Two risk score models for predicting incident Type 2 diabetes in Japan

Y. Doi, T. Ninomiya, J. Hata, Y. Hirakawa, N. Mukai, M. Iwase, Y. Kiyohara

Research output: Contribution to journalArticlepeer-review

34 Citations (Scopus)

Abstract

Aims Risk scoring methods are effective for identifying persons at high risk of Type2 diabetes mellitus, but such approaches have not yet been established in Japan. Methods A total of 1935 subjects of a derivation cohort were followed up for 14years from 1988 and 1147 subjects of a validation cohort independent of the derivation cohort were followed up for 5years from 2002. Risk scores were estimated based on the coefficients (β) of Cox proportional hazards model in the derivation cohort and were verified in the validation cohort. Results In the derivation cohort, the non-invasive risk model was established using significant risk factors; namely, age, sex, family history of diabetes, abdominal circumference, body mass index, hypertension, regular exercise and current smoking. We also created another scoring risk model by adding fasting plasma glucose levels to the non-invasive model (plus-fasting plasma glucose model). The area under the curve of the non-invasive model was 0.700 and it increased significantly to 0.772 (P<0.001) in the plus-fasting plasma glucose model. The ability of the non-invasive model to predict Type2 diabetes was comparable with that of impaired glucose tolerance, and the plus-fasting plasma glucose model was superior to it. The cumulative incidence of Type2 diabetes was significantly increased with elevating quintiles of the sum scores of both models in the validation cohort (P for trend <0.001). Conclusions We developed two practical risk score models for easily identifying individuals at high risk of incident Type2 diabetes without an oral glucose tolerance test in the Japanese population.

Original languageEnglish
Pages (from-to)107-114
Number of pages8
JournalDiabetic Medicine
Volume29
Issue number1
DOIs
Publication statusPublished - Jan 2012

All Science Journal Classification (ASJC) codes

  • Internal Medicine
  • Endocrinology, Diabetes and Metabolism
  • Endocrinology

Fingerprint

Dive into the research topics of 'Two risk score models for predicting incident Type 2 diabetes in Japan'. Together they form a unique fingerprint.

Cite this