The Cancer Therapy That Killed Three People Who Didn't Have Cancer
Three patients with autoimmune diseases died in Novartis CAR-T trials, forcing the company to pause eight studies and the entire field to confront an uncomfortable question: can cancer's most powerful therapy ever be safe enough for diseases that aren't trying to kill you?
Three patients walked into a clinical trial for autoimmune disease. They weren't dying. They had conditions like lupus and rheumatoid arthritis: painful, chronic, life-altering, but not immediately fatal. The treatment they received was a souped-up version of one of oncology's most powerful weapons. None of them walked out alive.
On August 24, 2026, Novartis quietly paused eight clinical trials of its experimental CAR-T therapy, rapcabtagene autoleucel (rap-cel), across autoimmune and neurological diseases. The reason: three patient deaths linked to a severe, life-threatening immune reaction. The company is now conducting a comprehensive review with independent safety boards and regulators.
This isn't just a Novartis problem. It's a reckoning for an entire field that's been racing to repurpose cancer's most aggressive therapy for diseases that don't kill you.
A Cancer Weapon Pointed at a Different Enemy
To understand why this matters, you need to understand what CAR-T therapy actually is. Doctors take a patient's own immune cells (T-cells), genetically engineer them to hunt down specific targets, and infuse them back into the body. In cancer, those targets are tumor cells. The therapy has produced near-miraculous remissions in patients with blood cancers who had run out of options.
But CAR-T comes with serious baggage. The most well-known risk is cytokine release syndrome (CRS), essentially an immune system wildfire where the engineered cells activate so aggressively that the body's inflammatory response spirals out of control. There's also neurotoxicity: confusion, seizures, encephalopathy. And infections, which are now recognized as the leading cause of non-relapse death in CAR-T patients. Treatment-related mortality across published studies sits around 1%.
In cancer, that tradeoff can make sense. You're staring down a disease that will kill you. A 1% chance of dying from the treatment is a price many patients will pay. It's like defusing a bomb: the risk of cutting the wrong wire is real, but the bomb is going to explode anyway.
Autoimmune disease is a fundamentally different equation. Lupus, rheumatoid arthritis, multiple sclerosis: these are terrible diseases. They steal quality of life. But they don't typically kill you in months. The bomb is ticking, sure, but it's on a much longer fuse. So the tolerance for treatment-related death drops to nearly zero.
Get tomorrow's biotech intelligence before your competitors.
Join thousands of biotech professionals who start their day with our free, daily briefing.
What Novartis Was Trying to Do
Novartis had big ambitions for rap-cel. The company wasn't just testing it in one autoimmune disease; it was running eight separate studies across lupus, rheumatoid arthritis, vasculitis, multiple sclerosis, myasthenia gravis, systemic sclerosis, Sjögren's disease, and idiopathic inflammatory myopathies.
The underlying idea was compelling: an "immune reset." Instead of managing autoimmune symptoms with drugs patients take for decades, CAR-T could theoretically wipe out the misbehaving immune cells and let the body rebuild from scratch. Early academic reports in diseases like lupus had shown patients achieving drug-free remission, a word that autoimmune patients almost never get to hear. The excitement was real.
Novartis positioned this as part of a broader shift in its immunology strategy, moving from symptom control toward something closer to a functional cure. The company had been building out its autoimmune pipeline with other approaches too, including the antibody ianalumab (bolstered by its 2024 MorphoSys acquisition) and the BTK inhibitor remibrutinib. But CAR-T was the crown jewel, the moonshot.
Then three patients died.
The Killer in the Room
Novartis attributed the deaths to a specific complication: immune effector cell-associated hemophagocytic syndrome, a rare but recognized CAR-T toxicity where the immune overreaction becomes so extreme that the body essentially starts destroying its own blood cells. Think of it as CRS's more dangerous cousin.
This wasn't an unknown risk. Hematologists who treat lymphoma and leukemia patients with CAR-T have seen it before. But context matters enormously. When a terminally ill cancer patient develops this complication, the clinical calculus, while tragic, is different. When someone with lupus dies from a therapy that was supposed to help them live better (not just longer), the moral weight is crushing.
Importantly, Novartis's cancer trials were not paused. Only the autoimmune and neurological studies were halted. That distinction tells you everything about the risk-benefit math: the same side effect is tolerable in one context and unacceptable in another.
The Ripple Effect Across the Field
Novartis isn't the only company in this space, and the fallout is already spreading. Bristol Myers Squibb paused its own autoimmune CAR-T studies as a precaution after observing inflammatory events, though the company characterized those events as temporary and reversible, with a safety profile still consistent with other CAR-T therapies.
Analysts at William Blair noted that the specific toxicity (IEC-HS) is rare but serious, and cautioned against extrapolating Novartis's experience to every autoimmune CAR-T program. They also raised an intriguing question: both Novartis's rap-cel and BMS's therapy use rapid manufacturing processes. Could the speed of production be relevant? It's unproven, but it's the kind of variable that safety reviews will now scrutinize.
But even amid the setback, some analysts noted that the event could actually strengthen interest in alternative B-cell-depleting therapies that might be cheaper and simpler to administer.
The competitive landscape tells a story of its own. Kyverna Therapeutics has treated 100 patients across multiple autoimmune diseases with its KYV-101 program. Cabaletta Bio is running Phase 1/2 trials in lupus, inflammatory myositis, and systemic sclerosis. Cartesian Therapeutics is taking a different route entirely with an mRNA-based CAR-T approach in myasthenia gravis. All of these companies will now face sharper questions from regulators, investors, and patients.
The Real Question Nobody Wants to Answer
The autoimmune CAR-T field is at an inflection point, and the core tension is deceptively simple: how much risk is acceptable when the disease won't kill you?
In oncology, CAR-T earned its place by saving patients who had no other options. The early era was defined by dramatic responses alongside fear of fatal CRS and neurotoxicity. Over time, doctors got better at managing those acute risks. Attention shifted to longer-term concerns like infections and, more recently, secondary cancers (which prompted FDA boxed warnings). The therapy matured.
Autoimmune CAR-T is still in its early era. Most programs are in Phase 1 or Phase 1/2. The diseases being targeted (lupus, systemic sclerosis, vasculitis) are serious and often refractory to existing treatments. Patients in these trials aren't mild cases; they've typically failed multiple lines of therapy. Their quality of life is genuinely terrible.
But "terrible quality of life" and "terminal illness" are different categories. The entire history of medicine is built around calibrating interventions to the severity of the disease. You don't prescribe chemotherapy for a headache. The question is whether CAR-T, even with improvements, can ever be safe enough for autoimmune patients, or whether the field needs fundamentally different engineering to make cell therapy work outside of cancer.
What Happens Next
Novartis says it's sharing information with regulators and continuing to monitor patients who've already received rap-cel. The timing of when the three deaths occurred hasn't been disclosed, which adds an uncomfortable layer of uncertainty. Were these recent? Months ago? The lack of clarity is notable.
The practical analyst takeaway, per multiple coverage outlets, is that this signal will slow enthusiasm and raise scrutiny for autoimmune CAR-T broadly. But it doesn't automatically invalidate the entire approach. The adverse event is recognized in CAR-T more broadly, and the data are still program-specific. Different manufacturing processes, different cell constructs, and different dosing strategies could yield very different safety profiles.
For Novartis specifically, the autoimmune strategy isn't dead. The company still has ianalumab and remibrutinib in its immunology pipeline, both of which don't carry CAR-T's inflammatory risks. The portfolio-level thesis is intact, even if the most ambitious piece just hit a wall.
But for the field as a whole, three deaths have changed the conversation. The promise of an immune reset, of telling autoimmune patients they might achieve remission without lifelong drugs, is still one of the most exciting ideas in medicine. Now it comes with an asterisk that won't be easy to erase.
The race to bring CAR-T to autoimmune disease was always going to produce a moment like this. Cancer therapies earn their brutal side effects through brutal diseases. When you point that same weapon at something less lethal, the margin for error doesn't shrink. It vanishes.
4 min read
The MyoKardia Gang Just Pulled Off a $400M Comeback
The team that built MyoKardia (acquired by Bristol-Myers Squibb for $13.1 billion) just took their new cardiac biotech public in a $400 million upsized IPO. Three late-stage drugs, $554 million in prior venture funding, and a stock that popped on day one: Kardigan is either biotech's next blockbuster or its most expensive sequel.
