Gene editing in the clinic

• ClinicalTrials.gov lists 262 ¹ interventional gene-editing studies; 131 ² are active and 11 ³ are Phase 3.
• 133 ⁴ started between January 2023 and September 2026: 54 ⁵ with a China site and 54 ⁶ with a US site.
• Base-editing studies number 35 ⁷ of the new starts, and in vivo liver-editing studies 10 ⁸.
• Cancer (36 ⁹) and sickle cell disease or beta-thalassaemia (18 ¹⁰) are the largest indication groups since 2023.

Gene-editing trial starts rose from 14 ¹¹ in 2020 to 27 ¹² in 2023 and 40 ¹³ in 2024, then 35 ¹⁴ in 2025. The numbers are small because each programme is expensive and manufacturing-intensive.

China's rise is the main geographic change. Studies with a China site went from 3 ¹⁵ in 2020 to 13 ¹⁶ in 2023 and 14 ¹⁷ in 2025, matching the United States (15 ¹⁸ in 2025). Chinese base-editing companies such as CorrectSequence, and hospital-led programmes in thalassaemia, drive much of this. Global cell and gene therapy spending reached $5.9 billion in 2023 ¹⁹, still a small share of medicine spending.

Edited cell therapies. CRISPR- or TALEN-edited CAR-T and TCR-T cells appear in 42 ³⁰ studies started since 2023. Editing is used to make allogeneic products, to knock out checkpoints or to add safety switches.

Base editing. Base editors change a single DNA letter without cutting both strands. They appear in 35 ⁷ new studies, more than any other single technique, including Beam's programmes and several Chinese thalassaemia programmes. In 2025 a team in Philadelphia treated an infant with a personalised base-editing therapy for CPS1 deficiency, a rare urea-cycle disorder (NEJM), showing that bespoke editing is feasible.

Ex vivo stem-cell editing for sickle cell disease and beta-thalassaemia appears in 18 ³¹ new studies. In vivo liver editing with lipid nanoparticles appears in 10 ⁸, including Intellia's Phase 3 programmes in transthyretin amyloidosis and hereditary angioedema and CRISPR Therapeutics' lipid-lowering programmes.

Cancer is listed in 36 ⁹ gene-editing studies started since 2023, mostly for edited cell therapies. Sickle cell disease and beta-thalassaemia follow with 18 ¹⁰, and inherited liver and metabolic diseases with 17 ³². Inherited eye disease accounts for 5 ³³.

The shift from blood disorders to liver disease matters commercially. Ex vivo editing needs stem-cell collection, chemotherapy conditioning and a hospital stay, which limits uptake. A one-time infusion that edits liver cells could treat common conditions such as high cholesterol, and not only rare diseases.

The FDA approved Casgevy (exagamglogene autotemcel), the first CRISPR-based medicine, for sickle cell disease on 8 December 2023 (FDA announcement). It later added transfusion-dependent beta-thalassaemia. No in vivo gene-editing therapy is approved yet.

Large companies are buying rather than building. Eli Lilly completed its acquisition of Verve Therapeutics in July 2025 (Lilly) to gain its in vivo base-editing programmes for cardiovascular disease.

• In vivo delivery beyond the liver is the key problem. Lipid nanoparticles reach the liver well. Muscle, brain, lung and blood-stem-cell delivery in vivo remain open, and whoever solves one owns a platform.
• Base and prime editing are the preferred tools. Their safety profile, with no double-strand breaks, suits common diseases where the risk tolerance is low.
• Pick diseases where one treatment replaces a lifetime of therapy. The economic case is strongest where a one-time edit replaces chronic, costly treatment, such as lipid disorders, hereditary angioedema and haemoglobin disorders.
• China is competitive in thalassaemia and base editing. Local disease prevalence and fast investigator-initiated studies give Chinese companies data sooner. Partnerships are a natural route to global markets.

Every trial count is the totalCount of one ClinicalTrials.gov API v2 query, retrieved on 26 September 2026. Each numbered citation opens a record whose source link re-runs the query. The track includes interventional studies whose record mentions CRISPR, Cas9 or Cas12, base or prime editing, zinc-finger nucleases, TALENs or a named gene-editing product.

Counts are small. Approach bars overlap: an edited CAR-T made with a base editor counts in both. Site-country counts overlap, and China-only registrations are missing. Approvals, the 2025 personalised treatment and acquisitions are based on FDA and company announcements and published reports; the global spending figure is cited to its publisher.