Breath biomarkers of insulin resistance in pre-diabetic Hispanic adolescents with obesity

Breath biomarkers of insulin resistance in pre-diabetic Hispanic adolescents with obesity

Mohammad Sharif Khan, Wake Forest Baptist health 

Poster PDF


A quarter of the world's adults are Insulin Resistance (IR), a condition associated with cellular imbalance between insulin and blood glucose metabolism. Approximately 70% of people with IR develop type-2 diabetes within 15-20 years. Despite progress on diabetic management tools, there is a lack of a diagnostic tool to track the development of IR prior to the onset of clinical type-2 diabetes symptoms. In this study, we report data from pre-diabetic Hispanic adolescents with obesity, age 13-17 years (males=18, females=10). Random forest machine learning was used to identify a set of important breath volatile metabolites. We developed a model based on the important breath metabolites (set of 10) that shows a high correlation with the Homeostatic Model Assessment of Insulin Resistance or HOMA-IR (R = 0.93, p <0.001). This breath-based IR prediction model correlates with quantitative insulin sensitivity check index or QUICKI (R=-0.72, p <0.001) and the fasting insulin resistance index or FTRI (R=-0.85, p <0.001). In addition, we observed a strong correlation of the breath-based IR with the Fasting Blood Insulin (mU/L) and Fasting bold glucose (mg/DL) level (R=-0.91, p <0.001 and R=-0.61, p <0.001, respectively). The predicted IR values did not correlate well with the Total Cholesterol (mg/dL), Triglycerides (m/dL) and HDL (mg/dL) levels. Identified metabolites limonene and undecane clusters individuals with and without insulin resistance significantly (p =0.0037 and p =0.0023, respectively). The limonene found in this study have been reported previously in breath biomarkers of liver cirrhosis and non-alcoholic fatty liver disease which suggests that our identification in pre-diabetic adolescents could be evidence of early signs of liver disease. The results of this pilot study show feasibility of a simple and non-invasive breath-based tool which could be developed further to monitor and track IR, potentially extending the time of onset of the type-2 diabetics.  
Keywords: insulin resistance; breath volatiles; homa-ir; metabolic biomarker 



1 comment
  • This is very interesting topic for me, personally and professionally! Can you name some other VOCs that you found to be related to IR, along the limonene? Additionally, did you measure concentration levels of relevant VOCs in IR patients' breath samples or maybe have noticed a relative abundance difference among participants with and without IR?

    Thank you!

    Comment actions Permalink

Please sign in to leave a comment.

Didn't find what you were looking for?

New post
Download Complete Guide