A dementia risk score specific to patients with type 2 diabetes was derived and validated by a recent study.
Researchers used longitudinal data from the Kaiser Permanente Northern California Diabetes Registry, including almost 30,000 diabetics age 60 and older, to create a prediction model for development of dementia over 10 years. The model was then validated in a separate cohort of more than 2,000 similar patients from Washington State. Results were published by The Lancet Diabetes and Endocrinology on Aug. 20.
The researchers developed a risk score including the factors that most predicted development of dementia in the studied patients: microvascular disease, diabetic foot, cerebrovascular disease, cardiovascular disease, acute metabolic events, depression, older age and less education. Point values were assigned to each factor, and patients were stratified into 14 risk categories, to create what the researchers called the type 2 diabetes-specific dementia risk score (DSDRS).
Patients with the lowest scores had a 10-year dementia risk of 5.3% compared to 73.3% in the highest-risk group. The C-statistic (a statistical measure indicating the predictive strength of a model, with 0.5 being the same as chance and values approaching 1.0 indicating a stronger predictive ability) for the DSDRS was 0.733 to 0.744, which is higher than other commonly used scores like the Framingham score, the researchers noted. The C-statistic for age alone was almost as high, however. This may lead to uncertainty about the value of the score, an accompanying editorial acknowledged, but use of the full score would show, for example, that a 60-year-old with diabetes complications and cardiovascular disease has similar dementia risk to an 80-year-old without those comorbidities.
Because the score uses easily gathered information, it may be useful in primary care to identify the diabetes patients who should be watched most vigilantly for cognitive deterioration and protected from hypoglycemia (which has been associated with cognitive impairment by other research), the study authors concluded. Ideally, this first dementia risk score for diabetes patients could eventually lead to development of measures to predict dementia earlier in life, which would allow more preventive action and motivate lifestyle changes, the accompanying editorial concluded.