What do early human trials tell us about CRISPR safety?
The strongest evidence for CRISPR safety in humans comes from a small phase 1 trial of NTLA-2001, an in vivo CRISPR therapy for transthyretin amyloidosis. In six patients, only mild adverse events were observed, and serum transthyretin protein—the misfolded protein causing disease—dropped by a mean of 52% at the lower dose and 87% at the higher dose [5]. This suggests that when delivered via lipid nanoparticles (LNPs) to the liver, CRISPR can be both safe and highly effective in the short term.
However, the field is still young. A 2022 review noted that while over 900 investigational gene therapies are in clinical studies, long-term safety data are lacking, and durability of effect remains uncertain—for example, factor VIII levels declined after 12 months in a hemophilia A gene therapy trial [1]. Similarly, CAR-T cell therapies using CRISPR-edited cells have shown high remission rates (77–80%) in leukemia but also carry risks of cytokine release syndrome (26% grade ≥3) and neurotoxicity (12% grade ≥3) [1].
What are the main safety concerns with CRISPR?
The biggest safety concern is off-target effects—unintended edits elsewhere in the genome that could cause cancer or other harm. A 2025 review emphasized that despite many detection tools, there are no standardized guidelines for measuring off-target activity, leading to inconsistent safety assessments across studies [3]. A 2023 review echoed this, noting that off-target effects remain a major hurdle for clinical translation [6].
Delivery method matters greatly. Viral vectors (like AAV) are efficient but carry risks of insertional mutagenesis and immune reactions; for instance, dosing was suspended in two AAV trials in 2021 due to deaths and liver tumors in preclinical studies [1]. In contrast, non-viral delivery using lipid nanoparticles (LNPs) shows lower immunogenicity and better safety profiles, as seen in the NTLA-2001 trial [4][5]. Messenger RNA delivery of CRISPR components also reduces off-target editing compared to plasmid DNA, because Cas9 is only expressed transiently [2].
Is CRISPR safe for use in embryos or germline editing?
CRISPR editing of the human germline (sperm, eggs, embryos) is far from safe for clinical use. A 2022 review highlighted major hurdles including loss of heterozygosity and mosaicism (where only some cells are edited), and noted that ethical constraints and a shortage of human material limit research [7]. Currently, no clinical applications of human germline gene editing are considered acceptable due to these unresolved safety and ethical issues.
For somatic (non-reproductive) cells, the safety profile is better but still evolving. For example, base editing using mRNA in muscle stem cells achieved 90% on-target editing efficiency without detectable off-target effects in one study, but the authors stressed that thorough unbiased off-target evaluation is needed before clinical trials [2].
Sources used in this answer
Clinical gene technology in Australia: building on solid foundations
Gene therapies including CRISPR are advancing rapidly, but long-term safety, durability, and immune responses remain open questions; AAV vector trials have been suspended due to deaths and liver tumors.
Messenger RNA brings gene editing a step closer to treat muscular dystrophies
mRNA delivery of adenine base editors in muscle stem cells achieved up to 90% on-target editing efficiency without detectable off-target effects, but thorough unbiased off-target evaluation is still needed before clinical trials.
Off-target effects in CRISPR-Cas genome editing for human therapeutics: Progress and challenges
Off-target genotoxicity is a major concern for CRISPR therapies; despite many detection tools, lack of standardized guidelines leads to inconsistent safety assessments.
Lipid Nanoparticles for Delivery of CRISPR Gene Editing Components
Lipid nanoparticles (LNPs) are attractive non-viral vectors for CRISPR delivery due to low immunogenicity and high efficiency, but targeting non-liver tissues remains challenging.
CRISPR-Cas9 In Vivo Gene Editing for Transthyretin Amyloidosis
In a phase 1 trial of NTLA-2001 for transthyretin amyloidosis, only mild adverse events occurred, and serum TTR protein dropped by 52% (low dose) and 87% (high dose) at day 28.
Off-target effects in CRISPR/Cas9 gene editing
Off-target effects remain a major hurdle for clinical translation of CRISPR/Cas9; many methods have been developed to detect and reduce them, but challenges persist.
CRISPR/Cas gene editing in the human germline
Human germline gene editing faces hurdles like loss-of-heterozygosity and mosaicism, and ethical constraints prevent clinical applications at this time.