A Single Injection Silenced a Disease Gene for 72 Weeks. The Cost of Not Doing It: $450,000 a Year.
Head-to-head clinical data from China's ART001 and Intellia's nex-z reveal that both in-vivo CRISPR programs hit 90%+ TTR knockdown for ATTR amyloidosis, but a statistical power calculation shows the apparent safety gap is a mirage of sample size, and the real story is what a one-time injection means for patients spending half a million dollars a year to stay alive.
92%. That is how much transthyretin protein disappeared from the blood of three patients in Suzhou, China, 72 weeks after a single intravenous infusion of a CRISPR-based drug called ART001. No booster shots, no follow-up infusions, just one needle and one hour, and the gene that had been poisoning their organs with misfolded protein was functionally silenced. Those results, published in Frontiers in Medicine on April 15, 2026, are the longest-duration in-vivo CRISPR data ever reported from a clinical trial. Ten patients, zero serious adverse events, zero infusion-related reactions.
Four thousand miles west, the same approach nearly killed somebody.
Intellia Therapeutics and Regeneron have been running the largest in-vivo gene editing trials in history: MAGNITUDE and MAGNITUDE-2, testing their drug nexiguran ziclumeran (nex-z) in roughly 650 patients with ATTR amyloidosis affecting the heart and peripheral nerves. In October 2025, one patient developed grade 4 liver transaminase elevation and increased bilirubin after receiving nex-z. That patient was hospitalized and died on November 5, 2025. In response, the FDA placed clinical holds on both trials. MAGNITUDE-2, the polyneuropathy trial, had its hold lifted on January 28, 2026. MAGNITUDE, the cardiomyopathy trial, remains partially on hold.
Two programs, same target, same mechanism, opposite safety profiles. Headlines write themselves: China's CRISPR program is safer. But the headlines are wrong, and the math explains why.
The Mirage in the Safety Data
Intellia's adverse event rate for grade 4 liver toxicity is less than 1% of the 450-plus patients dosed. Call it 0.2%, approximately one in every 500 patients dosed. Now run the binomial calculation on ART001's 10-patient trial: the probability of seeing zero events from a 0.2% base rate across 10 patients is (1 - 0.002)^10 = 98.0%. ART001's perfect safety record is not evidence that ART001 is safer. It is evidence that 10 patients cannot detect a 1-in-500 event. A trial this small would need to enroll roughly 1,500 patients to have 95% power to detect a 0.2% adverse event rate, which means ART001's safety profile is statistically consistent with nex-z's and provides no basis for distinguishing between the two drugs on safety.
What ART001's data does show is the absence of infusion-related reactions, a finding that differs from nex-z trials where some patients experienced transient cytokine responses during infusion. Both drugs use lipid nanoparticle delivery to shuttle Cas9 mRNA and a guide RNA into hepatocytes, but their LNP formulations likely differ in lipid composition, PEGylation density, and manufacturing process. Whether ART001's cleaner infusion profile reflects a genuinely better LNP or simply the luck of 10 draws from the same distribution remains an open question that only larger trials can answer.
The Clinical Comparison Nobody Built
No published analysis has placed these two programs side by side with matched metrics. Here is what the data shows:
| Metric | ART001 (Accuredit, China) | Nex-z (Intellia/Regeneron, US) |
|---|---|---|
| Delivery vehicle | LNP with Cas9 mRNA + sgRNA | LNP with Cas9 mRNA + sgRNA |
| Target | TTR gene, hepatocytes | TTR gene, hepatocytes |
| Development stage | Investigator-initiated trial (10 patients) | Phase 3 (~650 enrolled, ~450 dosed) |
| Max TTR knockdown | 92% at 1.0 mg/kg (72 weeks) | ~90% (Phase 1 data, sustained) |
| Dose range tested | 0.05 to 1.0 mg/kg (6 cohorts) | 0.1 to 1.0 mg/kg |
| Durability | 72 weeks and counting | Sustained (Phase 1 follow-up ongoing) |
| Serious adverse events | 0 of 10 (0%) | Grade 4 liver toxicity in <1%, 1 death |
| Infusion reactions | 0 of 10 | Some transient cytokine responses |
| Regulatory status | IIT in China | Phase 3 in US/EU, partial FDA hold |
The efficacy numbers are strikingly close. Both programs achieve roughly 90% reduction in circulating TTR protein from a single dose, delivered through nearly identical molecular machinery. The convergence is not coincidental, because both programs descended from the same foundational work: Intellia's 2021 New England Journal of Medicine publication was the first demonstration that in-vivo CRISPR could safely edit genes in living humans, and ART001's investigators cite it directly.
The $450,000-a-Year Problem
ATTR amyloidosis is a disease that kills through accumulation. Misfolded transthyretin protein deposits in the heart, nerves, and gastrointestinal tract over years and decades, progressively destroying organ function. Hereditary ATTR (ATTRv) affects an estimated 10,186 patients across 36 countries, according to epidemiological surveys. Wild-type ATTR (ATTRwt), the non-hereditary form associated with aging, is massively underdiagnosed: autopsy studies suggest 10 to 25 percent of heart failure patients over 80 have amyloid deposits, pointing to a potential US patient population of 300,000 to 500,000 if properly screened.
Current treatment options are expensive, permanent commitments that never end.
Patisiran, marketed as Onpattro by Alnylam Pharmaceuticals, is an RNA interference drug that temporarily silences TTR production. It requires intravenous infusion every three weeks for the rest of the patient's life. Annual cost: approximately $450,000, according to the Institute for Clinical and Economic Review. Tafamidis, sold as Vyndaqel by Pfizer, stabilizes the TTR protein to prevent misfolding rather than reducing its production. It is an oral daily pill. Annual cost: approximately $225,000. Both drugs work. Neither cures. Stop taking them, and the protein accumulates again.
A one-time CRISPR injection changes the arithmetic completely.
Consider a patient diagnosed with ATTRv at age 50, facing 20 to 30 years of treatment. Lifetime patisiran cost at $450,000 per year over 20 years: $9 million. Lifetime tafamidis cost over the same period: $4.5 million. If a one-time CRISPR therapy is priced at $500,000 to $1 million, consistent with precedents set by Casgevy's $2.2 million price tag for sickle cell and Zolgensma's $2.1 million for spinal muscular atrophy, it breaks even against patisiran in 13 to 27 months. Against tafamidis, breakeven arrives in 27 to 53 months. Every month after that, the health system saves money, and the patient never sits for another infusion.
Economically, the math is even more dramatic for ATTRwt patients diagnosed in their 60s or 70s. Shorter treatment horizons compress the savings window, but the convenience argument accelerates: elderly patients on patisiran must travel to infusion centers every three weeks, manage IV access, and tolerate premedication regimens. One injection and done is not just cheaper. It is kinder.
Strongest Counterargument
The strongest case against celebrating these results is that in-vivo CRISPR gene editing is irreversible, and we have no 5-year or 10-year safety data for either program. Patisiran's RNA interference mechanism wears off in weeks if you stop dosing, which is a feature disguised as a bug. If a patient develops an unexpected side effect from TTR suppression, you simply stop treatment and TTR levels return to baseline within a month. CRISPR edits the genome permanently. If long-term TTR suppression below 10% of normal causes an unforeseen problem, perhaps related to TTR's role in carrying retinol-binding protein and thyroxine, there is no undo button.
This is not hypothetical. TTR carries about 15% of circulating thyroxine and is the sole carrier of retinol-binding protein in cerebrospinal fluid. Clinical trials have monitored vitamin A levels and thyroid function and found no clinically significant changes so far, but "so far" is 72 weeks for ART001 and roughly two years for nex-z Phase 1 patients. Decades of 90%+ TTR suppression in a 50-year-old patient is uncharted territory, and the honest answer is that nobody knows whether it is safe at that timescale. Every patient in these trials is betting their genome on the answer being yes.
Limitations
ART001's trial enrolled 10 patients at a single center in China with an open-label design and no control arm. Its 92% TTR knockdown figure comes from the highest dose cohort of three patients, a sample that cannot establish population-level efficacy with statistical confidence. The Frontiers in Medicine publication reports safety and tolerability as the primary endpoint, not efficacy. The trial registration (ChiCTR2400081216) lists it as an investigator-initiated trial, not a registration trial required for regulatory approval. We have not independently verified the reported data and rely entirely on the published paper and its supplementary materials.
Our cost-per-year comparisons use list prices reported by ICER and publicly available drug pricing databases. Actual reimbursed costs vary by payer, geography, and negotiated discounts, and could differ from list prices by 20 to 50 percent. The breakeven calculation assumes constant pricing and does not account for generic or biosimilar competition that could reduce patisiran costs within the next decade, as Alnylam's composition-of-matter patents begin expiring.
The epidemiological estimates for ATTRwt prevalence (300,000 to 500,000 in the US) are extrapolated from autopsy studies and likely overestimate the number of patients who would be symptomatic enough to seek treatment. Actual diagnosed prevalence is far lower.
What You Can Do
If you or a family member has ATTR amyloidosis: Ask your cardiologist or neurologist whether genetic testing has confirmed the specific TTR mutation. Hereditary ATTR (ATTRv) is the form most likely to qualify for CRISPR trials. Intellia's MAGNITUDE-2 trial is enrolling again after the FDA lifted its clinical hold in January 2026. ART001 is currently limited to China.
If you are a cardiologist or internist: Screen heart failure patients over 65 for ATTR, especially those with unexplained left ventricular hypertrophy and preserved ejection fraction. Technetium-99m pyrophosphate scintigraphy is non-invasive and widely available. The diagnostic gap between estimated prevalence and actual diagnoses suggests hundreds of thousands of patients remain unidentified.
If you invest in biotech: Watch Intellia's MAGNITUDE trial closely. The FDA's partial hold on the cardiomyopathy arm is the single largest regulatory risk to in-vivo CRISPR commercialization. If the hold lifts with modified dosing or monitoring protocols, nex-z has a clear path to being the first in-vivo CRISPR drug approved in the US. ART001's regulatory path in China is less transparent but faster: China's NMPA has approved gene therapies with smaller datasets than any Western regulator would accept. The ART001 paper is open access.
The Bottom Line
Two countries built the same molecular weapon against the same disease and fired it into patients within months of each other. That weapon works. A single injection knocked TTR protein levels down by 90% or more in both programs, and ART001's data shows that suppression holding at 72 weeks with no sign of wearing off. One patient died in the American trial. Nobody died in the Chinese trial. Those facts are both true and misleading, because the Chinese trial was 45 times smaller and lacked the statistical power to detect rare events. What matters is not safety but who gets to market first with a one-time cure for a disease that currently costs $225,000 to $450,000 every single year to manage, for every remaining year of the patient's life. For the roughly 10,000 patients worldwide who know they carry a hereditary TTR mutation and the hundreds of thousands who do not yet know they have the wild-type form, the answer to that question is worth a great deal more than the injection itself.
Sources: Frontiers in Medicine, April 15, 2026, Jiang et al.; NeurologyLive, Jan 28, 2026; ICER patisiran pricing assessment; ClinicalTrials.gov NCT06128005 (MAGNITUDE-2); Alnylam Pharmaceuticals (Onpattro prescribing information); Pfizer (Vyndaqel prescribing information).