Home / Resources / Articles / Which Index is the Best Predictor of Type 2 Diabetes? 

Which Index is the Best Predictor of Type 2 Diabetes? 

May 8, 2021
 
Editor: Steve Freed, R.PH., CDE

Author: Destiny Reed, PharmD. Candidate, Florida A&M College of Pharmacy and Pharmaceutical Sciences

Twenty insulin sensitivity indices compared to determine the superior predictor of type 2 diabetes development in Japanese Americans

Insulin resistance plays an essential role in type 2 diabetes’s pathogenesis and is a known risk factor for this condition. Insulin sensitivity is commonly assessed using either the hyperinsulinemic-euglycemic clamp, the gold standard, or the frequently sampled intravenous glucose tolerance test (FSIVGTT). These methods can be challenging; therefore, insulin sensitivity indices seem to be a more practical method for assessing insulin sensitivity. Though studies have compared the indices’ ability to predict type 2 diabetes in white, African Americans, Hispanic Americans, and other ethnicities, no such studies have been conducted among the Asian population.   

 

This study analyzed data from The Japanese American Community Diabetes Study. This prospective epidemiological study examined the etiology and pathogenesis of type 2 diabetes in 2nd or 3rd generation Japanese American adults with 100% Japanese ancestry. The purpose of this analysis was to compare 20 insulin sensitivity indices to determine which would be a better predictor of type 2 diabetes development in Japanese Americans. Insulin sensitivity indexes were included in the analysis if they had the following: fasting samples only (fasting c-peptide, insulin, and glucose); 2-h and fasting samples; or fasting, 30, 60, and 120 min samples during a 75-g OGTT. 

This analysis included individuals between 34 and 76 years of age. Patients were included in the study if they had no history of diabetes or use of antihyperglycemic agents or insulin, had an FBG <126 mg/dL, and had a 2-h plasma OGTT <200 mg/dL at baseline. Patients were followed for 10-11 years, with followup examinations at 5-6 years and 10-11 years. Patients were diagnosed with type 2 diabetes at follow-up if their FBG was ≥126 mg/dL, their 2-h plasma OGTT was ≥200 mg/dL, or if the patient was taking antihyperglycemic agents or insulin.  

Insulin sensitivity indices were analyzed using multiple logistic regression models. They were assessed as continuous data in the logistic regression models. The indices were also evaluated by being categorized into quintiles, and the lowest and highest odds quintiles were compared. The area under the receiver operator characteristic curve (AUROC), integrated discrimination improvement (IDI), and the category free net reclassification improvement (cfNRI) were calculated to provide the extent to which the indices improved the prediction of type 2 diabetes in the patient population. 

Four hundred eighteen patients were included in the analysis. Ninety-five diagnoses of type 2 diabetes were identified at the end of the 10-11 year followup period. Patients who developed diabetes had lower insulin sensitivity across all indices, higher BMIs, FBG, and 2-h glucose levels, were older and more likely to have a family history of type 2 diabetes. 

The Cederholm index, which included fasting, 30 min, 60 min, and 2-h sample post-OGTT, had the highest difference between the lowest and highest quintile cumulative incidence (58.3%). The lowest quintile displayed the highest cumulative incidence (59.5%), and the highest quintile demonstrated the lowest cumulative incidence (1.2%). The second-highest difference between the lowest and highest quintile cumulative incidence was seen in the Gutt index (51.2%), which included only fasting and 2-h samples post-OGTT. The Cederholm and Gutt indices had the highest multiple-adjusted odds ratios between their quintiles of incident diabetes. 

When evaluating the extent to which the indices improved the prediction of type 2 diabetes, the IDI, cfNRI, and AUROC were highest in the Cederholm and Gutt indices Stumvoli index and the 2-h ISI, both indices based on 2-h and fasting samples. There was no significant difference between the Cederholm and Gutt indices (p=1.000). 

This analysis found the Cederholm and Gutt insulin sensitivity indices to be superior in predicting type 2 diabetes in Japanese Americans. The IDI, cfNRI, and AUROC demonstrated that these indices’ ability to predict type 2 diabetes is the same, though the Gutt index is based on fewer measurements. Some limitations to this study include that the population was exclusively Japanese American and may not be generalizable to other races/ethnicities. Previous research has shown that the Gutt index was the best predictor of type 2 diabetes incidence in African American, Hispanic, and white individuals, which may validate applying the results to other races/ethnicities. Since the Gutt index requires fewer samples, it may be preferred as an alternative when the FSIVGTT or hyperinsulinemic-euglycemic clamp cannot be used to assess insulin sensitivity as a risk factor for type 2 diabetes. 

Practice Pearls: 

  • The Cederholm and Gutt insulin sensitivity indices were superior in predicting type 2 diabetes incidence in Japanese Americans. 
  • Though the indices based on fasting, 30 min, 60 min, and 2-h samples incorporated more measurements, they were not superior to indices based on 2-h and fasting samples alone. 
  • Gutt index may be preferred in assessing insulin resistance due to it requiring fewer samples than the Cederholm index. 

 

Onishi, Yukiko et al. “Comparison Of Twenty Indices Of Insulin Sensitivity In Predicting Type 2 Diabetes In Japanese Americans: The Japanese American Community Diabetes Study”. Journal Of Diabetes And Its Complications 

 

Destiny Reed, PharmD. Candidate, Florida A&M College of Pharmacy and Pharmaceutical Sciences