Overview
Intermediate hyperglycaemia, often called prediabetes, sits between normal glucose regulation and type 2 diabetes. It is not a single disease state. Some people have only mild changes in fasting glucose, while others already show more significant abnormalities after a glucose load or in HbA1c, the marker that reflects average blood sugar over the previous two to three months.
A new analysis from the ELSA-Brasil cohort suggests that staging intermediate hyperglycaemia more precisely may improve diabetes prevention. Instead of using a simple yes-or-no label of prediabetes, the researchers tested whether combining fasting plasma glucose with 1-hour plasma glucose from an oral glucose tolerance test, or with HbA1c, could better identify who is at highest risk of progressing to type 2 diabetes.
The main message is straightforward: fasting glucose plus 1-hour glucose provided a more detailed and clinically useful risk stratification than fasting glucose plus HbA1c. Adding a simple clinical risk score before ordering laboratory tests also reduced unnecessary testing while improving specificity.
Why staging prediabetes matters
For years, prediabetes has been used as a broad category to flag people at increased risk of diabetes. But that label can hide important differences. Two people may both meet a prediabetes threshold, yet one may progress slowly while the other is already close to diabetes and may also have a higher burden of cardiovascular and metabolic risk.
This matters because prevention efforts are most effective when they are targeted. Lifestyle interventions, weight loss, physical activity, and, in selected patients, medication can delay or prevent diabetes. To use these tools efficiently, clinicians need better ways to identify who is most likely to benefit.
HbA1c is widely used because it is convenient and does not require fasting. However, it may miss some people whose main problem is post-meal hyperglycaemia, the rise in glucose after eating or after a glucose challenge. The 1-hour plasma glucose measurement, taken during an oral glucose tolerance test, may capture this earlier metabolic impairment more effectively.
How the ELSA-Brasil study was designed
The investigators used data from 1,174 middle-aged and older Brazilian adults enrolled in the Rio Grande do Sul center of the ELSA-Brasil cohort. The analysis covered data collected from 2017 to 2024, and participants were followed for an average of 5.31 years.
The team built staging systems based on previously recommended thresholds for mild and moderate hyperglycaemia:
Mild hyperglycaemia by fasting plasma glucose: 5.6 to 6.0 mmol/L
Moderate hyperglycaemia by fasting plasma glucose: 6.1 to 6.9 mmol/L
Mild hyperglycaemia by 1-hour plasma glucose: 6.7 to 8.5 mmol/L
Moderate hyperglycaemia by 1-hour plasma glucose: 8.6 to 11.5 mmol/L
They also created a separate staging framework that combined fasting plasma glucose with HbA1c:
Mild HbA1c elevation: 39 to 41 mmol/mol, or 5.7% to 5.9%
Moderate HbA1c elevation: 42 to 46 mmol/mol, or 6.0% to 6.4%
In addition, the researchers tested whether a clinical risk score, with a cutoff of at least 10% probability of developing diabetes over 10 years, could be used first to decide who should proceed to laboratory testing.
To understand the biological meaning of the stages, they examined two important features at baseline:
Insulin responsiveness, measured by the insulin secretion-sensitivity index-2, or ISSI-2
Estimated risk of complications, using Whitehall subphenotype clusters
They then tracked new diabetes cases identified by self-report and glucose testing over follow-up.
What the study found
The fasting glucose plus 1-hour glucose approach divided participants into three stages with a clear metabolic gradient. As the stage advanced, insulin responsiveness fell and the proportion of people with a high estimated risk of complications increased.
This is important because it shows that the staging system was not just a statistical exercise. It aligned with worsening underlying physiology.
The risk of developing diabetes also rose steadily across stages. In crude analyses, the relative risk of incident diabetes reached 15.4 in stage 3. After adjustment for relevant factors, the relative risk remained high at 11.4. In practical terms, people in the highest stage by this schema were far more likely to develop diabetes during follow-up than those in the lowest stage.
The prognostic performance was also strong. The fasting glucose plus 1-hour glucose schema achieved 89.1% sensitivity and 53.7% specificity for predicting incident diabetes. High sensitivity means the model captured most future diabetes cases, while moderate specificity means some people were still flagged who did not go on to develop diabetes.
When a clinical risk score was used first, specificity improved to 60.3%, and the need for laboratory testing fell by 27.1%. This is a meaningful operational advantage, especially in settings where access to glucose testing or oral glucose tolerance testing is limited.
How HbA1c compared
The investigators also tested a staging system that combined fasting plasma glucose with HbA1c. Because fewer participants had abnormal HbA1c values, this approach naturally created fewer people in the higher stages and produced less pronounced risk separation.
Compared with the fasting glucose plus 1-hour glucose schema, the fasting glucose plus HbA1c schema showed lower relative risks and weaker prognostic performance. In other words, it was less effective at distinguishing between people at modest risk and those at much higher risk of future diabetes.
Still, the HbA1c-based approach performed better than a simple binary diagnosis of prediabetes. That finding is important because it suggests that staging, even with HbA1c, may be more informative than using a single cutoff.
However, the study indicates that HbA1c alone, or HbA1c added to fasting glucose, may not capture the full spectrum of intermediate hyperglycaemia as well as 1-hour glucose does.
What ISSI-2 and complication risk tell us
ISSI-2 is a measure that combines how well the pancreas secretes insulin with how sensitive the body is to insulin. A lower ISSI-2 suggests worsening beta-cell function and poorer metabolic compensation.
As the staging advanced, ISSI-2 declined, which supports the idea that the stages reflect true progression along the diabetes pathway.
The Whitehall subphenotype clusters are a way of estimating the likelihood of diabetes-related complications. The increasing frequency of a high-risk pattern across stages suggests that the most advanced hyperglycaemia stages may already be linked to broader metabolic harm, not just future diabetes alone.
Taken together, these biomarkers suggest that intermediate hyperglycaemia is not a uniform condition. Some people are earlier in the disease process, while others are already showing metabolic features associated with greater risk.
Clinical implications
This study has several practical implications.
First, it supports the idea of staging prediabetes rather than treating it as one undifferentiated category. A staged approach may help clinicians decide who needs intensive lifestyle support, closer monitoring, or additional testing.
Second, the 1-hour plasma glucose measurement appears to add important predictive value. Although the oral glucose tolerance test is more time-consuming than fasting glucose or HbA1c alone, it may uncover risk that would otherwise be missed.
Third, using a simple clinical score before laboratory testing could make prevention strategies more efficient. If a person’s estimated 10-year diabetes risk is low, clinicians may avoid unnecessary testing. If the score is high, laboratory confirmation and staging become more worthwhile.
Fourth, the findings reinforce the importance of individualized prevention. People in the highest stage may benefit from earlier and more intensive intervention, whereas those in lower stages may need monitoring and lifestyle counseling but not the same level of urgency.
What this means for patients and clinicians
For patients, the study underscores that “prediabetes” is not a fixed or harmless state. It is a warning sign, but the degree of risk can vary widely. A person with mildly elevated fasting glucose may have a very different outlook from someone whose glucose rises sharply after a glucose load.
For clinicians, the study suggests that relying only on fasting glucose or HbA1c may miss an important portion of at-risk individuals. Where feasible, adding a 1-hour glucose measurement during an oral glucose tolerance test could improve risk assessment.
That said, this does not mean everyone needs more testing. In settings with limited resources, a clinical risk score may help identify who is most likely to benefit from laboratory evaluation.
Limitations to keep in mind
As with any observational cohort study, this analysis cannot prove that one test or stage causes better outcomes. It shows association and predictive value, not direct treatment benefit.
The study was conducted in a Brazilian cohort of middle-aged and older adults, so results may not fully generalize to younger populations or to settings with different ethnic, genetic, or healthcare characteristics.
Also, while 1-hour plasma glucose showed strong performance here, it is not yet as universally used in routine care as fasting glucose or HbA1c. Implementation would require practical workflows, clinician familiarity, and local validation.
Bottom line
The ELSA-Brasil study suggests that staging intermediate hyperglycaemia with fasting glucose plus 1-hour glucose provides a more refined and clinically useful way to predict type 2 diabetes than fasting glucose plus HbA1c. The 1-hour glucose approach better separated people by metabolic severity, identified future diabetes more accurately, and aligned with worsening insulin function and higher complication risk.
Adding a simple clinical risk score before laboratory testing improved specificity and reduced testing burden. Overall, the findings support a more nuanced, staged model of diabetes prevention rather than a single prediabetes label.
If confirmed in other populations, this approach could help clinicians target prevention efforts more precisely, detect high-risk patients earlier, and use healthcare resources more efficiently.
