What safety concerns led Novartis and Bristol Myers Squibb to temporarily halt or pause enrollment in multiple autoimmune-disease trials of their autologous CD19 CAR-T therapies—Novartis’s rapcabtagene autoleucel (rap-cel), after three cases of potentially fatal immune effector c

The pauses reflect a serious but still incompletely explained immune-toxicity signal in autoimmune CAR-T—not proof that rapid manufacturing caused it. Novartis identified three IEC-HS cases with rap-cel and stopped autoimmune development/enrollment while continuing its oncology program; BMS paused zola-cel enrollment after new transient inflammatory events were found in routine review, against the background of one earlier IEC-HS case. 121314

• IEC-HS: Immune effector cell-associated hemophagocytic syndrome is a delayed, dysregulated hyperinflammatory syndrome related to macrophage activation/HLH after cellular therapy. It can involve persistent fever, striking ferritin elevation, cytopenias, liver dysfunction, coagulopathy, and potentially multiorgan failure. It is considered potentially fatal; however, a reliable mortality percentage for these small autoimmune CAR-T datasets has not been publicly established. The key issue is severity, not a demonstrated death rate in either current program.

• Novartis / rap-cel: The company’s autoimmune pause followed three IEC-HS cases across its broad rheumatology/neurology program, including lupus, systemic sclerosis, myasthenia gravis and multiple sclerosis. Oncology studies were not included in the halt. 1213 Novartis is reviewing the individual cases and the pooled safety, pharmacokinetic and cell-expansion data to determine risk factors and mitigation measures before deciding whether—and on what protocol terms—to restart enrollment. Public reporting does not establish a restart date. 1214

• BMS / zola-cel: BMS described the newly observed events as inflammatory, transient and reversible, but paused new enrollment voluntarily while reviewing the full data set. 14 The earlier IEC-HS event raises the importance of determining whether the new events represent a related biological spectrum or a separate, manageable inflammatory phenomenon; that causal link has not been demonstrated publicly.

• Why rapid manufacturing is a plausible hypothesis, not a conclusion: Novartis’s T-Charge process is designed to avoid prolonged ex-vivo culture and preserve less-differentiated, less-exhausted T cells. 13 BMS’s NEX T platform has a similar strategic aim: faster manufacture and a fitter, more proliferative product. Such cells may expand more vigorously after infusion; greater expansion/persistence can improve B-cell depletion and efficacy but can also plausibly amplify cytokine-driven inflammation. No company has publicly shown that T-Charge or NEX T caused the IEC-HS cases, so the concern remains mechanistic and under review rather than proven.

• Competitors: Cabaletta’s rese-cel has, in its reported autoimmune experience, shown mostly absent or low-grade CRS and very little ICANS, including data supporting outpatient administration; the company reported no dose-limiting toxicities or ICANS in an early preconditioning-free cohort, with one transient grade 1 CRS case. 911 Miltenyi’s zorpo-cel program is supported by the published CASTLE basket study, which reported generally acceptable safety with one dose-limiting toxicity in its first phase. 5 Neither public data set has produced an analogous broad enrollment pause for IEC-HS. That is encouraging but not a head-to-head safety advantage: populations, doses, conditioning, follow-up and manufacturing differ, and rare toxicities often emerge only with broader exposure. 5

• Cellares/Breyanzi issue: This was separate from zola-cel’s clinical safety review. BMS ended its $380 million Cellares agreement—originally intended to expand manufacture of the approved CAR-T Breyanzi in the U.S., EU and Japan—after concluding that Cellares’ automated Cell Shuttle manufacturing platform could not meet BMS’s requirements for Breyanzi production. 78 It was a manufacturing-platform/scale-up fit problem, not an announced patient-safety finding involving Breyanzi.