Identifying the metabolic profile of Hashimoto's thyroiditis from the METHAP clinical study

Hashimoto's thyroiditis (HT), one of the most common autoimmune diseases and the leading cause of hypothyroidism, is linked to metabolic and cellular dysfunctions that contribute to disease aetiopathogenesis. This case-control study aimed to identify potent metabolic biomarkers of HT employing machine learning techniques. 62 euthyroid patients with HT and 58 healthy individuals were included from the metabolic biomarkers in Hashimoto's thyroiditis and psoriasis (METHAP) clinical trial. Quantification of 73 metabolites was performed using gas-chromatography/mass spectrometry in plasma and urine samples of fasted participants. Changes in the tricarboxylic acid cycle, carbohydrate, neurotransmitter, microbiome and lipid metabolism were identified in the HT group. Ordinary least squares and beta regression modeling associated the presence of HT with methylmalonic acid, 4-hydroxyphenylpyruvic acid, palmitic acid, palmitoleic acid, myristoleic acid and total saturated fatty acids, adjusting for confounders. Artificial neural network analysis had good predictive accuracy with an AUC of 0.8, while debiased sparse partial correlation network analysis identified metabolite-metabolite interactions distinct for HT. These findings provide insights into novel biomarkers associated with HT, and we discuss their biological relevance and clinical significance. Hashimoto's thyroiditis is associated with mitochondrial dysfunction, micronutrient decreased bioavailability, microbiome imbalances, and carbohydrate and fatty acids dysfunctional metabolism.