Insulin

1. Overview

Insulin is an endogenous 51-amino-acid peptide hormone secreted by pancreatic beta cells within the islets of Langerhans. It is the principal anabolic hormone of vertebrate metabolism, coordinating postprandial nutrient disposal and long-term fuel homeostasis across liver, skeletal muscle, adipose tissue, and the central nervous system. The molecule consists of two chains, an A-chain of 21 residues and a B-chain of 30 residues, covalently linked by two interchain disulfide bonds (Cys A7-Cys B7 and Cys A20-Cys B19) and stabilized by a single intrachain disulfide bond (Cys A6-Cys A11) [24][26].

Insulin holds a singular place in medical history. In the summer of 1921, Frederick Banting and Charles Best, working in John Macleod's laboratory at the University of Toronto, produced the first pancreatic extract capable of reversibly correcting hyperglycemia in depancreatized dogs. Following refinement of the purification process by James Collip, the extract was administered on January 11, 1922 to 14-year-old Leonard Thompson, who was dying of diabetic ketoacidosis at Toronto General Hospital. The first dose produced only a modest response and caused a sterile abscess. A second, more purified preparation administered on January 23, 1922 dramatically lowered blood glucose, resolved ketonuria, and restored the patient. Leonard Thompson survived until 1935 [24]. Banting and Macleod were awarded the 1923 Nobel Prize in Physiology or Medicine; Banting shared his portion with Best, and Macleod shared his with Collip.

For the next six decades, clinical insulin was obtained by acid-ethanol extraction from bovine and porcine pancreata. In 1978, Genentech scientists led by David Goeddel reported the first recombinant human insulin expressed in Escherichia coli from chemically synthesized A-chain and B-chain genes, with subsequent chain combination [25]. Humulin R, the first recombinant DNA-derived therapeutic protein, was approved by the FDA in 1982, marking the beginning of the biotechnology era of medicine.

Contemporary insulin therapy comprises an ecosystem of molecular species engineered to match specific physiological roles:

• Regular human insulin and NPH (neutral protamine Hagedorn), representing the native hexamer and a protamine-zinc suspension respectively;
• Rapid-acting analogs (lispro, aspart, glulisine, ultra-rapid lispro, and faster aspart) designed to dissociate more quickly from hexamers after subcutaneous injection;
• Long-acting basal analogs (glargine U100, glargine U300, detemir, degludec) with isoelectric precipitation, fatty-acid albumin binding, or soluble multihexamer depot formation;
• Ultra-long once-weekly analogs (icodec, efsitora alfa) which extend pharmacology to the seven-day dosing interval;
• Fixed-ratio GLP-1 combinations (IDegLira, iGlarLixi) that couple basal insulin with a GLP-1 receptor agonist in a single pen;
• Inhaled prandial insulin (Afrezza, Technosphere insulin) with Tmax of 12-15 minutes after oral inhalation;
• Biosimilar basal insulins (Basaglar, Rezvoglar, Semglee for glargine) that expanded access through interchangeable designations beginning in 2021.

Molecular Weight

~5808 Da (human insulin monomer)

Structure

Two-chain peptide: A-chain (21 aa) and B-chain (30 aa) linked by two interchain disulfide bonds (A7-B7, A20-B19) with one intrachain disulfide (A6-A11)

Biosynthesis

Preproinsulin (110 aa) to proinsulin (86 aa) to insulin plus C-peptide via PC1/PC3, PC2, and carboxypeptidase E in beta-cell secretory granules

Receptor

Insulin receptor (INSR), a receptor tyrosine kinase; signals through IRS-1/2, PI3K-AKT, and Ras-MAPK pathways

Key Effectors

GLUT4 translocation (skeletal muscle, adipose), glycogen synthase activation, lipogenesis, protein synthesis, suppression of hepatic gluconeogenesis

Analog Modifications

Lispro (B28 Lys to Pro swap), aspart (B28 Pro to Asp), glulisine (B3 Asn to Lys, B29 Lys to Glu), glargine (A21 Asn to Gly, +2 Arg at B30), detemir (B29 myristoyl, B30 deletion), degludec (B29 hexadecanedioyl via L-gamma-Glu, B30 deletion), icodec (A14, B16, B25 substitutions plus C20 fatty diacid at B29)

FDA Status

Approved; first therapeutic protein 1922; recombinant human insulin (Humulin, Genentech/Lilly) 1982; analogs from 1996 (lispro); once-weekly icodec (Awiqli) approved 2026

WADA Status

Prohibited at all times (S4.5) except for athletes with documented insulin-dependent diabetes with TUE

2. Biosynthesis

2.1 Preproinsulin to Proinsulin

Insulin biosynthesis begins with the INS gene on chromosome 11p15.5, transcribed in pancreatic beta cells and translated on ribosomes bound to the endoplasmic reticulum [26]. The primary translation product is preproinsulin, a 110-amino-acid precursor comprising an N-terminal signal peptide (24 residues), followed by the B-chain (30 residues), the connecting C-peptide (31 residues), and the A-chain (21 residues).

During co-translational translocation into the ER lumen, the signal peptide is cleaved by signal peptidase, yielding proinsulin. Proinsulin folds with the assistance of ER chaperones (BiP/GRP78, PDI) and forms its three disulfide bonds through oxidative protein folding. Correctly folded proinsulin is a single-chain polypeptide with the structural topology of mature insulin, the C-peptide acting as an internal tether linking the C-terminus of the B-chain to the N-terminus of the A-chain [26].

2.2 Granule Formation and Proteolytic Processing

Properly folded proinsulin is transported via the Golgi apparatus to immature secretory granules, where it co-packages with zinc ions to form stable hexamers around a central pair of Zn2+ atoms. Acidification of the granule interior by a vacuolar H+-ATPase activates the granule-resident endoproteases prohormone convertase 1/3 (PC1/PC3) and prohormone convertase 2 (PC2), which cleave proinsulin at the dibasic junctions flanking the C-peptide (Arg31-Arg32 between B-chain and C-peptide, and Lys64-Arg65 between C-peptide and A-chain). Carboxypeptidase E (CPE) trims the resulting C-terminal basic residues, yielding mature insulin and free C-peptide in a 1:1 stoichiometry [26].

Mature granules contain rhombohedral microcrystals of Zn2+-insulin hexamers plus soluble C-peptide. Glucose-stimulated insulin secretion releases both species into the portal circulation in equimolar amounts, which is why serum C-peptide is used clinically as a marker of endogenous beta-cell function unaffected by exogenous insulin administration.

2.3 Regulated Secretion

Glucose enters beta cells through the GLUT2 (SLC2A2) or GLUT1/GLUT3 transporter and is phosphorylated by glucokinase, the rate-limiting step that functions as the beta-cell 'glucose sensor'. Glycolytic and mitochondrial oxidative metabolism raises the ATP:ADP ratio, which closes ATP-sensitive K+ channels (KATP; composed of SUR1/ABCC8 and Kir6.2/KCNJ11). Membrane depolarization activates voltage-gated L-type Ca2+ channels, and the resulting Ca2+ influx triggers exocytosis of insulin-containing granules. Incretin hormones (GLP-1, GIP) amplify glucose-stimulated insulin secretion through Gs-coupled receptors elevating cAMP and activating PKA and EPAC2.

3. Receptor and Signal Transduction

Insulin exerts its effects by binding to the insulin receptor (INSR), a disulfide-linked homodimer of alpha-beta subunits belonging to the receptor tyrosine kinase superfamily [27]. Two isoforms, INSR-A (without exon 11, predominant in fetal tissues and tumors) and INSR-B (with exon 11, predominant in adult liver, muscle, and adipose), differ modestly in binding kinetics. The closely related IGF-1 receptor (IGF1R) forms heterodimers with INSR and contributes to insulin-IGF signaling crosstalk.

Insulin binding causes a conformational change that unlocks the transmembrane beta-subunit kinase activity, leading to trans-autophosphorylation of tyrosines in the activation loop of one beta-subunit by the other. Activated INSR then phosphorylates insulin receptor substrate proteins (IRS-1, IRS-2, IRS-3, IRS-4) on multiple tyrosines, creating docking sites for SH2-domain-containing signaling proteins [27].

Two major signaling branches emerge:

PI3K-AKT-GLUT4 branch (metabolic actions). IRS tyrosines recruit the p85 regulatory subunit of class IA phosphoinositide 3-kinase (PI3K), activating the p110 catalytic subunit to generate PIP3 at the plasma membrane. PIP3 recruits PDK1 and mTORC2, which phosphorylate AKT at Thr308 and Ser473 respectively, fully activating the kinase. AKT phosphorylates and inactivates AS160/TBC1D4, a Rab-GAP whose inactivation permits Rab8A/Rab10/Rab13-dependent fusion of GLUT4-containing storage vesicles with the plasma membrane in skeletal muscle and adipose tissue, increasing glucose uptake. AKT also phosphorylates glycogen synthase kinase 3 (GSK3) to activate glycogen synthesis, phosphorylates FoxO1 to suppress hepatic gluconeogenic gene expression (PEPCK, G6Pase), and activates mTORC1 via TSC1/TSC2 inhibition to stimulate protein synthesis [27].

Ras-MAPK branch (mitogenic and transcriptional actions). IRS tyrosines also recruit GRB2-SOS, activating Ras and the RAF-MEK-ERK cascade, which mediates insulin's effects on cell proliferation, gene expression, and differentiation.

Physiological net effects include stimulation of hepatic glycogen synthesis, suppression of hepatic gluconeogenesis and glycogenolysis, stimulation of skeletal muscle glucose uptake and glycogen synthesis, stimulation of adipose glucose uptake and de novo lipogenesis, suppression of adipose lipolysis, stimulation of protein synthesis and suppression of proteolysis across tissues, suppression of ketogenesis, and central actions on appetite and autonomic outflow. Insulin resistance, a hallmark of type 2 diabetes, reflects impaired PI3K-AKT signaling with relative preservation of MAPK signaling, contributing to the dysmetabolic and atherogenic phenotype.

4. Pharmacologic Insulins: Structural and Pharmacokinetic Design

4.1 Regular Human Insulin and NPH

Regular (soluble) human insulin is the unmodified recombinant hormone, formulated as a zinc-stabilized hexamer at neutral pH. After subcutaneous injection, hexamers must dissociate into dimers and monomers before absorption across capillary endothelium, producing an onset of 30-60 minutes, a peak at 2-4 hours, and a duration of 6-8 hours. This slow prandial coverage results in poor postprandial control and late hypoglycemia risk.

NPH (neutral protamine Hagedorn) insulin is a crystalline suspension of insulin with protamine and zinc, which dissolves slowly at the injection site to yield an intermediate profile with onset 1-2 hours, peak 4-10 hours, and duration 10-16 hours. NPH is inexpensive and remains widely used globally, but its pronounced peak contributes to nocturnal hypoglycemia when used as basal coverage.

4.2 Rapid-Acting Analogs

Insulin lispro (Humalog, Lilly, 1996) was the first commercial insulin analog. Reversing the natural ProB28-LysB29 sequence to LysB28-ProB29 disrupts the hydrophobic dimer interface, weakening hexamer self-association. After subcutaneous injection, lispro dissociates rapidly from hexamers to monomers, producing an onset of 10-15 minutes, peak at 30-90 minutes, and duration of 3-5 hours. Peak serum concentrations are approximately 3-fold higher, Tmax is approximately 4-fold faster, and duration is approximately half that of regular insulin [17].

Insulin aspart (Novolog/NovoRapid, Novo Nordisk, 1999) substitutes aspartate for proline at B28, introducing negative charge that destabilizes the dimer. Pharmacokinetic and pharmacodynamic profiles are similar to lispro [18].

Insulin glulisine (Apidra, Sanofi, 2004) substitutes lysine for asparagine at B3 and glutamate for lysine at B29, with surfactant-free formulation, producing comparable rapid-acting kinetics.

Ultra-rapid insulin lispro (Lyumjev) and faster insulin aspart (Fiasp) add treprostinil and citrate/niacinamide excipients, respectively, to accelerate local vasodilation and absorption, shifting Tmax approximately 5 minutes earlier and reducing 1-hour postprandial glucose excursions.

4.3 Long-Acting Basal Analogs

Insulin glargine (Lantus, Sanofi, 2000) carries an A21 Asn-to-Gly substitution and two additional arginines at the C-terminus of the B-chain (B31, B32), shifting the isoelectric point from pH 5.4 to 7.0. Formulated at acidic pH (4.0) for solubility, glargine precipitates as amorphous microcrystals after subcutaneous injection into physiologic pH tissue, forming a depot that releases monomers slowly. The resulting pharmacokinetic profile is relatively peakless with a duration of approximately 20-24 hours [19]. Glargine U300 (Toujeo, Sanofi, 2015) uses the same molecule at a 3-fold higher concentration, producing a smaller injection depot with smaller surface area and thus further prolonged and flatter kinetics (duration 30-36 hours) [29].

Insulin detemir (Levemir, Novo Nordisk, 2005) is modified by deletion of threonine B30 and myristoylation (C14 fatty acid acylation) of LysB29. The fatty acid promotes self-association into dihexamers at the injection site and reversible binding to serum albumin in circulation. Duration is approximately 14-24 hours, typically requiring twice-daily administration.

Insulin degludec (Tresiba, Novo Nordisk, 2013) carries the same B30 deletion but is acylated with hexadecanedioic acid (C16 diacid) linked via an L-gamma-glutamate spacer to LysB29. In the pharmaceutical solution (phenol, zinc) degludec is a dihexamer; upon subcutaneous injection and phenol diffusion it forms soluble multihexamer chains that slowly dissociate into monomers. Albumin binding of the fatty-diacid modification provides additional buffering. Terminal half-life is approximately 25 hours and duration exceeds 42 hours, supporting flexible once-daily dosing with steady-state trough-to-peak ratios well below those of glargine [6][23].

4.4 Once-Weekly Analogs

Insulin icodec (Awiqli, Novo Nordisk, FDA-approved 2026) is engineered for once-weekly subcutaneous administration. Three amino acid substitutions (TyrA14-to-Glu, TyrB16-to-His, PheB25-to-His) reduce insulin receptor binding affinity and proteolytic degradation, and a C20 fatty diacid linked via a gamma-Glu-two-oxoethoxy acetate linker to LysB29 provides strong, reversible albumin binding [30]. Circulating icodec is more than 99% albumin-bound with a terminal half-life of approximately 196 hours (~8 days). After subcutaneous injection, the albumin-bound depot slowly releases active insulin that reaches target tissues, producing a flat pharmacokinetic profile over seven days. The ONWARDS 1-6 phase 3 program evaluated icodec in over 4,000 adults with type 2 and type 1 diabetes [10][11][12][13][14][16][22].

Insulin efsitora alfa (Eli Lilly, QWINT program) uses a different strategy. Efsitora is a fusion protein combining a single-chain insulin variant with a human IgG2 Fc domain, conferring long half-life through neonatal Fc receptor (FcRn)-mediated recycling, the same mechanism that extends the half-life of therapeutic antibodies. In QWINT-2 (T2D insulin-naive), efsitora reduced HbA1c by 1.34% vs 1.26% with degludec over 52 weeks, meeting non-inferiority and showing favorable time-in-range [15]. The QWINT phase 3 program spans insulin-naive T2D (QWINT-1 vs glargine, QWINT-2 vs degludec), switch from basal insulin (QWINT-3), basal-bolus T2D (QWINT-4), and T1D (QWINT-5).

4.5 Fixed-Ratio GLP-1 Combinations

IDegLira (Xultophy, Novo Nordisk, 2016) combines 100 U/mL insulin degludec with 3.6 mg/mL liraglutide in a single pen (1 dose step equals 1 U degludec plus 0.036 mg liraglutide). The DUAL program (DUAL I-X) established superior HbA1c reduction, lower hypoglycemia than degludec alone, and weight neutrality or modest weight loss vs weight gain with basal insulin alone [8].

iGlarLixi (Soliqua, Sanofi, 2016) combines insulin glargine U100 with lixisenatide. LixiLan-L demonstrated greater HbA1c reduction, higher rates of reaching HbA1c below 7.0%, and 1.4 kg weight advantage vs glargine alone [9]. Both fixed-ratio combinations exploit the complementary mechanisms of basal insulin (fasting glucose control) and prandial GLP-1 action (postprandial glucose and weight).

4.6 Inhaled Insulin (Afrezza)

Technosphere insulin (Afrezza, MannKind, FDA-approved 2014) delivers regular human insulin adsorbed onto fumaryl diketopiperazine (FDKP) microparticles by oral inhalation. FDKP particles dissolve at alveolar pH, releasing insulin that is absorbed rapidly across the pulmonary epithelium [20]. Tmax is 12-15 minutes, making Afrezza the fastest-acting approved insulin. It is indicated for prandial glycemic control in adults with type 1 or type 2 diabetes but requires baseline and serial pulmonary function monitoring (spirometry at baseline, 6 months, then annually) and is contraindicated in chronic lung disease (asthma, COPD) due to a risk of acute bronchospasm.

4.7 Biosimilars and Interchangeables

Insulin biosimilars gained regulatory pathways in the US through the Biologics Price Competition and Innovation Act amendments. Basaglar (insulin glargine, Lilly), Semglee (insulin glargine, Mylan/Biocon, first interchangeable biosimilar insulin in 2021), Rezvoglar (Lilly), and Admelog (insulin lispro, Sanofi) have expanded competitive options, reducing costs and increasing access.

5. Physiology and Glucose Homeostasis

5.1 Normal Insulin Secretion

Basal insulin secretion (approximately 0.5-1.0 U/hour in healthy adults) maintains fasting glucose through suppression of hepatic glucose production. Prandial secretion follows a biphasic pattern: a first-phase spike within 2-10 minutes of meal ingestion from pre-docked granules, followed by a sustained second phase over 1-2 hours as newly synthesized insulin is released. Total daily endogenous secretion is approximately 0.5-1.0 U/kg in healthy adults.

5.2 Type 1 Diabetes

Type 1 diabetes (formerly juvenile-onset or insulin-dependent) results from autoimmune destruction of pancreatic beta cells, typically in genetically susceptible individuals (HLA-DR3/DR4, INS VNTR, PTPN22, CTLA4) after environmental triggering. Autoantibodies to insulin, GAD65, IA-2, and ZnT8 precede clinical presentation by months to years. Absolute insulin deficiency necessitates lifelong exogenous insulin replacement; untreated diabetic ketoacidosis is fatal within weeks to months. The landmark DCCT established that intensive insulin therapy delays microvascular complications [1], and long-term EDIC follow-up demonstrated cardiovascular protection persisting decades after randomization, a phenomenon termed 'metabolic memory' or 'legacy effect' [3].

5.3 Type 2 Diabetes

Type 2 diabetes is characterized by peripheral insulin resistance coupled with progressive beta-cell dysfunction. Insulin therapy is ultimately required in many patients as beta-cell function declines. UKPDS 33 established that intensive glycemic control (achieved with sulphonylureas or insulin) reduces microvascular complications in newly diagnosed T2D [2]. The ACCORD trial, however, showed that overly aggressive targeting of HbA1c below 6.0% in long-standing T2D with cardiovascular disease increased mortality, establishing the importance of individualized targets [4]. ORIGIN demonstrated that early basal insulin in high-risk dysglycemia is cardiovascular-neutral (neither harmful nor beneficial) and reduces progression to overt diabetes [5].

5.4 Gestational and Type 3c Diabetes

Insulin is the preferred pharmacotherapy for gestational diabetes not controlled by diet and exercise, given its lack of placental transfer at therapeutic doses and extensive safety record. Type 3c diabetes (pancreatogenic, e.g., post-pancreatectomy, chronic pancreatitis, cystic fibrosis-related diabetes) typically requires insulin due to loss of both insulin and glucagon secretion.

6. Landmark Clinical Trials

6.1 DCCT (1993) and EDIC (2005-)

The Diabetes Control and Complications Trial (DCCT, 1983-1993) randomized 1,441 patients with type 1 diabetes to intensive insulin therapy (three or more daily injections or continuous subcutaneous insulin infusion, guided by 4-7 daily self-monitored glucose measurements) vs conventional therapy (1-2 daily injections). Over a mean 6.5 years, intensive therapy achieved HbA1c of approximately 7.2% vs 9.1% with conventional, and reduced the risk of clinically significant retinopathy by 76% (primary-prevention cohort) and slowed progression by 54% (secondary-intervention cohort), with parallel reductions in nephropathy and neuropathy [1]. The cost was a 3-fold increase in severe hypoglycemia and modest weight gain.

The Epidemiology of Diabetes Interventions and Complications (EDIC) observational follow-up showed that intensive therapy conferred a 42% reduction in any cardiovascular disease event and a 57% reduction in nonfatal MI, stroke, or cardiovascular death at 17 years, long after HbA1c differences between groups had converged to approximately 8.0% in both [3]. This 'metabolic memory' established the durable benefit of early glycemic control in type 1 diabetes.

6.2 UKPDS 33 and 34 (1998)

The UK Prospective Diabetes Study enrolled 5,102 patients with newly diagnosed type 2 diabetes across 23 UK centers from 1977 to 1991. UKPDS 33 randomized 3,867 patients to intensive control with sulphonylureas (chlorpropamide, glibenclamide, or glipizide) or insulin, targeting fasting plasma glucose below 108 mg/dL, vs conventional control (diet alone, medication added only if FPG exceeded 270 mg/dL or symptoms appeared) [2]. Over a median 10 years, median HbA1c was 7.0% intensive vs 7.9% conventional, and any diabetes-related endpoint was reduced by 12% (P equals 0.029), driven primarily by a 25% reduction in microvascular endpoints. Macrovascular events were not significantly reduced in the main trial, though a post-trial observational legacy effect emerged over the subsequent decade.

UKPDS 34 showed that metformin monotherapy in overweight patients produced additional reductions in macrovascular events beyond glycemic control, establishing metformin as first-line therapy.

6.3 ACCORD (2008)

The Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial enrolled 10,251 type 2 diabetes patients with either established cardiovascular disease (age 40-79) or additional cardiovascular risk factors (age 55-79) and randomized them to intensive (HbA1c target below 6.0%) or standard (HbA1c 7.0-7.9%) glycemic control. The intensive arm was terminated early at 3.5 years due to a 22% increase in all-cause mortality (HR 1.22; 95% CI 1.01-1.46) without reduction in major cardiovascular events [4]. Severe hypoglycemia requiring medical assistance was 3-fold more frequent in the intensive arm. The trial fundamentally reshaped glycemic targets, prompting guidelines to recommend individualized HbA1c goals (e.g., 6.5-7.0% for young, healthy patients; 7.0-8.0% or higher for older patients with multiple comorbidities or hypoglycemia risk).

6.4 ORIGIN (2012)

The Outcome Reduction with an Initial Glargine Intervention (ORIGIN) trial randomized 12,537 patients with impaired fasting glucose, impaired glucose tolerance, or early type 2 diabetes (plus cardiovascular risk factors) to insulin glargine titrated to fasting plasma glucose 95 mg/dL or below, or standard care, over a median 6.2 years [5]. ORIGIN was the first dedicated cardiovascular outcomes trial of a basal insulin. Insulin glargine was cardiovascular-neutral (HR 1.02; 95% CI 0.94-1.11) and reduced new-onset type 2 diabetes by 28%. Severe hypoglycemia was uncommon (approximately 1 event per 100 patient-years) but more frequent than standard care, and modest weight gain (1.6 kg) was observed with glargine.

6.5 DEVOTE (2017)

DEVOTE directly compared the cardiovascular safety of insulin degludec U100 vs insulin glargine U100 in 7,637 type 2 diabetes patients at high cardiovascular risk over a median 1.99 years [7]. Degludec was non-inferior for three-point MACE (HR 0.91; 95% CI 0.78-1.06) and reduced severe hypoglycemia by 40% (rate ratio 0.60; 95% CI 0.48-0.76) and nocturnal severe hypoglycemia by 53%. The trial established degludec as cardiovascular-safe with a hypoglycemia advantage over glargine U100, influencing formulary and guideline positioning.

6.6 BEGIN Basal-Bolus Type 1 (2012)

The BEGIN Basal-Bolus Type 1 phase 3 trial randomized 629 type 1 diabetes patients 3:1 to insulin degludec or insulin glargine, each with prandial insulin aspart, over 52 weeks [6]. Degludec was non-inferior for HbA1c and reduced overall confirmed hypoglycemia by 25% and nocturnal hypoglycemia by 25%. Results supported degludec's flat, ultra-long pharmacokinetic profile in the setting of tight T1D control.

6.7 ONWARDS Program: Once-Weekly Icodec

The ONWARDS phase 3a program (six trials, over 4,000 participants) evaluated once-weekly insulin icodec across the diabetes spectrum [22]:

• ONWARDS 1 (NEJM 2023, N=984): Insulin-naive T2D, icodec vs glargine U100 over 52 weeks. Icodec reduced HbA1c by 1.55% vs 1.35% (ETD -0.19%), with comparable hypoglycemia [10].
• ONWARDS 2 (Lancet D and E 2023, N=526): Switch from basal insulin, icodec vs degludec over 26 weeks. HbA1c 7.20% vs 7.42% (statistically superior); slightly higher hypoglycemia with icodec [11].
• ONWARDS 3 (JAMA 2023, N=588): Insulin-naive T2D, icodec vs degludec over 26 weeks. Icodec superior for HbA1c (ETD -0.22%); level 2/3 hypoglycemia numerically higher but low absolute rates [12].
• ONWARDS 4: Basal-bolus T2D, icodec vs glargine U100 both with bolus aspart. Non-inferior HbA1c.
• ONWARDS 5 (Annals 2023, N=1,085): Insulin-naive T2D, icodec plus dosing-guide app vs standard-care once-daily analogs over 52 weeks. HbA1c 1.68% vs 1.31% (superiority P equals 0.009); improved treatment satisfaction [13].
• ONWARDS 6 (Lancet 2023, N=582): Type 1 diabetes basal-bolus, icodec vs degludec both with mealtime aspart over 52 weeks. Non-inferior HbA1c but statistically higher level 2/3 hypoglycemia with icodec, with time below 54 mg/dL at the 1.0% threshold [14].

A 2024 participant-level meta-analysis of ONWARDS 1-5 confirmed a pooled HbA1c advantage for icodec of -0.17% (95% CI -0.24 to -0.10) with similar overall hypoglycemia rates [16].

The FDA issued a Complete Response Letter in July 2024 citing concerns about hypoglycemia in the type 1 indication (driven by ONWARDS 6). Novo Nordisk resubmitted the BLA in September 2025 for type 2 diabetes only, and the FDA approved Awiqli (insulin icodec-abae, 700 U/mL) for adults with T2D in March 2026, making it the first once-weekly basal insulin approved in the US [28][30]. Awiqli was already approved in the EU (March 2024), Canada, Australia, Japan, Switzerland, and China for T2D (and T1D outside the US) prior to the US approval.

6.8 QWINT Program: Once-Weekly Efsitora

The QWINT phase 3 program evaluated insulin efsitora alfa (Lilly) against daily basal analogs. QWINT-2 (NEJM 2024, N=928, insulin-naive T2D) demonstrated non-inferior HbA1c reduction vs degludec (1.34% vs 1.26%) over 52 weeks with 45 additional minutes per day of time-in-range and low hypoglycemia rates (0.58 vs 0.45 combined severe or clinically significant events per patient-year) [15]. QWINT-1 showed non-inferiority to glargine U100 in insulin-naive T2D; QWINT-3 tested switch from basal insulin; QWINT-4 evaluated basal-bolus T2D; and QWINT-5 addressed type 1 diabetes. Efsitora is expected to seek FDA approval on the basis of this program.

6.9 Fixed-Ratio GLP-1 Combinations

DUAL I (N=1,663) compared IDegLira vs degludec alone vs liraglutide alone in insulin-naive T2D on metformin over 26 weeks, with a 26-week extension [8]. IDegLira reduced HbA1c by 1.9% vs 1.4% with degludec and 1.3% with liraglutide, with 81% reaching HbA1c below 7.0%, weight neutrality, and lower hypoglycemia than degludec alone. The DUAL II-X trials extended these findings to various T2D populations and comparators.

LixiLan-L (N=736) compared iGlarLixi vs insulin glargine in T2D inadequately controlled on basal insulin and metformin [9]. iGlarLixi produced 1.1% HbA1c reduction vs 0.6% with glargine, with 55% vs 30% reaching HbA1c below 7.0% and a 1.4 kg weight advantage. LixiLan-O showed similar benefits in insulin-naive T2D.

7. Automated Insulin Delivery (Artificial Pancreas)

Continuous glucose monitoring (CGM) coupled to insulin pumps via algorithmic control has transformed type 1 diabetes management. The first FDA-approved hybrid closed-loop system, Medtronic MiniMed 670G (2016), combined Guardian 3 CGM with a PID-based algorithm that automated basal insulin delivery (auto-mode) while still requiring manual meal boluses [21]. The pivotal trial (124 participants, 3 months) reduced HbA1c from 7.4% to 6.9%, increased time in range (70-180 mg/dL) from 67% to 72%, and reduced time below 70 mg/dL by 44% with no severe hypoglycemia or DKA events [21].

Subsequent systems include:

• Medtronic MiniMed 770G/780G (advanced hybrid closed-loop with automatic correction boluses and adjustable targets to 100 mg/dL);
• Tandem t:slim X2 with Control-IQ (model-predictive control with Dexcom G6/G7);
• Insulet Omnipod 5 (tubeless patch pump with integrated algorithm and Dexcom G6/G7);
• Beta Bionics iLet (bionic pancreas with meal-announcement and weight-based initialization, approved 2023);
• DIY systems (OpenAPS, Loop, AndroidAPS) developed by the #WeAreNotWaiting community, with OpenAPS algorithm forming the basis of the commercial Tidepool Loop.

Real-world registry and trial evidence consistently shows these systems improve time-in-range by 10-15 percentage points vs sensor-augmented pump therapy without increasing hypoglycemia, and outcomes are further improved with faster-acting insulin analogs. Dual-hormone (insulin plus glucagon) closed-loop systems remain investigational.

8. Smart Insulin Concepts

Glucose-responsive ('smart') insulins aim to auto-regulate insulin activity in response to ambient glucose, potentially eliminating hypoglycemia risk. Approaches under investigation include:

• Phenylboronic acid-containing insulins that reversibly bind diols (glucose) to release insulin at hyperglycemia;
• Concanavalin A-based hydrogels that release encapsulated insulin when glucose displaces bound mannose residues;
• Glucose-responsive MK-2640 (Merck) and similar candidates that use mannosyl modifications and the mannose receptor to internalize and sequester the analog at euglycemia;
• Oral and transdermal formulations using nanoparticles, ionic liquids, and microneedle arrays.

None have yet achieved regulatory approval, though several candidates remain in early-phase clinical trials. The recent MK-2640 program was discontinued after a phase 1 study did not show sufficient glucose responsiveness to warrant advancement.

9. Dosing in Research and Practice

9.1 Type 1 Diabetes (basal-bolus)

Total daily insulin (TDI) typically 0.4-1.0 U/kg/day in adults, higher during puberty (up to 1.5 U/kg/day) and the 'honeymoon' phase requiring less. Standard distribution is approximately 50% basal (glargine, degludec, detemir, or icodec in T2D) and 50% prandial (lispro, aspart, glulisine, or their faster variants) divided at meals. Prandial doses are calculated using carbohydrate-to-insulin ratios (I:C, typically 1 unit per 8-15 g carbohydrate) and correction factors (ISF, typically 1 unit reduces glucose by 30-60 mg/dL), individualized by total daily dose using rules such as the '500 rule' (500 / TDI equals grams of carbohydrate per unit) and '1800 rule' for short-acting analogs.

9.2 Type 2 Diabetes

Basal initiation. Start basal insulin (glargine U100/U300, degludec, detemir, icodec, or NPH) at 10 U/day or 0.1-0.2 U/kg/day, titrated by 2 U every 3 days until fasting plasma glucose is 80-130 mg/dL, pausing or reducing if hypoglycemia occurs.

Intensification. If HbA1c remains above target despite basal doses exceeding approximately 0.5 U/kg/day with at-goal fasting glucose, options include:

• Add GLP-1 receptor agonist (typically preferred due to weight and CV benefit);
• Switch to fixed-ratio GLP-1 combination (IDegLira, iGlarLixi);
• Add prandial insulin to the largest meal (basal-plus) and intensify to full basal-bolus if needed;
• Consider premixed insulin for patients preferring simpler regimens.

Once-weekly basal. With FDA approval of icodec (Awiqli) in 2026 and anticipated approval of efsitora, once-weekly basal insulin is now an option for T2D, with starting dose of approximately 70 U weekly (or loading doses equivalent to a daily basal total) titrated by 20 U weekly to fasting plasma glucose 80-130 mg/dL.

10. Safety and Adverse Effects

10.1 Hypoglycemia

Hypoglycemia is the most common, most feared, and most dose-limiting adverse effect of insulin therapy. The American Diabetes Association classifies hypoglycemia as:

• Level 1 (alert): glucose less than 70 mg/dL but 54 mg/dL or above;
• Level 2 (clinically significant): glucose less than 54 mg/dL;
• Level 3 (severe): cognitive impairment requiring external assistance.

Severe hypoglycemia rates in the DCCT were 62 per 100 patient-years with intensive therapy vs 19 with conventional [1]. ACCORD saw a 3-fold increase with intensive T2D therapy [4]. DEVOTE demonstrated a 40% reduction in severe hypoglycemia with degludec vs glargine U100 at similar HbA1c [7]. Hypoglycemia unawareness develops with recurrent episodes, particularly with long-standing T1D, tight control, beta-blockers, alcohol, and autonomic neuropathy. Counseling on recognition, glucagon rescue (auto-injector, nasal spray), and 'hypoglycemia holidays' with temporarily relaxed targets can restore awareness.

10.2 Weight Gain

Insulin therapy is commonly associated with weight gain of 2-6 kg in the first year, driven by reduced glucosuria (restoring caloric retention), anabolic effects, and defensive eating to prevent hypoglycemia. Analog insulins (particularly degludec, detemir) and GLP-1 co-therapy minimize this effect. Detemir is modestly less weight-promoting than glargine or NPH.

10.3 Injection-Site Reactions and Lipodystrophy

Lipohypertrophy (fatty lump from repeated injection at the same site) is common (up to 30-50% of chronic insulin users), impairs insulin absorption unpredictably, and drives variability. Rotation of injection sites and use of shorter (4 mm pen) needles reduce risk. Lipoatrophy (localized fat loss) is rare with modern analogs.

10.4 Hypersensitivity and Immunogenicity

IgE-mediated hypersensitivity to insulin itself is rare with current recombinant human and analog products but has been reported, sometimes responding to desensitization. Local reactions to excipients (protamine, metacresol, phenol) are more common. Anti-insulin antibodies form in a subset of patients but rarely cause clinical insulin resistance in the modern era.

10.5 Insulin Edema and Hypokalemia

Initiation or aggressive intensification of insulin can precipitate fluid retention and peripheral edema, typically self-limited. Insulin drives potassium into cells, and intensive insulin therapy (especially IV during DKA) can cause hypokalemia requiring monitoring and supplementation.

10.6 Diabetic Retinopathy

Rapid reduction of HbA1c, particularly by more than 2 percentage points within 3-6 months in patients with pre-existing retinopathy, can cause transient worsening ('early worsening'). Ophthalmologic evaluation prior to and during intensive initiation is recommended.

10.7 Cancer Risk

Early observational data raised concerns about glargine and cancer, but the ORIGIN trial and subsequent meta-analyses have not confirmed a causal association [5]. Current evidence supports insulin's oncologic safety at clinical doses.

10.8 Intentional Misuse and Factitious Hypoglycemia

Insulin has been used as a tool of self-harm, factitious illness (Munchausen syndrome), and homicide. Evaluation of unexplained hypoglycemia in a person without diabetes includes measurement of insulin, C-peptide, proinsulin, beta-hydroxybutyrate, and sulfonylurea screen. Disproportionately high insulin with suppressed C-peptide suggests exogenous insulin administration.

11. Regulatory and Historical Timeline

• 1921-1922: Banting and Best isolate pancreatic extract; Collip purifies; Leonard Thompson treated January 1922 [24].
• 1923: Nobel Prize awarded to Banting and Macleod; Eli Lilly commercializes insulin from bovine/porcine pancreas.
• 1936: NPH insulin (Hagedorn) introduced, providing intermediate-acting basal coverage.
• 1955: Sanger completes the amino acid sequencing of insulin (Nobel Prize 1958).
• 1969: Hodgkin solves the three-dimensional crystal structure of insulin.
• 1978-1982: Goeddel/Genentech produce recombinant human insulin in E. coli [25]; Humulin approved 1982.
• 1996: Insulin lispro (Humalog) approved, the first insulin analog [17].
• 1999-2000: Insulin aspart (Novolog) and insulin glargine (Lantus) approved [18][19].
• 2005: Insulin detemir (Levemir) approved.
• 2013: Insulin degludec (Tresiba) approved [6].
• 2014: Afrezza (inhaled insulin) approved [20]; first MiniMed hybrid closed-loop system (670G) approved 2016 [21].
• 2015-2016: Glargine U300 (Toujeo), IDegLira (Xultophy), iGlarLixi (Soliqua) approved.
• 2021: Semglee becomes first interchangeable biosimilar insulin.
• 2024: EMA approves insulin icodec (Awiqli) for T1D and T2D in EU; FDA Complete Response Letter for US.
• 2026: FDA approves Awiqli for T2D, the first once-weekly basal insulin in the US [28][30].

12. Athletic Misuse and WADA Status

Insulin is prohibited in sport at all times (S4.5, Hormone and Metabolic Modulators, WADA Prohibited List) except for athletes with insulin-requiring diabetes holding a therapeutic use exemption (TUE). Misuse in bodybuilding and endurance sport exploits the anabolic and glycogen-loading effects but carries substantial risk of severe or fatal hypoglycemia in non-diabetic users. Detection in anti-doping programs is challenging because recombinant human insulin is identical to endogenous hormone, but rapid-acting and long-acting analogs are detectable by mass spectrometry due to their sequence modifications.

13. Clinical Evidence Summary

14. Related Peptides

See also: Semaglutide, Tirzepatide, Liraglutide, Dulaglutide, Glucagon, C-Peptide, Amylin, Pramlintide

15. References

1. [1] DCCT Research Group (1993). The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. New England Journal of Medicine. DOI PubMed
2. [2] UK Prospective Diabetes Study (UKPDS) Group (1998). Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet. DOI PubMed
3. [3] Nathan DM, Cleary PA, Backlund JY, et al. (DCCT/EDIC Study Research Group) (2005). Intensive diabetes treatment and cardiovascular disease in patients with type 1 diabetes. New England Journal of Medicine. DOI PubMed
4. [4] ACCORD Study Group (Gerstein HC, Miller ME, Byington RP, et al.) (2008). Effects of intensive glucose lowering in type 2 diabetes. New England Journal of Medicine. DOI PubMed
5. [5] ORIGIN Trial Investigators (Gerstein HC, Bosch J, Dagenais GR, et al.) (2012). Basal insulin and cardiovascular and other outcomes in dysglycemia. New England Journal of Medicine. DOI PubMed
6. [6] Heller S, Buse J, Fisher M, et al. (2012). Insulin degludec, an ultra-longacting basal insulin, versus insulin glargine in basal-bolus treatment with mealtime insulin aspart in type 1 diabetes (BEGIN Basal-Bolus Type 1). Lancet. DOI PubMed
7. [7] Marso SP, McGuire DK, Zinman B, et al. (DEVOTE Study Group) (2017). Efficacy and safety of degludec versus glargine in type 2 diabetes. New England Journal of Medicine. DOI PubMed
8. [8] Gough SC, Bode B, Woo V, et al. (2015). One-year efficacy and safety of a fixed combination of insulin degludec and liraglutide (IDegLira) in patients with type 2 diabetes: results of a 26-week extension to a 26-week main trial. Diabetes, Obesity and Metabolism. DOI PubMed
9. [9] Aroda VR, Rosenstock J, Wysham C, et al. (2016). Efficacy and safety of LixiLan, a titratable fixed-ratio combination of insulin glargine plus lixisenatide, in type 2 diabetes inadequately controlled on basal insulin and metformin: the LixiLan-L randomized trial. Diabetes Care. DOI PubMed
10. [10] Rosenstock J, Bain SC, Bajaj HS, et al. (ONWARDS 1) (2023). Weekly icodec versus daily glargine U100 in type 2 diabetes without previous insulin. New England Journal of Medicine. DOI PubMed
11. [11] Philis-Tsimikas A, Asong M, Franek E, et al. (ONWARDS 2) (2023). Switching to once-weekly insulin icodec versus once-daily insulin degludec in individuals with basal insulin-treated type 2 diabetes (ONWARDS 2). Lancet Diabetes and Endocrinology. DOI PubMed
12. [12] Lingvay I, Asong M, Desouza C, et al. (ONWARDS 3) (2023). Once-weekly insulin icodec vs once-daily insulin degludec in adults with insulin-naive type 2 diabetes: the ONWARDS 3 randomized clinical trial. JAMA. DOI PubMed
13. [13] Bajaj HS, Bergenstal RM, Christoffersen A, et al. (ONWARDS 5) (2023). Once-weekly insulin icodec with dosing guide app versus once-daily basal insulin analogues in insulin-naive type 2 diabetes (ONWARDS 5): a randomized trial. Annals of Internal Medicine. DOI PubMed
14. [14] Russell-Jones D, Babazono T, Cailleteau R, et al. (ONWARDS 6) (2023). Once-weekly insulin icodec versus once-daily insulin degludec as part of a basal-bolus regimen in individuals with type 1 diabetes (ONWARDS 6). Lancet. DOI PubMed
15. [15] Bajaj HS, Aberle J, Davies M, et al. (2024). Once-weekly insulin efsitora alfa versus once-daily insulin degludec in adults with type 2 diabetes without previous insulin (QWINT-2). New England Journal of Medicine. DOI PubMed
16. [16] Rosenstock J, Bajaj HS, Lingvay I, et al. (2024). Once-weekly insulin icodec compared with daily basal insulin analogues in type 2 diabetes: participant-level meta-analysis of the ONWARDS 1-5 trials. Diabetes, Obesity and Metabolism. DOI PubMed
17. [17] Keen H (1997). Insulin lispro: a fast-acting insulin analog. Drugs. DOI PubMed
18. [18] Mudaliar SR, Lindberg FA, Joyce M, et al. (1999). Insulin aspart (B28 asp-insulin): a fast-acting analog of human insulin: absorption kinetics and action profile compared with regular human insulin in healthy nondiabetic subjects. Diabetes Care. DOI PubMed
19. [19] Rosenstock J, Park G, Zimmerman J (U.S. Insulin Glargine [HOE 901] Type 1 Diabetes Investigator Group) (2000). Basal insulin glargine (HOE 901) versus NPH insulin in patients with type 1 diabetes on multiple daily insulin regimens. Diabetes Care. DOI PubMed
20. [20] Goldman-Levine JD, Patel K (2015). Technosphere inhaled insulin (Afrezza). Annals of Pharmacotherapy. DOI PubMed
21. [21] Duvvuri S, Wang E, Lim HK (2019). MiniMed 670G hybrid closed loop artificial pancreas system for the treatment of type 1 diabetes mellitus: overview of its safety and efficacy. Expert Review of Medical Devices. DOI PubMed
22. [22] Kaur P, Mishra SK, Mithal A (ONWARDS 1-6 design paper) (2023). Rationale and design of the phase 3a development programme (ONWARDS 1-6 trials) investigating once-weekly insulin icodec in diabetes. Diabetes, Obesity and Metabolism. DOI PubMed
23. [23] Heise T, Nosek L, Bottcher SG, Hastrup H, Haahr H (2012). Ultra-long-acting insulin degludec has a flat and stable glucose-lowering effect in type 2 diabetes. Diabetes, Obesity and Metabolism. DOI PubMed
24. [24] Hoffmann-La Roche / Banting FG, Best CH, Collip JB, Campbell WR, Fletcher AA (1922). Pancreatic extracts in the treatment of diabetes mellitus. Canadian Medical Association Journal.
25. [25] Goeddel DV, Kleid DG, Bolivar F, et al. (1979). Expression in Escherichia coli of chemically synthesized genes for human insulin. Proceedings of the National Academy of Sciences USA. DOI PubMed
26. [26] Steiner DF (2004). The proinsulin C-peptide: a multirole model. Experimental Diabesity Research. DOI PubMed
27. [27] White MF (2002). IRS proteins and the common path to diabetes. American Journal of Physiology - Endocrinology and Metabolism. DOI PubMed
28. [28] FDA (2026). FDA approves first once-weekly basal insulin (Awiqli, insulin icodec-abae) for type 2 diabetes. FDA News Release.
29. [29] Riddle MC, Bolli GB, Ziemen M, et al. (EDITION 1) (2014). New insulin glargine 300 units/mL versus glargine 100 units/mL in people with type 2 diabetes using basal and mealtime insulin: glucose control and hypoglycemia in a 6-month randomized controlled trial (EDITION 1). Diabetes Care. DOI PubMed
30. [30] Syed YY (2024). Insulin icodec: first approval. Drugs. DOI PubMed