Gene therapy is not risk-free, but for most approved treatments, the benefits have been judged to outweigh the dangers for people with serious or life-threatening diseases. The safety picture varies significantly depending on the type of gene therapy, how it’s delivered, and the dose involved. Some patients experience mild, manageable side effects, while rare but severe complications, including organ failure and death, have occurred in both clinical trials and post-approval use.
How Gene Therapy Delivery Affects Risk
The way a gene therapy reaches your cells is one of the biggest factors in how safe it is. Most gene therapies use modified viruses (called viral vectors) to carry new genetic instructions into cells. There are two broad approaches, and they carry different levels of risk.
In “ex vivo” therapy, doctors remove your cells, modify them in a lab, verify the changes, and then return them to your body. This approach allows greater control and verification of genetic modifications before anything goes back into you, making it potentially safer and more precise. CAR-T cell therapies for certain blood cancers work this way.
In “in vivo” therapy, the viral vector is injected directly into your body, either into the bloodstream or a specific tissue. This offers less control over what happens once the vector is inside you. Your immune system may react to the viral vector itself, sometimes aggressively. The gene therapy Luxturna, injected directly into the eye for an inherited form of blindness, carries different risks than Zolgensma, which is delivered intravenously to treat spinal muscular atrophy in infants.
Liver Toxicity: The Most Common Serious Risk
When gene therapy is delivered through the bloodstream, the liver absorbs a large share of the viral vectors. This makes liver toxicity the most common serious adverse event following intravenous gene therapy. Symptoms typically appear between one and four weeks after injection, and doctors monitor liver enzyme levels closely during this window to catch problems early.
Most patients who develop liver inflammation recover after treatment with steroids or other immune-suppressing drugs. But in some cases, the damage is severe enough to cause acute liver failure. In early 2025, the FDA requested that Sarepta Therapeutics suspend distribution of Elevidys, a gene therapy for Duchenne muscular dystrophy, after three deaths linked to acute liver failure. The FDA also placed clinical holds on multiple related gene therapy trials using the same viral vector platform and revoked the platform’s technology designation, citing insufficient evidence that it could be safely used across multiple products.
That action illustrates an important reality: gene therapy safety is still being defined in real time. Even after a product reaches the market, new safety signals can emerge that change the risk calculus.
Immune Reactions and Cytokine Release
Your immune system can treat the viral vector or the new protein it produces as a foreign invader. This triggers an inflammatory response that ranges from mild fever to a dangerous condition called cytokine release syndrome (CRS), where the immune system floods the body with inflammatory signals.
CRS is especially common with CAR-T cell therapies. In studies of patients with leukemia or lymphoma receiving CAR-T treatment, 43 to 100% experienced some degree of cytokine release. Most cases are mild, grade 1 or 2, and managed with fever reducers, fluids, and supplemental oxygen. Severe cases (grade 3 or 4) are life-threatening and require intensive care. Fortunately, effective treatments exist: corticosteroids and a targeted drug that blocks a key inflammatory signal can usually bring severe CRS under control.
Beyond CRS, the body can also mount a longer-term immune response. Empty viral capsids (essentially virus shells without the therapeutic gene inside) and editing proteins left over from the manufacturing process can trigger both immediate and delayed immune reactions. These residual impurities are a known challenge in gene therapy manufacturing. Safety events linked to pre-existing immunity to viral vectors have been associated with high doses, and doctors use strategies like steroid pretreatment or filtering antibodies from the blood before treatment to reduce this risk.
Off-Target Effects With Gene Editing
Newer gene therapies that use gene-editing tools like CRISPR don’t just add a gene. They cut and modify existing DNA. The concern here is precision: the editing tool might cut in the wrong place, creating unintended changes elsewhere in the genome. These “off-target effects” could theoretically disrupt important genes, potentially leading to problems like uncontrolled cell growth.
Researchers have developed increasingly sensitive methods to detect off-target edits, and so far, clinical trials of CRISPR-based therapies (like Casgevy for sickle cell disease) have not reported major safety problems from unintended edits. But the technology is still young, and detecting rare off-target events requires long-term monitoring that is ongoing.
Why Long-Term Monitoring Matters
One of the unique challenges of gene therapy is that its effects are meant to be permanent or very long-lasting. That’s the appeal, but it also means side effects could emerge years later. The FDA requires long-term follow-up plans for gene therapy products that present ongoing risks. These extended assessments continue monitoring patients well past the active treatment period, specifically watching for delayed adverse events that might not appear for months or years.
Not every gene therapy requires this extended follow-up. The FDA asks manufacturers to perform a risk assessment based on the type of vector, the target tissue, and whether the new genetic material integrates into the patient’s chromosomes (which carries a higher long-term risk than vectors that remain separate from the genome). For products that do integrate, one historical concern has been “insertional mutagenesis,” where the new gene lands in a spot that activates a cancer-promoting gene. Early gene therapy trials in the 2000s saw cases of leukemia caused by this mechanism, and it drove major changes in vector design that have made newer products significantly safer.
Somatic vs. Germline: A Critical Distinction
Every gene therapy currently approved or in clinical trials targets somatic cells, the ordinary cells that make up your tissues. Changes to somatic cells affect only you. They are not passed to your children. There is broad agreement among researchers and bioethicists that somatic gene therapy can be an ethical approach to treating disease, and the safety framework for evaluating it is well established.
Germline gene editing is a completely different matter. This would modify reproductive cells (eggs or sperm) or embryos, meaning changes would be inherited by future generations. The clinical use of germline gene editing is prohibited in many countries. The American Society of Gene & Cell Therapy’s position reflects a universal international consensus: it is neither safe nor effective at this time to use gene editing on germline cells to prevent disease in a yet-unborn person. The unknowns are simply too large, and any error would have multigenerational consequences. The World Health Organization is developing global standards for governance of germline editing, but clinical use remains off the table for the foreseeable future.
Putting the Risks in Context
Gene therapy is typically reserved for serious, often life-threatening conditions where existing treatments are inadequate. A child with spinal muscular atrophy, a patient with a severe inherited blood disorder, or someone going blind from a genetic eye disease faces a baseline of significant suffering or shortened life. The risks of gene therapy are real, but they’re weighed against that baseline.
Every gene therapy that reaches the market has gone through the FDA’s approval process, which is designed to verify that a treatment’s benefits justify its risks for a specific patient population. That doesn’t mean every product will prove safe in the long run. The Elevidys situation in 2025 shows that the FDA is willing to pull products and halt trials when new safety signals emerge. The system is imperfect but active.
For patients considering gene therapy, the practical safety picture depends on the specific product, the condition being treated, your own health status, and whether you have pre-existing antibodies to the viral vector being used. The risk profile of a one-time injection into the eye is fundamentally different from a high-dose intravenous infusion or an immune cell therapy that triggers CRS in the majority of recipients. There is no single answer to “is gene therapy safe” because the category now spans dozens of distinct products with very different risk profiles.