Novartis: Three deaths put CAR-T expansion into autoimmune diseases on hold

Why it matters: The promise of CAR-T therapy is no longer limited to cancer. After transforming treatment for several blood cancers, the technology is increasingly being explored as a potential way to reset the immune system in severe autoimmune diseases.

Backstory: Novartis is one of the companies at the forefront of that expansion. Its autologous CD19 CAR-T therapy rapcabtagene autoleucel (rap-cel, YTB323) is designed to eliminate CD19-positive B cells. The rationale is fundamentally different from conventional immunosuppression: rather than continuously suppressing an overactive immune system, CAR-T could potentially remove a population of dysfunctional immune cells and allow the immune system to rebuild.

Zoom in: The approach has generated considerable excitement. At the EULAR 2026 meeting, Novartis presented early data from its AUTOGRAPH studies showing clinically meaningful improvements and rapid, deep B-cell depletion in patients with severe, treatment-refractory idiopathic inflammatory myopathies and diffuse cutaneous systemic sclerosis. At the time, the company described the safety profile as manageable.

Yes, but: The latest development puts that optimism into a more complicated context. At the end of August, Novartis temporarily halted screening, randomization and treatment administration across eight autoimmune and neurological studies of rap-cel after three patients died following serious immune-related events.

• According to Novartis, the deaths were associated with immune effector cell-associated hemophagocytic syndrome (IEC-HS), a severe systemic inflammatory complication associated with immune-cell therapies.
• The company has not disclosed when the deaths occurred or detailed the individual cases. It is conducting a comprehensive review of the safety data and is working with independent safety boards and regulators. Patients who have already received treatment will continue to be monitored according to the study protocols.
• The affected program spans a broad range of indications, including systemic lupus erythematosus and lupus nephritis, systemic sclerosis, ANCA-associated vasculitis and myositis, as well as rheumatoid arthritis, Sjögren’s disease, generalized myasthenia gravis and multiple sclerosis.

How it works: CAR-T therapy involves collecting a patient’s own T cells, genetically engineering them to recognize a specific target and reinfusing them into the patient. Once back in the body, the engineered cells can expand and attack their target.

• In autoimmune disease, the target is not a tumor cell. Novartis’ approach targets CD19-positive B cells, which play a central role in antibody production and autoimmune responses. The goal is to remove the dysfunctional B-cell population and potentially produce a deeper and longer-lasting reset of the immune system. That is what makes CAR-T so attractive for severe autoimmune disease — and also what makes its safety profile particularly important.

Reality check: The technology’s major toxicities are already well known in oncology. CAR-T cells can trigger powerful systemic immune activation, including cytokine release syndrome and other inflammatory complications. IEC-HS represents a particularly severe form of immune dysregulation that can lead to organ damage.

• The challenge becomes even more pronounced when CAR-T moves into chronic autoimmune diseases. In advanced cancer, patients and physicians may accept substantial treatment risks because the underlying disease can be immediately life-threatening. In autoimmune disease, patients may be severely ill but are often expected to live for decades with their condition. The acceptable risk threshold is therefore fundamentally different.

The big picture: The development also raises broader questions for the emerging autoimmune CAR-T field. Bristol Myers Squibb has reportedly paused similar CAR-T studies in autoimmune diseases as a precaution after observing immune-related adverse events. Unlike the Novartis program, the reported BMS events were not fatal. Both companies are targeting CD19-positive B cells with the aim of resetting the immune system.

• That parallel does not mean that CAR-T is failing as a technology. CAR-T remains an established treatment platform in several blood cancers, where it has produced durable responses in some patients. Novartis is also continuing its rap-cel cancer studies, according to reports. Rather, the events highlight a central question for the next phase of cell therapy: Can the depth of immune-system intervention that makes CAR-T potentially transformative in autoimmune disease be achieved with a risk profile acceptable for patients with chronic disorders?
• The therapeutic upside is substantial. Case reports and early clinical studies have suggested that a single CAR-T treatment can produce prolonged remission in some patients with otherwise difficult-to-treat autoimmune disease. In April, Nature reported on a patient with three rare autoimmune diseases who remained symptom-free and off medication for more than a year after CAR-T treatment. But anecdotal successes cannot resolve the population-level safety question.

What’s next: For Novartis, the immediate priority will be to determine whether the three fatal IEC-HS events share identifiable risk factors. Investigators will need to assess whether the events are linked to particular patient populations, disease severity, lymphodepletion regimens, CAR-T exposure or other aspects of the treatment protocol.

• The breadth of the pause suggests that Novartis is taking a program-wide approach rather than investigating a single trial in isolation. The company has said it is reviewing the evolving clinical and safety data across the program in cooperation with independent safety committees and regulators.
• The outcome could have implications well beyond rap-cel. If specific risk factors or mitigation strategies can be identified, the current pause could ultimately lead to more selective patient recruitment, modified treatment protocols or tighter monitoring rather than termination of the program. If the deaths point to a more fundamental limitation of CD19 CAR-T in autoimmune disease, however, the setback could slow one of the most closely watched expansions of cell therapy beyond oncology.

Bottom line: For now, the message is neither that CAR-T has failed nor that the autoimmune opportunity has disappeared. It is that reprogramming the immune system is a powerful intervention — and the biology that makes it effective can also make it dangerous.