错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Discovery and validation of plasma, saliva and multi-fluid plasma–saliva metabolomic scores predicting insulin resistance and diabetes progression or regression among Puerto Rican adults

  • Danielle E. Haslam,
  • Liming Liang,
  • Kai Guo,
  • Marijulie Martínez-Lozano,
  • Cynthia M. Pérez,
  • Chih-Hao Lee,
  • Evangelia Morou-Bermudez,
  • Clary Clish,
  • David T. W. Wong,
  • JoAnn E. Manson,
  • Frank B. Hu,
  • Meir J. Stampfer,
  • Kaumudi Joshipura,
  • Shilpa N. Bhupathiraju

摘要

Aims/hypothesis

Many studies have examined the relationship between plasma metabolites and type 2 diabetes progression, but few have explored saliva and multi-fluid metabolites.

Methods

We used LC/MS to measure plasma (n=1051) and saliva (n=635) metabolites among Puerto Rican adults from the San Juan Overweight Adults Longitudinal Study. We used elastic net regression to identify plasma, saliva and multi-fluid plasma–saliva metabolomic scores predicting baseline HOMA-IR in a training set (n=509) and validated these scores in a testing set (n=340). We used multivariable Cox proportional hazards models to estimate HRs for the association of baseline metabolomic scores predicting insulin resistance with incident type 2 diabetes (n=54) and prediabetes (characterised by impaired glucose tolerance, impaired fasting glucose and/or high HbA1c) (n=130) at 3 years, along with regression from prediabetes to normoglycaemia (n=122), adjusting for traditional diabetes-related risk factors.

Results

Plasma, saliva and multi-fluid plasma–saliva metabolomic scores predicting insulin resistance included highly weighted metabolites from fructose, tyrosine, lipid and amino acid metabolism. Each SD increase in the plasma (HR 1.99 [95% CI 1.18, 3.38]; p=0.01) and multi-fluid (1.80 [1.06, 3.07]; p=0.03) metabolomic scores was associated with higher risk of type 2 diabetes. The saliva metabolomic score was associated with incident prediabetes (1.48 [1.17, 1.86]; p=0.001). All three metabolomic scores were significantly associated with lower likelihood of regressing from prediabetes to normoglycaemia in models adjusting for adiposity (HRs 0.72 for plasma, 0.78 for saliva and 0.72 for multi-fluid), but associations were attenuated when adjusting for lipid and glycaemic measures.

Conclusions/interpretation

The plasma metabolomic score predicting insulin resistance was more strongly associated with incident type 2 diabetes than the saliva metabolomic score. Only the saliva metabolomic score was associated with incident prediabetes.

Graphical Abstract