High Blood Sugar Signals Rising Dementia Risk in Older Adults with Diabetes, Meta-Analysis Finds

Older adults living with diabetes have long been told to watch their blood sugar, but a new systematic review suggests the stakes may extend well beyond the cardiovascular system. A research team from Shandong Xiehe University in Jinan, China, pooled data from 33 studies covering 16,926 participants aged 60 and over to ask a deceptively simple question: do routine blood biomarkers predict who will develop cognitive impairment or dementia? Their answer, published in BMC Geriatrics, is a qualified yes, with important caveats about how strong the evidence really is.

The researchers focused on four biomarkers that clinicians already measure in everyday practice. Glycated hemoglobin, or HbA1c, reflects average blood glucose over roughly two to three months and is the standard yardstick of long-term diabetic control. Fasting plasma glucose, or FPG, captures the sugar level at a single point in time after an overnight fast. The triglyceride-glucose index, known as the TyG index, is a calculated value derived from fasting triglycerides and glucose that serves as a convenient proxy for insulin resistance. Adiponectin, by contrast, is a hormone secreted by fat tissue that generally improves insulin sensitivity, so higher levels might be expected to protect rather than harm the brain.

To be included in the analysis, studies had to enroll older adults with diabetes who were free of cognitive impairment or dementia at baseline, then follow them over time or assess cognition at a single point. The team searched PubMed, Embase, the Cochrane Library, and Web of Science for literature published between 2000 and March 13, 2026. They pooled effect sizes using random-effects models, which assume that the true association may vary from study to study, and for fasting glucose they went a step further, applying restricted cubic splines to model a non-linear dose-response relationship between glucose levels and dementia risk.

The headline finding concerns HbA1c. Across 24 studies reporting cognitive impairment, involving 8,473 participants, elevated HbA1c was associated with a 22 percent increase in the odds of cognitive impairment, with an odds ratio of 1.22. For dementia, drawn from 6 studies and 4,677 participants, the association was stronger: an odds ratio of 1.60, meaning roughly a 60 percent increase in odds. These are modest but clinically meaningful effect sizes, and they align with the broader hypothesis that chronic hyperglycemia damages cerebral vasculature and neurons through mechanisms including oxidative stress, inflammation, and impaired clearance of amyloid proteins.

However, the authors themselves urge caution on the HbA1c result. Funnel plot asymmetry, a significant Egger’s test with a t-statistic of 5.38 and a P value below 0.01, and a trim-and-fill analysis imputing 13 missing studies all pointed to substantial publication bias, the tendency for studies finding strong associations to be published more readily than null results. When such bias is present, pooled estimates can be inflated, so the true association between HbA1c and cognitive decline may be weaker than the headline numbers suggest. The research team explicitly states that the pooled estimate should be interpreted with caution.

Fasting glucose told a more nuanced story. Across 6 studies of dementia with 3,177 participants, higher FPG was associated with a 36 percent increase in dementia odds. But the dose-response analysis revealed something more interesting: the relationship was non-linear, with a P value for non-linearity of 0.008 and an overall P value below 0.001. At moderate glucose elevations the risk rose gently, but above roughly 10.5 millimoles per liter the curve steepened sharply, with the relative risk reaching 1.51 at that threshold and 1.77 at 12.5 millimoles per liter. This pattern suggests there may be a tipping point beyond which each additional increment of fasting glucose carries disproportionately greater danger for the aging brain. The caveat, however, is stark: the dose-response analysis rested on only 3 studies, a very limited sample size.

The TyG index and adiponectin results are even more preliminary. A single study with 308 participants linked higher TyG index values to cognitive impairment, with an odds ratio of 2.24, the largest effect size in the review, and another study with 739 participants examined dementia. Adiponectin, measured in just 2 studies with 251 participants, emerged as a protective factor, associated with a 33 percent reduction in the odds of cognitive impairment. The authors are careful to frame both findings as exploratory and hypothesis-generating rather than confirmatory, and no study in the pool examined adiponectin in relation to dementia at all.

Heterogeneity, the statistical term for variability among studies, was substantial across the analyses, with I-squared values ranging from 58.7 to 69.5 percent and P values below 0.05. In practical terms, this means the individual studies differed considerably in their populations, biomarker cutoffs, cognitive assessment methods, and follow-up durations, which limits confidence that any single pooled number captures a universal truth. Random-effects models partially accommodate this variability, but they cannot eliminate the underlying diversity in how the studies were conducted.

Why does any of this matter? Diabetes affects hundreds of millions of people worldwide, and older adults with the condition already face elevated dementia risk compared with their non-diabetic peers. If simple, inexpensive blood tests could flag which patients are drifting toward cognitive decline, clinicians could intensify monitoring, optimize glucose management, and potentially intervene earlier with lifestyle and pharmacological strategies. The TyG index is particularly attractive from a screening standpoint because it requires only two values from a standard metabolic panel, making it nearly free to calculate in any clinic.

The review’s authors conclude that elevated HbA1c, fasting glucose, and TyG index levels were associated with increased risk of cognitive impairment and dementia in elderly diabetic patients, with modest effect sizes overall, while adiponectin showed a possible protective signal. They call for more high-quality cohort investigations to explore the long-term associations, particularly given the thin evidence base behind the TyG and adiponectin findings and the publication bias clouding the HbA1c estimate. For now, the message for patients and clinicians is one of informed vigilance rather than alarm: keeping blood sugar in check remains fundamental to diabetes care, and the brain may be among the organs that benefit most.

Subject of Research: Associations between glycemic and metabolic biomarkers and cognitive impairment and dementia in elderly patients with diabetes

Article Title: Association of glycemic and metabolic biomarkers with cognitive impairment and dementia among elderly patients with diabetes: a systematic review and a dose-response meta-analysis

Article References: Liu, F., Zhang, P., Wu, G., Zhai, S., Zhang, J., Yu, X., Shen, C., & Wang, R. (2026). Association of glycemic and metabolic biomarkers with cognitive impairment and dementia among elderly patients with diabetes: a systematic review and a dose-response meta-analysis. BMC Geriatrics. https://doi.org/10.1186/s12877-026-08513-4

Image Credits: AI Generated

DOI: 10.1186/s12877-026-08513-4

Keywords: dementia, cognitive impairment, diabetes, HbA1c, fasting plasma glucose, TyG index, adiponectin, meta-analysis, geriatrics, blood sugar, insulin resistance, systematic review