Is there a glycemic threshold for mortality risk?

OBJECTIVE: To determine whether there are thresholds for fasting and for 2-h glucose above which the risk of death from all causes and from coronary heart disease (CHD) increases. RESEARCH DESIGN AND METHODS: We studied 23-year mortality data from the Paris Prospective Study of the 7,018 men, aged 44-55 years, who were not known as diabetic at the baseline examination. The effect of glucose concentrations on mortality was studied using the observed relative risks and an age-adjusted Cox proportional hazards model. RESULTS: For all causes of death, there were J-shaped relationships with both fasting and 2-h glucose concentrations, and the lowest observed death rates were in the intervals centered on 5.5 mmol/l for fasting glucose and 5.0 mmol/l for 2-h glucose. The death rates for CHD were low in this population: for fasting glucose, the hazards ratio was best modeled by a positive linear relationship; for 2-h glucose, it was modeled by a J-shaped curve and the lowest observed death rate was in the interval centered on 6.0 mmol/l. CONCLUSIONS: In the Paris Prospective Study, there were no clear thresholds for fasting or 2-h glucose concentrations above which mortality sharply increased; in the upper levels of the glucose distributions, the risk of death progressively increased with increasing fasting and 2-h glucose concentrations.