Joint Pain, Back Pain and Aches on a GLP-1: What the Trials Counted

Aches, stiff joints and a sore back are a common thing to notice a few weeks into a GLP‑1, and a common thing to blame on it. The trials did count them. In most of them, back pain and joint pain were about as common on the drug as on placebo, and no FDA label for these medicines lists either one. The one kind of pain the labels do list is in the skin, and it is worth knowing about if you are on a high semaglutide dose.

What the trials counted

Each trial posts its adverse-event tables on ClinicalTrials.gov, including the everyday complaints that never reach a label. Back pain and joint pain (arthralgia) show up in them, and on placebo nearly as often as on the drug.

Read the rows and a pattern emerges. On semaglutide in STEP 1, back pain was identical to placebo. On tirzepatide, it was about one percentage point higher, with no rise from the lowest dose to the highest. On Ozempic, and in the knee-arthritis trial, joint pain was actually less common on the drug. Across the placebo-controlled records that listed these terms, none showed back or joint pain climbing with the dose.^[5][6][7]

The exception needs saying plainly. In STEP UP, the trial behind the newer 7.2 mg Wegovy dose, back pain was reported by 5.1% on 7.2 mg and 7.0% on 2.4 mg against 2.0% on placebo. The placebo group was small, 201 people, and joint pain on 7.2 mg was no higher than on placebo.^[8] It is one trial against several, but it is the one that tested the highest dose.

What the labels list, and what they leave out

Search the full text of the current labels for Wegovy, Zepbound, Mounjaro, Ozempic, Rybelsus, Saxenda and Foundayo, including the Medication Guides, and the words back pain, arthralgia, myalgia and joint do not appear.^[1][2] That fits the trial tables: the Wegovy and Zepbound labels list reactions seen in at least 2% of people and more often than on placebo, and back and joint pain did not clear that bar.

Two caveats. The Mounjaro and Ozempic labels only table reactions seen in at least 5% of people, so their silence says less.^[4] And the closest labeled entries are not aches as such. Zepbound’s fatigue row, which includes malaise, the general feeling of being unwell, ran 5% to 7% against 3% on placebo.^[2] Saxenda’s trial in teenagers is the only one on any current label with a limb-pain row: pain in the arms or legs, 4% against 2.4%.^[3] If tiredness is the bigger part of it, why GLP-1 drugs make you tired covers what is and is not known.

The pain the label does list: skin that hurts

Wegovy’s label has a reaction called dysesthesia: altered skin sensation, and by the label’s own definition it includes pain of the skin and allodynia, where a light touch hurts. It is uncommon at the standard dose, 2% against 1% on placebo. At 7.2 mg it was reported by 22% of people, against 6% on 2.4 mg in the same trials and almost no one on placebo. On the Wegovy tablet it was 5% against 0%.^[1] The STEP UP registry adds two related entries at 7.2 mg that barely appeared on placebo: sensitive skin in 7.1% and heightened skin sensitivity in 5.5%.^[8]

The label says it rose with the dose and with drug levels in the blood. Among the 288 people who had it on 7.2 mg, 18% had not recovered by the end of the trial, and 17 of the 38 who recovered and had the dose raised again got it back. It cleared faster when the prescriber acted, by pausing, lowering or stopping the dose.^[1] A 2025 case series described four people on semaglutide 2.4 mg whose skin tenderness tracked the timing of their doses.^[20] On tirzepatide the reaction is rare: 0.2% to 0.4% on Zepbound against 0.1%.^[2] So if everything seems to ache and your skin hurts to touch on a high semaglutide dose, that is a labeled, dose-related reaction to raise with your prescriber, not something to push through. How long GLP-1 side effects last covers why this one does not follow the usual settling pattern.

Pain can also go the other way

Where a trial measured pain on a proper scale, it improved or held steady. In STEP 9, in people with knee osteoarthritis, knee pain fell 41.7 points on semaglutide against 27.5 on placebo on the WOMAC pain score, so the drug’s own share was about 14 points; the number behind that headline is covered in its own article.^[9] It is not a given, though. In a 52-week trial of liraglutide after a weight-loss diet, knee pain did not improve compared with placebo, a difference of 0.9 points.^[11] On whole-body pain, tirzepatide improved the bodily-pain score of a standard quality-of-life questionnaire more than placebo in SURMOUNT-2.^[12]

What is plausibly going on

If the trials do not show the drug causing aches, the aches still have causes. Here is what the evidence supports for each common suspect, from measured to guessed.

• A sore injection site. Measured and drug-specific. On Zepbound, injection-site reactions, which include pain at the site, ran 6% to 8% against 2% on placebo; on Wegovy, 1.4% against 1%.^[2][1] Where semaglutide and tirzepatide differ on this is covered separately.
• Losing muscle along with fat. Real, but not special to the drug. In SURMOUNT-1’s body-scan substudy, about a quarter of the weight lost was lean tissue on tirzepatide and on placebo alike.^[13] A 2026 review of 35 trials found none that measured objective physical function, so nobody has linked the loss to pain.^[14] Muscle loss on a drug versus on a diet and muscle volume versus muscle function go further.
• Gout. Uric acid falls on these drugs, by about 0.95 mg/dL on tirzepatide 15 mg against 0.18 on placebo, mostly through the weight loss.^[17] A pooled analysis of 22 outcome trials found no significant effect on gout risk.^[15] The idea that rapid weight loss sets off flares comes from gastric bypass, where a third of patients with previous gout had an attack after surgery and none without it did.^[16] Uric acid and two answers has the detail.
• Arthritis. Across 43 randomized trials with 100,488 people, the drugs were not linked to gout, rheumatoid arthritis, osteoarthritis or disc problems.^[18] GLP-1 drugs and rheumatoid arthritis covers the one study on that disease.
• Fluids, vitamin D and protein. Often blamed, never tested. A PubMed search pairing each of them with muscle, joint or bone pain in GLP-1 users found no studies at all. Eating much less, not drinking enough and low vitamin D are all worth having checked, because they matter for their own reasons, but nothing yet shows they explain aches on these drugs.

One more, on the Wegovy label only: in SELECT, hip and pelvis fractures were more common on Wegovy than placebo in women (1% against 0.2%) and in people 75 and older (2.4% against 0.6%).^[1] That is why new hip or pelvic pain in those groups deserves a mention to a prescriber. What was measured on bone is covered separately.

What helps

No randomized trial has tested a GLP‑1 against back pain as an outcome; a search of PubMed limited to randomized trials returned nothing, and the one registered trial has not started recruiting. So there is no drug-specific remedy to recommend. Exercise is the best-studied companion to the drug, and what exercise adds on a GLP-1 sets out what the trials found. For aches that are new, spreading or getting worse, the useful move is to describe them to your prescriber with the timing: when they started, and whether they followed a dose increase.