Skip to content
PeptideAI
Approved drugblood sugarmuscle growth

Insulin analogues

The archetypal therapeutic peptide hormone family, spanning rapid-acting mealtime analogues through to once-weekly basal insulin icodec.

Also known as insulin lispro, insulin aspart, insulin glargine, insulin degludec, insulin icodec, Humalog, Novolog, Lantus, Levemir, Tresiba, Toujeo, Awiqli, Lyumjev, Fiasp

Approved drug — Licensed by a major regulator for human use, with phase-3 trial data behind it.

A century of clinical use and the most thoroughly studied therapeutic peptide in existence, with modern analogues supported by extensive phase 3 programmes (ONWARDS for icodec, BEGIN for degludec, EDITION for glargine U300). The evidence is not in question; the risk is entirely in the dosing.

How it works

All insulin analogues are the same hormone with the same receptor pharmacology; what differs is the pharmacokinetics engineered into them. Rapid-acting analogues (lispro, aspart, glulisine) carry substitutions that prevent hexamer formation so they absorb in 10-20 minutes and cover meals. Basal analogues use different tricks: glargine shifts the isoelectric point so it precipitates at subcutaneous pH, detemir and icodec use fatty-acid acylation for albumin binding, and degludec forms soluble multihexamer chains that dissolve slowly. Insulin icodec extends albumin binding far enough for weekly dosing. Downstream, receptor activation triggers PI3K-Akt signalling driving GLUT4 translocation, glycogen synthesis, lipogenesis and mTOR-mediated protein synthesis - the last of which is why insulin appears in bodybuilding contexts, and why it kills people who misuse it.

Targets: Insulin receptor, IGF-1 receptor (weak cross-reactivity)

Dosing

ProtocolDoseFrequencyRoute
Basal-bolus regimen (illustrative only)Basal at a fixed time each day; rapid-acting 0-15 minutes before eating.—basal once or twice daily plus a bolus with each mealsubcutaneous
Weekly basal insulin icodecSame day each week.—once weeklysubcutaneous
  • · Insulin is dosed in units, not micrograms, and the correct dose is entirely individual - typically starting around 0.2-0.4 units/kg/day total in type 1 diabetes and titrated on glucose readings. There is no generic dose that is safe to publish, and copying someone else's numbers is how people end up in a coma.
  • · Awiqli was FDA-approved in March 2026 for type 2 diabetes. Doses are in units and are titrated weekly; the long duration means a dosing error takes a week to clear rather than a day.

Titration

Titrate basal insulin on fasting glucose, typically by 2-4 units every 3 days, and bolus insulin on post-meal readings. Weekly insulins need slower, more patient titration because a change takes 2-3 weeks to fully express.

Cycling

Replacement therapy for life in type 1 diabetes; long-term in insulin-requiring type 2. Not something to cycle.

Work out your exact syringe units →

Pharmacology

Half-life
Ranges enormously by analogue: rapid-acting insulins act over 3-5 hours, glargine and degludec over 24-42 hours, and insulin icodec has a half-life of roughly a week.
Onset
Rapid-acting analogues start working in 10-20 minutes; basal analogues take 1-4 hours to onset and days to reach steady state.
Routes
subcutaneous, intravenous, inhaled
Molecule
Recombinant two-chain 51-amino-acid peptide hormone analogues
Sequence length
51 amino acids

Handling

Diluent
Not applicable - supplied as sterile solutions in vials, cartridges and pens
Lyophilised
Not applicable.
Reconstituted
Unopened vials and pens refrigerated at 2-8 C until expiry; once in use, most are good for 28 days at room temperature (degludec 56 days).
Light sensitive
Yes — keep it out of the light

Mixing

Never dilute insulin except with the manufacturer's specific diluent under clinical direction.

Side effects

  • very commonHypoglycaemia— The defining risk. Severe hypoglycaemia causes seizure, coma and death, and non-diabetic misuse for muscle-building has killed people.
  • very commonWeight gain— Anabolic and antilipolytic effects plus defensive eating against lows.
  • very commonLipohypertrophy at injection sites— Caused by repeatedly injecting the same spot; it makes absorption erratic, which then causes unexplained lows.
  • commonHypokalaemia— Insulin drives potassium intracellularly - clinically important in DKA management and in overdose.
  • uncommonInjection-site reactions and rare allergy
  • uncommonPeripheral oedema on initiation— Usually transient.

Do not use if

  • Hypoglycaemia at the time of administration.
  • Non-diabetic use for muscle-building - there is no dose that is both effective and safe, and fatal hypoglycaemia in this context is well documented.
  • Inhaled insulin specifically is contraindicated in asthma and COPD.

Combining it

  • cautionsemaglutide — GLP-1 agonists substantially reduce insulin requirements; doses need cutting at initiation to avoid hypoglycaemia.
  • cautionpramlintide — Boxed-warning-level interaction - mealtime insulin must be halved when pramlintide is started.
  • cautionigf-1-lr3 — Both drive glucose into cells; the combination is a well-known cause of severe hypoglycaemia in performance contexts.
  • conflictglucagon — Directly opposing hormones; glucagon is the rescue for insulin excess.

What to monitor

  • · Continuous glucose monitoring or frequent fingersticks - non-negotiable.
  • · HbA1c every 3 months.
  • · Injection-site rotation and inspection for lipohypertrophy.
  • · Potassium in acute or high-dose settings.

Legal status

FDA- and EMA-approved, prescription-only. Human insulin is available without prescription in some US states, which is a major contributor to misuse injuries.

References

  • ONWARDS 1-6 phase 3 programme for insulin icodec (trial)
  • ADA Standards of Care in Diabetes, insulin therapy chapter (guideline)
  • Awiqli, Tresiba, Lantus and Humalog US prescribing information (label)

Mechanism in depth

The insulin receptor is a receptor tyrosine kinase, not a GPCR, which makes it mechanistically unlike everything else in this class. Binding triggers autophosphorylation and recruitment of insulin receptor substrate proteins, which branch into two arms: the PI3K-Akt arm handles the metabolic effects (GLUT4 translocation to the membrane in muscle and fat, glycogen synthase activation, suppression of hepatic gluconeogenesis, inhibition of hormone-sensitive lipase), and the MAPK arm handles mitogenic effects. That branching is why insulin analogue design has always had to watch IGF-1 receptor affinity - a molecule that over-activates the mitogenic arm is a theoretical cancer risk, and it is the reason one early analogue programme was abandoned. The pharmacological story of the last thirty years is entirely about controlling absorption rate rather than about receptor pharmacology, because all these molecules do the same thing at the receptor. Making insulin faster means destabilising the hexamer so it disperses more quickly from the depot. Making it slower means the opposite: precipitation at physiological pH (glargine), soluble multihexamer chains that disassemble slowly (degludec), or albumin binding (detemir, degludec, icodec). Icodec is the extreme version, combining albumin binding with substitutions that slow receptor-mediated clearance, giving a weekly injection. The clinical consequence of a weekly basal insulin is worth stating plainly: it is convenient, and it means a dosing error takes a week to wash out rather than a day. The ONWARDS programme showed non-inferior or superior HbA1c with icodec but the hypoglycaemia picture is the thing to watch, particularly in type 1 diabetes.

What usually goes wrong

Insulin causes more deaths and hospital admissions than every other compound in this class combined, and almost none of it is about the drug being wrong - it is about arithmetic. The specific patterns: giving a rapid-acting analogue and then not eating; not reducing insulin when starting a GLP-1 agonist or a dual agonist; not reducing insulin as renal function declines, because injected insulin is renally cleared; injecting repeatedly into a lipohypertrophic site and getting erratic absorption, then correcting for the erratic absorption; confusing a rapid-acting pen with a basal pen, which is why they are colour-coded and why a U-500 concentrated insulin in a U-100 syringe is a five-fold overdose; and alcohol, which suppresses gluconeogenesis and causes delayed hypoglycaemia hours later. Weekly icodec adds a new version of the same problem: a dose error takes a week to leave the body. Human insulin being available without a prescription in some US states is a genuine contributor to injury.

Bloodwork worth running

MarkerWhenWhy it matters
Continuous glucose monitoring - time in range, time below range, and glucose variabilityContinuously.This has replaced HbA1c as the primary tool for anyone on insulin, and it is not close. HbA1c averages away the hypoglycaemia that actually harms people.Act if: More than 4% of time below 3.9 mmol/L (70 mg/dL), or any time below 3.0 mmol/L (54 mg/dL), means the regimen is too aggressive regardless of what the HbA1c says.
HbA1cEvery 3 months.Still the standard long-term measure and the basis of most outcome evidence, but it must be read alongside CGM data.Act if: A low HbA1c with frequent hypoglycaemia is a worse result than a slightly higher one without it.
PotassiumDuring acute illness, dose changes in renal impairment, or DKA treatment.Insulin drives potassium into cells. This matters acutely in DKA management and in anyone with renal impairment or on potassium-affecting drugs.Act if: Potassium under 3.5 mmol/L needs correcting before aggressive insulin dosing.
Creatinine and eGFREvery 6 months, and after any unexplained pattern of hypoglycaemia.Injected insulin is substantially renally cleared, so falling kidney function raises insulin exposure and causes unexplained hypoglycaemia. This is one of the most under-recognised causes of recurrent lows in long-standing diabetes.Act if: A falling eGFR means the insulin dose almost certainly needs reducing.
Injection site examination for lipohypertrophy (clinical)Every review.Repeated injection into the same site produces fatty lumps that absorb insulin erratically. This is a leading cause of unexplained glucose variability and it is found by feeling the abdomen, not by any blood test.Act if: Any palpable lump means rotate sites away from it - and expect a substantial fall in insulin requirement when you do, which is itself a hypoglycaemia risk.
Anti-insulin antibodiesOnly in unexplained severe insulin resistance or wildly erratic response.Rare with modern analogues but a real cause of erratic, unexplainable insulin action.Act if: Positive high titres warrant specialist input.

Pharmacokinetics

Time to steady state
3 days
Crosses blood-brain barrier
partial
Accumulates
Yes — doses stack before steady state
Metabolism
Receptor-mediated internalisation and degradation by insulin-degrading enzyme in liver, kidney and muscle. Acylated analogues (detemir, degludec, icodec) additionally bind albumin, which is their protraction mechanism.
Elimination
Renal and hepatic proteolysis.

Receptor targets

  • Insulin receptor (INSR), receptor tyrosine kinase — Analogues are designed to match human insulin's affinity closely; deviations are a design failure rather than a feature

    PI3K-Akt signalling driving GLUT4 translocation, glycogen synthesis, suppression of hepatic glucose output and inhibition of lipolysis. This is the metabolic arm and it is what insulin is for.

  • IGF-1 receptor — Low cross-reactivity by design; specific ratios vary by analogue and were not verified here

    Mitogenic signalling through the MAPK arm. Deliberately minimised in analogue design because excess IGF-1 receptor activity is a theoretical proliferative risk.

Trials

  • ONWARDS 1 Phase 3a · n=984 · 78 weeks · 2023

    Once-weekly insulin icodec produced greater HbA1c reduction than once-daily glargine U100 in insulin-naive type 2 diabetes, with a higher rate of level 2 and 3 hypoglycaemia.

What to expect, and when

Rapid-acting analogues start working in 10-20 minutes, peak at 45-90 minutes and last 3-5 hours. Ultra-rapid formulations shift that earlier by roughly 5-10 minutes. Regular human insulin takes 30-60 minutes to onset and lasts 6-8 hours. Glargine and detemir onset over 1-4 hours with a duration of about 24 hours; glargine U300 and degludec run longer, degludec beyond 42 hours. Icodec has a half-life of roughly one week and takes several weeks to reach steady state, which is why a loading approach is used at initiation.

Stacking and comparisons

Insulin plus any GLP-1 agonist is the combination that matters most here, and it is both genuinely useful and the most common source of serious harm in this entire class. GLP-1 agonists and dual agonists reduce insulin requirement substantially - a 2.0-2.4% HbA1c reduction from tirzepatide arriving over eight weeks means the previous insulin dose is now wrong by a lot. The rule is to cut prandial insulin by 20-50% at initiation and reduce basal as glucose falls, with continuous glucose monitoring rather than fingersticks during the transition. Fixed combinations exist precisely because getting the ratio right is hard: Xultophy (degludec plus liraglutide), Soliqua (glargine plus lixisenatide) and IcoSema (icodec plus semaglutide). Insulin plus a sulfonylurea is redundant and hypoglycaemogenic - stop the sulfonylurea. Insulin plus an SGLT2 inhibitor carries a euglycaemic diabetic ketoacidosis risk that is easy to miss because the glucose looks normal. Beta-blockers mask the adrenergic warning symptoms of hypoglycaemia. And anyone on insulin should have glucagon rescue in the house and someone trained to use it.

Against every incretin here: insulin is the only compound in this class that lowers glucose regardless of what the glucose is, which is exactly why it is both the most powerful and the most dangerous. GLP-1 agonists are glucose-dependent and essentially cannot cause hypoglycaemia alone; insulin has no such safety mechanism. Within the family: degludec and glargine U300 have flatter profiles and less nocturnal hypoglycaemia than glargine U100; icodec trades daily convenience for a week-long error window. The most important comparative point is directional - as incretin therapy has become more effective, the correct move in type 2 diabetes has increasingly been to reduce or remove insulin rather than to add to it.

Rough cost

$35–$300/month. In the US, manufacturer cap programmes from Lilly, Novo and Sanofi put many insulins at around 35 per month out of pocket, and Medicare caps insulin copays at 35. Without any of those, list prices for analogue insulins run several hundred dollars per month. Market observation from mid-2026, not verified pricing.

Genuinely uncertain

  • This is a family entry rather than a single molecule, so most pharmacokinetic fields are left null - any single number would be wrong for most of the family.
  • No individual insulin label was fetched and verified in this session; the analogue modification descriptions are from standard published structural descriptions and were not verified residue-by-residue.
  • IGF-1 receptor affinity ratios differ by analogue and were not verified.
  • The 984-participant figure for ONWARDS 1 was not individually re-verified against the paper.
  • Cost figures are market observations and US-specific, not verified pricing.

Papers