Summary: In at-risk older adults aged 60 to 77 without previously diagnosed diabetes, a secondary analysis of the FINGER lifestyle trial found that higher baseline dysglycaemia, particularly oral glucose tolerance test (OGTT) measures, was associated with less favourable two-year changes in cognition and hippocampal volume. Higher baseline HbA1c showed mixed and partly paradoxical associations, including more favourable changes in processing speed and cortical thickness. These are observational associations, not causal effects, and no effect sizes or p-values were reported in the abstract.
PICO Summary
| Element | Detail |
|---|---|
| Population | At-risk older adults aged 60 to 77 without dementia (N = 1259) enrolled in the FINGER 2-year multidomain lifestyle randomised controlled trial in Finland; 1025 without previously diagnosed diabetes underwent OGTT. Brain MRI was available for 132 participants, and amyloid (PiB)-PET and FDG-PET for 47. |
| Intervention (exposure) | Baseline glucose-related markers as continuous exposures: OGTT-derived measures, fasting glucose, HbA1c, and the triglyceride-glucose (TyG) index, assessed against 2-year change in cognition and neuroimaging. This is a within-trial secondary analysis, not a randomised glucose-lowering intervention. |
| Comparison | Lower versus higher levels of each baseline glycaemia and insulin-resistance marker, and comparison of which marker best detected subtle abnormalities. The trial randomisation (lifestyle intervention versus general health advice) is a separate axis and did not drive these associations. |
| Outcome | Higher baseline dysglycaemia, particularly OGTT measures, was associated with less favourable 2-year changes in multiple cognitive measures and in hippocampal volume. Higher baseline TyG index was associated with higher amyloid accumulation and decline in brain glucose metabolism. Higher baseline HbA1c was associated with more favourable changes in processing speed and cortical thickness (mixed direction). No significant intervention-control differences in glycaemia change were seen, and baseline dysglycaemia did not modify the previously reported intervention benefits. The abstract reported no effect sizes, 95% confidence intervals, p-values, or ARR/NNT. |
Expert Commentary
This is a secondary, observational analysis nested within a randomised prevention trial, and it should be read as hypothesis-generating rather than confirmatory. The central signal is consistent and biologically plausible: among at-risk older adults without diagnosed diabetes, higher baseline dysglycaemia, especially post-load OGTT measures, was associated with worse two-year cognitive trajectories and with hippocampal volume loss, and a higher TyG index tracked with amyloid accumulation and reduced brain glucose metabolism. The verdict is that OGTT-derived measures appear more sensitive than HbA1c for detecting early glucose-related brain risk, but causation cannot be inferred and the magnitude is unquantified. One limitation deserves weight: the neuroimaging subgroups were very small, with MRI in only 132 and PET in 47 participants, so the brain findings are fragile and the paradoxical HbA1c result, where higher HbA1c tracked with seemingly favourable cortical and processing-speed changes, may reflect confounding, reverse causation, or chance rather than a true protective effect. Can I use this with my patients? Not yet as a screening change, but it is a reasonable prompt to consider an OGTT, rather than relying on HbA1c alone, when an older at-risk patient has borderline glycaemia and early cognitive concern. Larger, adequately powered imaging cohorts should test whether OGTT-guided detection translates into actionable prevention.
References
Lorenzo T, Ngandu T, Lehtisalo J, et al. Associations of prediabetes, diabetes and glucose-related markers with cognition and neuroimaging in a 2-year multidomain lifestyle randomised controlled trial. Diabetes Metab Res Rev. 2025;41(5):e70053. doi:10.1002/dmrr.70053
