects of acute hyperglycemia on cognitive performance [23, 24].

Overall, the observed differences in cognitive performance between groups were primarily driven by alterations in sustained attention and timeliness, whereas the association with impulsivity became attenuated after adjusting for relevant clinical covariates.

Interestingly, no significant associations were observed between the frequency of L1 hypoglycemia and MOXO-CPT performance. This finding suggests that mild hypoglycemic episodes, even when occurring relatively frequently, may not be sufficient to produce measurable alterations in attentional functioning. These results support the concept of a potential threshold effect, whereby more severe hypoglycemic events are more likely to be associated with neurocognitive consequences [9, 25].

Taken together, the results of this study suggest that acute metabolic disturbances in children and adolescents with T1D may selectively affect specific domains of attentional functioning rather than leading to global cognitive impairment. This dissociation indicates that different metabolic disturbances may influence distinct cognitive processes to varying degrees.

Importantly, this study distinguishes between the attentional and executive function profiles of CSH and DKA, demonstrating domain-specific rather than global cognitive alterations. These findings also support the concept that the relationship between T1D and cognitive functioning is likely bidirectional. While acute metabolic disturbances may contribute to cognitive dysfunction, attentional deficits may also interfere with effective diabetes self-management, potentially increasing the risk of subsequent metabolic complications. However, because of the cross-sectional design, the direction of these associations cannot be determined.

Methodological considerations and the use of MOXO-CPT

The MOXO-CPT offers several advantages over traditional continuous performance tests, such as the Test of Variables of Attention TOVA or the Conners CPT, due to its incorporation of realistic visual and auditory distractors, which enhance ecological validity [15]. Previous studies have demonstrated that the inclusion of environmental distractors improves sensitivity in detecting attentional dysfunction, particularly in pediatric populations [15, 26]. Unlike many CPTs, MOXO-CPT assesses four independent cognitive domains, including hyperactivity, which is often not captured by other tools [15]. Its multi-phase structure allows evaluation of sustained attention over time and the detection of attentional fatigue [27]. Prior research has demonstrated the utility of MOXO-CPT in pediatric populations with chronic diseases, including T1D and chronic kidney disease, supporting its validity and sensitivity in detecting subtle cognitive alterations [27, 28]. Importantly, while MOXO-CPT is less commonly used than traditional neuropsychological batteries such as WISC or CANTAB, its design, incorporating multimodal distractors, may provide greater sensitivity for detecting subtle attentional impairments under ecologically valid conditions. This may be particularly relevant in children with T1D, where attentional dysfunction is often mild and domain-specific [29, 30].

Potential mechanism

The distinct cognitive profiles observed in the present study are consistent with previous evidence suggesting that different metabolic disturbances may affect cognitive functioning through partially different mechanisms [31]. DKA has been associated with alterations in brain development and executive functioning [7, 11, 32, 33, 34, 35], whereas hypoglycemia primarily affects neuronal metabolism and brain regions involved in attention and cognitive processing [20, 21]. These differences may partly explain the distinct patterns observed in our cohort, with clinically significant hypoglycemia predominantly associated with sustained attention and timeliness, and DKA mainly associated with impulsivity. Future longitudinal studies combining neuropsychological assessment with neuroimaging and biological markers may help clarify the mechanisms underlying these associations.

Clinical and educational implications

From a clinical and neuropsychological perspective, the observed differences are highly relevant to the everyday functioning of children with T1D. Attentional functioning and timeliness are crucial for effective academic performance, glucose monitoring, and treatment planning, whereas impulsivity and hyperactivity impact response inhibition and the ability to make informed therapeutic decisions. The present findings suggest that children with a history of DKA may be particularly vulnerable to difficulties in self-regulation and behavioral control, while those experiencing CSH may be more prone to reduced alertness, slower processing speed, and greater attentional instability. Taken together, these results highlight the need for a more targeted assessment of attentional and executive functioning that considers the predominant type of metabolic disturbance when evaluating cognitive functioning in children and adolescent living with T1D. Overall, these findings emphasize the dominant role of clinically significant hypoglycemia in shaping attentional dysfunction in pediatric T1D. From a clinical perspective, these findings support the need for targeted cognitive screening in children with T1D, particularly in those with a history of CSH or DKA. Early identification of domain-specific cognitive difficulties may improve diabetes self-management, treatment adherence, and long-term outcomes.

Strengths and Limitations

The strengths of this study include the use of an objective, standardized computerized assessment (MOXO-CPT), a relatively large cohort of children and adolescents with T1D, and the analysis of real-world clinical data derived from routine care. The application of a sensitive tool targeting multiple domains of attentional functioning enabled a nuanced characterization of cognitive profiles associated with distinct metabolic exposures.

Several limitations should also be acknowledged. First, the cross-sectional design limits causal inference and precludes conclusions regarding longitudinal cognitive trajectories. Second, information on the frequency and severity of metabolic events was based on medical records and self-reports, which may introduce reporting bias. Third, this was a single tertiary diabetes center study, which may limit the generalizability of the findings to other populations and healthcare settings. Fourth, several potential confounding factors, including pubertal status, body mass index (BMI), eating disorders and psychosocial factors, were not assessed. Finally, although MOXO-CPT is a validated and sensitive measure of attentional functioning, the use of a single computerized assessment tool limits the evaluation of broader neuropsychological domains, and comprehensive neuropsychological batteries may identify additional or more subtle cognitive deficits. Moreover, the lack of neuroimaging data prevents direct evaluation of the structural or functional neural correlates underlying the observed cognitive differences. Additionally, the relatively small subgroup sizes in certain analyses may have limited statistical power.

Future directions

Future studies should focus on longitudinal assessment of cognitive trajectories in children with T1D and integrate neuropsychological findings with neuroimaging and continuous glucose monitoring metrics. Such approaches may help to better understand the mechanisms underlying domain-specific cognitive alterations and identify patients at increased risk of clinically relevant impairment.

Conclusions

This study demonstrates that acute metabolic disturbances in children and adolescents with T1D are associated with distinct profiles of attentional functioning rather than generalized cognitive dysfunction. A history of DKA was primarily related to increased impulsivity, suggesting difficulties in behavioral inhibition and self-regulation, whereas CSH was associated with poorer sustained attention, slower timeliness, and greater attentional instability, suggesting potential difficulties in maintaining alertness and efficient information processing during everyday tasks. Adequate cognitive functioning is essential for the ability of children and adolescents with T1D to recognize metabolic threats, evaluate risk, and respond appropriately to glycemic disturbances. The distinct attentional profiles identified in this study support the consideration of individual cognitive capacities in diabetes education and clinical care, particularly in patients with a history of metabolic instability.

Finally, the MOXO- CPT appears to be a cost-effective and clinically useful tool for assessing attentional functioning in pediatric patients with T1D, allowing identification of domain-specific cognitive profiles relevant to everyday disease management.

These findings support the need for domain-specific neuropsychological assessment and reinforce the necessity for longitudinal monitoring of cognitive functioning in pediatric T1D to optimize both safety and long-term disease management. Identification of domain-specific attentional difficulties may support individualized diabetes education and improve long-term metabolic safety in pediatric patients with T1D.

Supplementary material

Below is the link to the supplementary material.

Author contributions

P.W. conceived and designed the study, developed the methodology, collected, and curated the data, performed the statistical analyses, prepared the visualizations, administered the project, wrote the original draft of the manuscript, and participated in manuscript review and editing. E.N. contributed to the study conceptualization and methodology, provided supervision and validation, and participated in manuscript review and editing. M.S.-K. contributed to the study conceptualization and methodology, participated in data collection and data curation, provided supervision and validation, and participated in manuscript review and editing. K.A.M. participated in data collection and data curation and contributed to manuscript review and editing. P.F. contributed to the study conceptualization and methodology, provided supervision and validation, and participated in manuscript review and editing. A.K. provided supervision and validation and participated in manuscript review and editing. All authors reviewed and approved the final manuscript.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics statement

Ethical approval for this study was obtained from the Bioethics Committee at Poznan University of Medical Sciences, Poland (No.1140/19). The study was carried out in compliance with relevant guidelines and regulations. Written informed consent was obtained from all participants aged ≥16 years and their parents or legal guardians, and from the parents or legal guardians of participants younger than 16 years.

Competing interests

The authors declare no competing interests.