Glycated hemoglobin measurement and prediction of cardiovascular disease

The Emerging Risk Factors Collaboration. Glycated hemoglobin measurement and prediction of cardiovascular disease. JAMA. 2014;311(12):1225-1233.

A1C measurement provided little benefit for CVD risk prediction among subjects without diabetes or CVD history in an analysis of data from prospective cohort studies.

This analysis assessed whether adding A1C measurement to conventional CV risk models improves prediction of CVD risk. Prediction of risk with fasting, random, and postload glucose were also examined. Data were analyzed from 294,998 middle-aged and older subjects without known diabetes or CVD history. The primary outcome was first-onset CVD (defined as fatal or nonfatal CHD event or any stroke) in subjects aged ≥40 years. The main outcome measures were risk discrimination (C-index) for CVD outcomes and reclassification of subjects across 10-year CVD risk categories.

CVD outcomes
Over 9.9 years median follow up:

  • 20,840 fatal and nonfatal CVD outcomes (13,237 CHD; 7,603 stroke) in 294,998 subjects.
  • J-shaped associations were seen between all glycemic measures (A1C and fasting, random, postload glucose) and CVD risk after adjustment for conventional CVD risk factors
    • A slight change in HRs was seen after adjustment for total cholesterol, triglycerides or eGFR; this change was attenuated after adjustment for HDL-C or C-reactive protein.

Click on slide thumbnail to view larger. All slides available for download in the slide library.
Association of A1C With CVD Outcomes in Subjects With No Baseline Diabetes or CVD History   
Association of A1C With CVD Outcomes in
Subjects With No Baseline Diabetes or CVD History
 

CVD risk prediction 

  • Small changes in measure of risk discrimination that assesses how well a model can separate subjects who do/do not develop CVD (C-index) seen after glycemic measures added to CVD risk prediction models
  • Greatest change in C-index seen with A1C; improvement provided by A1C was equal to or better than other glycemic measurements, as shown in the table below
  • Despite addition of glycemic information to CVD risk prediction models, no significant improvements in net reclassification of 10-year CVD risk: net reclassification of subjects across 10-year risk categories improved by 0.42 (-0.63 to 1.48) with addition of A1C.

 

Glycemic measure Change in C-index when glycemic measure added to
conventional risk factors (95% CI)
A1C 0.0018 (0.003 to 0.0033); P<0.05
Fasting glucose 0.0013 (0.0007 to 0.0018); P<0.001
Random glucose 0.0005 (-0.0002 to 0.0013)
Postload glucose 0.0004 (-0.0001 to 0.0009)

 

 

Click on slide thumbnail to view larger. All slides available for download in the slide library.
A1C Measurement: Little Benefit for CVD Risk Prediction Among Subjects Without Diabetes or CVD   
A1C Measurement: Little Benefit for CVD Risk Prediction
Among Subjects Without Diabetes or CVD
 


The CVD models in the study  included age, sex, smoking, systolic BP level, total cholesterol, and HDL-C. Mean study levels: A1C=5.37%, fasting glucose=96 mg/dL, random glucose=99 mg/dL, postload glucose=125 mg/dL.  

CHD=coronary heart disease; eGFR=estimated glomerular filtration rate; HR=hazard ratio

   

May 2014 

This overview was created by KnowledgePoint360 Group, LLC, and was not associated with funding via an educational grant or a promotional/commercial interest.  

Previous Article
Changes in diabetes-related complications in the United States, 1990-2010  

Next Article
Bariatric surgery versus intensive medical therapy for diabetes--3-year outcomes    

Back to
Recent Diabetes Articles 

BG Footer

 

 

Last Modified: 8/4/2014