Glucose Levels Linked to Dementia, Even in Nondiabetics

Gregory A. Nichols, PhD


November 14, 2013

Glucose Levels and Risk of Dementia

Crane PK, Walker R, Hubbard RA, et al
N Engl J Med. 2013;369:540-548

Study Summary

The Adult Changes in Thought (ACT) study included randomly selected dementia-free members of Group Health Cooperative (GHC) who were 65 years of age or older. The sample was limited to 2067 participants who had at least 1 follow-up visit, had been enrolled in GHC for at least 5 years before study entry, and had at least 5 measurements of glucose or glycated hemoglobin over the course of 2 or more years before study entry.

Participants were invited to return at 2-year intervals for the purpose of identifying incident cases of dementia, which were assessed using the Cognitive Abilities Screening Instrument and additional evaluation to confirm low scores. Fasting and random glucose levels and glycated hemoglobin levels were transformed to compute a time-varying estimate of average glucose levels at baseline and subsequent 5-year intervals. Diabetes was defined by receipt of at least 2 antidiabetic medications.

Data were analyzed with Cox regression models stratified by diabetes status, controlling for age; sex; education level; exercise level; blood pressure; and presence of coronary artery disease, smoking, atrial fibrillation, and treatment for hypertension.

The median age of participants was 75 years, and 59% were women. Over a median follow-up period of about 7 years, dementia developed in 25.4% of participants, including those who did not have diabetes and those who had developed diabetes by the end of follow-up.

Among those without diabetes, the adjusted hazard ratio for dementia for patients with an average glucose level of 115 mg/dL compared with 100 mg/dL was 1.18. Among participants with diabetes, the adjusted hazard ratio for dementia for patients with an average glucose level of 190 mg/dL compared with 160 mg/dL was 1.40. Although it was not statistically significant, participants with diabetes and an average glucose level of 150 mg/dL also had a higher hazard ratio compared with 160 mg/dL.


This elegantly conducted study demonstrated that an increased risk for dementia was associated with higher glucose levels in populations with and without diabetes. The approach, which combined different measurements into a single composite estimate of glucose exposure, was innovative and allowed for more inclusive long-term follow-up.

Stratifying by diabetes did not allow for an overall assessment of the association across the full spectrum of glucose levels, but because diabetes is defined by hyperglycemia and could thus have dominated the association, that was probably a necessary decision. More disappointing was the definition of diabetes, because it depended on medication use only. GHC certainly has the data to more accurately identify diabetes and has done so for many years.[1]

This is potentially extremely important because the misclassification of people with diabetes as being without diabetes, who presumably would have been at the upper end of the glucose range for that stratum, could have driven those results. Furthermore, although it was not significant, the analysis of the diabetes stratum showed a possible U-shaped relationship between glucose levels and dementia. The misclassification probably resulted in the exclusion of individuals with diabetes who were at the lower end of the glucose range for that stratum.

Because recent analyses have suggested a U-shaped association between A1c and cardiovascular disease and mortality,[2,3] similar relationships between glucose and other outcomes take on considerable meaning. Nevertheless, this is an important study that deserves attention. Further research is certainly needed to determine whether lowering glucose levels, even in the nondiabetic range, will reduce dementia risk.



Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.