Prediabetes Open Controversies in the Field

Why Prediabetes Is Scientifically Contested

The definition of prediabetes rests on numerical thresholds chosen partly by committee consensus, not purely by biological tipping points. This means the entire epidemiology of the condition shifts when a guideline committee moves a cut-off by even 0.1 percentage points on HbA1c.

Prediabetes is not a single pathophysiological entity. Some people carry elevated fasting glucose alone, some have isolated impaired glucose tolerance on oral glucose challenge, and some meet both criteria. These phenotypes carry different risks and may respond to different interventions, yet most clinical guidelines fold them into one label. That folding has generated decades of argument.

The Scale of the Problem

The CDC estimates that 98 million American adults, roughly 38% of the adult population, had prediabetes in 2021. [1] Only about 20% of them knew it. Whether this reflects a true epidemic or a definitional artifact is exactly what the field debates.

The ADA defines prediabetes as an HbA1c of 5.7 to 6.4%, a fasting plasma glucose of 100 to 125 mg/dL, or a 2-hour post-load glucose of 140 to 199 mg/dL during an oral glucose tolerance test. [2] The WHO sets the HbA1c floor higher, at 6.0%, which alone excludes tens of millions of people from the diagnosis. [3]

Why the Gap Between ADA and WHO Matters Clinically

A person with an HbA1c of 5.8% is, by ADA criteria, a prediabetic who may be counseled, screened, and potentially medicated. By WHO criteria, that same person is normal. The downstream consequences, psychological, financial, and clinical, are substantial. A 2019 BMJ analysis estimated that adopting the lower ADA threshold in countries that previously used WHO criteria could triple the number of adults classified as having prediabetes. [4]

Controversy 1: Are the Diagnostic Thresholds Evidence-Based?

The short answer is: partially. The lower boundary of the ADA's HbA1c range (5.7%) was set to maximize sensitivity for future diabetes risk, not to identify a threshold where complications begin. The WHO's threshold (6.0%) was set closer to where retinopathy data suggested a biological inflection point.

Where the Retinopathy Argument Stands

Retinopathy risk rises steeply above an HbA1c of roughly 6.5%, the standard T2D cut-off. Below that, the data are less clean. A 2011 analysis published in Diabetes Care found that microvascular changes detectable below 6.5% HbA1c were modest and their clinical significance unclear. [5] The WHO used this reasoning to set its floor at 6.0%, arguing that values below that carry negligible organ-level risk even if they predict future diabetes.

The Cardiovascular Risk Counterargument

ADA proponents argue that waiting for 6.0% misses an important cardiovascular prevention window. A 2019 meta-analysis of 129 prospective cohorts (N=10,036,945) published in The BMJ found that impaired fasting glucose and impaired glucose tolerance each independently associated with a 13% higher risk of cardiovascular events compared with normoglycemia. [6] That signal exists across the 5.7 to 6.0% range, which is the ADA's core justification for the lower cut-off.

The WHO would respond that association is not causation, and that treating a glycemic number rather than underlying cardiometabolic risk factors misattributes the intervention target.

Controversy 2: Does the Prediabetes Label Itself Cause Harm?

Several researchers argue that medicalizing borderline glycemia creates anxiety, overtreatment, and unnecessary healthcare costs without proportionate benefit. This is the overdiagnosis argument, and it has grown louder since the 2010 ADA threshold revision.

Evidence for Labeling Harm

A 2014 paper in JAMA Internal Medicine by Yudkin and Montori argued explicitly that the lowered ADA cut-off pathologizes normal variation and that the label "prediabetes" was partly a marketing construct that expanded the potential market for diabetes drugs. [7] They noted that 40% of people labeled as prediabetic will spontaneously revert to normoglycemia within 5 years without any intervention, a fact that rarely appears in patient counseling.

Evidence Against This View

The Diabetes Prevention Program (DPP, N=3,234) showed that people with impaired fasting glucose or impaired glucose tolerance, the two phenotypes that constitute prediabetes, did progress to T2D at a clinically meaningful rate in the placebo arm (11 cases per 100 person-years). [8] The lifestyle intervention reduced that rate by 58% and metformin by 31%, both P<0.001 vs. Placebo. Those are not trivial numbers. Withholding a label could also withhold access to structured prevention programs that insurers require a diagnosis code to cover.

The tension here is genuine: the same threshold that grants insurance coverage for prevention also creates 15 years of unnecessary worry for the 40% who would have reverted anyway. A risk-stratified labeling approach, using the label only when HbA1c exceeds 6.0% or when multiple cardiometabolic risk factors coexist, has been proposed in the literature but has not been adopted by any major guideline body as of early 2025.

Controversy 3: Is Prediabetes Truly Reversible, and What Does "Reversal" Mean?

The word "reversal" appears frequently in patient-facing content but has no agreed clinical definition. HbA1c below 5.7% after a period of elevated glucose? Sustained normoglycemia at 12 months? At 5 years? The absence of a standard definition makes it impossible to compare studies.

Regression Rates and Their Limitations

A 2017 systematic review in PLOS Medicine (N=circa 3,000 across studies) found that 33 to 59% of people with impaired glucose tolerance returned to normoglycemia over follow-up periods ranging from 1 to 11 years. [9] But many of those same people later progressed again. Glycemic status in the prediabetes range appears to fluctuate, which raises the question of whether "reversal" means cure or temporary reprieve.

The Obesity Connection

Because adiposity drives insulin resistance, weight loss reliably improves glycemia. The LOOK AHEAD trial found that intensive lifestyle intervention producing roughly 8.6% body weight loss at 1 year moved HbA1c meaningfully toward normal in participants with T2D. [10] Extrapolating to prediabetes is reasonable but not directly proven in equivalently powered trials. GLP-1 receptor agonists such as semaglutide 2.4 mg (STEP-1, N=1,961) produced 14.9% mean weight loss at 68 weeks, and post-hoc glycemic data from STEP trials suggest many participants with baseline prediabetes entered normoglycemic ranges. [11] Whether that represents durable metabolic change or pharmacologically suppressed glucose is an open question the field has not resolved.

Controversy 4: Should Metformin Be Prescribed Routinely for Prediabetes?

This is the most pharmacologically contentious debate in the prediabetes field. The ADA's 2024 Standards of Medical Care state that metformin "may be considered" for adults with prediabetes who are under age 60, have a BMI of at least 35 kg/m², or have a history of gestational diabetes. [2] That hedged language reflects genuine disagreement about whether the evidence justifies routine prescribing.

What the DPP Data Actually Show

In the DPP, metformin 850 mg twice daily reduced T2D incidence by 31% vs. Placebo over 2.8 years. [8] That effect was most pronounced in participants aged 25 to 44 and those with BMI above 35. In participants over age 60, metformin performed no better than placebo. The DPPOS (DPP Outcomes Study), which followed participants for 15 years, found the lifestyle group maintained a 27% reduction in T2D risk while the metformin group maintained an 18% reduction, suggesting lifestyle remains superior even in the long run. [12]

The Case Against Routine Prescribing

The AACE 2023 guidelines take a more conservative position than the ADA. They prioritize lifestyle modification as the primary intervention and recommend pharmacotherapy only when lifestyle goals are not achievable or sustained. [13] Their argument: most people with prediabetes who achieve 5 to 7% body weight loss through diet and exercise will reduce their T2D risk by an amount comparable to or exceeding metformin's benefit, without the GI side effects that cause roughly 5% of patients to discontinue the drug.

The Case for Earlier Prescribing

Some endocrinologists argue that the real-world adherence to lifestyle interventions is so poor that metformin should be the default rather than the fallback. A 2021 retrospective cohort study in Diabetes Care found that fewer than 4% of eligible prediabetes patients were prescribed metformin despite meeting ADA criteria for drug consideration, and that those who did receive it had measurably lower T2D conversion rates at 3 years. [14]

The ADA's 2024 guideline states: "Metformin therapy for prevention of type 2 diabetes should be considered in those with prediabetes, especially for those with BMI <35 kg/m², younger adults, and those with higher baseline HbA1c." [2] That sentence structure leaves substantial room for clinical judgment, which is a polite way of saying the evidence is not strong enough for a Class I recommendation.

Controversy 5: Cardiovascular Risk, Who Is Actually High-Risk?

Prediabetes is associated with higher cardiovascular disease (CVD) risk, but quantifying that risk and deciding how aggressively to manage it is unsettled. The 2019 BMJ meta-analysis cited earlier (N=10,036,945) put the hazard ratio for CVD at 1.13 (95% CI 1.06 to 1.21) for impaired fasting glucose and 1.20 (95% CI 1.12 to 1.29) for impaired glucose tolerance. [6] A 1.13 to 1.20 hazard ratio is real but modest compared to the 2 to 4-fold CVD risk associated with established T2D.

Does Treating Prediabetes Reduce CVD Events?

No randomized trial has demonstrated that intervening on prediabetes itself reduces hard cardiovascular outcomes (myocardial infarction, stroke, or cardiovascular death). The DPP was not powered for CVD events. The DPPOS found no statistically significant difference in CVD event rates between the lifestyle and metformin arms after 15 years. [12]

Implications for Statin and Aspirin Prescribing

Because the CVD risk attribution is uncertain, clinicians disagree about whether prediabetes alone should lower the treatment threshold for statin therapy. The American Heart Association and American College of Cardiology 2019 Guideline on Primary Prevention lists diabetes as a risk-enhancing factor for statin consideration but does not specifically address prediabetes as an independent indication. [15] That gap leaves clinicians using pooled cohort equations that do not include glycemic status, potentially undertreating some prediabetic patients with otherwise borderline 10-year ASCVD scores.

Controversy 6: Racial and Ethnic Disparities in Diagnosis

HbA1c is not equally accurate across racial groups. Black Americans have higher HbA1c values than White Americans at the same mean plasma glucose, likely due to differences in red blood cell turnover and glycation rates. This means HbA1c-based thresholds may overdiagnose prediabetes in Black adults and underdiagnose it in some Asian populations where diabetes-level complications occur at lower BMI thresholds.

A 2021 analysis in JAMA Internal Medicine found that among participants with identical fasting glucose values, Black adults had HbA1c values 0.15 to 0.4 percentage points higher than White adults. [16] Applying a uniform HbA1c cut-off of 5.7% without accounting for this difference misclassifies a non-trivial number of people.

The ADA's 2024 Standards acknowledge this limitation but do not yet provide race-specific thresholds, instead noting that clinicians should use fasting plasma glucose or oral glucose tolerance testing when HbA1c accuracy is in question. [2] The WHO has not updated its guidance on this issue as of the date of this article.

Controversy 7: GLP-1 Receptor Agonists, Should They Be Used for Prediabetes Prevention?

GLP-1 receptor agonists are approved for T2D management and, in the case of semaglutide 2.4 mg, for chronic weight management. Neither semaglutide nor any other GLP-1 agonist carries an FDA indication for prediabetes prevention as a standalone condition. [17]

What the Trial Data Show

The STEP-1 trial (N=1,961) demonstrated 14.9% mean body weight reduction at 68 weeks with semaglutide 2.4 mg vs. 2.4% with placebo. [11] Baseline HbA1c in that trial was 5.7 to 6.4% in a significant proportion of participants. A pre-specified analysis found that 84.1% of participants with prediabetes at baseline who received semaglutide reverted to normoglycemia at 68 weeks vs. 47.8% with placebo (P<0.001). [11]

The SELECT trial (N=17,604), published in NEJM in 2023, showed that semaglutide 2.4 mg reduced major adverse cardiovascular events by 20% in adults with established CVD and overweight or obesity but without diabetes. [18] A meaningful portion of that population would have met prediabetes criteria, though SELECT was not designed to test that subgroup specifically.

Why No Prediabetes Indication Exists Yet

The FDA requires a trial powered for a prediabetes-specific clinical endpoint, likely T2D incidence, to grant an indication. No such trial has been completed for a GLP-1 agonist. Cost-effectiveness concerns are also substantial. Metformin costs roughly $4, $10 per month for generic 850 mg BID dosing. Semaglutide 2.4 mg (Wegovy) listed at approximately $1,300 per month before insurance in 2024. Whether the superior glycemic and weight outcomes justify that cost differential for a condition that 40% of people exit spontaneously is a question payers and guideline bodies have not resolved.

Where Major Guidelines Currently Stand

No single table captures the full divergence between the ADA, WHO, AACE, and USPSTF on prediabetes, but the following summarizes the most clinically relevant disagreements as of 2025.

The USPSTF recommends screening for prediabetes and T2D in adults aged 35 to 70 who are overweight or obese, with a Grade B recommendation, meaning there is moderate certainty of substantial net benefit. [19] They do not specify a preferred screening test.

The ADA recommends screening beginning at age 35 (or earlier with risk factors), with repeat testing every 3 years if normal. [2] They accept HbA1c, fasting plasma glucose, or 2-hour OGTT as equivalent.

The AACE recommends lifestyle intervention as the primary treatment and pharmacotherapy (metformin) only if lifestyle goals are unmet. [13] They do not endorse GLP-1 agonist use for prediabetes outside of obesity-indicated use.

The Endocrine Society has not issued a standalone prediabetes guideline as of 2025 but its member statements generally align with the ADA's broader criteria while emphasizing individualized risk assessment.

The ADA's 2024 Standards of Medical Care state: "The progression from prediabetes to type 2 diabetes is not inevitable." [2] That sentence is now quoted by both sides of the overdiagnosis debate, which itself illustrates how contested this territory remains.