A 15-Person CRISPR Trial Cut LDL by Half and Kept It There for a Year

September 5, 2026 · Science · Biomarkers

What Was Reported

One-year follow-up data on CTX310, a CRISPR-Cas9 therapy that edits the ANGPTL3 gene, was presented at the European Society of Cardiology Congress 2026 and published as a letter in the New England Journal of Medicine on September 4, 2026. Fifteen adults with lipid disorders that had not responded to existing medication received one intravenous infusion. At the highest dose, 0.8 mg/kg, LDL cholesterol was 52.5% lower and triglycerides 47.8% lower than baseline a year later, according to MedicalXpress. The trial was run through Cleveland Clinic with Luke Laffin as lead author.

Those figures rest on the phase 1a trial published in the same journal in November 2025. It was open-label, single-arm and conducted at six sites in Australia, New Zealand and the United Kingdom. Fifteen adults received a single infusion at one of five dose levels between 0.1 and 0.8 mg/kg. Six had familial hypercholesterolemia, six had mixed dyslipidemia, two had moderate to severe hypertriglyceridemia and one had non-familial hypercholesterolemia. Median age was 53, 87% were men, and 40% had established atherosclerotic cardiovascular disease. Median baseline LDL was 155 mg/dL and median baseline triglycerides 192 mg/dL.

What the Numbers Do and Do Not Support

CRISPR Therapeutics, which develops CTX310, gives the one-year figures at the top dose as a mean 53% LDL reduction with a maximum of 84%, and a mean 48% triglyceride reduction with a maximum of 78%, alongside a mean 79% reduction in circulating ANGPTL3. The means and the maximums are doing different work. A maximum describes the single participant who responded best out of a group small enough to fit around a table. In a trial this size, the mean is the number worth carrying forward and the maximum is close to anecdote.

The published per-dose results make the same point from another direction. Mean LDL reduction was 39.2% at 0.6 mg/kg, 21.0% at 0.7 mg/kg and 48.9% at 0.8 mg/kg. ANGPTL3 reduction ran 32.7%, 79.7% and 73.2% across those same three doses. A dose-response curve that moves up, down and up again is not evidence that 0.7 mg/kg is a bad dose. It is what happens when three or four people receive each level. The trial establishes that the effect is real and large at the top of the range tested, and does not yet establish its size with any precision.

The design limits what any of it can mean. There was no control group, so nothing separates the drug's effect from anything else that changed for fifteen people who had just enrolled in a closely monitored trial. For an outcome like LDL, which responds to diet and to starting or restarting other medication, that gap matters more than it would for a harder endpoint. The investigators named the small size and the mixed patient population as limitations themselves, and said long-term follow-up in much larger samples is needed to detect late-emerging or uncommon safety problems.

Why a Permanent Edit Is a Different Proposition

Every lipid drug we have covered here shares one failure mode: people stop taking it. The evidence for lipid lowering is strong and the delivery of it is leaky. Trials that successfully raise the share of patients on treatment keep finding smaller clinical differences than the biology predicts. RADICAL PC 2 raised statin use from 40% to 63% and found no difference in cardiovascular events over nearly six years.

A one-time edit removes that failure mode by removing the choice. It also removes the ability to stop if something goes wrong later, which is the trade nobody can price yet. A daily drug with an unexpected long-term effect can be discontinued. An edited gene cannot be un-edited, and the safety database supporting the decision currently consists of fifteen people followed for a year. Those two facts are the whole story of why this result is genuinely important and why it is nowhere near a clinic.

ANGPTL3 is also worth understanding as a target rather than a test. It is a protein that restrains the clearance of triglycerides and LDL from the blood. Suppressing it lets that clearance run faster, which is why a single edit moves both numbers at once. ANGPTL3 is not on the panels sold by the platforms we review, and adding it would not tell a customer anything they could act on.

What This Changes for Someone Reading a Lipid Panel

Nothing about what to measure, and something about what a number represents. LDL and ApoB remain the best-evidenced things a lipid panel reports, and our ApoB guide covers why particle count separates people whose standard panels look alike. Our lipid panel guide covers what the standard panel does and does not capture, and Lp(a) covers the one lipid marker that is genuinely fixed for life and worth measuring once.

A result like this describes a possible future in which an elevated LDL is treated once rather than managed for life. That future is not here. A reader with an elevated LDL today has well-established options, and the decision about them belongs with a clinician who can see the rest of the record: existing diagnoses, family history, other medications, and what has already been tried. Nobody should defer that conversation on the strength of a fifteen-person phase 1 trial.

There is also a testing-cadence implication worth naming, because it cuts against the way platforms tend to sell repeat panels. If a therapy sets a lipid level durably, the case for frequent retesting weakens while the case for long-horizon monitoring grows. The open question moves from whether the number is still down to whether anything else has changed in the years since. A quarterly panel is not the instrument for that question.

What Would Change This Read

A controlled trial would change it most. Everything on this page comes from fifteen people compared against their own baselines. CRISPR Therapeutics reports a phase 1b with a fixed flat-dose regimen already running, with an update expected in the second half of 2026 focused on people with severe hypertriglyceridemia. Numbers from a design with a comparison group would establish how much of the reduction is the edit.

Longer safety follow-up would change it second. The reported plan to monitor participants for 15 years exists because the risk profile of a permanent genomic change cannot be characterized in twelve months. Until several years of that accumulates across a larger group, the size of the LDL reduction is the least uncertain thing about this therapy.

Related Coverage

Two evolocumab trials reported the same day split on what a lower LDL buys over different time horizons. The STAREE trial in adults over 70 lowered LDL by 48 mg/dL, cut cardiovascular events, and left its primary endpoint unchanged. A Stanford analysis tied LDL increases on low-carb diets to a genetic score. The through-line across all of them is that lipid markers move readily and the outcomes they predict are counted on a much slower clock. A therapy that claims to move a marker permanently will need years rather than a single year before it can be judged. The biomarkers hub collects the markers these conversations start from.

Sources

• MedicalXpress, "CRISPR gene-editing therapy safely and continuously lowers cholesterol and triglycerides, first-in-human trial shows," September 4, 2026, medicalxpress.com (accessed September 5, 2026). Source of the 52.5% LDL and 47.8% triglyceride mean reductions at 12 months at the highest dose, the CTX310 and ANGPTL3 identification, the Cleveland Clinic and Luke Laffin attribution, the 0.1 to 0.8 mg/kg dose range, the New England Journal of Medicine letter (DOI 10.1056/NEJMc2609825) dated September 4 2026, the ESC Congress 2026 presentation, the 15-year planned monitoring period, and the stated intention to study the therapy in a larger group.
• Laffin LJ, Nicholls SJ, Scott RS, et al., "Phase 1 Trial of CRISPR-Cas9 Gene Editing Targeting ANGPTL3," New England Journal of Medicine 393(21), published online November 8, 2025, DOI 10.1056/NEJMoa2511778; abstract retrieved via Europe PMC (accessed September 5, 2026). Source of the serious adverse events in two participants (13%), the spinal disk herniation, the sudden death 179 days after treatment at the 0.1 mg/kg dose, the infusion-related reactions in three participants (20%), and the per-dose mean ANGPTL3 reductions of 32.7%, 79.7% and 73.2% at 0.6, 0.7 and 0.8 mg/kg.
• Cleveland Clinic Consult QD, "CRISPR-Cas9 Gene Editing for Refractory Dyslipidemia Shows Safety and Preliminary Efficacy," November 10, 2025, consultqd.clevelandclinic.org (accessed September 5, 2026). Source of the open-label phase 1a design across six sites in Australia, New Zealand and the United Kingdom, the diagnosis breakdown, the median age of 53 and 87% male, the 40% with clinical atherosclerotic cardiovascular disease, the median baseline LDL of 155 mg/dL and triglycerides of 192 mg/dL, the five dose levels, the per-dose mean LDL reductions of 39.2%, 21.0% and 48.9%, the transient aminotransferase elevation, the statement that no serious adverse events or dose-limiting toxicities were deemed related to CTX310, and the investigators' stated limitations on small size, heterogeneous population and the need for long-term follow-up in larger samples.
• CRISPR Therapeutics, "CRISPR Therapeutics Presents Phase 1a Data for CTX310 Demonstrating Deep and Durable ANGPTL3 Editing, Triglyceride and LDL Lowering at ESC Congress 2026," August 30, 2026, distributed via BioSpace (accessed September 5, 2026). Source of the sponsor identification, the five named dose levels, the one-year mean reductions of 79% ANGPTL3 (maximum 89%), 48% triglycerides (maximum 78%) and 53% LDL (maximum 84%) at 0.8 mg/kg, the statement that all 15 participants completed at least one year of follow-up at data cutoff, the absence of treatment-related serious adverse events and of Grade 3 or higher transaminase changes, the one allergic reaction and three Grade 2 infusion-related reactions, and the ongoing phase 1b fixed flat-dose trial with an update expected in the second half of 2026. This is the sponsor's own release and is cited as such.