Fasting insulin

Almost everything else on this list is, at bottom, a way to do this. Losing fat — especially the visceral fat around your organs — is what drove the only A-grade hard outcomes here: the Diabetes Prevention Program cut new diabetes by 58% with lifestyle change,⁸ and DiRECT put established type 2 diabetes into remission in nearly half of people through weight loss alone.⁹ The honest caveat: preventing diabetes is proven; buying heart protection is not — Look AHEAD lost the weight, improved the metabolics, and still found no reduction in cardiovascular events.¹⁰ If you’re already lean, this lever isn’t yours — skip to the A-tier.

Muscle is your largest glucose sink, and building it improves insulin sensitivity independent of weight loss — a network meta-analysis puts the HOMA-IR effect around d −0.25, larger with heavier, longer training.¹⁵ Two or three sessions a week is the highest-leverage thing most desk-bound people aren’t doing.

One of the larger single dietary levers: a meta-analysis of low-carb trials found a pooled HOMA-IR drop of about SMD −0.66.¹³ But be honest about the spread — it’s heterogeneous, and the advantage all but vanishes when both groups also exercise.¹⁴ You don’t need full keto; cutting refined starch, sugary drinks, and ultra-processed carbs captures most of it.

Roughly on par with resistance training for insulin sensitivity, and combining the two is the strongest option of all (combined training: HOMA-IR MD −0.52).¹⁵ The popular “4 × 45 min of Zone 2 a week” dose is a sensible convention, not a trial-derived prescription — the point is regular, sustained aerobic work.

Underrated because it’s free. A single night of short sleep cut insulin sensitivity about 20% in healthy volunteers,¹⁷ and a week at five hours a night did the same chronically.¹⁸ These are small mechanistic studies, hence the C — but the direction is unambiguous, and you can act on it tonight.

A 10–15 minute walk after meals flattens the post-meal glucose spike — and three short walks beat one long one for 24-hour glucose.¹⁶ The direct fasting-insulin data is thinner (hence C), but the effort is near-zero and the metabolic logic is sound. Easiest habit on the page.

For someone with real obesity, these produce weight loss — and insulin improvement — that lifestyle often can’t match, and SELECT showed a 20% drop in cardiovascular events.¹² The grade is A for that population (SELECT enrolled people with prior heart disease); for a lean primary-prevention reader it’s a B-grade extrapolation. Lift while you take it — some of the loss is muscle. Roughly $1,000+/month branded.

Cheap, old, and proven: it cut new diabetes 31% in the DPP⁸ and all-cause mortality 36% in overweight type 2 diabetics in UKPDS 34.¹¹ Tier B only because it’s a prescription and that hard-outcome evidence sits in at-risk groups, not healthy people chasing an optimal number. A reasonable conversation if you’re prediabetic or have PCOS.

Don’t dismiss it: the meta-analytic HOMA-IR effect (−0.85) is genuinely meaningful, bigger than most supplements manage.¹⁹ The catch isn’t the effect size — it’s that the trials are in patients, product quality varies wildly between bottles, there are no hard-outcome data, and the safety record is far thinner than metformin’s. A real add-on, not “nature’s metformin.”

A reasonable, low-risk option in PCOS, where insulin resistance is common — but be straight about the evidence: the synthesis behind the 2023 International PCOS Guideline rated the metabolic data “limited and inconclusive.”²⁰ It’s a conditional, shared-decision recommendation, and there’s little reason to expect it to help insulin outside PCOS.

The internet’s favourite metabolic hack does blunt a single meal’s glucose — but it doesn’t fix your baseline. A 2025 meta-analysis found HOMA-IR unchanged and fasting insulin actually rose slightly.²¹ Fine as a condiment; useless as an insulin intervention.

Cinnamon’s effect on fasting insulin is inconsistent; chromium helps mainly if you’re deficient, and most people aren’t; exogenous ketones raise blood ketones without improving insulin sensitivity. Spend the money and attention on the A-tier instead.