Background:
Type 2 diabetes (T2D) is characterized by profound metabolic abnormalities. Current glycemic indicators have limitations in identifying early metabolic alterations.
Objective:
To identify novel metabolic predictors of T2D in American Indians participating in the Strong Heart Family Study.
Methods:
Among 2,129 participants who had normal fasting glucose (NFG) at baseline (2001-2003) and also attended clinical examination after 5-year follow-up (2006-2009), 142 developed incident T2D, 514 developed incident impaired fasting glucose (IFG), and 1,473 remained to be NFG. The current analysis included all incident cases of T2D (n=142), 146 incident IFG (randomly selected from 514 participants with incident IFG) and 144 NFG (randomly selected from 1,473 participants with NFG at both visits). Baseline plasma metabolites were detected by high-resolution LC/MS. The prospective association of each metabolite with risk for T2D or IFG was investigated using weighted Cox’s hazard regression with frailty model, adjusting for sex, study center, age, BMI, renal function, fasting glucose and fasting insulin at baseline. Multiple testing was corrected by Bonferroni correction (significance level 2.8х10-6).
Results:
Thirty-nine metabolites from several major fuel sources, including sugar amino acids, amino acids, lipids, alkaloids, alkylamines, carboxylic acids, steroids, and aromatic homomonocylic/heteropolycyclic compounds, significantly predicted future risk of T2D (10 metabolites), or IFG (27 metabolites), or both (2 metabolites). Of these, N1,N12-diacetylspermine and betanidin, respectively, were the strongest predictors for increased (HR=4.59, 95% CI, 2.55-8.24, P=3.49х10-7) and decreased risk of T2D (HR=0.38, 95% CI, 0.28-0.52, P=4.64х10-10). The corresponding strongest predictors for IFG were hexanoic acid (HR=2.34, 95% CI, 1.84-2.98, P=3.15х10-12) and l-palmitoylcarnitine (HR=0.26, 95% CI, 0.18-0.37, P=1.14х10-13), respectively. Two metabolites, betanidin and dopamine, significantly predicted future onset of both T2D (HR=0.38, 95% CI, 0.28,0.52, P=4.64х10-10 for betanidin; HR=2.48, 95% CI, 1.71-3.58, P=1.42х10-6 for dopamine) and IFG (HR=0.52, 95% CI, 0.43,0.62, P=1.35х10-12 for betanidin; HR=2.24, 95% CI, 1.73,2.89, P=5.79х10-10 for dopamine). Multiple unknown compounds were also independently associated with risk of T2D, IFG or both.
Conclusions:
This study identifies both novel and known metabolic alterations associated with risk of diabetes in American Indians, an ethnic group suffering from disproportionately high rates of T2D. The incomplete overlapping of metabolic profiles between T2D and IFG highlights differential metabolic states of diabetes development. Our results not only provide novel insights in disease pathogenesis but also valuable data on potential new targets for risk prediction and treatment.
Longitudinal Plasma Lipidome and Risk of Type 2 Diabetes in a Large Sample of American Indians With Normal Fasting Glucose: The Strong Heart Family Study
