Is CRISPR gene editing safe for treating genetic diseases in humans?

What do clinical trials tell us about CRISPR safety?

The most direct evidence comes from human trials, and the results so far are encouraging but limited. In a landmark 2021 phase 1 study of 6 patients with hereditary transthyretin amyloidosis, a single infusion of the CRISPR therapy NTLA-2001 caused only mild adverse events and reduced the disease-causing protein by an average of 87% at the higher dose [2]. This was the first in-human study of a systemically delivered CRISPR therapy, and its clean safety profile in a small group is a strong signal. Similarly, in trials for sickle cell disease and beta-thalassemia, which led to the FDA approval of Casgevy in 2023, all 22 patients for whom data is available experienced resolution of major symptoms with a good safety record [6]. These trials show that for blood and liver disorders, CRISPR can be both effective and well-tolerated in the short term.

However, these are early-stage studies with small numbers of patients (6 to 22), and they focus on short-term safety—typically 28 days to a few months. The authors of these studies explicitly note that long-term safety, including the risk of unintended genetic changes, remains an open question [1][3]. So while the immediate safety data is positive, it is not yet definitive.

What are the main risks that could make CRISPR unsafe?

The biggest safety concern is off-target editing—when CRISPR cuts DNA at the wrong place, potentially causing harmful mutations. A 2023 review notes that off-target effects can include small insertions or deletions, as well as larger structural changes like inversions and translocations [7]. A 2024 review of detection methods confirms that these unwanted edits are a 'vital concern' for therapy development, but also highlights that new tools have been developed to predict and minimize them [4]. For example, in zebrafish studies, most predicted off-target sites had very low actual editing frequencies (less than 1%) [8], suggesting that careful design can reduce risk.

A rarer but more alarming risk is 'CRISPRthripsis'—a phenomenon where CRISPR-induced DNA breaks lead to massive chromosome shattering, similar to what is seen in some cancers. A 2022 paper warns that this can occur when chromosomal truncations create micronuclei and chromosome bridges, leading to extensive rearrangements [11]. The authors note that this risk is influenced by factors like the specific guide RNA and cell type, and that strategies exist to minimize it, but it remains a concern for clinical applications [11]. This is not something seen in the human trials to date, but it underscores that CRISPR is not a perfectly precise tool.

How do delivery methods and long-term effects affect safety?

Getting CRISPR components into the right cells is a major safety challenge. The CRISPR system (Cas9 protein and guide RNA) can be delivered as DNA, RNA, or protein, and each method has different risks. Viral delivery, while efficient, can cause immune reactions and has limited capacity for large genetic payloads [10]. Non-viral carriers, such as lipid nanoparticles (used in the amyloidosis trial [2]), are safer but less efficient. A 2022 review emphasizes that developing better carriers is essential to protect the CRISPR system from degradation, target specific tissues, and reduce immunogenicity [9]. The choice of delivery method directly impacts safety.

Long-term safety is still unknown. Because CRISPR makes permanent changes to DNA, any off-target effects could persist for a patient's lifetime. For germline editing (changes that could be inherited), the ethical and safety concerns are even greater, as unintended changes could be passed to future generations [1][5]. The authors of a 2024 review on CRISPR therapies for blood and muscle disorders note that ongoing trials are specifically monitoring for 'long-term efficacy, potential genotoxicity, and adverse immune reactions' [3]. Until we have years of follow-up data from larger patient groups, the full safety profile of CRISPR therapies will remain incomplete.

About These Sources

This answer is built on 11 peer-reviewed studies — published from 2021 to 2025, 4 from 2024 or later, 8 in Q1 journals, collectively cited 2,455 times — selected as the most relevant from 11 studies that passed quality screening, drawn from 73 papers retrieved from a database of over 500 million.

Sources used in this answer

Related research questions