Insulin

A two-chain, 51-amino-acid protein hormone secreted by pancreatic beta cells and the first peptide drug ever given to a patient, in 1922; dozens of recombinant human and analog products now hold FDA approval for diabetes.

Also referenced as: Human Insulin, Insulin Human, Regular Insulin, Insulin Glargine, Insulin Lispro, Insulin Aspart, Insulin Degludec, Insulin Icodec, Humulin, Novolin, Lantus, Humalog, NovoLog, Tresiba

Also appears in: Hormone

COAPublic product evidenceSearch the public certificate ledger for this compoundNo exact compound records are currently indexed under this profile name.→

---

What is Insulin?

Insulin is a genuine peptide hormone — at the upper end of the range, where the word “peptide” shades into “small protein.” Mature human insulin is two chains held together by disulfide bonds: an A chain of 21 residues and a B chain of 30, joined by two interchain bridges (A7–B7 and A20–B19) with a third bridge internal to the A chain (A6–A11), totaling 51 amino acids and about 5.8 kDa. It is not synthesized in that form. Beta cells make preproinsulin, cleave the signal sequence to proinsulin, and then excise the connecting C-peptide with prohormone convertases and carboxypeptidase E, releasing insulin and C-peptide into the circulation in equal molar amounts — a detail that turns out to matter forensically, and is covered under Key considerations.

It is the founding peptide drug. Pancreatic extract was given to a patient in Toronto in January 1922, and Banting and Best published the underlying work that year (Journal of Laboratory and Clinical Medicine, 1922;7:251–266). Insulin was also the first protein whose complete amino acid sequence was determined — Sanger and Tuppy worked out the B chain first (Biochemical Journal, 1951;49:463–490), Sanger and Thompson the A chain in 1953, and the disulfide assignments completed the structure in 1955. And it was the first medicine of any kind produced by recombinant DNA: FDA approved Humulin in October 1982.

Its place in the research-peptide market is unusual and worth stating precisely. Insulin is not a catalog compound. No vendor on this site’s pricing board sells it — the listings that surface under a search for “insulin” are insulin syringes, an injection supply. Insulin does not need a gray channel to be obtainable, because in most U.S. states human insulin is sold at pharmacy counters without a prescription. It appears in this reference set because it circulates in bodybuilding and physique contexts as pharmacy-purchased or gray-import vials used outside any clinical indication, and because the compounds that surround it in vendor catalogs — IGF-1 variants, growth hormone secretagogues — are routinely confused with it.

How it works

• Receptor binding and kinase activation. The insulin receptor is a disulfide-linked α2β2 tyrosine kinase that exists as a pre-formed, autoinhibited dimer. Cryo-EM of the fully liganded complex shows up to four insulins binding one receptor dimer across two structurally distinct site types; that occupancy collapses the autoinhibited Λ-shaped conformation into an active “T”-shaped dimer, apposing the intracellular kinase domains for trans-autophosphorylation (Uchikawa et al., eLife, 2019;8:e48630). Phosphotyrosines then recruit IRS proteins, driving PI3K–AKT signaling and a parallel Ras–MAPK arm (Haeusler, McGraw & Accili, Nature Reviews Molecular Cell Biology, 2018;19:31–44).
• Glucose disposal via GLUT4 trafficking. In skeletal muscle and adipose tissue, AKT-dependent phosphorylation of trafficking regulators such as AS160/TBC1D4 releases GLUT4 storage vesicles for exocytosis, raising plasma-membrane transporter density and with it glucose uptake (Leto & Saltiel, Nature Reviews Molecular Cell Biology, 2012;13:383–396).
• Suppression of hepatic glucose output and of lipolysis. Insulin lowers blood glucose at least as much by restraining production as by promoting uptake, acting both directly on hepatocytes and indirectly by suppressing adipose lipolysis and the substrate flux that feeds gluconeogenesis (Petersen & Shulman, Physiological Reviews, 2018;98:2133–2223). Inhibition of lipolysis occurs at lower insulin concentrations than glucose disposal requires — a point that undercuts the assumption that insulin’s actions can be dosed apart from one another.
• On muscle protein, it is anti-catabolic rather than anabolic in humans. A systematic review and meta-analysis of insulin-infusion studies (44 studies reviewed, 25 pooled, 173 participants) found no effect of insulin on muscle protein synthesis and a significant reduction in muscle protein breakdown (Abdulla et al., Diabetologia, 2016;59:44–55). The dose-response saturates early: raising serum insulin from 5 to 30 mU/L halved leg protein breakdown, with no further inhibition at 72 or 167 mU/L and synthesis rates unchanged throughout (Greenhaff et al., American Journal of Physiology-Endocrinology and Metabolism, 2008;295:E595–E604).

Research status

Insulin has the deepest clinical record of any peptide in this reference set — roughly a century of continuous therapeutic use, and an evidence base that is unambiguous about both what it does and what it costs.

Glycemic control and complications. The DCCT randomized 1,441 people with type 1 diabetes to intensive versus conventional insulin therapy and followed them a mean of 6.5 years. In the primary-prevention cohort, intensive therapy cut the adjusted mean risk of developing retinopathy by 76% (95% CI 62–85%); in the secondary-intervention cohort it slowed retinopathy progression by 54%. Across both cohorts, microalbuminuria fell 39%, albuminuria 54%, and clinical neuropathy 60%. The chief adverse event was a two- to threefold increase in severe hypoglycemia (DCCT Research Group, New England Journal of Medicine, 1993;329:977–986). In type 2 diabetes, UKPDS 33 randomized 3,867 newly diagnosed patients and reported over ten years a median HbA1c of 7.0% on intensive therapy versus 7.9% on conventional, a 25% reduction in microvascular endpoints (95% CI 7–40, p=0.0099), and no significant effect on macrovascular disease. Major hypoglycemic episodes ran 1.8% per year in the insulin arm versus 0.7% on conventional treatment, and the insulin arm gained 4.0 kg (UK Prospective Diabetes Study Group, Lancet, 1998;352:837–853).

Cardiovascular outcomes are neutral, not protective. ORIGIN randomized 12,537 people with dysglycemia and cardiovascular risk factors to insulin glargine targeting normal fasting glucose or to standard care, with a median 6.2 years of follow-up. Both coprimary cardiovascular composites were flat (hazard ratios 1.02 and 1.04), cancers showed no difference, severe hypoglycemia rose from 0.31 to 1.00 events per 100 person-years, and median weight increased 1.6 kg versus a 0.5 kg fall on standard care (ORIGIN Trial Investigators, New England Journal of Medicine, 2012;367:319–328). This trial is frequently cited in both directions; what it actually settled is that basal insulin neither raises nor lowers cardiovascular risk over six years.

Failed and stalled programs. Oral insulin has repeatedly not worked. Oramed’s ORMD-0801 Phase 3 trial (ORA-D-013-1, 710 patients with type 2 diabetes) missed its primary endpoint of HbA1c change at 26 weeks and its secondary fasting-plasma-glucose endpoint, results announced in January 2023, after which the company said it expected to stop work in that indication. Separately, oral insulin as an immunologic intervention rather than a glucose-lowering one also failed: the TrialNet study randomized 560 autoantibody-positive relatives of people with type 1 diabetes to 7.5 mg/day oral insulin or placebo, and in the 389-participant main study group found no significant difference in time to diabetes (hazard ratio 0.87, P=0.21) over a median 2.7 years (Writing Committee for the Type 1 Diabetes TrialNet Oral Insulin Study Group, JAMA, 2017;318:1891–1902). Inhaled insulin has a mixed record of its own: Exubera was approved in 2006 and withdrawn by Pfizer in October 2007 after poor uptake, while Afrezza was approved in June 2014 and carries a boxed warning for acute bronchospasm, contraindicated in asthma and COPD.

Once-weekly basal insulins — approved for type 2, unresolved for type 1. ONWARDS 6 (NCT04848480) randomized 582 adults with type 1 diabetes to once-weekly icodec or once-daily degludec alongside mealtime aspart. Icodec met non-inferiority on HbA1c at week 26 (treatment difference 0.05 percentage points), but combined clinically significant or severe hypoglycemia was significantly higher: 19.9 versus 10.4 events per patient-year of exposure, rate ratio 1.9 (95% CI 1.5–2.3), p<0.0001 (Lancet, 2023;402:1636–1647). QWINT-5 found the same pattern for Lilly’s once-weekly efsitora alfa in type 1 diabetes — non-inferior HbA1c reduction, higher level 2 and level 3 hypoglycemia than degludec (Lancet, 2024;404:1132–1142). FDA issued a Complete Response Letter for icodec in July 2024 citing manufacturing questions and the type 1 indication; Awiqli (insulin icodec-abae) was ultimately approved in the United States on March 26, 2026 for adults with type 2 diabetes only.

Performance and physique use has no controlled evidence. There is no randomized trial supporting insulin as a muscle-building or performance agent in people without diabetes. The human protein-turnover literature described above points the other way, and the record of non-clinical use consists of case reports and narrative reviews rather than trials (Evans & Lynch, British Journal of Sports Medicine, 2003;37:356–357; Holt & Sönksen, British Journal of Pharmacology, 2008;154:542–556; a 30-year-old bodybuilder presenting in hypoglycemic coma from covert insulin injection, Journal of Emergency Medicine, 2019;56:279–281).

Common dosage forms

• 10 mL vials at U-100 (100 units per mL), the standard concentration for human insulin and most analogs, drawn with dedicated unit-marked insulin syringes.
• 3 mL prefilled disposable pens and cartridges, the dominant modern presentation, again typically U-100.
• Concentrated presentations for people requiring large volumes: U-200 (lispro, degludec), U-300 (glargine), and U-500 regular human insulin. Awiqli, the once-weekly product, is supplied at 700 units per mL. Concentration is a labeled property of each product and is not interchangeable between them.
• Pharmacologic classes by time-action profile rather than by molecule alone: rapid-acting analogs, short-acting regular human insulin, intermediate-acting NPH, long- and ultra-long-acting basal analogs, fixed-ratio premixes such as 70/30, and fixed-ratio combinations with a GLP-1 receptor agonist.
• Inhalation powder cartridges (Afrezza), and pump reservoirs of rapid-acting analog for continuous subcutaneous infusion.
• No oral form exists. Insulin is degraded by gastric and intestinal proteases and is not absorbed intact; the oral formulations discussed under Research status remain investigational or discontinued, and any capsule or oral spray marketed as insulin does not deliver the hormone.
• Insulin is dosed in international units, not milligrams — roughly 28.8 units per milligram of human insulin by the USP standard. This is the reverse of nearly every other compound in this reference set and is a common source of arithmetic error.

This section describes product formats only and is not dosing guidance.

Key considerations

• Insulin misuse kills people, and the mechanism is not exotic. Insulin lowers blood glucose whether or not the user has diabetes, and in someone without diabetes there is no hyperglycemia to absorb the dose — the entire effect is unopposed. Severe hypoglycemia produces disorientation, seizures, unconsciousness, permanent neurologic injury, and death, and it can develop faster than a person can self-treat. Even under medical supervision the burden is large: insulin-related hypoglycemia and errors accounted for an estimated 97,648 U.S. emergency department visits per year, with 29.3% resulting in hospitalization and severe neurologic sequelae documented in an estimated 60.6% of visits (Geller et al., JAMA Internal Medicine, 2014;174:678–686). Outside medicine, the case literature is coma and near-death (Evans & Lynch, 2003; Journal of Emergency Medicine, 2019;56:279–281). Two compounding factors make self-administration worse: repeated hypoglycemia blunts the counterregulatory and symptomatic warning response, a documented syndrome called hypoglycemia-associated autonomic failure (Cryer, New England Journal of Medicine, 2013;369:362–372); and long-acting basal analogs cannot be withdrawn once injected, so an overdose has to be outlasted rather than reversed. Nothing in the human evidence base offers a benefit to weigh against this. Insulin does not raise muscle protein synthesis in humans, and what suppression of breakdown it produces is already maximal at near-physiological concentrations (Abdulla et al., 2016; Greenhaff et al., 2008).
• Approved, and also unusually easy to obtain. Insulin products have been FDA-approved for a century, and on March 23, 2020 they transitioned from drug applications to biologics licenses under the Public Health Service Act, which opened the biosimilar and interchangeable pathway — Semglee (insulin glargine-yfgn) became the first interchangeable biosimilar of any kind in July 2021. Practically, regular human insulin and NPH are stocked behind the U.S. pharmacy counter and sold without a prescription in most states, while the modern analogs remain prescription-only. There is no “research use only” channel here and no legal gray zone to navigate; accessibility, not scarcity, is what makes non-clinical use common.
• Prohibited in sport at all times. Insulins and insulin-mimetics appear on the WADA Prohibited List under S4.4.2, in the Metabolic Modulators group of section S4 (Hormone and Metabolic Modulators), and are not classed as specified substances; athletes with diabetes require a Therapeutic Use Exemption. The 2025 list added named insulin-mimetic examples (S519, S597) alongside insulin itself.
• Exogenous insulin is identifiable after the fact. Because beta cells release insulin and C-peptide in equimolar amounts, injected insulin produces the diagnostic signature of high insulin with suppressed C-peptide. This is standard in the workup of unexplained hypoglycemia and in forensic investigation, so covert use is not undetectable in a clinical or medicolegal setting even when routine anti-doping assays are limited.
• Naming and product confusions are frequent and consequential. Insulin is not IGF-1: insulin-like growth factor 1 is a separate 70-amino-acid peptide with its own receptor, profiled separately on this site, and “insulin-like” in a compound name signals sequence homology rather than identity. Human insulin and analog insulins are distinct molecules with different onset and duration and are not interchangeable unit-for-unit. Brand pairs that differ by two letters — Humalog and Humulin, NovoLog and Novolin — carry entirely different time-action profiles, and product mix-ups are a documented category of outpatient adverse event in national surveillance data (Pharmacoepidemiology and Drug Safety, 2021;30:573–581). U-100, U-200, U-300, U-500, and 700 units/mL presentations look similar and are not.