Since my last post, the Lp(a) drug I was watching failed, the cancer test I was looking forward to missed its primary endpoint, and I fell down the stairs and fractured my seventh thoracic vertebra. So it’s time for some updates.

Don’t worry. It’s not all bad. Progress on lowering LDL cholesterol to prevent cardiovascular disease and lowering amyloid-beta plaques to slow or even prevent Alzheimer’s disease is going fast.

PCSK9 Inhibitors

Lower-Risk Groups

PCSK9 inhibitors were already a smashing success in the secondary prevention of CVD as of my last writeup. Since then, the VESALIUS-CV trial of primary prevention found that Repatha (evolocumab) causes a 25% reduction in the composite of heart attacks, strokes, and deaths from heart disease. The heart attack reduction specifically was 36%.

Statins, previously the most powerful class of drugs for primary prevention of CVD, have three notable side effects. As a class, they’re possibly the most well-studied drugs in history, so we can detect these side effects even though they’re uncommon. At high intensity, they raise new diabetes diagnoses from 3.5% to 4.8% of patients per year, mostly in people whose blood sugar was already near the diagnostic cutoff. They cause a small excess of mostly mild muscle pain, about 11 extra reports per 1,000 patients per year, mainly in the first year. And they cause slightly more abnormal liver tests, about 1 per 1,000 patients per year, but those tests don’t precede more serious liver harm.

Repatha looks cleaner on all three. In my view, adding this new trial to the pool of evidence on side effects, helping us build towards the certainty we have in statins, is its main benefit. We already knew from the FOURIER trial that Repatha didn’t raise the risk of new diabetes and the high end of the confidence interval was already low enough to rule out harm as large as that seen in statins (hazard ratio 1.05, 95% CI 0.94 to 1.17). Meta-analyses of the earlier trials also looked good: one of non-statin cholesterol drugs as a group found no excess muscle or liver problems, and a network meta-analysis gave PCSK9s a lower risk of diabetes and liver dysfunction than statins. VESALIUS-CV fit that pattern, with serious adverse events in 28.2% of patients on Repatha and 28.9% on placebo.

Oral Versions

In July, the FDA approved Lipfendra (enlicitide), a once-daily PCSK9 pill. It lowered LDL by nearly 60% in two trials, close to what the injectables do. Trials with tens of thousands of patients for this and competing pills are ongoing, so we’ll learn a lot more in the next few years.

I’m not sure if the pills will make much difference, though. Pills are good for people who hate shots, but the success of GLP-1s suggests that group may not be large. I had expected the pills to be cheaper to manufacture, but that may be wrong. A year of Repatha is 3.6 grams, with a manufacturing cost of around $100 per gram for similar drugs, or $360 per year. A year of Lipfendra is 7.3 grams, with a manufacturing cost of around $50 per gram for similar drugs, or $365 per year.

Gene Editing

Gene editing, which should last a lifetime, is the most exciting version of PCSK9 inhibition to me. A phase 1 trial of VERVE-102 showed no safety problems and a 60% reduction in LDL cholesterol lasting through 12 months of follow-up. CRISPR Therapeutics is editing a different gene, ANGPTL3. Its drug CTX310 lowered LDL by 53% and triglycerides by 48%, also lasting through a year of follow-up and with no safety problems.

I expect gradual progress over decades, following the path of statins. These drugs are being tested now in secondary prevention and I expect those studies to be followed by more in primary prevention. When statins were first studied in primary prevention, the standard risk calculators looked forward only 10 years, but we now have 30-year risk calculators, which were recently adopted as part of the 2026 US guidelines for statins. In the future, men might get inoculated against heart disease at 30, and women, who get it about a decade later, at 40.

What would these gene therapies cost to manufacture? I couldn’t find a good estimate, so here’s a rough one of my own. VERVE-102’s top dose is 1 mg/kg. Moderna’s original COVID shot had 100 micrograms of mRNA, and one estimate put its production cost at $23 billion for 8 billion doses, or about $2.90 a dose. Scaling up by RNA mass gives about $2,400 per treatment for an average 84 kg adult, so they could be much cheaper than a lifetime of pills or injections.

Lp(a) Reduction

Lp(a) is a genetically determined risk factor for CVD independent of LDL cholesterol. In September, pelacarsen, the first Lp(a)-lowering drug to finish an outcomes trial, failed. It lowered Lp(a) but didn’t reduce cardiovascular events.

Maybe Lp(a) isn’t causal after all. That would be a surprise, because genetic studies link high Lp(a) to heart disease, and genetics correctly predicted that PCSK9 inhibitors would work. Maybe the patients were too healthy, with mean LDL of 65 mg/dL. Or maybe the drug didn’t lower Lp(a) enough. The trial targeted patients with Lp(a) of 70 mg/dL or higher and pelacarsen lowered Lp(a) by 80% in phase 2, so it drove a smaller absolute drop in Lp(a) than is expected in upcoming trials.

Four more outcomes trials are targeting patients with higher baseline Lp(a) and testing three more drugs, two of which lower Lp(a) by over 90%. If all the trials fail, that’s likely the end of Lp(a) as a causal risk factor for CVD.

Multi-Cancer Early Detection Tests

GRAIL’s Galleri blood test had its big trial readout this year and missed its primary endpoint. In the NHS-Galleri trial of about 142,000 people, three annual rounds of screening didn’t reduce combined stage III or IV cancer diagnoses. Stage IV diagnoses alone were 14% lower (incidence rate ratio 0.86, 95% CI 0.74 to 1.00), but that was a secondary endpoint.

What went wrong? The trial may have been too short. GRAIL reports that the stage IV reduction grew with each round: 9%, then 22%, then 26%. In February, GRAIL said it “plans to extend the trial’s follow up period by 6-12 months.” It’s also possible that the test isn’t sensitive enough yet. It caught 55% of the 12 prespecified cancers and 31% of all cancers. The company has shared early results for newer versions of the test, so hopefully it will keep improving.

Galleri will likely still get FDA approval and Medicare coverage (not financial advice!). An FDA advisory panel voted 7 to 2 in favor of the test and a new law allows Medicare to cover FDA-approved MCED tests. I hope companies will still feel a financial incentive to run trials to show their tests reduce late-stage cancer diagnosis and mortality.

Amyloid Beta

Blood Tests

Before the first amyloid PET tracer was approved in 2012, it wasn’t practical to run a trial of a disease-modifying therapy in AD. There wouldn’t have been a way to know if the patients had AD until they died. Nine years later, the first disease-modifying drug was approved, which is practically instantaneous for a novel drug.

But PET scans are expensive and hard to access. Blood tests will improve that. The FDA cleared the first one in May 2025, and four are now cleared. That should make diagnosis more widespread and trials simpler. The TRAILBLAZER-ALZ 3 prevention trial screened 63,000 people using a blood test to find people with amyloid before symptoms appear, which would have been impractical with PET scans.

Therapies

Three amyloid antibodies have been approved: Aduhelm (aducanumab) in 2021, Leqembi (lecanemab) in 2023, and Kisunla (donanemab) in 2024. The newer ones are better than the first.

Aduhelm got off to a rocky start. Its benefit was weak, it had significant side effects, and it cost $56,000 a year. A congressional investigation found that the FDA and the company had inappropriately collaborated during the approval process. Medicare restricted coverage, and the company discontinued the drug in 2024.

The news stories at the time left the public with the impression that the lead investigators studying the amyloid-beta hypothesis were not only wrong but also jerks. They may be jerks, but they were right. In separate trials, Leqembi and Kisunla slowed clinical decline by 27% and 29% over 18 months by reducing amyloid-beta plaques. Kisunla can be stopped once plaques are cleared, and Leqembi can now be started with a weekly injection at home, so the drugs are becoming more convenient as well.

The best evidence that amyloid removal is doing the work is the pattern across trials. A 2023 meta-analysis of eight immunotherapy trials found plaque load a year before the end of the trial predicted the final clinical benefit (Spearman r = 0.85). A 2024 meta-analysis found similar results and predicted the results would grow over time.

Upcoming amyloid-beta drugs look even better. In early trials of trontinemab, 91% of patients were amyloid-negative at six months, with a brain swelling rate of 1%, compared to 35% in Aduhelm. Two phase 3 trials are due to finish in 2028.

Phase 3 trials are also moving earlier in the disease. Three trials are testing people with no symptoms but a high load of amyloid beta. AHEAD 3-45 is testing Leqembi and TRAILBLAZER-ALZ 3 is testing Kisunla, with results expected in a year, and PrevenTRON is testing trontinemab with results expected in 2032.

In a way, AD is following the path of CVD. Widely available tests allowed us to refine our understanding of what causes the disease (LDL in one case, amyloid beta in the other) and then develop drugs to reduce the cause. Those drugs were tested first in the latest-stage, highest-risk patients and then in progressively earlier-stage, lower-risk patients and got more powerful and safer over time.

Osteoporosis

After my bone fracture, I was diagnosed with low bone density for my age, which seems to be nothing but an insurance-billing-code difference from osteoporosis. I promise I didn’t throw myself down the stairs just to get to try the cool new osteoporosis drugs for myself, but here we are anyway.

I started on Forteo (teriparatide), one of the more powerful options, but had to go off of it due to high calcium excretion (I took that test three times and got wildly different results, so I’m not sure going off the drug is the right choice, but my doctor only has so much time in the day to listen to her patients’ weird theories).

The other powerful drugs that are approved in men work along the same pathway that raises calcium excretion. The most powerful drug of all, Evenity (romosozumab), isn’t approved in men, so insurance won’t cover it. That’s frustrating, because the 2016 BRIDGE trial already showed that Evenity raised bone density by 11% in a year in men. I’m not sure why the company never pursued approval in men.

So I’ve fallen back to the OG osteoporosis drug, Fosamax (alendronate).

The pipeline doesn’t look much better for men. The next notable addition is EB613, an oral tablet of Forteo. That would spare people a daily injection, but it’s the same drug that gave me trouble. I’ll be on the lookout for more powerful drugs approved in men, but that market is pretty small, so there doesn’t appear to be much commercial interest.

Overall, Lp(a) reduction and MCED tests have progressed slower than I’d hoped, but LDL reduction and amyloid-beta reduction are going faster than I expected.