What is a person’s risk of developing diabetes in 10 years? A tool created by researchers at the Federal University of Rio Grande do Sul, Porto Alegre, Brazil; the Federal University of Espírito Santo, Vitória, Brazil; the Federal University of Minas Gerais, Belo Horizonte, Brazil; and the Oswaldo Cruz Foundation, Rio de Janeiro, Brazil, aims to answer that question.

Developed using data from 15,105 adults enrolled in the Longitudinal Study of Adult Health (ELSA-Brasil) — a multicenter cohort study that followed employees of universities and research institutions in six Brazilian states for an average of 7.4 years — the Brazilian Diabetes Risk Score (BrDMrisc) also shows more favorable diagnostic properties than other diabetes screening strategies, according to a study published in Revista de Saúde Pública (Journal of Public Health).

The calculator uses 12 clinical and laboratory characteristics: age, BMI, waist circumference, hypertension, ethnicity, sex, family history of diabetes, fasting plasma glucose, 2-hour plasma glucose after an oral glucose tolerance test, glycated hemoglobin, triglycerides, and high-density lipoprotein cholesterol. Although the score is based on these 12 characteristics, it is not necessary to have information for all of them to calculate risk.

“We have 27 different mathematical models that allow risk to be estimated with whichever variables are available for the individual and/or clinician. In practice, this means we can estimate 10-year risk with only fasting plasma glucose, for example, or only clinical data,” said Paula Andreghetto Bracco, PhD, statistician and biomedical scientist and assistant professor at the Federal University of Rio Grande do Sul and the study’s first author.

The tool’s versatility allows different approaches to be used. Risk can be initially calculated using clinical data available during the consultation. If risk appears high, laboratory tests can be ordered to refine the estimate, combining those results with clinical characteristics in a new calculation.

In an interview with Medscape Medical News, Bracco said that no single cutoff is defined to identify high risk, but the tool allows comparison of an individual’s estimated risk with that of the more than 15,000 ELSA-Brasil participants.

The researchers used a 20% risk cutoff point over 10 years as a reference — a value close to the cumulative risk observed in patients with impaired glucose tolerance in major diabetes prevention trials, which showed a 58% reduction in disease incidence with lifestyle intervention.

BrDMrisc is available on the ELSA-Brasil website. However, Bracco noted the group is still in contact with other cohorts to perform external validation.

Examples of Use

The article illustrates the tool’s versatility. In the first example, a 44-year-old White woman with a BMI of 28, a waist circumference of 88 cm, hypertension, but no family history of diabetes would have an 11% 10-year risk based on clinical data alone. Adding a fasting plasma glucose of 113 mg/dL would raise her recalculated risk to 35%. Based on the 20% cutoff used in the study and the risk observed in prevention trials, a 35% risk would justify preventive intervention, whether through structured lifestyle changes or pharmacotherapy.

The study also shows how risk changes continuously. For a 50-year-old White man with a BMI of 28, a waist circumference of 100 cm, and no hypertension or family history, the estimated risk would be 10% with a fasting plasma glucose of 100 mg/dL, but it would jump to 26% with a fasting glucose of 110 mg/dL and to 63% with a fasting glucose of 123 mg/dL.

Another example demonstrates the impact of BMI on risk. For a 50-year-old White man with a waist circumference of 100 cm and no hypertension or family history, the 10-year risk would range from 10% with a BMI of 23 to 30% with a BMI of 40.

Advantages in Screening

Currently, other strategies are used to screen for diabetes, including fasting plasma glucose (≥ 100 mg/dL) and the Finnish Diabetes Risk Score (FINDRISC) questionnaire.

According to Bracco, the main limitation of screening based on a single test, such as fasting glucose, is that anyone above the cutoff is classified the same way. When we add continuous values into the calculations, we allow risk to differ between a patient with a fasting glucose of 105 mg/dL and one with 115 mg/dL.

Regarding the FINDRISC questionnaire, the author noted that although it has been translated and adapted for use in Latin America, it was developed using data from a country with economic, social, and cultural differences from Brazil, and no study has reported external validation of its effectiveness here.

In that sense, for application in our clinical context, BrDMrisc has a considerable advantage because it was developed with entirely Brazilian data, the researcher said, noting that the ELSA-Brasil sample included diverse segments of Brazilian society in terms of sex, age, ethnicity, and education, among others.

She also pointed out that the meaning of a “positive” fasting glucose varies significantly depending on other factors, such as BMI. According to the article, depending on clinical characteristics, a person with a fasting plasma glucose of 100 mg/dL could have a 10-year risk ranging from 6% to 20%. With a fasting glucose of 110 mg/dL, that risk could range from 16% to 46%.

Bracco and colleagues found, for example, that the BrDMrisc calculation based on clinical data plus fasting plasma glucose compared with detection using fasting glucose alone (≥ 100 mg/dL) identified a more manageable fraction of the population as high risk (20.0% vs 40.6%), and those identified had about twice the risk of developing diabetes (28.5% vs 17.1%).

Compared with FINDRISC, which identifies only 8.4% of the population as high risk but detects just 31.8% of future cases, BrDMrisc strikes a better balance: it identifies 20% of the population as high risk while maintaining a sensitivity of 63.1% for detecting future cases.

“Using a score such as FINDRISC followed by laboratory testing improves interpretation of a positive result. However, a better risk estimate is obtained by combining all available factors into the risk calculation, as BrDMrisc does,” the author said.

Another advantage of the new tool, according to the researcher, is that its output is immediately interpretable — the estimated 10-year risk of developing diabetes. “Knowing a positive FINDRISC result or a single elevated fasting glucose does not provide this risk estimate,” she said.

What Other Experts Think

According to Rodrigo Moreira, MD, PhD, director of the Diabetes Mellitus Department at the Brazilian Society of Endocrinology and Metabolism, BrDMrisc emerges as another option. “This is an initial study that simply tested this calculator in one population. It appears to work well and may help us identify patients at higher risk of diabetes.”

However, he added that more studies are needed. “We use the calculator, we identify the patient at higher risk, and now what is the next step? That is a question that remains unanswered.”

“We need multiple studies that truly evaluate the impact of this tool, especially in changing the prognosis. That is, if I use this tool, identify patients, and treat them, will that reduce the risk of diabetes?” Moreira asked.

He said this is an initial study of a calculator that could be useful in the future, but further research is needed to learn how to use it and understand what its impact will be in our clinical practice and prevention, especially to establish treatments that prevent or delay the onset of diabetes.

This story was translated from Medscape’s Portuguese edition.