Retatrutide lowered systolic blood pressure by about 6.8 mmHg and cut three of four blood-fat markers across its clinical trials, according to a meta-analysis pooling the drug's randomized data. The good cholesterol, HDL, did not move at all.
The analysis, published June 29 in High Blood Pressure and Cardiovascular Prevention ↗, gathers the randomized controlled trials of retatrutide ↗ run so far and combines their blood-pressure and lipid results into single pooled estimates. The authors searched PubMed, the Cochrane Library, Scopus, and ClinicalTrials.gov, then pooled the trials with a random-effects model. Retatrutide, Eli Lilly's experimental triple agonist also known as LY3437943, hits three gut and metabolic hormone receptors at once: the GLP-1 ↗ and GIP ↗ receptors that tirzepatide ↗ and semaglutide ↗ already exploit between them, plus the glucagon ↗ receptor that adds a third lever over fat and energy use. It is the receptor breadth that makes its effects beyond weight worth measuring.
What pooled together
Blood pressure came out tight and consistent. Across trials, retatrutide cut systolic pressure, the top number, by 6.79 mmHg (95 percent confidence interval 5.23 to 8.36) and diastolic pressure, the bottom number, by 2.46 mmHg. The heterogeneity statistic, which measures how much trials disagreed, was low for both (an I-squared of 20 percent for systolic and 0 percent for diastolic), meaning the trials largely told the same story. A systolic drop near 7 mmHg is in the range a dedicated blood-pressure pill delivers.
The lipid picture was beneficial but noisier. Total cholesterol fell by 21.88 mg/dL, LDL-C (the so-called bad cholesterol) by 13.10 mg/dL, and triglycerides (a blood fat tied to metabolic risk) by 40.90 mg/dL. HDL-C, the protective cholesterol fraction, changed by essentially nothing: a pooled difference of 0.01 mg/dL with a p-value of 0.98, statistical language for no detectable effect.
Where the noise is
The lipid numbers carry a caveat the single figures hide. The total-cholesterol estimate had an I-squared of 84 percent, which is high. That means the included trials disagreed sharply on how much cholesterol retatrutide removes, so the pooled minus-22 is a center of mass across scattered results rather than a number any one trial would reliably reproduce. Triglycerides and LDL-C were moderately heterogeneous (53 and 47 percent). Blood pressure was the cleanest signal in the analysis; cholesterol was the messiest.
There is a larger limit. These are surrogate markers, the numbers on a lab panel, not cardiovascular events. Lower blood pressure and lower LDL predict fewer heart attacks and strokes across populations, but a meta-analysis of metabolic readouts from weight-loss and diabetes trials is not the same as a trial that counts the events themselves. The retatrutide outcomes trials that would do that are still running. And the pool is small: retatrutide has only a handful of completed randomized trials, so each one pulls hard on the combined estimate.
How it fits the triple-agonist story
This sharpens a picture peptidemodel has been tracking on the retatrutide card ↗. Earlier this month we covered how retatrutide hit its diabetes numbers while carrying a small cardiac imbalance ↗ into the same readout, a reminder that the cardiovascular profile of the triple agonist is still being assembled piece by piece. The blood-pressure and lipid effects here are the favorable side of that ledger: the glucagon arm of the molecule, the receptor that separates retatrutide from the GLP-1 and dual-agonist drugs, is associated with shifts in fat metabolism, and a clean systolic drop plus falling LDL and triglycerides is the kind of cardiometabolic bonus the field hopes the triple agonists deliver.
The honest reading is split. On blood pressure, the signal is consistent and real. On cholesterol, the direction is right but the trials disagree enough that the pooled number should be read as provisional. And on HDL, the one marker that did not move, the meta-analysis is unambiguous. The good cholesterol stayed exactly where it started.