Skip to content
PeptideAI
Human RCTblood sugar

Insulin efsitora alfa

Eli Lilly's once-weekly basal insulin, a single-chain insulin fused to an antibody Fc fragment, giving an unusually flat profile across the whole week.

Also known as basal insulin Fc, BIF, LY3209590, efsitora, LY3209590

Human RCT — Randomised controlled trials in humans, but not an approved product for this use.

Five completed phase 3 QWINT trials show HbA1c reduction non-inferior to daily basal insulin in type 2 diabetes. As of mid-2026 it remains investigational, with an FDA decision expected around July 2026; it is not yet marketed anywhere.

How it works

Efsitora is a single-chain insulin variant, engineered so the A and B chains are joined by a peptide linker, fused to a human IgG2 Fc fragment. The Fc portion is bound and recycled by the neonatal Fc receptor in endothelial and immune cells, the same salvage pathway that gives therapeutic antibodies their long lives, which stretches the half-life to about 17 days. Reduced insulin receptor affinity slows receptor-mediated clearance further. The practical signature is an exceptionally flat pharmacodynamic profile, with a peak-to-trough ratio far lower than any daily basal insulin, so there is little day-to-day sawtooth within the weekly interval. The QWINT phase 3 programme showed non-inferior HbA1c reduction versus daily glargine and degludec in type 2 diabetes, with fixed-dose escalation regimens that avoid the complexity of unit-by-unit titration.

Targets: Insulin receptor, Neonatal Fc receptor recycling, Hepatic glucose output

Dosing

ProtocolDoseFrequencyRoute
Type 2 diabetes, weekly basal insulin (investigational regimen from the QWINT programme)Same day each week.—once weeklysubcutaneous
  • · Dosed in insulin units. The trials used a fixed-dose escalation approach, starting at 100 units weekly in insulin-naive type 2 diabetes and stepping between preset dose levels based on fasting glucose rather than titrating unit by unit. Final labelled dosing depends on the approved product information.

Titration

Fixed-dose steps at weekly to fortnightly intervals in trials. As with icodec, the extremely long half-life means over-eager titration produces hypoglycaemia days later.

Cycling

Intended as continuous lifelong basal therapy, like any other basal insulin.

Work out your exact syringe units →

Pharmacology

Half-life
About 17 days, which is the longest of any insulin developed to date.
Onset
Glucose lowering starts within days, but steady state takes 3 to 5 weeks.
Routes
subcutaneous
Molecule
Fusion protein of a single-chain insulin variant and a human IgG2 Fc domain

Handling

Diluent
Not applicable. Developed as a prefilled pen solution.
Lyophilised
Not applicable.
Reconstituted
Refrigerated storage before use is expected, in line with other insulin pens. Confirm against the final label once marketed.
Light sensitive
Yes — keep it out of the light

Side effects

  • commonHypoglycaemia— Rates were comparable to daily basal insulin in type 2 diabetes trials, but higher than degludec in the type 1 QWINT-5 study.
  • commonWeight gain
  • commonInjection-site reactions
  • uncommonHypokalaemia

Do not use if

  • Hypoglycaemia at the time of intended dosing
  • Diabetic ketoacidosis
  • Situations with rapidly changing insulin requirements, given a 17-day half-life
  • Hypersensitivity to the product

Combining it

  • redundantinsulin-icodec — Direct competitor in the weekly basal insulin category; you would use one or the other.
  • synergytirzepatide — Weekly incretin plus weekly basal insulin is the obvious pairing; basal dose typically needs reducing when the incretin is added.

What to monitor

  • · Fasting glucose or continuous glucose monitoring during titration
  • · HbA1c every 3 months
  • · Hypoglycaemia episodes and their timing within the weekly interval

Legal status

Investigational. Not approved or commercially available as of July 2026; regulatory submissions are under review in the US and elsewhere.

References

  • QWINT-1 through QWINT-5 phase 3 trials of insulin efsitora alfa (trial)
  • Eli Lilly investor disclosures on the efsitora phase 3 programme and regulatory submissions (other)

Mechanism in depth

Efsitora solves the same problem as icodec by a completely different route, and the contrast is instructive. Icodec uses albumin as a passive depot. Efsitora uses an active salvage pathway: the neonatal Fc receptor in vascular endothelium and immune cells binds IgG Fc domains at the acidic pH of the endosome, diverts them away from lysosomal destruction and releases them back into circulation at neutral extracellular pH. That is the mechanism that gives monoclonal antibodies their three-week half-lives, and grafting an Fc domain onto insulin borrows it wholesale. Layered on top is the single-chain insulin design, which links the A and B chains and substantially reduces insulin receptor affinity, slowing the receptor-mediated internalisation that normally clears insulin fast. The result is a 17-day half-life and a pharmacodynamic profile flatter than anything else in the class, with very little within-week sawtooth. The downstream biology is ordinary insulin biology once the receptor is engaged. The clinically distinctive feature of the QWINT programme was the dosing approach: rather than titrating unit by unit, the trials used fixed-dose escalation between preset dose levels driven by fasting glucose, which suits a molecule whose exposure changes so slowly that fine adjustment is meaningless. That is a genuinely different way to think about insulin dosing and, if it survives into the label, it may simplify basal insulin management considerably.

What usually goes wrong

The 17-day half-life is the whole risk story, and it is longer than icodec's. A dose given today is still contributing meaningfully in two and a half weeks. That makes efsitora the least forgiving insulin ever developed in terms of error recovery, and it makes patient selection critical: erratic eating, frequent illness, alcohol excess, unpredictable exercise and poor hypoglycaemia awareness are all much more consequential than they would be on daily basal. Over-titration is the other major hazard, because steady state takes three to five weeks and any escalation made before then is stacking on a rising baseline. The type 1 QWINT-5 study showed higher hypoglycaemia than degludec, which is the same pattern icodec showed in type 1, and it points to the same conclusion: weekly basal insulin suits type 2 diabetes with residual endogenous insulin secretion and stable requirements, and does not suit type 1. And a straightforward practical caution: as of mid-2026 this product is investigational and not marketed anywhere, so anything sold under that name outside a clinical trial is not the real thing.

Titration ladder

  1. —Week 1, insulin-naive type 2 diabetes (QWINT regimen) — Dosed in insulin units, not micrograms. The trials started at 100 units once weekly in insulin-naive type 2 diabetes.
  2. —Weekly to fortnightly steps — Fixed-dose escalation between preset dose levels based on fasting glucose, rather than unit-by-unit titration. This is a deliberate simplification suited to a molecule whose exposure changes over weeks.
  3. —Weeks 3-5 — Steady state takes 3 to 5 weeks. As with icodec, over-eager escalation produces hypoglycaemia days to weeks later, and with a 17-day half-life the lag is even longer and the correction even slower.
  4. —Final labelled dosing — Depends on the approved product information, which did not exist at the time of writing. Treat all of the above as the investigational regimen.

Bloodwork worth running

MarkerWhenWhy it matters
Fasting glucose, or continuous glucose monitoringDaily during titration, ideally by CGM.The titration variable in the QWINT fixed-dose escalation approach, and the practical safety monitor.Act if: The trials stepped between preset dose levels based on fasting glucose rather than titrating unit by unit. Any hypoglycaemia in the preceding interval steps the dose down rather than holding it.
Time below range on CGMContinuously during titration and reviewed at each step.For a 17-day half-life drug, hypoglycaemia is the safety metric that matters and it cannot be corrected by withholding the next dose.Act if: Any time below 3.0 mmol/L, or more than 4 percent below 3.9 mmol/L, means stepping down.
HbA1cEvery 3 months.The registration endpoint across the QWINT programme.Act if: Interpret alongside time below range rather than alone.
Serum potassiumBaseline and periodically in at-risk patients.Standard insulin consideration, relevant in diuretic use, cardiac disease and renal impairment.Act if: Below 3.5 mmol/L needs replacement.
Timing of hypoglycaemia within the weekly intervalDocumented at every episode.A specific and underappreciated monitoring point for weekly insulins: knowing whether hypos cluster in the first days after injection or later in the week tells you whether the problem is the dose or the profile.Act if: Clustering early in the interval suggests the dose is too high; clustering late suggests something else is going on, such as reduced intake or added therapy.

Pharmacokinetics

Time to steady state
28 days
Crosses blood-brain barrier
no
Accumulates
Yes — doses stack before steady state
Metabolism
The Fc domain is handled by the same FcRn salvage pathway that gives therapeutic antibodies their long half-lives. The insulin moiety is degraded conventionally once it escapes that cycle.
Elimination
Proteolytic catabolism after FcRn-mediated recycling is exhausted, plus receptor-mediated internalisation.

Receptor targets

  • Insulin receptor — Deliberately reduced by the single-chain architecture; numeric affinity not resolved this session.

    Conventional receptor tyrosine kinase signalling: suppression of hepatic glucose output, peripheral glucose uptake, inhibition of lipolysis. Reduced affinity slows receptor-mediated clearance.

  • Neonatal Fc receptor (FcRn) — The IgG2 Fc domain binds FcRn at endosomal pH

    Not a pharmacological target but the half-life mechanism: endosomal capture and recycling back to circulation rather than lysosomal degradation, giving roughly 17 days.

  • IGF-1 receptor — Characterised in development but not resolved this session

    Standard mitogenicity consideration for any engineered insulin.

Trials

  • QWINT-1 Phase 3, randomised · 2025

    Weekly fixed-dose insulin efsitora in type 2 diabetes without previous insulin therapy. HbA1c reduction with a simplified fixed-dose escalation regimen rather than unit-by-unit titration.

  • QWINT-2 Phase 3, randomised · 2024

    Insulin efsitora versus degludec in type 2 diabetes without previous insulin treatment. Non-inferior HbA1c reduction.

  • QWINT-3 Phase 3, randomised · 2025

    Once-weekly efsitora versus once-daily degludec in adults with type 2 diabetes already on basal insulin. Non-inferior HbA1c reduction.

  • QWINT-4 Phase 3, randomised · 2025

    Once-weekly efsitora versus once-daily glargine U100 in adults with type 2 diabetes on basal and prandial insulin. Non-inferior HbA1c reduction.

What to expect, and when

Glucose lowering begins within days of the first injection but steady state takes three to five weeks, longer than icodec. The pharmacodynamic profile within each weekly interval is exceptionally flat, so there is little day-to-day variation once steady state is reached, which is the drug's main claimed advantage. Any dose change takes three to five weeks to fully express. Washout after stopping takes a similar period.

Stacking and comparisons

The natural pairing is with a weekly incretin, tirzepatide or semaglutide, giving a genuinely once-weekly injectable regimen for type 2 diabetes. As with any basal, the basal dose typically needs reducing when the incretin is started or escalated, and with a 17-day half-life that reduction has to be anticipated well in advance rather than made in response to a hypo. Metformin and SGLT2 inhibitors continue. Sulfonylureas are usually stopped at initiation because the combined hypoglycaemia risk is hard to manage against an irreversible depot. Corticosteroid courses are genuinely difficult: you cannot ramp a weekly basal up and down around a five-day prednisolone course, so short-term daily basal or prandial cover alongside is the practical answer. The same applies to any planned surgery or acute illness.

Against insulin icodec: the direct competitor, and the comparison comes down to engineering philosophy. Icodec uses reversible albumin binding for a roughly one-week half-life; efsitora uses FcRn recycling for roughly 17 days and a flatter profile. Efsitora's fixed-dose escalation approach is a genuine simplification if it survives into the label, whereas icodec uses conventional unit titration. Neither has demonstrated superiority over daily basal on HbA1c beyond non-inferiority, and both showed more hypoglycaemia than degludec in type 1 diabetes. Against daily basal analogues: five phase 3 trials show non-inferior HbA1c in type 2 diabetes, so the case rests entirely on the reduction from 365 to 52 injections a year and on what that does for acceptance and adherence, not on better glycaemic control. Against no basal insulin: the same argument as icodec, that clinical inertia around starting insulin in type 2 diabetes is a genuine problem and a weekly injection lowers the barrier.

Rough cost

Not marketed, so there is no price. An FDA decision was expected around mid-2026 with submissions under review elsewhere.

Genuinely uncertain

  • No approved label existed at the time of writing, so tmax, bioavailability, volume of distribution, protein binding and numeric clearance are all null.
  • The 28-day time to steady state is the midpoint of the reported three to five week range.
  • The 17-day half-life is consistently reported in the QWINT literature but I did not resolve a primary pharmacokinetic publication this session.
  • Trial enrolments and durations were not confirmed against the papers, so all are null except where a fetched result stated them.
  • QWINT-5, the type 1 diabetes trial referenced in the Core record, was not resolved this session and is therefore not listed as a verified trial here.
  • Regulatory status and final labelled dosing were unresolved at the time of writing, and the fixed-dose escalation regimen described is the investigational one.

Papers