Skip to content
GLP Loss
← Research
Evidence

GLP-1 and Cholesterol: Which Column Actually Moves

At the highest labeled semaglutide dose, LDL moved 1.3% and 3.8% in the wrong direction against placebo while triglycerides fell 15.4% and 25.0% in the same two trials. More weight loss bought no LDL reduction at all.

Owen Castellanos10 min read
LDL and triglycerides do not move togetherRelative difference from placebo, same trial, same tableSemaglutide 7.2 mg, obesityLDL plus 1.3 percent. Triglycerides minus 15.4 percent.Semaglutide 7.2 mg, type 2 diabetesLDL plus 3.8 percent. Triglycerides minus 25.0 percent.Tirzepatide 15 mg, type 2 diabetesLDL minus 3.0, interval crossing zero. Triglycerides minus 24.8.The effect that reproduces most reliablyHappens after a meal, where no routine panel looksA fasting draw is the wrong instrument for the strongest effect.

The semaglutide label now carries a 7.2 mg dose, the largest weight reduction in the program. At week 72 its LDL cholesterol result against placebo was +1.3% in adults with obesity and +3.8% in adults with obesity and type 2 diabetes — the wrong direction, at the dose that takes off the most weight. In those same two trials, triglycerides fell 15.4% and 25.0% relative to placebo.[1] Losing more weight bought more triglyceride reduction and no LDL reduction at all, which is the single most useful thing to know before a lipid panel is used to judge whether one of these drugs is working.

The tirzepatide table prints the same split with intervals attached

In adults without type 2 diabetes at week 72, the LDL relative difference from placebo was −2.9% (95% CI, −6.6 to 0.9) at 5 mg, −4.0% (95% CI, −7.5 to −0.5) at 10 mg and −5.5% (95% CI, −8.9 to −2.0) at 15 mg. Triglycerides in the same three arms fell 16.5% (95% CI, −21.2 to −11.4), 19.3% (95% CI, −23.9 to −14.4) and 24.9% (95% CI, −29.1 to −20.4).[2]

The lowest dose fails to separate from placebo on LDL and separates decisively on triglycerides. Non-HDL cholesterol — the column that counts every atherogenic particle rather than one of them — behaved like the triglycerides, falling 5.8%, 7.2% and 9.6%, each interval clear of zero, and HDL rose 7.7% to 9.9% across the doses.[2]

In type 2 diabetes the LDL effect disappears entirely

The same label prints the diabetes trial beside it, and the direction inverts. LDL rose from baseline in every arm: by 7.4% on placebo, 1.8% on tirzepatide 10 mg and 4.1% on 15 mg. The relative differences from placebo were −5.2% (95% CI, −10.1 to 0.1) and −3.0% (95% CI, −8.4 to 2.6) — both intervals containing zero.[2]

Triglycerides in the same patients fell 24.6% (95% CI, −30.0 to −18.7) and 24.8% (95% CI, −30.3 to −18.9).[2] One column in that table cannot be distinguished from no treatment; the column beside it moves by a quarter with an interval nowhere near zero. Anyone told that a GLP-1 will handle their cholesterol is being sold the second column under the name of the first, and what that means for a statin already in the regimen is worked through in the statin article.

Across the class more broadly, a network meta-analysis of 76 randomized trials covering 15 agents and 39,246 participants with type 2 diabetes put the pooled LDL change against placebo at −0.16 mmol/L (95% CI, −0.30 to −0.02) and total cholesterol at −0.48 mmol/L (95% CI, −0.84 to −0.11).[3] Roughly six milligrams per deciliter of LDL, in a population where the trials that change outcomes move it by fifty.

The triglyceride gain is rented on the same terms as the weight

The tirzepatide randomized withdrawal study puts a number on that. Over the 36-week open-label lead-in, triglycerides fell 40.4% from baseline. Participants were then randomized to continue or to switch to placebo, and over the next 52 weeks triglycerides rose 13.5% in the withdrawn group against a further fall of 4.8% in those who continued — a relative difference of 16.1% (95% CI, −21.7 to −10.0). Non-HDL cholesterol moved the same way, at a relative difference of 8.1% (95% CI, −11.3 to −4.8).[2]

The column that responds is also the column that rebounds. Nothing here describes a durable change to a lipid profile; it describes a change that holds while the injections continue, on the same terms as everything else in the stopping article.

Is any of it independent of the weight loss

For triglycerides, the honest answer is a qualified yes, and one study answers it directly. A post hoc lipoprotein analysis of a 26-week phase 2b trial measured apolipoprotein B, apolipoprotein C-III, preheparin lipoprotein lipase and nuclear magnetic resonance particle counts in patients with type 2 diabetes randomized to tirzepatide, dulaglutide or placebo. Tirzepatide dose-dependently lowered apoB and apoC-III and raised preheparin lipoprotein lipase; at 10 and 15 mg it reduced large triglyceride-rich lipoprotein particles, small LDL particles and the lipoprotein insulin resistance score against both placebo and dulaglutide.[4]

The result that answers the question is the last one. At 26 weeks, change in apoC-III — not body weight — was the best predictor of the change in triglycerides on tirzepatide, explaining up to 22.9% of the variability, and the apoC-III reduction was larger in patients who started with high triglycerides than in those who did not.[4] ApoC-III inhibits lipoprotein lipase, so lowering it clears triglyceride-rich particles faster. That is a mechanism, in people with diabetes, in a post hoc analysis, and it is the strongest available evidence that part of the triglyceride effect is not simply weight coming off.

A companion metabolomic analysis of the same 259 participants found branched-chain amino acids, glutamate and 2-hydroxybutyrate falling against baseline and placebo, with triglycerides and diglycerides lowered against baseline, dulaglutide and placebo, biased toward shorter and more saturated species. Its authors state that the improvements in triglycerides and lipoprotein markers were only partially attributable to weight loss.[5] Partially is the operative word, and nobody has quantified the rest.

The strongest lipid effect is one no routine panel measures

A 2026 review of incretin effects on lipid metabolism states the conclusion plainly: the most reproducible lipid-related effect of these drugs is the attenuation of postprandial lipemia, including reductions in triglycerides and in apolipoprotein B48-containing particles, which are the lipoproteins the intestine assembles after a meal — while effects on fasting lipid panels are generally less impressive and more heterogeneous across agents and populations.[6]

Every number quoted above came from a fasting draw. The measurement people are given is the one on which the drug performs worst. Human kinetic and stable-isotope studies in that review point to both reduced intestinal lipoprotein output and faster clearance, with the apoC-III and lipoprotein lipase axis implicated alongside adipose clearance and intestinal lymphatic transport.[6]

That postprandial literature is small enough to be counted study by study, and counting it is more useful than summarizing it. A 2022 review that did so found exenatide reducing postprandial triglyceride in 4 of 6 studies and apolipoprotein B-48 in 3 of 3; liraglutide reducing triglyceride in 2 of 2 and apoB-48 in 3 of 3; lixisenatide in one study; semaglutide in one study; and dulaglutide reducing apoB-48 in one study.[7] The molecule most people are prescribed has a single postprandial study behind the class’s most reproducible lipid effect, and tirzepatide is not in that count at all.

Whether the lipids explain the cardiovascular benefit

They may not, and the best test of that came from the outcome trial itself. A prespecified analysis of 17,604 participants examined whether the reduction in major adverse cardiovascular events tracked the adiposity changes that drive the lipid changes. In the semaglutide group there was no linear trend linking weight loss at week 20 to subsequent event risk, and an estimated 33% of the observed benefit was mediated through waist circumference reduction (HR 0.86; 95% CI, 0.77 to 0.97 after adjustment for time-varying waist change).[8]

The subgroup that reverses is in the placebo arm. There, lower baseline waist circumference predicted lower event risk, baseline body weight did not (HR 0.99; 95% CI, 0.97 to 1.01; P = 0.28), and weight loss was paradoxically associated with increased event risk — because unintended weight loss in established cardiovascular disease marks illness rather than improvement.[8] The same exposure carries opposite meanings in the two arms of one trial. What that trial did and did not show is set out in the outcome trial article, and the event arithmetic in the heart benefits article.

The 2026 lipid review reaches the matching verdict from the mechanistic side: the lipid changes point in an antiatherosclerotic direction, but current evidence does not permit an estimate of how much they contribute to the established cardiovascular protection of this class.[6]

What to do with a panel that looks unchanged

An unchanged LDL on a GLP-1 is the expected result, not a sign the drug is failing, and it is not a reason to stop a lipid-lowering prescription that was started for a different reason. A falling triglyceride and a falling non-HDL are the columns that reliably respond, and the hepatic side of the same triglyceride story is covered in the liver article.

Every figure here came from branded product at labeled doses in randomized trials with central laboratories. Most sellers covered here dispense compounded semaglutide or tirzepatide, which is not FDA-approved and is not reviewed by the FDA for safety, efficacy or quality before it is dispensed; no lipid substudy has been run on a compounded preparation, and a program listed on the semaglutide board advertising cholesterol benefits is quoting trials of a different product.

Frequently asked

Do GLP-1 drugs lower cholesterol?
They lower triglycerides reliably and LDL cholesterol barely. On the tirzepatide weight label the LDL relative difference from placebo ran from 2.9% at the lowest dose, on an interval containing zero, to 5.5% at the highest, while triglycerides fell 16.5% to 24.9% across the same arms. At the highest labeled semaglutide dose, LDL moved 1.3% and 3.8% in the wrong direction against placebo.
Why does LDL barely move when so much weight comes off?
Nothing in the labels explains it, and the pattern is consistent enough to be a property of the class rather than a quirk. In the type 2 diabetes trial on the tirzepatide label, LDL rose from baseline in every arm including placebo, and neither dose separated from placebo. A network meta-analysis of 76 trials in 39,246 people put the pooled LDL reduction at 0.16 mmol/L.
Is any of the lipid change independent of weight loss?
For triglycerides, partly. In a post hoc analysis of a 26-week trial, the change in apolipoprotein C-III rather than the change in body weight was the best predictor of triglyceride change on tirzepatide, explaining up to 22.9% of the variability. A companion metabolomic analysis of the same participants concluded that the lipoprotein improvements were only partially attributable to weight loss.
What is the postprandial effect and why does it matter?
A 2026 review concludes that the most reproducible lipid effect of this class is reduced lipemia after a meal, including lower triglycerides and lower apolipoprotein B48-containing particles, while fasting panels respond less and less consistently. A standard cholesterol test is a fasting draw, so the routine measurement misses the effect that reproduces best.
Should a statin be stopped if the panel improves?
Nothing in this evidence supports that. The cardiovascular outcome trial enrolled a population overwhelmingly already on lipid-lowering therapy, and a prespecified analysis found no linear trend linking early weight loss to subsequent event risk, with an estimated 33% of the benefit mediated through waist circumference reduction. The 2026 mechanistic review states that current evidence cannot estimate how much the lipid changes contribute to the cardiovascular protection.

Sources

  1. [1] Novo Nordisk Pharmaceutical Industries, LP (2026). WEGOVY (semaglutide) injection, solution — Clinical Studies 14, Table 17: anthropometry and cardiometabolic parameters at Week 72 in Studies 8 and 9 DailyMed, U.S. National Library of Medicine. Source
  2. [2] Eli Lilly and Company (2026). ZEPBOUND (tirzepatide) injection, solution — Clinical Studies 14.1, Table 3: cardiometabolic parameters at Week 72 in Studies 1 and 2, and Table 7: Study 4 following randomized withdrawal DailyMed, U.S. National Library of Medicine. Source
  3. [3] Yao H, Zhang A, Li D, et al. (2024). Comparative effectiveness of GLP-1 receptor agonists on glycaemic control, body weight, and lipid profile for type 2 diabetes: systematic review and network meta-analysis. BMJ. PMID 38286487
  4. [4] Wilson JM, Nikooienejad A, Robins DA, et al. (2020). The dual glucose-dependent insulinotropic peptide and glucagon-like peptide-1 receptor agonist, tirzepatide, improves lipoprotein biomarkers associated with insulin resistance and cardiovascular risk in patients with type 2 diabetes. Diabetes Obes Metab. PMID 33462955
  5. [5] Pirro V, Roth KD, Lin Y, et al. (2022). Effects of Tirzepatide, a Dual GIP and GLP-1 RA, on Lipid and Metabolite Profiles in Subjects With Type 2 Diabetes. J Clin Endocrinol Metab. PMID 34608929
  6. [6] Syed-Abdul MM, Lewis GF (2026). Incretin-based therapies: Effects on plasma lipids and potential mechanisms. Pharmacol Ther. PMID 42250734
  7. [7] Novodvorský P, Haluzík M (2022). The Effect of GLP-1 Receptor Agonists on Postprandial Lipaemia. Curr Atheroscler Rep. PMID 35080714
  8. [8] Deanfield J, Lincoff AM, Kahn SE, et al. (2025). Semaglutide and cardiovascular outcomes by baseline and changes in adiposity measurements: a prespecified analysis of the SELECT trial. Lancet. PMID 41138739

Where to get it

Best GLP-1 injections

Every injectable seller we can verify, with the price each one publishes and an honest read of what the trials measured.

Compare providers →

More in Evidence