Prediabetes: What to Eat to Reverse It Before It Becomes Type 2 Diabetes
By Dino Pohilj, Founder, RealFoods
What prediabetes actually means
The diagnostic labels fasting glucose and HbA1c are lagging indicators. By the time fasting glucose rises above 100 mg/dL, the pancreas has typically been overproducing insulin for years to compensate for growing cellular resistance. The glucose appears normal on tests until it does not, because insulin is quietly covering the gap.
Dr. Jason Fung, nephrologist and author of The Diabetes Code, argues that this overproduction of insulin is the disease, and that elevated fasting glucose is merely the symptom that becomes detectable late in the process. We argue this reframe matters practically: it means the intervention window for dietary reversal begins well before the fasting glucose crosses the 100 mg/dL threshold, and it remains open into the early prediabetes range.
The most sensitive early test is fasting insulin, not fasting glucose. A fasting insulin above 10-12 uIU/mL, combined with borderline fasting glucose, is a stronger indicator of insulin resistance than either measurement alone. We recommend asking your clinician for a fasting insulin test at your next routine bloodwork if you have any of the risk factors: family history of type 2 diabetes, abdominal weight gain, post-meal fatigue, carbohydrate cravings, or elevated triglycerides.
Why the standard advice often fails
The typical prediabetes advice from a general practitioner is "lose weight, eat less, exercise more." This guidance is not wrong, but it is incomplete in a way that matters. The reason the same calorie deficit produces different outcomes in different people is that the foods creating the deficit have different hormonal impacts.
A 1,600-calorie day built on whole-wheat bread, low-fat yogurt, fruit juice, and granola produces a very different insulin environment than the same 1,600 calories built on eggs, vegetables, salmon, olive oil, and legumes. The first keeps insulin chronically elevated despite the deficit. The second allows insulin to fall low enough for fat oxidation and insulin sensitivity restoration to occur.
Dr. Robert Lustig, professor of paediatric endocrinology at UCSF, sources the same argument in his clinical research: the hormonal environment produced by food determines whether the body can heal insulin resistance, and that environment is driven by the type of carbohydrate, not the total calorie count. We argue the clinical implication is clear: optimise for insulin response, not calorie balance.
The foods that accelerate prediabetes
- Liquid sugar in all forms. Fruit juice, soda, sweetened coffee drinks, flavoured oat milk, energy drinks. These deliver glucose and fructose without fibre, producing immediate insulin spikes with no satiety signal. They are the single fastest-acting dietary driver of insulin resistance.
- Refined grains eaten without protein or fat. White bread, instant oats, white rice, crackers, pasta in large portions. The glycaemic impact of these foods is dramatically lower when eaten with protein, fat, and fibre, but very high when eaten in isolation.
- Ultra-processed breakfast foods. Most commercial cereals, granolas, and breakfast bars are refined starch plus sugar plus a health claim. A bowl of granola with skim milk and orange juice can produce an insulin response equivalent to several slices of white bread.
- Low-fat flavoured dairy. The fat reduction is offset by added sugar to maintain palatability, resulting in a higher glycaemic impact than the full-fat version. Full-fat Greek yogurt without added fruit is consistently a lower-insulin-impact option than low-fat fruit yogurt.
- Snacking between meals. Every time insulin rises, the fat-burning signal switches off. Frequent snacking, even on "healthy" foods, keeps insulin elevated across the day and prevents the recovery windows that allow insulin sensitivity to improve.
Ready to stop guessing what to eat?
The 5-step reversal protocol
Step 1: Get your baseline numbers
Request fasting glucose, HbA1c, and fasting insulin at your next blood draw. If your fasting insulin is above 10-12 uIU/mL, you have insulin resistance even if your fasting glucose looks borderline. This baseline is your starting point and your progress measurement tool.
Step 2: Find and remove your highest-scoring meals
Photograph your meals for one week with RealFoods. Identify the meals scoring above 6 on the Weight Impact Scale. These are the meals producing the insulin spikes that are maintaining your insulin resistance. Replacing even one high-scoring meal per day with a low-scoring alternative creates a meaningful cumulative reduction in the daily insulin load.
Step 3: Restructure meal composition
Eat protein and fat before carbohydrates at every meal. Fill half the plate with non-starchy vegetables. Pair any remaining carbohydrates (rice, bread, fruit, legumes) with adequate protein and fat to blunt the glucose curve. Jessie Inchauspé's research has demonstrated that this sequencing alone reduces post-meal glucose peaks by 36-73% with no change in total food quantity.
Step 4: Move after meals
Skeletal muscle absorbs glucose without insulin when contracting. A 10-minute walk after eating produces a meaningful reduction in post-meal blood sugar. The DPP demonstrated that 150 minutes of moderate weekly activity was equivalent to metformin in reducing diabetes risk. Walking after meals is the most accessible delivery mechanism for that activity.
Step 5: Keep a daily average WI Score below 5
The Weight Impact Score translates the complex interaction of carbohydrate type, protein, fat, and fibre into a single per-meal number. A daily average below 5 corresponds to the dietary glycaemic load target that research has associated with meaningful improvements in insulin sensitivity. This is the number to track, not calories.
What the landmark trials actually found
The Finnish Diabetes Prevention Study (Tuomilehto et al., NEJM 2001) randomised 522 middle-aged adults with impaired glucose tolerance to a lifestyle intervention or control. After four years, the lifestyle group had a 58% lower incidence of type 2 diabetes. The intervention focused on weight reduction, reduced fat intake, increased fibre, and increased physical activity. Notably, participants who achieved at least four of the five targets had zero cases of diabetes develop during the trial.
The U.S. Diabetes Prevention Program (Knowler et al., NEJM 2002) replicated this: 3,234 participants with impaired fasting glucose were randomised to lifestyle intervention, metformin, or placebo. Lifestyle intervention reduced diabetes incidence by 58%; metformin reduced it by 31%. Lifestyle beat medication.
We argue the mechanism behind both results is insulin suppression through lower dietary glycaemic load, not the calorie reduction itself. The food-composition changes recommended in both trials all work by reducing the insulin response per meal. That is the mechanism the Weight Impact Score is designed to target directly.
Where medication fits
Metformin reduces hepatic glucose production and modestly improves insulin sensitivity. GLP-1 receptor agonists (semaglutide, tirzepatide) reduce appetite, slow gastric emptying, and produce meaningful improvements in glycaemic control. Both are appropriate for some people with prediabetes, particularly those with a strong family history or elevated cardiovascular risk. We do not argue against them. We argue that the food environment that produced the insulin resistance remains the primary lever, and that medication works best when the dietary insulin load is also being managed. Medication modulates the signal at the margins. Food produces or removes it at the source.
Continue reading: what insulin resistance is and how it develops, how to lower insulin naturally without medication, or how blood sugar drives stubborn weight gain.
