Neuropathic phenotypes of type 1 diabetes are related to different signatures of magnetic resonance spectroscopy-assessed brain metabolites
Tine Maria Hansen, Suganthiya Santhiapillai Croosu, Johan Røikjer, Carsten Dahl Mørch, Niels Ejskjær, Jens Brøndum Frøkjær
Abstract
The study aimed to investigate brain metabolites in type 1 diabetes and the associations with disease characteristics. We explored the metabolic profiles predicting different neuropathic phenotypes using multiple linear regression analyses. We compared brain metabolites in 55 adults with type 1 diabetes (including painful diabetic peripheral neuropathy (DPN), painless DPN, without DPN) with 20 healthy controls. Proton magnetic resonance spectroscopy measurements (N-acetylaspartate (NAA), glutamate (glu), myo-inositol (mI), and glycerophosphocholine (GPC) were obtained in ratios to creatine (cre)) from the parietal region, anterior cingulate cortex and thalamus. The overall diabetes group revealed decreased parietal NAA/cre compared to healthy controls (1.41 ± 0.12 vs. 1.55 ± 0.13,p < 0.001) and increased mI/cre (parietal: 0.62 ± 0.08 vs. 0.57 ± 0.07,p = 0.025, cingulate: 0.65 ± 0.08 vs. 0.60 ± 0.08,p = 0.033). Reduced NAA/cre was associated with more severe DPN (all p ≤ 0.04) whereas increased mI/cre was associated with higher hemoglobin A1c (HbA1c) (p = 0.02). Diabetes was predicted from decreased parietal NAA/cre, increased parietal ml/cre, and decreased thalamic glu/cre. DPN was predicted from decreased parietal NAA/cre and increased GPC/cre. Painful DPN was predicted from increased parietal GPC/cre and thalamic glu/cre. Specific metabolic brain profiles were linked to the different phenotypes of diabetes, DPN and painful DPN. Assessment of metabolic profiles could be relevant for detailed understanding of central neuropathy in diabetes.