Insulin Sensitivity — A Sardine Protocol Dossier

TL;DR

The insulin-sensitivity story is the strongest evidence base in the Sardine Protocol's bundle of mechanisms. Repeated short fasts in healthy adults produce clamp-measured insulin sensitivity improvements of approximately 15–20% across two-week protocols ¹. In pre-diabetic adults, early time-restricted feeding (independent of weight loss) improves insulin sensitivity, blood pressure, and oxidative stress markers across five-week protocols ². In type-2 diabetic adults, sustained ketogenic dietary patterns produce T2D remission/reversal at scale ^{3 4}. The Twin-Cycle Hypothesis ⁵ and the DiRECT trial ⁶ provide the integrative framework for why short-term substantial calorie deficit produces durable insulin-sensitivity remodeling. The popularly cited "57% insulin drop in 48 hours" claim could not be traced to a specific primary study in our search; the actual measured magnitudes in the primary literature are more modest and the dossier walks through what we believe its real provenance is.

What we mean by insulin sensitivity

Insulin is a hormone secreted by pancreatic beta cells in response to elevated blood glucose, certain amino acids, and incretin signals (GLP-1, GIP). Insulin signals to peripheral tissues — primarily liver, skeletal muscle, and adipose — to take up glucose, suppress hepatic glucose output, and store substrate as glycogen and triglyceride. Insulin sensitivity is the measure of how much glucose disposal a given concentration of insulin produces. Insulin resistance is the inverse — high circulating insulin producing limited glucose disposal.

Several measurements operationalize this:

• Fasting insulin. Cheap, widely available, but noisy. Truly insulin-sensitive adults typically have fasting insulin under 7 µIU/mL; clinical insulin resistance typically presents above 12–15 µIU/mL.
• HOMA-IR. Calculated from fasting glucose × fasting insulin / 405. Crude but widely used. Below 1.0 is sensitive; above 2.5 is likely insulin resistant.
• OGTT 2-hour insulin and Kraft pattern. A 2-hour glucose-tolerance test with insulin measurement at multiple timepoints. More informative than fasting alone.
• Hyperinsulinemic-euglycemic clamp. Gold standard. Insulin is infused at a fixed concentration while glucose is titrated to maintain euglycemia; the glucose infusion rate measures whole-body insulin-mediated glucose uptake. The Halberg 2005 study ¹ uses this method.

Insulin resistance precedes type-2 diabetes by years to decades. The Twin-Cycle Hypothesis ⁵ frames T2D as a state in which excess fat accumulation in liver and pancreas produces a self-reinforcing loop of impaired insulin secretion and increased gluconeogenesis. Removing the fat via substantial calorie deficit can reverse the loop, often dramatically — a mechanistic foundation for the DiRECT trial ⁶ demonstrating ~46% T2D remission rates at 12 months in an intensive-weight-management arm.

What the evidence says (the public preview cuts here)

What actually happens to insulin and insulin sensitivity during a short fast:

Hour 0–12 (post-prandial → fed-fasted transition): Insulin falls as glucose is cleared. By hour 8–12, fasting insulin reaches a normal trough.

Hour 12–48 (fed-fasted switch → ketosis onset): Liver glycogen depletes; gluconeogenesis ramps up. Insulin remains low, glucagon rises. Counter-regulatory hormones modestly elevate. Adipose lipolysis accelerates. The substrate biology is mapped in detail in the classical starvation work ⁷ and confirmed across the modern intermittent-fasting literature ^{8 9}.

Hour 48–120 (sustained fasted state): Insulin remains low; glucose stabilizes around 70–90 mg/dL via gluconeogenesis. Ketones rise to nutritional ketosis range. Substrate for tissues is now predominantly fatty acid and ketone, not glucose.

Does a single fast improve insulin sensitivity? This is where the literature is most often misread. The careful answer:

• A few studies, including the Halberg 2005 paper ¹, show that repeated short fasts (alternate-day or every-other-day 20-hour fasts) over 15 days improve clamp-measured insulin sensitivity by ~16% in healthy young men. This is one of the cleanest pieces of evidence we have.
• Sutton 2018 ² extends this with a tightly controlled supervised-feeding trial: eating-window timing alone (early time-restricted feeding, 6-hour window with last meal before 3pm) improved insulin sensitivity in pre-diabetic men, isolated from weight loss, over 5 weeks. This is one of the strongest demonstrations that the biology is not solely about calorie deficit.
• Heilbronn 2005 ¹⁰ and Catenacci 2016 ¹¹ compare alternate-day fasting against continuous caloric restriction; effects are broadly comparable for insulin-sensitivity endpoints when total energy deficit is matched.
• Klein and Wolfe 1992 ¹² shows that carbohydrate restriction reproduces much of the substrate biology of fasting, suggesting the insulin-sensitivity gains during a sardine fast come substantially from the very-low-carbohydrate substrate pattern as well as from the calorie deficit.

For T2D specifically:

• Borgundvaag 2021 ¹³ meta-analyzed intermittent fasting in T2D; effects on A1c are generally small to moderate and broadly comparable to other dietary interventions producing similar weight loss.
• The Virta Health 2-year cohort ³ is the strongest case for a sustained ketogenic-based intervention producing durable T2D remission/reversal at scale, though it is a non-randomized cohort, not an RCT.
• Westman 2008 ⁴ is one of the early RCTs of low-carbohydrate eating in T2D, with strong effects on A1c and insulin requirement reduction.
• The DiRECT trial ⁶ demonstrates that substantial calorie deficit (in a non-ketogenic, low-calorie liquid-formula intervention) produces T2D remission, independent of mechanism — supporting the Twin-Cycle hypothesis ⁵.

The omega-3 contribution:

The Akinkuolie 2011 meta-analysis ¹⁴ and the Cochrane T2D omega-3 review ¹⁵ identify a small but real omega-3-driven insulin-sensitivity effect, more pronounced in insulin-resistant populations and at chronic dosing > 8 weeks. This is the most speculative piece of the sardine-protocol-specific insulin story.

The "57% insulin drop in 48 hours" claim:

This number is widely repeated in popular fasting content, often attributed (loosely) to Jason Fung. We could not trace it to a specific primary clamp or AUC study in our search. Halberg 2005's clamp data shows ~16% improvement in insulin-mediated glucose uptake after 15 days of every-other-day fasting; this is an order of magnitude smaller and on a different timescale than "57% in 48 hours." The most likely origin is either (a) an extrapolation from fasting-insulin trough data combined with a particular baseline assumption, (b) a conflation of fasting insulin level decline with insulin-mediated-glucose-uptake change (different things), or (c) a specific patient case-study being generalized. We have not been able to identify a single peer-reviewed study that establishes "57% in 48 hours" as a measured population-level effect, and we treat round dramatic numbers in popular fasting content as motivational rather than literal.