Normal-weight Non-Hispanic Asian adults have a 25.6% prevalence of dysglycemia — prediabetes or diabetes — compared to 16.5% among normal-weight Non-Hispanic White adults. That gap is striking on its own. But set it against this: obese Non-Hispanic White adults have a dysglycemia prevalence of 42.8%. The distance between a thin Asian American and an obese white American, on the most consequential metabolic measure in clinical medicine, is just 17.2 percentage points.

BMI-based screening thresholds were built on population data that skewed heavily white. The consequence isn't theoretical.

Normal Weight on the Scale, Elevated Risk in the Blood

Among normal-weight adults, Non-Hispanic Asian adults had a mean HbA1c of 5.58%, compared to 5.37% for Non-Hispanic White adults. That 0.21 percentage point difference sounds small, but it sits right at the threshold that separates routine monitoring from clinical concern. At the population level, it translates directly into the dysglycemia gap: 25.6% versus 16.5%, a 55% higher relative prevalence despite identical BMI ranges.

The triglyceride picture reinforces this. Normal-weight Non-Hispanic Asian adults had mean triglycerides of 107.3 mg/dL, versus 91.0 mg/dL for normal-weight Non-Hispanic White adults. HDL ran in the opposite direction: 56.8 mg/dL for Asian adults versus 61.4 mg/dL for white adults. Lower HDL and higher triglycerides together signal insulin resistance even when body weight looks unremarkable. A patient who clears a BMI-based screening threshold may still be accumulating cardiovascular and metabolic risk that goes unmeasured.

The diabetes component alone is telling. Among normal-weight Non-Hispanic Asian adults, 7.1% already have diabetes, compared to 2.0% among normal-weight Non-Hispanic White adults. That's more than a threefold difference within the same BMI category, in a population that standard clinical algorithms would classify as low-risk.

The Metabolic Risk Burden Hiding Behind a Normal BMI

Dysglycemia is one marker. The broader metabolic picture is worse. Among normal-weight Non-Hispanic Asian adults, only 42.9% had zero metabolic risk markers out of four assessed (elevated HbA1c, elevated triglycerides, low HDL, and elevated waist circumference). More than a third, 36.3%, had at least two.

The distribution breaks down this way: 20.8% had exactly one risk marker, 22.8% had exactly two, 7.8% had exactly three, and 5.7% had all four. In a population that would pass a standard BMI screen, more than one in three carries a metabolic risk profile that would prompt clinical attention if it appeared in a heavier patient. That's not a rounding error in the data. It's a structural gap in how risk is assessed.

The overweight category makes the pattern even sharper. Overweight Non-Hispanic Asian adults had the highest dysglycemia prevalence of any BMI group shown for Asians: 45.8%, with a mean HbA1c of 5.64%. For context, obese Non-Hispanic White adults had a dysglycemia prevalence of 42.8%. An Asian American who is merely overweight by standard BMI criteria carries more dysglycemia burden than a white American who is clinically obese.

What the Screening Gap Costs

The American Diabetes Association has recommended lower BMI thresholds for diabetes screening in Asian Americans (23 kg/m² instead of 25 kg/m²) since 2015. But implementation in clinical practice has been inconsistent, and the NHANES data here reflects the real-world result of that inconsistency.

Consider the 7.1% diabetes prevalence among normal-weight Non-Hispanic Asian adults. Some share of those individuals were likely undiagnosed before the survey, because they never met the BMI criteria that would have triggered a screening order. Undiagnosed diabetes means delayed treatment, more time accumulating end-organ damage, and higher downstream costs. The screening tool didn't fail to catch a rare edge case. It failed to catch a population where one in fourteen normal-weight adults already has diabetes.

The sample for the metabolic risk marker analysis is 94 adults, small enough to warrant some caution on the precise percentages. But the direction is consistent across every measure in this dataset: HbA1c, triglycerides, HDL, and composite risk all point the same way. A BMI of 22 means something different depending on who's standing on the scale.

Given that 25.6% of normal-weight Non-Hispanic Asian adults already meet the dysglycemia threshold, what proportion of those with diabetes were previously undiagnosed precisely because their BMI kept them off the screening list?