Autoimmune CAR-T Hits a Wall: What Novartis and BMS Pauses Mean for Cell Therapy

A Pause That Goes Deeper Than One Molecule

Novartis and Bristol Myers Squibb did not simply stop a single study. They pulled multiple autoimmune CAR-T programs off the table at roughly the same time, each citing distinct but related safety findings. That simultaneity is the signal.

Novartis halted eight trials of rap-cel on August 24 after three cases of immune effector cell-associated hemophagocytic syndrome — a rare, systemic inflammatory condition that can trigger organ failure and death. Bristol Myers paused enrollment in zola-cel studies after detecting transient inflammatory events during routine surveillance. Neither company has disclosed a direct link between the two signals. But the overlapping timing and the shared therapeutic area suggest something broader is at play.

The stakes are high. Autoimmune CAR-T was the industry’s most promising pivot from oncology, carrying the promise of treating conditions like lupus and multiple sclerosis with a one-time infusion that could reset the immune system. If the platform falters now, the ripple effects will be felt across a pipeline that several smaller biotechs have already bet their valuations on.

Why These CAR-Ts Are Different

Rap-cel and zola-cel were built differently from the CAR-T products that already treat blood cancers. The older therapies — Kymriah, Yescarta and their kin — are personalized, made from a patient’s own T cells, and manufactured over weeks. The new generation uses accelerated production workflows designed to shrink turnaround time and make the approach viable for chronic autoimmune diseases, where repeated infusions would be impractical and the patient population is far larger than the oncology base.

The speed-up may be coming with a cost. William Blair analyst Sami Corwin noted that the faster expansion process could be driving increased cell proliferation — and therefore heightened inflammatory risk. That does not mean the technology is flawed. It means the dose-response relationship may look different outside cancer, where the target antigen landscape, tumor burden, and inflammatory environment are entirely distinct.

This is not the first time a CAR-T platform has behaved unexpectedly when moved into a new indication. But the autoimmune space is unusually unforgiving. Patients with lupus or myasthenia gravis are not critically ill in the same way leukemia patients are. Their baseline risk tolerance is lower, and the therapeutic window is narrower. A cytokine storm that might be survivable in a heavily pretreated cancer patient could be catastrophic in someone whose organs are already stressed by chronic inflammation.

The Fatal Signal

Novartis confirmed that complications from the three IEC-HS cases led to fatal outcomes. The company described the hold as temporary and said it is engaged with regulators while reviewing the full safety dataset. BMS characterized its findings as transient and reversible and said it expects to resume testing soon.

The asymmetry matters. Novartis is dealing with deaths. BMS is dealing with events that resolved. Both pauses are voluntary, which is standard practice when a sponsor encounters unexpected toxicity. But the public framing already tells a story: Novartis will face more rigorous questioning from the FDA before any restart, and the data it produces in the coming weeks will set the tone for the entire indication.

IEC-HS is not a well-characterized reaction in the autoimmune CAR-T context. It has been reported in oncology trials, but at low frequency. Its appearance multiple times in a single program — and with fatal consequence — suggests either a patient susceptibility factor, a manufacturing variable, or a dosing threshold that has not yet been identified. Novartis has not specified which. Until it does, the FDA is likely to demand a more granular safety analysis than it would for a routine hold.

Who Gets Hit Next

The immediate victims are the patients enrolled in the paused studies. Novartis was evaluating rap-cel in lupus, myasthenia gravis, multiple sclerosis and several other immunology and neuroscience indications. BMS was preparing to begin a systemic sclerosis trial this month. Those patients lose access to a therapy that, in early studies, produced treatment-free remission in a subset of participants with disease that had resisted every conventional option.

The secondary victims are the competitors. Cabaletta Bio, which runs a late-stage myositis trial, and Kyverna Therapeutics, which is pursuing stiff person syndrome and myasthenia gravis, will watch these pauses closely. Both companies are smaller and more exposed. A regulatory chill that makes the FDA more conservative about autoimmune CAR-T approvals will affect them disproportionately. Kyverna was expected to file for approval in stiff person syndrome by year-end. That timeline is now clouded.

Investors will recalibrate. The autoimmune CAR-T narrative has been one of the most compelling in biotech over the past three years. It promised a category expansion that could multiply the addressable market for CAR-T tenfold. The pause does not kill the thesis, but it introduces a risk premium that the market has not priced in.

What the FDA Will Do

The agency has never approved a CAR-T for an autoimmune indication. It is writing the rulebook as it goes. A fatal signal from two different sponsors in the same therapeutic area will make that rulebook stricter.

Expect the FDA to ask for more preclinical mechanistic data, more rigorous patient selection criteria, and longer monitoring periods before re-enrollment. It may also push for centralized safety review boards rather than leaving toxicity assessments to individual sites. The goal will not be to stop development — the agency has every incentive to see a first-in-class autoimmune CAR-T approved — but to ensure the first product that reaches patients does so with a safety profile that justifies the label.

That is a higher bar than oncology required, and it is one the industry underestimated.

The Bigger Picture

The pause is a setback, not a terminus. Novartis has other rap-cel programs in cancer that are continuing. BMS has expressed confidence in zola-cel. The biology is sound — autologous CAR-T can induce deep remissions in autoimmune disease, and the academic literature supports further investment.

But the industry needs to confront an uncomfortable truth: CAR-T was optimized for cancer, and cancer patients tolerate toxicity that autoimmune patients cannot. Shifting the indication requires more than a new target antigen. It requires rethinking manufacturing, dosing, monitoring and patient selection from the ground up. The faster production workflows that made autoimmune CAR-T commercially attractive may also be the source of the new toxicity signal.

The companies that treat this pause as a calibration moment — not a circuit breaker — will emerge stronger. Those that assume the science is proven and the path forward is clear will pay the price when the FDA demands evidence they have not yet produced.

Autoimmune CAR-T is not dead. It is just no longer easy.