Plasma metabolite profiling identifies non-diabetic CP patients with metabolic alterations progressing to prediabetes prior to HbA1c

Ketavarapu Vijayasarathy; Addipilli Ramunaidu; Nagarjunachary Ragi; Pallerla Pavankumar; Venu Simhadri; Suvidha Manne; Sannapaneni Krishnaiah; Mohsin Aslam; Rupjyoti Talukadar; Ch Venkataramana Devi; G V Rao; Ramars Amanchy; D Nageshwar Reddy; Prabhakar Sripadi; Mitnala Sasikala

doi:10.14309/ctg.0000000000000704

Plasma metabolite profiling identifies non-diabetic CP patients with metabolic alterations progressing to prediabetes prior to HbA1c

Clin Transl Gastroenterol. 2024 Apr 25. doi: 10.14309/ctg.0000000000000704. Online ahead of print.

Authors

Affiliations

¹ Translational Research Center, Asian Healthcare Foundation, AIG Hospitals, Mindspace Road, Gachibowli, Hyderabad - 500032, Telangana, India.
² Centre for Mass Spectrometry, Department of Analytical & Structural Chemistry, CSIR-Indian Institute of Chemical Technology, Tarnaka, Hyderabad - 500007, Telangana, India.
³ Asian Institute of Gastroenterology, Somajiguda, Hyderabad - 500032, Telangana, India.
⁴ AIG Hospitals, Mindspace Road, Gachibowli, Hyderabad - 500032, Telangana, India.
⁵ Department of Biochemistry, University College of Science, Osmania University, Hyderabad - 500007, Telangana, India.
⁶ Department of Applied Biology, CSIR-Indian Institute of Chemical Technology, Tarnaka, Hyderabad 500007, Telangana, India.

PMID: 38661171
DOI: 10.14309/ctg.0000000000000704

Abstract

Introduction: Diabetes (T3cDM) secondary to chronic pancreatitis (CP) arises due to endocrine dysfunction and metabolic dysregulations. Currently, diagnostic tests are not available to identify patients who may progress from normoglycemia to hyperglycemia in CP. We conducted plasma metabolomic profiling to diagnose glycemic alterations early in the course of disease.

Methods: Liquid chromatography-tandem mass spectrometry was employed to generate untargeted, targeted plasma metabolomic profiles in CP patients, controls (n=445) following TRIPOD guidelines. Patients were stratified based on glucose tolerance tests following ADA guidelines. Multivariate analysis was performed using PLS-DA to assess discriminatory ability of metabolites among stratified groups. COMBIROC, logistic regression were employed to derive biomarker signatures. AI-ML tool(Rapidminer) was employed to verify these preliminary results.

Results: Ceramide, lysophosphatidylethanolamine, phosphatidylcholine, lysophosphatidic acid, phosphatidylethanolamine, carnitine and lysophosphatidylcholine discriminated T3cDM CP patients from healthy controls with AUC 93%(95%CI 0.81-0.98, p<0.0001), integration with pancreatic morphology improved AUC to 100%(95%CI 0.93-1.00, p<0.0001). Lysophosphatidic acid, phosphatidylinositol and ceramide discriminated non-diabetic CP with glycemic alterations (pre-diabetic CP);AUC 66% (95% CI 0.55-0.76, p=0.1),integration enhanced AUC to 74%,(95% CI 0.55-0.88,p=0.86). T3cDM was distinguished from pre-diabetic by lysophosphatidic acid, phosphatidylinositol and sphinganine(AUC 70%; 95%CI 0.54-0.83,p=0.08), integration improved AUC to 83% (95%CI 0.68-0.93,p=0.05). CombiROC cutoff identified 75% and 78% prediabetes in validation 1 and 2 cohorts. Random forest algorithm assessed performance of integrated panel demonstrating AUC of 72% in predicting glycemic alterations.

Discussion: We report for the first time that a panel of metabolites integrated with pancreatic morphology detects glycemia progression prior to HbA1c in CP patients.

Abstract

Grants and funding