Novartis and Bristol Myers Squibb Halt Autoimmune CAR-T Trials: What the Three Deaths Actually Show

In late August 2026, Novartis halted screening, enrollment, and dosing across eight clinical trials testing its experimental CD19-directed CAR-T cell therapy rapcabtagene autoleucel ("rap-cel") in autoimmune diseases including lupus, systemic sclerosis, rheumatoid arthritis, myasthenia gravis, and multiple sclerosis, after three patients died from a severe inflammatory complication. Bristol Myers Squibb followed within days, pausing enrollment in its own CD19 CAR-T candidate, zolacabtagene autoleucel ("zola-cel"), as a precaution. This article covers what actually happened, what is and isn't yet known about the cause, and what it means for the fast-growing effort to bring cancer-derived CAR-T technology into autoimmune medicine — a story where genuine promise and genuine risk both belong in the same sentence.

What Happened

On August 24, 2026, Novartis halted eight Phase 1/2 and Phase 2 trials of rap-cel spanning systemic lupus erythematosus and lupus nephritis, systemic sclerosis, ANCA-associated vasculitis, idiopathic inflammatory myopathies, rheumatoid arthritis, Sjögren's disease, generalized myasthenia gravis, and both relapsing and non-active progressive multiple sclerosis. The company said three trial participants died of immune effector cell-associated hemophagocytic syndrome (IEC-HS), a recognized but uncommon complication of CAR-T therapy in which rapidly expanding, highly activated engineered T cells trigger an overwhelming systemic inflammatory response that can damage organs throughout the body. Novartis stated that patients already treated in the affected trials would continue to be monitored under protocol, and that its oncology CAR-T programs, including studies in chronic lymphocytic leukemia, were not affected. Bristol Myers Squibb subsequently paused enrollment in its own CD19 CAR-T trials for lupus, systemic sclerosis, and myositis, describing what it observed as "transient and reversible" inflammatory events rather than fatalities, and calling its pause a voluntary, precautionary step "out of an abundance of caution" while it reviewed safety data (Fierce Biotech, 2026; Endpoints News, 2026).

The Actual Evidence So Far

What is confirmed: three deaths, one shared adverse-event mechanism (IEC-HS), and a company-initiated halt at Novartis, alongside a separate, less severe safety signal at Bristol Myers Squibb that has not been linked to any deaths. Both companies' rap-cel and zola-cel programs use newer "rapid manufacturing" platforms — Novartis's T-Charge and BMS's NEX-T — designed to shorten the vein-to-vein time between collecting a patient's T cells and re-infusing the engineered product, in some cases to under two days instead of the one-to-two weeks typical of earlier-generation CAR-T manufacturing. Industry analysts, including William Blair's Sami Corwin and Jefferies' Roger Song, have pointed to this accelerated manufacturing as a plausible contributor: cells produced this way may be "younger" and expand more aggressively once inside the body, and autoimmune patients already have more baseline immune activation than most cancer patients, which could compound that expansion (BioPharma Dive, 2026). None of this is confirmed causation — it is informed speculation from analysts, not a finalized root-cause finding from either company or a regulator, and should be treated as such.

Important Caveats and Limitations

This is not evidence that CAR-T therapy for autoimmune disease is unsafe as a category, and it is not evidence that it is safe — the honest answer, based on current public information, is that it is unresolved. Several load-bearing facts keep this uncertain: neither company has publicly disclosed exact patient totals dosed across the affected trials, so the underlying event rate cannot be calculated from what has been published. No regulatory body's independent findings have been reported publicly as of this writing; Novartis has said it is "engaged" with regulators, but a formal FDA clinical hold has not been confirmed in public reporting. IEC-HS itself is not new to CAR-T medicine — it has been documented for years in oncology use of CAR-T for lymphoma and myeloma — but severe or fatal cases have been rare in that setting, and it is not yet established whether autoimmune-disease patients face a meaningfully higher risk than cancer patients do, or whether the manufacturing-platform hypothesis holds up. Other companies developing autoimmune CAR-T therapies, including Kyverna Therapeutics and Cabaletta Bio, have said their manufacturing processes differ meaningfully from the paused programs, but these are company statements, not independent confirmation of lower risk.

What's Next

Both companies say they are conducting a comprehensive review of clinical and safety data before deciding whether or how to resume their halted trials. TD Cowen analyst Phil Nadeau suggested there may be "limited read-through" to other companies' autoimmune CAR-T programs given differences in manufacturing protocol, while acknowledging the field as a whole will face closer scrutiny. Patient advocacy groups, including the Lupus Foundation of America, have been tracking the pause closely given how much hope autoimmune patients have placed in CAR-T as a potential deep, drug-free remission strategy — early, smaller academic studies of CD19 CAR-T in lupus had previously reported encouraging remission signals, which is part of why large pharmaceutical companies moved so quickly into this space. Whether this pause becomes a temporary manufacturing fix or a more fundamental reassessment of how fast "next-generation" CAR-T platforms should be pushed into non-cancer indications will likely become clearer only after both companies complete and disclose their safety reviews.

Bottom Line

Three deaths from a known but serious CAR-T complication have paused Novartis's entire autoimmune CAR-T pipeline and prompted Bristol Myers Squibb to pause its own trials as a precaution — a genuine and sobering safety signal that deserves to be reported plainly, not minimized. At the same time, this is not proof that autoimmune CAR-T therapy is broadly unsafe, that other companies' programs share the same risk, or that regulators have found a systemic problem: much of the causal story (manufacturing platform, patient selection, dosing) remains analyst speculation rather than confirmed fact. Patients and clinicians should treat this as an active, unresolved safety investigation until the companies and regulators report their findings.