One Insulin Shot a Week: What the Efsitora Approval Covers

Basal insulin, the slow background dose that keeps blood sugar steady between meals, has meant one injection a day for decades. In September 2026 the FDA approved a version that lasts a week. Eli Lilly's insulin efsitora alfa, sold as Onswik, is cleared for adults with type 2 diabetes, alongside diet and exercise. Lilly announced the approval on September 24 and says the pens will reach US pharmacies in the coming months. It is the second weekly insulin the FDA has approved for type 2 diabetes, and it had already been authorized in the European Union, Mexico and Japan.

The appeal is arithmetic: 52 injections a year instead of 365. Lilly's own framing is more than 300 fewer shots. The more interesting question is what the trials say about control and safety, and why the label draws a firm line at type 1 diabetes.

What QWINT tested

The approval rests on the QWINT phase 3 program: four trials with more than 3,400 adults with type 2 diabetes, comparing weekly efsitora with daily insulin glargine or insulin degludec, two widely used basal insulins. Each trial met its primary goal, which was noninferiority on HbA1c, the three month average of blood sugar. In plain terms, the weekly insulin was shown to be not meaningfully worse than daily insulin, not better.

QWINT-2, published in the New England Journal of Medicine, is the clearest example. It randomized 928 adults who had never used insulin. Over 52 weeks, HbA1c fell from 8.21% to 6.97% with efsitora and from 8.24% to 7.05% with degludec, a difference of 0.09 points that was not statistically significant. Time spent in the target glucose range was 64.3% with efsitora and 61.2% with degludec.

Low blood sugar is the risk that matters most with any insulin, and more so with one that stays in the body for a week. In QWINT-2, clinically significant or severe hypoglycemia occurred at 0.58 events per participant year with efsitora and 0.45 with degludec. The rate ratio of 1.30 had a confidence interval from 0.94 to 1.78, so the trial could not rule out a modest increase. No severe episodes were reported with efsitora, against six with degludec.

Why type 1 diabetes is excluded

QWINT-5, published in The Lancet, tested efsitora in 692 adults with type 1 diabetes, who also take mealtime insulin. HbA1c results were again comparable: a drop of 0.51 points with efsitora and 0.56 with degludec at 26 weeks. The safety picture was not. Over 52 weeks, clinically significant or severe hypoglycemia ran at 14.03 events per patient year with efsitora against 11.59 with degludec, and severe hypoglycemia, the kind that needs another person's help, affected 35 of 343 people (10%) on efsitora versus 11 of 349 (3%) on degludec. The highest rates came in the first 12 weeks, while doses were being adjusted.

The FDA's announcement states that the drug is not recommended for type 1 diabetes because of an increased risk of severe hypoglycemia. That is the trade at the heart of weekly insulin. A dose that lasts seven days cannot be trimmed on a day when someone eats less or exercises more. In the type 2 trials, that rigidity did not translate into more severe lows. In type 1, it did.

The other end of the day: faster mealtime insulin

Insulin innovation is also moving in the opposite direction, toward insulin that acts faster at meals. In June 2026, the French company Adocia and its Chinese partner Tonghua Dongbao presented full phase 3 results for BioChaperone Lispro, an ultra rapid formulation of the same molecule as Lilly's Humalog. The trial enrolled more than 1,000 people with type 2 diabetes in China and showed HbA1c lowering not worse than Humalog at 26 weeks, with better glucose control after each of the three daily meals, according to the companies. A full presentation followed at EASD on September 29.

The companies also report superior long term control in two subgroups, people on metformin and people with HbA1c below 8.5%, and positive topline phase 3 results in Chinese adults with type 1 diabetes, announced in October 2025. These are company reported results, and subgroup wins deserve caution. A marketing application in China is in preparation; there is no filing in the US or Europe.

Who stands to benefit

The people most likely to gain from a weekly insulin are the ones who struggle with daily injections or skip them: people starting insulin for the first time, older adults who rely on a caregiver, anyone for whom a daily shot is the barrier between prescription and use. QWINT-2 was run in exactly that first group, which is why it is the most relevant result for the approved label. For people already on GLP-1 drugs, the trial was also designed to check users and nonusers separately and found noninferiority in both.

Weekly insulin will not suit everyone, and the type 1 data are a reminder that convenience has a price when doses need daily flexibility. For a large group of adults with type 2 diabetes, though, removing 313 injections a year is not a small thing, and the evidence says control holds up. For another option that reached patients this year, see our piece on the daily GLP-1 pill tested against insulin glargine.