A daily pill lowers lipoprotein(a) by a third — modestly, but cheaply
Obicetrapib cut Lp(a) by 37% in high-risk patients, a pooled analysis finds, alongside a similar drop in LDL. It reaches a lipid the costly new injectables target, if far less deeply.
| Group | Value (%) |
|---|---|
| LDL cholesterol | 37 |
| Lipoprotein(a) | 37.3 |
| Apolipoprotein B | 21.3 |
Obicetrapib, a once-daily oral cholesterol drug, lowered lipoprotein(a) by 37.3% and LDL cholesterol by 37.0% against placebo in a pooled analysis of high-risk patients [s1]. That makes it the rare pill that touches Lp(a) — a largely genetic, hard-to-shift lipid that a wave of expensive injectable RNA drugs is built to attack — though it moves the marker far less than those agents and has not yet been shown to prevent a single heart attack [s1].
Lp(a) is an LDL-like particle whose level is set mostly by inheritance and barely budges with statins, diet or exercise. High levels raise the risk of heart attack, stroke and aortic stenosis, which is why drugmakers have raced to lower it. Obicetrapib was not designed for that job: it blocks cholesteryl ester transfer protein (CETP), a target pursued chiefly to cut LDL. Lowering Lp(a) is a side effect that this analysis set out to quantify [s1].
What the pooled analysis found
The study combined trials testing 12 weeks of obicetrapib 10 mg daily against placebo in 2,356 patients with either inherited high cholesterol or established cardiovascular disease [s1]. The group was high-risk and heavily pre-treated: median age 66, 82% with atherosclerotic disease, 27% with heterozygous familial hypercholesterolaemia, and 91% already on a statin [s1]. Median baseline LDL cholesterol was 92 mg/dL, apolipoprotein B 87 mg/dL, and Lp(a) 42.9 nmol/L [s1]. That most were already at guideline-range LDL on a statin matters: the reductions below are what obicetrapib added on top of existing therapy, not what it can do alone [s1].
Placebo-adjusted, obicetrapib cut LDL cholesterol by 37.0% (35 mg/dL), apolipoprotein B by 21.3% (20 mg/dL), and Lp(a) by 37.3% (14.9 nmol/L) [s1]. The Lp(a) effect was larger in relative terms among patients whose starting level was moderately raised — a 43.3% reduction, or 36.3 nmol/L, in those between 50 and 150 nmol/L [s1]. Patients with the highest baseline levels, above 150 nmol/L, saw a smaller percentage fall but a similar absolute drop, 32.3 versus 36.3 nmol/L [s1].
That last detail is the analysis's real argument. The authors note that people with only mildly elevated Lp(a) are unlikely to qualify for the RNA-targeted drugs now in late trials, which are being tested in patients with very high levels — so a pill delivering a meaningful absolute reduction across the milder range could fill a gap those injectables leave open [s1].
Safety over a year
A separate pooled safety analysis of two phase 3 trials followed 2,880 patients on obicetrapib or placebo for 365 days [s2]. Overall adverse-event rates were similar, 60.2% versus 62.0%, and events leading to discontinuation were slightly fewer on the drug, 4.1% versus 5.3% [s2]. There was no clinically significant blood-pressure change between groups and no difference in liver or muscle measures [s2]. A decline in kidney filtration occurred less often on obicetrapib, 6.7% versus 8.7% [s2]. Macular degeneration was reported once, in a single patient [s2]. That safety cohort of 2,880 was similar in makeup to the lipid analysis — mean age 64, 36% women, 82% with cardiovascular disease — but 35.8% also had diabetes, a group in whom the glucose effects of lipid drugs are worth watching; the authors reported no new safety signals over the year [s2].
What is missing
Two limits frame all of this. First, these are 12-week and one-year measures of blood lipids, not evidence that obicetrapib prevents cardiovascular events; that question is being tested in a separate, still-running outcomes trial, and no lipid drug should be judged a success on the marker alone [s1]. Second, both analyses pool trials run and funded by the drug's developer, NewAmsterdam Pharma, whose employees are among the authors [s1][s2] — peer-reviewed, but company-sponsored data all the same.
The Lp(a) reduction, while genuine, is also modest next to the competition. Injectable small-interfering-RNA drugs have cut Lp(a) by more than 90% in early trials, and they too have yet to prove they prevent events. Obicetrapib's case is not depth but reach and convenience: an oral drug, taken alongside statins, that nudges several atherogenic lipids at once, in a combination that genetic studies suggest may lower risk more than moving either LDL or Lp(a) alone. Whether that translates into fewer events, and how it is priced against both statins and the new injectables, are the questions still open. This article describes trial findings and is not medical advice.
Sources
- Obicetrapib and lipoprotein(a) levels in patients at high cardiovascular risk: a pooled analysis of trials — European Heart Journal , May 8, 2026
- Obicetrapib safety analysis: Pooled phase 3 clinical trial experience — American Journal of Preventive Cardiology , March 24, 2026
More on
Two siRNAs cut lipoprotein(a) by more than 90%. Nobody has shown it prevents anything yet
A first-in-human trial of Kylo-11 reports 97% reduction at 48 weeks after a single dose. A lepodisiran analysis reports near-total clearance of the oxidised phospholipids Lp(a) carries.
Should healthy adults get a coronary calcium scan? Guidelines disagree
It is not a routine test for everyone — the USPSTF finds the evidence insufficient. Cardiology guidelines back it for one group: intermediate-risk adults weighing whether to start a statin.
Scanning 16,808 adults with no known heart disease found plaque in more than half
The REACT study imaged people aged 18 to 70 with no known cardiovascular disease. Atherosclerosis was already present in some people in their twenties, and it was usually in the legs, not the heart.
PCBs and heart risk: what a predicted score can and cannot show
A Chinese cohort found serum PCBs associated with a higher calculated 10-year cardiovascular risk. The longitudinal results did not survive correction for multiple testing, and no events were counted.