Why Does My Ozempic Shot Leave a Lump? Its Label Never Says

A firm lump, knot or bump that has built up over weeks or months at a spot you keep injecting is common enough to have a name. In insulin users it affects between a third and half of people, and the name is lipohypertrophy. It is usually painless, it is not an infection, and it is not an emergency.

What it does do is make absorption from that spot unreliable, and that unpredictability is why it matters at all. The two changes with evidence behind them are rotating sites within the same area of your body and never reusing a needle.

The complication is thoroughly documented for insulin. For semaglutide, tirzepatide and research peptides it is not documented at all. Most of what you will read about injection lumps is insulin research being applied to drugs nobody has checked it against.

What is the lump most people mean?

Lipohypertrophy is the lump most people mean, and the insulin label gives a plain definition of it. Lantus describes subcutaneous insulin as having "resulted in lipoatrophy" and then defines the two forms: lipoatrophy is "(depression in the skin) or lipohypertrophy (enlargement or thickening of tissue) in some patients."¹

So the condition comes in two directions. Tissue either sinks away at the site, leaving a dent, or it thickens and builds outward, leaving a firm raised area that people describe as a knot, a bump or scar tissue. The patient leaflet on the same label puts both in everyday words as "skin thickening or pits at the injection site (lipodystrophy)."¹

None of those words describes the red, itchy, short-lived reaction people also get after an injection, where the swelling comes up fast and settles. The label treats that separately, listing injection site reactions "including redness, pain, itching, urticaria, edema, and inflammation" under allergic reactions.¹ Our guide to when an injection site reaction means stop and get seen covers the ones that need attention.

How common is it in insulin users?

Insulin users show prevalence running from roughly a third to more than half, depending on who is counted and how carefully. A study in Diabetes Therapy examining children and adults found that "Thirty-nine percent and 32% of children and adults had LH," and among insulin pump users specifically, "Of the pump users, 18 had LH (36%)."²

A larger cross-sectional study of 407 insulin users across three hospitals in northwest Ethiopia put the figure higher. It reports that "Insulin-induced lipodystrophy was observed in 225 cases (53.1%)," with grade 2 lipohypertrophy the most common form at 52.7% of cases and grade 3 lipoatrophy rare at 1.8%.³

Those are not small numbers for a complication many people have never heard of. They also come from populations injecting several times a day for years, a different exposure pattern from a weekly GLP-1 injection, and that difference matters when deciding how much of this transfers.

What does the insulin label actually say?

The insulin label states the rotation instruction and the reason for it in the same sentence. Its Highlights section says "Rotate injection sites to reduce the risk of lipodystrophy," and the administration section expands that to "rotate injection sites within the same region from one injection to the next to reduce the risk of lipodystrophy."¹

Both halves of that line carry weight, and the second half is the part people skip. Lantus does not simply say to move around. It says to move around within the same body region, because switching regions changes absorption on its own. And the stated purpose is prevention of a specific named complication.

The practical reading is narrow. Move around inside your abdomen, or inside one thigh, instead of alternating between abdomen and thigh, because the second pattern changes how fast the drug arrives as well as where it goes in. The patient leaflet on the same label states the consequence of not doing it in plain words, listing "skin thickening or pits at the injection site (lipodystrophy)" among the possible side effects.¹

Does a lump change how the drug is absorbed?

A lump makes absorption from that site less predictable. Unpredictable absorption is why any of this matters. The Diabetes Therapy paper states the consequence directly: "It results in an erratic absorption of insulin."²

The same paper reports, citing earlier work by Famulla and colleagues, that lipohypertrophy raises within-subject uptake variability roughly three to five fold.² A spread that wide means the same dose into the same lump behaves differently on different days.

The clinical signal in that study is consistent with it. "HbA1c was lower in children and adults with no LH," at a p value below 0.001.² Insulin dose itself showed no reliable difference between those with and without lipohypertrophy, in either children or adults, so the effect showed up in control instead of in how much people were taking.²

What should you actually change?

Two changes have evidence behind them, and both come straight out of the documents instead of from anyone's opinion. Rotate within one body region instead of between regions, and stop reusing needles.

The rotation instruction is on both labels. Lantus says to "rotate injection sites within the same region from one injection to the next to reduce the risk of lipodystrophy," and Ozempic says to "use a different injection site each week when injecting in the same body region."¹⁴ Neither says to alternate between your abdomen and your thigh. Those two sites absorb at different rates, so swapping between them adds a second source of variation on top of the first.

Needle reuse is the one with the largest number attached. Going past ten uses of a single syringe was associated with lipodystrophy at an adjusted odds ratio of 4.09 in the Ethiopian study, against 3.06 for six to ten uses, both at p below 0.001.³ Those figures come from a setting where reuse was routine, so the size may not carry over to someone using a fresh pen needle most of the time.

Neither change is a treatment for a lump you already have. They are both about not adding to it.

Will a lump go away on its own?

Nothing in either label or either study says how long a lump lasts or whether leaving a site alone reverses it. That silence needs saying plainly, because it is the question people ask first and the documents do not answer it.

What the insulin literature does establish is that the change sits in the tissue and not on the surface, so it gets diagnosed by feeling the site instead of looking at it.²³ A firm area that has taken months to build is not the same kind of thing as a welt that comes up in an hour and settles by morning.

For anything that is hot, spreading, increasingly painful or accompanied by fever, the timeline question does not apply and the answer is a clinician. Our guide on injection site reactions that need attention sets out those signs.

Does Ozempic's label mention any of this?

No. Ozempic's label does not mention it anywhere. The words lipodystrophy, lipohypertrophy and lipoatrophy appear nowhere in the prescribing information revised in October 2025.⁴

The only use of the word nodule in that label refers to thyroid nodules in the boxed warning, not to anything at an injection site.⁴ Whether the absence reflects a drug that does not cause the problem, or one where nobody has looked, the label does not say. One thing to hold in mind: the insulin figures come from people injecting several times a day for years, against once a week here, so a lower rate would be the expected finding even if the mechanism were identical.²³

What the label does report for the injection site is short. Under other adverse reactions it states that "injection site reactions (e.g., injection-site discomfort, erythema) were reported in 0.2% of OZEMPIC-treated patients."⁴ Two in a thousand, and the described reactions are discomfort and redness, not lumps.

Why does Ozempic tell you to rotate, then?

Ozempic tells you to rotate and gives no reason for it. Section 2.1 says to "Instruct patients to use a different injection site each week when injecting in the same body region," and the medication guide tells patients to "Change (rotate) your injection site with each injection."⁴

Set that beside the insulin label and only the reasoning differs. Lantus attaches a purpose. Ozempic tells you what to do and stops.

The instructions for use section is thinner still. At the injection step it says only "Choose your injection site and wipe the skin with an alcohol swab," with no rotation language at all.⁴ A person following the step-by-step pictures and not the full label would never encounter the advice, and the pictures are what most people follow after the first week.

Does rotating actually prevent it?

Whether rotating actually prevents lumps is unsettled. These two studies point in different directions. The Ethiopian study found failure to rotate associated with lipodystrophy at an adjusted odds ratio of 1.8, with a confidence interval of 1.02 to 3.19.³

The Diabetes Therapy study found the opposite pattern and said so plainly: "In our study, we found that the rotation routine was positively associated with the frequency of LH."² That association reached significance in children, at a p value of 0.026, and not in adults, at 0.287.²

A likely explanation is that people who already have lumps are told to rotate and start doing it, so the behavior follows the condition instead of preventing it. Cross-sectional studies cannot separate those, and neither paper claims to. Rotating sites is something the labels tell you to do for a sensible reason, without clean prospective evidence behind it.

Does reusing a needle matter?

Needle reuse shows the strongest association of any risk factor measured in these two studies. Using a syringe 6 to 10 times came out at 3.06 in the Ethiopian paper, confidence interval 2.00 to 4.25, and going past 10 times at 4.09, confidence interval 3.42 to 5.19, both at p below 0.001.³

Those are larger effects than rotation, insulin dose or glycemic control showed in the same analysis. The Diabetes Therapy study could not test the question at all. Its authors wrote that "As the rate of needle re-use was low, we were not able to establish its association with the prevalence of LH."²

One caution on reading across: the reuse figures come from a setting where syringes were being used many times over, a different practice from the single-use pens and syringes most readers here will have. The finding holds in that population and its size may not transfer.

Is a dent the same thing as a lump?

A dent and a lump are the two directions of the same condition. They are nowhere near equally common. The Ethiopian study graded what it found and reported that grade 2 lipohypertrophy, the raised form, accounted for 52.7% of cases, while grade 3 lipoatrophy, the sunken form, came to 1.8%.³

So the lump is the ordinary presentation by a wide margin and the dent is rare. That ratio has shifted over the decades, since lipoatrophy was far more common when insulins were less purified, and the modern picture is dominated by the thickened form.

Both forms appear in the same sentence of the Lantus label, named together there as the two ways the tissue changes.¹ A reader who finds a soft depression at an old site is looking at the same family of change, not a different problem.

What else raises the risk?

Three other factors reached significance in the Ethiopian analysis. Needle reuse was not the largest of them in every comparison. The study reports adjusted odds ratios for each, drawn from 407 insulin users across three hospitals between April and September 2023.³

Two of those sit differently from the others. Glycemic control and insulin dose per kilogram are plausibly consequences as much as causes, since lumps worsen control and people whose control is poor end up on higher doses. A cross-sectional design cannot tell which direction the arrow points, and the paper does not claim to.

One number in that table needs a warning label of its own. The rotation row appears in the Ethiopian abstract as "Failure to rotate (AOR = 1.8, p-value 0.004)" and in its own results section at a p value of 0.04, a tenfold difference, and the confidence interval of 1.02 to 3.19 sits almost on the no-effect line, fitting the weaker figure.³ So the results section looks like the reliable half and the abstract looks like the typo.

The Diabetes Therapy study looked at insulin dose too and found nothing. Children with lumps used a mean of 34.9 units a day against those without, p 0.246, and in adults the means were 73.8 and 69.7, p 0.765.² Two studies, two opposite answers on the same variable.

Is there any evidence for research peptides?

No source cited on this page addresses lumps from research peptide injection, and no label exists for those compounds because none is approved.¹⁴ BPC-157, TB-500 and the rest are injected subcutaneously in the same tissue by the same technique, so the question is reasonable and it remains unanswered.

The insulin mechanism is specific to insulin. Lipohypertrophy is attributed to repeated local exposure to a hormone that acts on fat cells. Peptides without that action would not be expected to produce the same tissue change by the same route.

So a firm lump at a peptide injection site is not automatically lipohypertrophy, and treating it as though it is means borrowing a diagnosis from a drug class with different properties.² I did not run a systematic search for peptide-specific case reports, so read this as a gap in the sources here and not as a finding that none exist. Our guide to what actually goes wrong with peptides covers the documented problems, and lumps are not among the characterized ones.

What else produces a lump at an injection site?

Several things other than lipohypertrophy can produce a lump. Insulin labels separate them explicitly. Lantus files redness, pain, itching, urticaria, edema and inflammation under local allergy, in a different place from its lipodystrophy entry.¹

Timing separates them in the documents, since the allergic and inflammatory reactions appear soon after an injection and settle, while lipodystrophy is described as a change in the tissue that develops at sites used repeatedly.¹

Anything hot, spreading, increasingly painful or accompanied by fever sits outside what either label describes as expected, and belongs with a clinician instead of an article. Neither label puts a timeframe on how long a lipodystrophic change persists or whether it resolves once a site is left alone, so that question has no answer in the documents either. Our guide on injection site reactions that need attention sets out those signs.

What would settle this?

Measuring it in GLP-1 users would settle most of it. Both insulin prevalence studies worked by examining and palpating injection sites in a clinic, an inexpensive design nobody appears to have run in people injecting semaglutide or tirzepatide weekly.²³

Until someone does, the position is that lipohypertrophy is documented, common and consequential in insulin users, absent from the Ozempic label entirely, and uncharacterized in research peptides.¹⁴ Our guides on where to inject and why rotation matters and why an injection itches or welts cover the adjacent questions.

A lump that has been building for months at one spot should be shown to a clinician who can feel it, since palpation is how the condition was diagnosed in every study cited here.

Common questions

What is the hard lump at my injection site?

Is a knot the same thing as a lump?

How common is lipohypertrophy?

Does Ozempic cause lipohypertrophy?

What does Ozempic's label say about injection site problems?

Why do both labels tell you to rotate?

Does rotation prevent lumps?

Does reusing needles cause lumps?

Do lumps affect how well the drug works?

Do people with lumps have worse blood sugar control?

Can BPC-157 or TB-500 cause lipohypertrophy?

Is a lump the same as an injection site reaction?

How do I know a lump is not something worse?

Where should I rotate to?

Is this documented for tirzepatide?

How is lipohypertrophy diagnosed?

What research would answer this for GLP-1 users?

Sources

1. Sanofi-Aventis. LANTUS (insulin glargine injection) for subcutaneous use, US prescribing information, revised 5/2019. Read at source as the FDA-hosted label PDF on accessdata.fda.gov. Quoted for the Highlights rotation instruction, the section 2.1 administration instruction, the section 6.1 lipodystrophy entry defining lipoatrophy and lipohypertrophy, the section 6.1 local allergy list, the section 17 patient wording, and the instructions for use rotation line. The Highlights cross-reference to section 2.2 against text printed in section 2.1 is noted above as printed. ↩1 ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12 ↩13 ↩14 ↩15 ↩16
2. Deeb A, Abdelrahman L, Tomy M, et al. Impact of Insulin Injection and Infusion Routines on Lipohypertrophy and Glycemic Control in Children and Adults with Diabetes. Diabetes Therapy, 2019, volume 10, pages 259 to 267. Read at source on the publisher's site. Quoted for the prevalence figures in children, adults and pump users, the HbA1c comparison, the erratic absorption statement, the rotation association with its p values, and the needle reuse limitation. The three to five fold variability figure is this paper's citation of Famulla et al., 2016, and that underlying paper was not retrieved. ↩1 ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12 ↩13 ↩14 ↩15 ↩16 ↩17 ↩18
3. Bazezew ZA, Zeleke TK, Negesse CT, et al. Lipodystrophy and associated factors among patients with diabetes receiving insulin therapy: a multicenter study in Ethiopia. Scientific Reports, 2026, volume 16, article 13560. Read at source on the publisher's site. Cross-sectional, 407 participants across three hospitals in northwest Ethiopia, conducted April 5 to September 30, 2023. Quoted for the 53.1% prevalence, the grade breakdown, the rotation odds ratio and the needle reuse odds ratios. The abstract and the results section disagree on the rotation p value, 0.004 against 0.04, and both figures are printed above. ↩1 ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12 ↩13 ↩14
4. Novo Nordisk. OZEMPIC (semaglutide) injection, for subcutaneous use, US prescribing information, Highlights revised 10/2025. Read at source as the FDA-hosted label PDF on accessdata.fda.gov, in two passes covering the full document of roughly 131,000 characters. Quoted for the injection site reaction incidence, the section 2.1 and medication guide rotation instructions, and the instructions for use injection step. The absence of lipodystrophy, lipohypertrophy and lipoatrophy from the label was established by searching both passes, not inferred from a partial read. ↩1 ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12
5. Eli Lilly and Company. ZEPBOUND (tirzepatide) injection, for subcutaneous use, US label via the FDA structured product labeling service. Not retrieved. Only the product data section was returned, and the clinical sections containing adverse reactions and administration instructions were not reached. No claim about tirzepatide appears on this page as a result, and the FAQ says so. ↩1