The fix was the wrapper
A single IV infusion made one permanent DNA edit in the liver and cut LDL cholesterol by up to 62% — but the new result is that swapping the delivery vehicle cleared the safety signal that halted the first attempt.
In 2023, the first human DNA base-editing trial worked and was halted in the same season: one infusion knocked out the cholesterol gene PCSK9 and dropped LDL by half, then a patient's liver enzymes and platelets crashed and enrollment stopped. The edit had done its job. The fat-bubble carrier that ferried it into the liver had not.
The edit had done its job. The carrier that ferried it into the liver had not.
The follow-up keeps the exact same edit and changes the package. A single infusion now lowers PCSK9 by up to 88% and LDL cholesterol by up to 62% — a 78 mg/dL drop at the top dose in patients who couldn't hit their numbers on maximum pills — with no dose-limiting toxicity across 35 people, and the effect holding for over a year in the earliest-treated handful. The headline is one-and-done cholesterol: one shot instead of a lifetime of statins or twice-yearly injections.
But the genuinely new fact isn't that a base editor lowers LDL — that was settled in 2023. It's that the thing that broke last time was the delivery, not the edit, and a redesigned carrier appears to have fixed it. The molecular scissors were never the risk; the wrapper was.
That reframes what to watch. The same property being sold as the feature — a permanent, one-time edit with no off-switch — is also the part that can't be taken back if a later edit lands in the wrong place. Every patient still needed steroid premedication just to tolerate the infusion. The pitch and the peril are the same gene.
The advance here is in the delivery vehicle, not the gene edit — and a permanent edit has no off-switch.