Plasma Metabolomic Profiling in 1391 Subjects with Overweight and Obesity from the SPHERE Study

Overweight and obesity have high prevalence worldwide and assessing the metabolomic profile is a useful approach to study their related metabolic processes. In this study, we assessed the metabolomic profile of 1391 subjects affected by overweight and obesity, enrolled in the frame of the SPHERE study, using a validated LC–MS/MS targeted metabolomic approach determining a total of 188 endogenous metabolites. Multivariable censored linear regression Tobit models, correcting for age, sex, and smoking habits, showed that 83 metabolites were significantly influenced by body mass index (BMI). Among compounds with the highest association, aromatic and branched chain amino acids (in particular tyrosine, valine, isoleucine, and phenylalanine) increased with the increment of BMI, while some glycerophospholipids decreased, in particular some lysophosphatidylcholines (as lysoPC a C18:2) and several acylalkylphosphatidylcholines (as PC ae C36:2, PC ae C34:3, PC ae C34:2, and PC ae C40:6). The results of this investigation show that several endogenous metabolites are influenced by BMI, confirming the evidence with the strength of a large number of subjects, highlighting differences among subjects with different classes of obesity and showing unreported associations between BMI and different phosphatidylcholines.

SEMIPARAMETRIC EFFICIENCY FOR CENSORED LINEAR REGRESSION MODELS WITH HETEROSKEDASTIC ERRORS

Using a simplified approach developed by Severini and Tripathi (2001), we calculate the semiparametric efficiency bound for the finite-dimensional parameters of censored linear regression models with heteroskedastic errors. Under an additional identification at infinity type assumption, we propose an efficient estimator based on a novel result from Lewbel and Linton (2002). An extension to censored partially linear single-index models is also presented.