Bimodality of 2-h Plasma Glucose Distributions in Whites

The Rancho Bernardo Study

  1. Juanjuan Fan, PHD1,
  2. Susanne J. May, PHD2,
  3. Yue Zhou, MS1 and
  4. Elizabeth Barrett-Connor, MD2
  1. 1Department of Mathematics and Statistics, San Diego State University, San Diego, California
  2. 2Department of Family and Preventive Medicine, University of California San Diego, La Jolla, California
  1. Address correspondence and reprint requests to Elizabeth Barrett-Connor, MD, Department of Family and Preventive Medicine, University of California, San Diego, 9500 Gilman Dr., MC 0607, La Jolla, CA 92093-0607. E-mail: ebarrettconnor{at}ucsd.edu

Abstract

OBJECTIVE—Several studies have shown a bimodal curve in the distribution of glucose in populations with a high prevalence of type 2 diabetes, but bimodality has not been reported among whites of Northern European ancestry. It is not clear whether this difference reflects the lower prevalence of diabetes, obscuring a second mode, or implies a more fundamental difference between whites and nonwhites. We investigate this issue by studying glucose distributions in older white patients.

RESEARCH DESIGN AND METHODS—A study of diabetes was conducted among older community-dwelling white residents of a suburban Southern California community between 1984 and 1987. Two-hour plasma glucose data were collected from 2,326 older white men and women aged 23–92. To investigate bimodality of glucose distributions, we fit unimodal and bimodal normal models to 2-h plasma glucose concentrations transformed by the Box-Cox family of transformations.

RESULTS—We found that the bimodal normal mixture model fit the data significantly better than the unimodal skewed distribution model for both sexes and all age-groups except those ≥80 years. The cut points separating the two modes were generally within the 11.1- to 13.6-mmol/l range.

CONCLUSIONS—The bimodality of glucose distributions among whites, combined with previous findings, indicates that this phenomenon may be universal. A smaller second mode in our study compared with other studies suggests that whites have diabetes susceptibility but may require more obesity to demonstrate it. With increasing obesity in the U.S., the predicted epidemic of diabetes may affect all ethnic groups including whites.

Footnotes

  • A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.

    • Accepted March 6, 2005.
    • Received July 7, 2004.
« Previous | Next Article »Table of Contents