Document resource
WHAT IS ALREADY KNOWN ON THIS TOPIC Previous research has produced conflicting evidence about cognitive performance in children with type 1 diabetes, with some studies suggesting impairments, others showing no results, leaving clinicians uncertain about the true cognitive profile and clinical significance of any differences.WHAT THIS STUDY ADDS The comprehensive review of 129 studies reveals that children with type 1 diabetes showing consistent patterns of worse outcome in cognitive skills especially in executive function, while academic performance and language skills remain largely preserved, particularly in large-scale studies.HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY These findings should guide targeted cognitive support focusing on executive function rather than broad educational interventions, while highlighting the need for research connecting cognitive assessments to real-world functional outcomes and informing clinical screening protocols for at-risk children.Introduction Type 1 diabetes (T1D) is a chronic autoimmune condition that poses long-term challenges for approximately 1.5 million children and adolescents worldwide. 1 2 T1D requires ongoing management that includes insulin therapy, continuous glucose monitoring and lifestyle adaptations.3 T1D may result in acute and life-threatening complications, including chronic hyperglycemia, diabetic ketoacidosis (DKA) and hypoglycemia.4 5These complications potentially impact cognitive skills, particularly during critical periods of brain development.6 Previous research highlighted the complex interplay between metabolic control and cognitive outcomes in children and adolescents with T1D.6 7 Persistent hyperglycemia contributes to inflammation, oxidative stress and microvascular damage that compromises brain health by attenuating neurogenesis and neuronal integrity.8 9 Similarly, frequent episodes of severe hypoglycemia can cause acute neurological dysfunction and long-term cognitive deficits.10 11 Both hyperglycemia and hypoglycemia have been associated with altered brain structure and cognitive impairments.6 11–13 In children and adolescents, these changes are of particular concern given that the developing brain is highly sensitive to metabolic imbalances.14 Some studies have suggested that children and adolescents with T1D may have lower cognitive scores, which could be associated with reduced language and academic performance.15 16 However, while certain studies report statistically significant cognitive differences between children with T1D and their non-T1D peers,17 18 others have found differences to be insignificant.19 20The inconsistency in findings between studies, combined with the variety of cognitive assessment tools used, necessitates a comprehensive review of existing literature to provide clearer insights into cognitive measures and outcomes in pediatric populations with T1D.This scoping review aims to document measures and outcomes used to assess cognitive skills in children with T1D and to examine the relationship between T1D and cognitive skills in pediatric populations.Methods Study design This study adheres to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses-Extension for Scoping Reviews (PRISMA-ScR) guidelines. 21 A scoping review was conducted to document and categorize studies on cognitive skills and brain imaging in children and adolescents with T1D in terms of assessment tools used and results documented. This approach enabled us to inclusively consider findings on T1D and cognition within various study designs. Meta-analyses to quantify differences between children with vs without T1D were not a primary focus of this review, but were performed where sufficient data were available.Protocol and search strategy The protocol was published on the Open Science Framework platform in January 2024. 22 A comprehensive literature search queried five databases: Medline (Ovid), Web of Science Core Collection (WOS), Cochrane Library (Cochrane Library), EMBASE (Ovid) and PsycINFO (EBSCO). The search incorporated targeted keywords and phrases related to cognitive skills, children and T1D. The search strategy was developed in collaboration with a medical librarian (online supplemental Table 2). The search was executed with no time restrictions and included articles published in English up to 31 December 2023.SP110.1136/bmjdrc-2025-005635.supp1Supplementary dataEligibility criteria Articles were included if they (1) described research of any original study design (including randomized or non-randomized controlled trial, retrospective cohort, prospective cohort, case–controls and cross-sectional studies), (2) included pediatric participants aged 0–19 years diagnosed with T1D and (3) assessed cognitive skills. Studies were excluded if they (1) were review articles or editorials, (2) only reported on participants with preterm birth before 34 weeks’ gestation, (3) only reported on participants with low birth weight (<2000 g), (4) only reported on participants with intellectual or learning disabilities, (5) only reported on samples with prior in-patient psychiatric treatment, (6) reported on participants who experienced neurologic events not related to diabetes or (7) had a cohort which explicitly reported a prevalence of Attention-Deficit/Hyperactivity Disorder (ADHD) greater than the prevalence rate in the USA (>11%). 23 This threshold reflects the expected background prevalence in children and was chosen to help distinguish T1D-related cognitive patterns from general population variability. Moreover, for studies presenting T1D versus control group comparisons, control groups had to consist of peers without any type of diabetes (eg, siblings, school-matched controls or community samples). These criteria were established in an attempt to exclude confounding factors that could affect cognitive assessments.Selection of sources of evidence The literature search results were screened in Covidence. 24 The title and abstract screening was conducted by DA and TS, independently. The overall inter-rater reliability between DA and other reviewers was assessed using Cohen’s Kappa, and moderate agreement (Kappa=0.63) was observed.25 Disagreements were resolved through discussions between reviewers to reach the consensus. Full-text screening was conducted by DA, which was deemed sufficient after a random selection of 20 studies was reviewed in duplicate by LvH, and there was 100% agreement between reviewers.Data extraction and synthesis Study characteristics, publication year, participant demographics, cognitive assessment methods and cognitive assessment outcomes in both numerical and non-numerical formats were extracted in Microsoft Excel by DA. In studies with multiple data collection timepoints, only baseline data were extracted. It became apparent during data extraction that one particular outcome, the Full Scale Intelligence Quotient (FSIQ) component of the Wechsler Intelligence Scale, would be suitable for meta-analyses to quantify differences between children with versus without T1D. The Wechsler Intelligence Scales are designed to assess a range of cognitive skills, including executive function, memory, learning and overall intelligence. FSIQ is a composite score derived from various subtests and provides a measure of overall mental ability or intelligence. Quantitative data for meta-analysis from relevant studies were independently extracted in duplicate by DA and LvH, with 94% agreement; any disagreements were resolved through discussion.Evidence synthesis This comprehensive review systematically identified and categorized cognitive assessment methods and outcomes from a wide range of research studies, as detailed in ( online supplemental Table 3. Many tests were found to evaluate multiple cognitive categories simultaneously, necessitating their classification under several categories. The primary synthesis of this review is presented as an evidence map, which organizes these categories and visually represents the relationships between these categories with T1D. For studies that included brain imaging data, the findings are presented as relevant differences between groups, as brain imaging does not allow for conclusions about whether outcomes are ‘better/worse’.Statistical analysis Among the included studies, the FSIQ component of the Wechsler Intelligence Scale demonstrated feasibility for meta-analysis. It has been demonstrated that the FSIQ, despite undergoing six editions with evolving components, maintains consistent overall scores, 26 which allows for direct comparison across studies using different editions of the FSIQ. Additionally, a subgroup analysis was conducted to examine the potential impact of different Wechsler editions (online supplemental Figure 1). In the meta-analysis, the mean difference (MD) represents the average difference in scores with negative values, indicating poorer performance in the T1D group. Each study’s MD is weighted by its inverse variance to account for variability across studies. Due to heterogeneity possibly due to differences in sample size and different Wechsler editions (I²=83%), a random effects model (DerSimonian-Laird method) was applied for all the meta-analyses. The significance is indicated by 95% CI and p<0.05. A funnel plot was created to assess publication bias (online supplemental Figure 2) and small-study effect was assessed with Egger’s test (online supplemental Figure 3). Sensitivity analysis was conducted by omitting each study one at a time (online supplemental Figure 4). To address heterogeneity, subgroup analyses based on era effect and economic development (based on World Bank classification)27 were conducted (online supplemental Figure 5 and online supplemental Figure 6).28 A quality score adapted from the Newcastle-Ottawa Scale (NOS) was used to evaluate the studies included in meta-analyses (online supplemental Table 4). The ‘meta’ package of R was used for all meta-analyses.28Results Our search protocol identified 2464 unique articles ( figure 1). After screening titles and abstracts, 2180 articles were excluded, and the full texts of 284 articles were reviewed. Ultimately, 129 articles met the inclusion criteria for data extraction. Sixteen studies were included in the meta-analysis. Full details of included articles are found in online supplemental Table 1. No articles were excluded due to the ADHD prevalence exclusion criterion.Figure 1PRISMA diagram of study selection. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.Overview of studies Across the 129 included studies, sample sizes ranged from small clinic-based cohorts of fewer than 20 participants to very large registry and population-based samples of several hundred thousand children and adolescents (eg, national education and health registries), with full details provided in online supplemental Table 1. Seven large-scale studies (sample size >1000) were included in the review. Most studies focused on school-aged children and adolescents, with mean ages typically between about 7 and 15 years; a minority included broader age ranges spanning early childhood through late adolescence (eg, 4–19 or 6–18 years), or more narrowly defined preschool or older adolescent samples. The geographical distribution of studies was predominantly in the USA and Canada (n=64, 48%), followed by Europe (excluding Turkey) (n=26, 20%) and Australia (n=15, 11%) (online supplemental Figure 7). The majority of the included articles described observational studies: 66 (51%) cross-sectional and 40 (31%) prospective cohort studies. The remaining studies comprised nine retrospective cohort studies, seven case–control studies, three non-randomized experimental studies, three randomized control trials and one case series. Cognitive skill assessments were categorized as: academic performance, language, executive function, memory and learning, and intelligence as well as brain imaging (online supplemental Table 2). Just around half of the included studies (n=75; 49%) incorporated a non-T1D comparison group, while the remainder evaluated only children and adolescents with T1D. Most studies contributed data to more than one domain, and the majority assessed three or more of these outcomes within the same cohort, meaning that individual studies appear multiple times in the evidence map (figure 2).Figure 2Evidence map illustrating the scope of research on cognitive skills in children and adolescents with type 1 diabetes by assessment category and year of publication. The evidence map in the form of a bubble plot shows the scope of available research on cognitive skills in children and adolescents with type 1 diabetes by assessment category (columns) and year of publication (rows). Each bubble represents a unique study; the same study may appear across multiple columns if a multitude of cognitive skill outcomes were reported. The bubble size indicates the sample size, with larger bubbles representing larger studies. Color-coding represents the directionality of between-group differences (where available), with orange indicating that the T1D group had significantly worse outcomes than the non-T1D group, and green indicating that there was no difference between the T1D group and non-T1D group. No study reported better outcomes in the T1D group vs non-T1D group. For brain imaging only, purple indicated that outcomes were significantly different between T1D and non-T1D groups, but whether the difference has functional implications is currently unclear. Across all categories, gray indicated that there was no comparison made between a T1D and a non-T1D group (eg, these studies may have descriptively reported cognitive skill outcomes within a T1D cohort, or correlated cognitive skills with clinical variables, such as age at T1D onset or frequency of DKA). DKA, diabetic ketoacidosis; T1D, type 1 diabetes.The evidence map (figure 2) shows the distribution of studies by categorization of cognitive skill assessments and year of publication as well as provides insight into the directionality of differences between T1D and non-T1D groups (where available) and the sample size of each study. The evidence map provides a visual summary of several key findings: (1) executive function was the most frequently studied outcome; (2) the volume of studies investigating cognitive skills in children and adolescents with T1D has increased markedly since 2010 and (3) where comparisons were available, the evidence-base was largely conflicting as to whether outcomes were worse in children and adolescents with T1D versus without T1D.Academic performance Academic performance was assessed in 29% of studies (n=37). Twenty-one studies compared children and adolescents with T1D to those without T1D. Among these, 7 studies reported worse results for groups with T1D, 15 16 29–33 and 15 found that results were similar between groups with and without T1D.20 34–47 Of the six large-scale studies, defined as having sample sizes of n≥1000, none found a significant difference between groups with and without T1D.34 39 42 44–46 The remaining 16 studies used assessments descriptively48 49 or lacked a non-T1D comparison group.50–63Language Language was assessed in 23% of studies (n=30). Twenty studies compared T1D to non-T1D groups. Seven studies reported worse results for groups with T1D, 15 29 31 64–67 and 13 found no differences between groups.30 34–36 39 40 44 68–73 Of the three large-scale studies, defined as having sample sizes of n≥1000, none found a significant difference between groups with and without T1D.34 39 44 Ten studies lacked a non-T1D comparison group.11 54 56 58 60 62 63 74–76Executive function Executive function was assessed in 77% of the included studies (n=101). Forty-eight studies compared executive function of children and adolescents with T1D to those without T1D. Thirty-one of these studies found that children and adolescents with T1D had worse executive function, 12 17 18 31 33 37 41 64 66–68 76–95 whereas 17 studies found no differences between the two groups.13 19 30 35 36 38 40 69–73 96–100 The remaining 53 studies assessed executive function for descriptive purposes101 102 or conducted research without a comparison group.10 11 49 50 53 55 58 60 62 63 74 75 103–141Memory and learning Memory and learning were concurrently assessed in 64% of studies (n=84). Forty-eight studies compared T1D to non-T1D groups. Thirty-one studies reported worse memory and learning scores in groups with T1D, 12 17 18 31 33 36 37 41 64–67 71 76–86 88 90–92 95 142 143 whereas 17 found no significant differences.13 19 30 35 38 40 68–70 72 73 89 96–100 Thirty-six studies did not include a non-T1D comparison group.10 11 49 50 54 55 60 62 63 73–75 103–106 108 109 111 113 116–119 122 123 125 127 128 130 132–135 137 144Intelligence Intelligence testing was conducted in 56% of studies (n=73). Thirty-seven studies compared intelligence test scores between children and adolescents with T1D and those without T1D. Twenty-two studies reported lower intelligence test scores among children and adolescents with T1D, 12 17 18 33 41 66 68 77–85 88 90 92 93 95 145 and 16 found no differences between T1D and non-T1D groups.13 19 20 30 35–37 40 69 72 89 96–98 100 146 The remaining 35 studies assessed intelligence descriptively,10 54 56 74 118 123 or did not include a comparison group.49 50 55 60 61 75 103 105–109 111 113 114 116 117 122 125 127 128 130 132–135 137 141 147Brain imaging Out of 16 studies that applied brain imaging modalities, including structural imaging 12 13 72 77 84 88 97 98 103 118 137 146 and functional imaging,36 96 132 143 15 compared results between individuals with and without T1D and reported different results for those with T1D.12 13 36 72 77 84 88 96–98 118 132 143 146 148 The remaining study explored differences in brain imaging related to DKA.103Meta-analysis The FSIQ score was identified as a suitable outcome for meta-analysis. Sixteen studies used this outcome to describe cognitive skills in children and youth with T1D and included non-T1D comparison groups ( figure 3).13 17 35 36 40 54 64 66 77 81–85 113 146 The overall meta-analysis sample (N=1594) consisted of 898 children and adolescents with T1D and 696 without T1D. Publication dates spanned from 1990 to 2023 with sample sizes ranging from 29 to 281 participants. Children with T1D demonstrated slightly lower FSIQ than non-T1D participants (overall MD −3.49, 95% CI (−6.16 to −0.82), p=0.010). The majority of studies showed negative MDs, indicating lower scores for T1D participants, with values ranging from −13.96 to 8.00. The subgroup analyses revealed that there was no significant difference according to economic development (online supplemental Figure 5) or by era (online supplemental Figure 6). The funnel plot showed slight asymmetry (online supplemental Figure 2), which may suggest potential publication bias favoring studies that reported larger negative MDs (ie, lower FSIQ scores in T1D participants); however, in Egger’s test, the intercept did not deviate significantly from zero (–0.2540 (SE=0.9386), t=−0.27, p=0.7907; online supplemental Figure 3), suggesting there was no statistical evidence of funnel plot asymmetry or publication bias.Figure 3Meta-analysis to quantify differences in Full Scale Intelligence Quotient (FSIQ) between children and adolescents with vs without type 1 diabetes (T1D).Risk of bias A quality score adapted from the NOS was used to evaluate the studies included in meta-analyses ( online supplemental Table 4). Cross-sectional studies generally showed moderate-to-low risk of bias, with several achieving high NOS scores, indicating good quality. All cohort studies received high NOS scores, reflecting robust methodologies and low bias risk. The single case–control study also demonstrated low bias risk with a high NOS score, confirming its strong design. Overall, these evaluations highlight a trend of good methodological quality across study types.Discussion This review is the first of its kind to comprehensively describe and categorize the range of cognitive skill assessments used to describe cognitive skills in children and adolescents with T1D. We identified over 80 different assessment modalities across five different cognitive skill categories, illustrating the complexities involved in assessing and interpreting research on this topic. When comparisons to non-T1D samples were possible, we generally identified a trend (>50% of studies) for worse outcomes for children and adolescents with T1D for executive function, memory and learning, and intelligence, whereas these patterns were less pronounced for studies focused on academic performance and Language.Academic performance and language Academic performance and language skills are discussed jointly in this review, as these were the only cognitive categories in which the majority of studies found that children and adolescents with T1D did not perform worse than their peers without T1D. Recent research, primarily involving large-scale studies of over 1000 participants, suggests the possible role of modern diabetes management strategies and advanced management devices in minimizing cognitive deficiencies associated with T1D. 149 Academic performance and language skills were the only cognitive skill categories that included studies with large-scale sample sizes (total n>1000), which could be due to the relative ease of using standardized educational data that are already collected in many settings.150 In addition, large-scale studies in pediatric T1D relating to academic skills are becoming more common, as findings may have significant implications for educational policies for these children.151It has been shown that academic performance positively correlates with gray matter density in the left dorsolateral prefrontal cortex.152 Studies on children and adolescents with T1D show greater activation and gray matter volume in prefrontal areas, possibly suggesting a compensatory mechanism.36 118 Language-relevant cortices include Broca’s area in the inferior frontal gyrus (IFG), Wernicke’s area in the superior temporal gyrus (STG) and parts of the middle temporal gyrus (MTG).153–156 In children and adolescents with T1D, a reduction in gray matter volume in the left MTG has been reported to correlate with lower language performance.13 However, increased functional connectivity between language-related areas of the right inferior temporal gyrus and the left orbital part of the IFG and increased gray matter volume in STG and MTG suggest possible compensatory adaptations.96 118These findings indicate that although T1D might be associated with structural changes affecting language, the brain might adapt to sustain function via compensatory mechanisms. Complications of T1D can potentially be associated with cognitive changes through various physiological mechanisms.6 However, there remains a gap in research on how serious acute and chronic T1D complications, such as the frequency and severity of hypoglycemic episodes or the number of moderate to severe DKA events, impact academic performance and language skills.Executive function Our review found that most studies investigating outcomes related to executive function, the most frequently studied category, consistently report that children and adolescents with T1D perform worse than their peers without T1D. Executive function, encompassing critical skills such as planning, decision-making, processing speed and problem-solving, was the most extensively studied cognitive domain in relation to T1D, possibly due to its bidirectional relationship with diabetes management. Research consistently demonstrates a relationship between executive function and glycemic control, suggesting that better executive functioning is associated with improved diabetes management and glycemic outcomes. 31 79 84 126 139 140 This connection is supported by evidence of compensatory changes in the prefrontal lobe, critical for executive function in children and adolescents with T1D.36 118 157 These changes include increased gray matter volume118 and heightened activation, particularly during tasks involving executive functions.36 However, young children with T1D may experience a significantly slower growth rate in overall brain volume, particularly in areas associated with executive functions.98Moreover, glycemic control is mediated by adherence to management in T1D.120 Poor executive function possibly leads to suboptimal T1D management, which in turn may impair cognitive processes.129 158 As a result, understanding and improving executive function in groups who experience these complications has become a priority rather than the comparison of executive function in children and adolescents with T1D to non-T1D peers. Recent literature on executive function has shifted to examining subgroups of individuals with T1D who experience complications. These studies have mostly found links between complications of T1D and executive function. DKA negatively correlates with executive function in children and adolescents with T1D,31 79 86 103 and hypoglycemic episodes can lead to poorer outcomes in this population.31 76 79 90 124 However, Perez et al found no significant association between executive function and glycemic control, possibly due to variations in illness duration (1–16 years), which may have biased the results.129These findings highlight the importance of developing targeted interventions to address executive function deficits in children and adolescents with T1D. Future research should investigate how enhancing executive functioning may improve diabetes management outcomes and help overcome the unique challenges faced by this population.Memory and learning Our review found that most studies investigating outcomes related to memory and learning generally report that children and adolescents with T1D perform worse than their peers without T1D. Memory and learning, the second most frequently studied cognitive skills category, is particularly sensitive to metabolic disturbances caused by T1D. Children and adolescents with T1D often report short-term memory loss following hypoglycemic events, highlighting the vulnerability of cognitive functions to metabolic changes 159 that can have lasting effects.124 160 161 However, some of the discrepant findings between studies could be explained by different types of memory being examined and the varied methodologies used in these studies. Memory is a complex cognitive category that includes several distinct types, such as short-term, long-term, working, spatial, visual, verbal and general memory. In children with T1D, these types of memory can be differentially affected by metabolic disturbances such as hypoglycemia and hyperglycemia.11 Severe hypoglycemic episodes have been linked to impairments in memory tasks, potentially due to their immediate impact on brain function.162 There is only one population-based study that examined memory and learning in T1D while accounting for glycemic complications. This study revealed that chronic hyperglycemia and even a single episode of DKA can cause immediate memory impairments.Furthermore, memory and learning are primarily associated with the hippocampus and medial temporal lobe of the brain.163 Researchers have reported no significant difference in hippocampal volume related to DKA unless HbA1c levels at diagnosis are considered.103 Also, reduced white matter volume, specifically in the middle temporal region, among children and adolescents with T1D, is reported.84 Overall, children and adolescents with T1D may experience changes in brain areas linked to memory and learning, and these changes can become worse with poor long-term management.Future research should prioritize longitudinal studies with large, diverse cohorts that differentiate between memory types and account for glycemic variables and T1D complications. These insights could inform targeted interventions, such as memory-focused cognitive training or metabolic monitoring protocols, to mitigate educational and functional challenges in children and adolescents with T1D.Intelligence Most studies related to intelligence reported worse outcome for children and adolescents with T1D compared with their peers without T1D. Intelligence is closely linked to other cognitive skill categories. 164 For example, one study found that when IQ was taken into account, there was no association between executive function and metabolic control.136 This underscores the importance of using intelligence as a descriptive baseline and adjustment factor in cognitive research on T1D. FSIQ was the most commonly used measure for evaluating intelligence in this population, which allowed us to perform meta-analyses on this outcome. The meta-analysis revealed a statistically significant lower FSIQ score (−3.49 points, 95% CI (−6.16 to −0.82); p=0.010) among children and adolescents with T1D compared with their peers without T1D. In the general population, Wechsler intelligence scales, including FSIQ, are standardized with a mean of 100 and an SD of 15 points.165 Based on a large body of evidence, a ‘minimally important difference’ has been suggested to commonly cluster around 0.5 SD,166 167 which in the case of the FSIQ would translate to 7.5 points. The observed MD of 3.49 points in our meta-analysis falls below this threshold, suggesting that this small difference may not have substantial implications for daily functioning or overall intellectual ability. However, this finding is consistent with previous meta-analyses showing small to moderate IQ reductions in the T1D group during key periods of brain development.168 169 Although FSIQ alone cannot predict whether a child will thrive,170 especially in the presence of significant variability among index scores,171 even modest population-level decreases in IQ shift the entire distribution of cognitive performance, thereby increasing the proportion of children at risk for adverse educational outcomes.172In the context of T1D, a decrease in IQ may be associated with T1D-related factors, including age of diabetes onset,173 diabetes duration168 and glycemic control.141 Early-onset diabetes disrupts peak periods of brain glucose metabolism and myelination, causing persistent structural abnormalities.6 173 Chronic hyperglycemia induces cerebral glucose hypometabolism and reduced gray and white matter volumes.12 174 Severe hypoglycemic episodes trigger dose-dependent neuronal death in the hippocampus and cortex.175 176 Longer diabetes duration creates cumulative neurodegenerative effects.168 Many studies included in this review reported on these T1D-specific variables descriptively, revealing substantial variability across studies in glycemic control (HbA₁c ranging from 7.4±0.5% to 13.0±2.0%), diabetes duration (newly diagnosed to 12.1±3.68 years), and age at onset (3.33±1.58 to 10.04±3.85 years; online supplemental Table 5). However, our meta-analyses were based on T1D versus non-T1D comparisons, precluding consideration of these variables statistically as they are only applicable to the T1D group. It is likely that the substantial heterogeneity observed in our meta-analyses (I²=83%) is at least in part attributable to these T1D-specific factors that our analytical approach could not account for. It is also of note that while most studies reported group mean FSIQ scores for children with and without T1D within the average range (90–110), some studies reported relatively high group means (>110). Cognitive test performance is known to be influenced by age-related factors such as educational attainment and reading comprehension,177 178 specific learning disabilities related to working memory and processing speed deficits179 as well as parental education and socioeconomic status.180 We were unable to account for these factors statistically in our meta-analyses based on the information that was available across studies. Finally, selection bias in individual studies may have inflated sample FSIQ estimates.Additional variability may have arisen from different versions of the assessment tools (multiple WISC editions), the economic development status of the country in which a study took place and a possible era effect, with the latter two possibly having implications for access to diabetes management technology. However, all subgroup analyses did not identify any significant differences according to these factors (online supplemental Figures 1, 5 and 6). As well, the leave-one-out sensitivity analyses confirmed that the pooled effect estimate was robust and not driven by any single influential study (online supplemental Figure 4).Brain imaging studies further link intelligence to structural and functional differences in brain regions. Intelligence is not localized to a single brain region but rather emerges from a network involving connections between frontal, parietal and cerebellar regions.181 These include positive correlations between IQ and gray matter volume in the left and right MTG13 97 and cerebral blood flow in the left postcentral gyrus.146Diabetes management and other potential mediators The mechanisms by which cognitive skills may be impacted by T1D include acute glucose fluctuations include disrupting neuronal energy supply, chronic hyperglycemia-induced oxidative stress and cerebral perfusion. 7 Evidently, tightly controlled blood glucose levels are critical for minimizing cognitive deficits.182 Diabetes technology, such as continuous glucose monitors, is highly effective for glycemic control in T1D183 and has also been associated with improved cognitive outcomes.149 Limited longitudinal research suggests that cognitive deficits may remain stable or potentially worsen over time in individuals with chronically elevated glycemic burden.88 98 Conversely, emerging evidence indicates that improved glycemic control through automated insulin delivery systems and intensive management protocols may mitigate cognitive decline and potentially improve cognitive outcomes.184 Future research should prioritize longitudinal cohort studies to determine whether cognitive differences persist, progress or improve with age and glycemic control optimization. Advocacy efforts must address socioeconomic disparities in technology access, as universal adoption could reduce neurodevelopmental risks while improving quality of life for children with T1D.Overall, academic performance and language skills are practical manifestations of other cognitive abilities.185 186 These categories are associated with memory, attention and problem-solving, which are essential for academic success and communication.185 186 Therefore, academic performance and language skills assessments provide an insight into how cognitive abilities are applied in daily life.187 188 The fact that we observed a trend towards lower scores in executive function, memory and learning, and intelligence in T1D versus non-T1D, but much more inconclusive patterns regarding academic performance and language skills between groups suggesting compensatory neural adaptations (eg, increased prefrontal lobe activation during cognitive tasks) may buffer real-world functioning.36 However, cognition in T1D is influenced by a variety of factors beyond glycemia, including genetic predispositions (eg, Apolipoprotein E epsilon 4 allele),189 environmental variables like socioeconomic status, school support and parental involvement,190 which are rarely systematically measured in current studies. Future research on cognitive skills in pediatric T1D should systematically collect data on diabetes technology use and other potential confounding factors, such as genetic, socioeconomic and family environment variables, to clarify causal relationships and identify effective strategies that help children with T1D thrive.Strengths and limitations This review is the first of its kind to provide a comprehensive overview of cognitive skills assessments used in children and adolescents with T1D and to provide a synthesis of cognitive skill outcomes between individuals with versus without T1D, including a meta-analysis for a commonly used intelligence test. By exploring various aspects of how T1D can relate to cognition, including brain structure adaptations and physiological activity, the review identifies research gaps and guides future research directions.However, there are several limitations. This review primarily concentrated on cross-sectional findings, which constrains our understanding of how cognitive development associated with T1D unfolds over time. Although this review identified potential moderating factors that could explain any inverse associations between T1D and cognitive skills, such as glycemic control,12 we were unable to perform specific analyses that could reveal mechanistic insight and identify subgroups at greater risk for cognitive challenges. As well, even though we did not exclude any articles due to a high prevalence of ADHD, many articles did not report on ADHD prevalence; it is, therefore, possible that any undisclosed high rates of ADHD in individual studies may have confounded results. In addition, given our focus on cognitive skills, brain imaging studies were only included in our review if they also included a cognitive skills assessment. It is possible that a broader inclusion criterion for brain imaging studies may have provided a more comprehensive picture of the neurological underpinnings associated with T1D. It is also of note that the majority of included research has been conducted in high-income countries, where there is greater access to the latest T1D management technology. Access to these devices correlates with better glycemic control and cognitive outcomes.182 191 Consequently, the focus of research in high-income countries potentially might under-represent cognitive variations in resource-limited settings. Finally, full-text screening was performed in duplicate for fewer than 10% of studies; however, high reviewer agreement (100%) suggests that the screening process was robust and unlikely to have introduced meaningful inclusion bias.Conclusion Although there is evidence suggesting lower scores in cognitive skills in children and adolescents with T1D compared with their peers without T1D, the practical implications of these findings remain unclear. More research is needed to fully understand the cognitive implications of T1D and to develop targeted and possibly individualized strategies to support affected individuals to thrive in their daily lives.