Bristol Myers Squibb has confirmed a patient death in its Phase 1 Breakfree-2 clinical trial of zolacabtagene autoleucel (zola-cel), an investigational CD19-directed CAR T cell therapy for multiple sclerosis — marking the first confirmed fatality in the program and raising hard questions about attribution, risk-benefit thresholds in non-terminal populations, and a shared manufacturing feature that links BMS’s safety events to three deaths in a parallel Novartis trial.

The company made the disclosure on October 23, 2026, in a BMS October 23 press release issued alongside its presentation of interim Breakfree-2 results at MSToronto 2026, the 10th Joint ACTRIMS-ECTRIMS Meeting in Toronto. The event occurred after the trial’s data cutoff date, which explains why it had not been previously reported.

What the Data Monitoring Committee Found — and What It Cleared

The patient was enrolled in Breakfree-2’s highly active relapsing MS (haRMS) cohort. After receiving a single infusion of zola-cel, the patient developed late-onset motor symptoms that failed to respond to intermittent corticosteroid treatment and intravenous immunoglobulin (IVIG). Clinicians then administered high-dose cyclophosphamide — a chemotherapy agent — to address those symptoms. The patient developed septic pneumonia and died the following day, 44 days post-infusion.

BMS was explicit about the limits of what it can conclude. “The motor symptoms were a late-onset side effect, and the patient had a series of other complications and was heavily pretreated prior to CAR T, confounding our ability to determine the root cause of the issue or draw further conclusions,” the company said. The patient’s CAR T cell expansion was described as consistent with that seen in other trial participants — meaning the death cannot be attributed to atypical cell behavior.

Across all Phase 1 zola-cel programs spanning 190 autoimmune Phase 1 patients, this remains the only Grade 5 (fatal) event on record. The independent data monitoring committee reviewed the Grade 5 event in full and cleared the MS cohorts to proceed once the program resumes.

Breakfree-2: Trial Design and Efficacy Signals

NCT06220201 Breakfree-2 registration is a multicenter, single-arm, dose-escalation Phase 1 study evaluating zola-cel in patients with highly active relapsing or progressive MS who had failed prior disease-modifying therapies — a median of two prior treatments per patient. The trial enrolled 66 participants across two dosing cohorts: Dose Level 1 (DL1) at 5×10⁶ CAR-positive T cells, and Dose Level 2 (DL2) at 10×10⁶ CAR-positive T cells. Both cohorts include haRMS and progressive MS (PMS) patients. The primary endpoint is safety.

Despite the disclosed fatality, the interim efficacy data presented at MSToronto 2026 carried meaningful signals. According to the MSToronto 2026 NeurologyLive report, among the 32 evaluable DL2 patients with at least 26 weeks of follow-up, six of seven haRMS patients assessed for no evidence of disease activity (NEDA-3) met that composite threshold — meaning no relapses, no MRI progression, and no disability worsening. In the progressive MS (PMS) cohort, 22 of 24 evaluable patients showed stable or improved disability, with no new lesions detected on MRI. All participants demonstrated robust zola-cel expansion and complete peripheral B-cell depletion following infusion.

A mechanistically significant finding from week 12 of the study: 7 of 10 patients who had baseline oligoclonal bands showed a reduction in those bands, with investigators noting this indicates B-cell depletion not only in peripheral circulation but within the central nervous system itself — providing a molecular basis for the NEDA-3 responses in haRMS patients.

On safety, approximately 42.4% of participants — 28 of 66 — experienced cytokine release syndrome, a known complication of CAR T therapy. All CRS events were Grade 1 or Grade 2; none were life-threatening, and the median time to resolution was two days. One patient in the PMS cohort developed a Grade 4 ICANS resolved event within four days. BMS described the overall safety profile as consistent with previously reported data from the broader zola-cel autoimmune program.

60 DL2 patients off therapy remain off disease-modifying therapies as of the data cutoff.

The Manufacturing Question: What Rapid Production Might Explain

Zola-cel is manufactured using BMS NEX-T platform description, a proprietary next-generation process designed to produce early-differentiated T cells in a compressed timeframe. The platform uses machine learning-informed process optimization to reduce turnaround time and cost of goods.

The rapid manufacturing feature may be more than a logistical advantage. In a hypothesis first surfaced by analysts at William Blair, according to a William Blair analyst note following the Novartis program’s deaths, fast T-cell production — shared by zola-cel and Novartis’s rapcabtagene autoleucel (rap-cel), which uses the T-Charge platform — could produce T cells that retain a higher expansion potential in patients, driving more aggressive proliferation and, potentially, more severe inflammatory responses. “Rapid manufacturing could be driving increased cell expansion and the reported toxicities,” the analysts wrote.

That hypothesis has not been confirmed. But it carries a significant implication for the competitive landscape: programs that use conventional slower manufacturing — including Autolus Therapeutics, Kyverna Therapeutics, Cabaletta Bio, and CRISPR Therapeutics — may have a structurally different safety profile in autoimmune patients. A CGTLive competitor analysis noted the broader field is watching for validation of this hypothesis. If it is confirmed, it would become a deciding criterion for how regulators, physicians, and patients evaluate competing programs.

BMS has also provided a more specific mechanistic handle on at least some of the inflammatory events in its program. In disclosing the Breakfree-2 results, the company said it has identified a manufacturing reagent change identified as a contributor to ICANS events observed in its separate lupus trial. That finding is significant precisely because it points to a specific, correctable cause rather than a fundamental biological problem with the therapy.

The Broader Context: Novartis Deaths, Voluntary Pause, and Risk-Benefit in Non-Terminal Patients

The Breakfree-2 disclosure does not exist in isolation. On August 24, 2026, Novartis paused eight clinical trials of rap-cel after three Novartis deaths from IEC-HS — a severe hyperinflammatory cascade in which the body’s immune system attacks its own organs. Those trials covered lupus, rheumatoid arthritis, vasculitis, multiple sclerosis, and myasthenia gravis; Novartis’s oncology trials were not included.

Within days of the Novartis disclosures, BMS announced a BMS September voluntary pause across its own zola-cel autoimmune program on September 2, 2026, citing transient inflammatory events detected during routine surveillance. BMS was explicit that the pause was precautionary and not prompted by IEC-HS events in its own cohorts.

Alfred Kim, MD, PhD, director of the Lupus Center at Washington University in St. Louis, told Healio at the time that the Novartis pause was the only defensible decision. In a Healio interview with Alfred Kim, Kim framed the risk-benefit question directly: autoimmune patients are often younger, are not facing imminent death, and increasingly have other treatment options — which means the acceptable risk threshold for fatal complications is structurally lower than in oncology. He added that he does not view the deaths as the end of cellular therapy for autoimmune disease, citing patients who have entered durable drug-free remission, but emphasized the field must get the safety profile right.

That framing matters for how the Breakfree-2 fatality will be evaluated. A patient with relapsed or refractory leukemia has few remaining options; a patient with highly active MS — even one who has failed multiple disease-modifying therapies — is managing a chronic disease with available alternatives. Deaths that carry acceptable risk in terminal oncology generate sharper regulatory and ethical scrutiny in chronic disease settings.

Jeffrey Cohen, Lead Investigator, and the MSToronto 2026 Presentation

Breakfree-2 is led by Jeffrey Cohen Cleveland Clinic lead, Jeffrey A. Cohen, MD, a professor of neurology at the Cleveland Clinic Lerner College of Medicine and director of experimental therapeutics at the Mellen Center for Multiple Sclerosis Treatment and Research. His leadership of the zola-cel trial reflects both Cleveland Clinic’s central role in the autoimmune CAR T field and the seriousness with which the MS research community is treating this program.

The MSToronto 2026 presentations also included data from TG Therapeutics, which TG Therapeutics azer-cel presentation at the same meeting delivered Phase 1 results for azercabtagene zapreleucel (azer-cel), an allogeneic CD19-directed CAR T candidate for progressive MS. Allogeneic programs do not use rapid manufacturing from the patient’s own cells, which may give them a different safety profile — a distinction that will receive increasing scrutiny as the field’s early data accumulates.

Path to Restart: FDA Interaction and What Must Be Resolved

BMY fell 2% on disclosure on the day of the October 23 announcement, a muted reaction that likely reflects both the disclosed efficacy data and the uncertainty around attribution.

BMS is working with the FDA to resume all zola-cel studies. The company’s path forward requires completing the causal review of the patient death, demonstrating to the FDA that the manufacturing reagent change identified in the lupus trial has been corrected, and presenting the full Breakfree-2 safety data — including the DMC’s clearance of the MS cohorts — in its regulatory package. Per BMS FDA interaction 2027 restart, BMS is targeting a return to enrollment in early 2027.

The company faces a dual burden: demonstrating that 1 fatality in 190 autoimmune patients is not a pattern, while simultaneously closing the attribution gap on the specific death. Unresolved attribution — “we cannot determine the root cause” — is not the same as exculpation. Regulatory reviewers will want a causal narrative, even if that narrative concludes the death was not zola-cel-related.

Clinical Trial Vanguard analysis observed: “The temptation is to read the death as a BMS failure. Across the zola-cel autoimmune programs, 190 patients have now been treated in Phase I with no other Grade 5 event.” That framing captures the tension the field is navigating — the rate is low, but the missing attribution is the open variable.

For patients with highly active or progressive MS who have exhausted available options, the Breakfree-2 efficacy signals are significant. Six of seven haRMS patients meeting NEDA-3 without ongoing therapy is a meaningful early indicator. Whether those signals hold at longer follow-up, and whether BMS can close the attribution gap on the death and satisfy the FDA for restart, will determine whether zola-cel advances into pivotal trials — or whether the autoimmune CAR T field’s risk-benefit reckoning produces a longer pause than anyone in the program is currently anticipating.

Frequently Asked Questions

What caused the death of the patient in the BMS zola-cel MS trial?

The patient, enrolled in the haRMS cohort of Breakfree-2, died 44 days after receiving a zola-cel infusion. The immediate medical cause was septic pneumonia that developed after high-dose cyclophosphamide was administered to address late-onset motor symptoms that had not responded to steroids or IVIG. BMS has stated it cannot definitively attribute the death to zola-cel itself, citing confounding factors including the patient’s prior heavy treatment history and the clinical decisions made in response to the motor symptoms. The attribution remains unresolved as of the October 2026 disclosure.

What is the rapid manufacturing hypothesis, and why does it matter for the entire CAR T autoimmune field?

Both BMS’s zola-cel and Novartis’s rap-cel — the programs that have now experienced the most serious safety events in autoimmune CAR T — share a rapid manufacturing approach that produces T cells in a compressed timeframe. Analysts at William Blair have hypothesized that this faster production may yield T cells with higher expansion potential in patients, driving more aggressive proliferation and, in some cases, severe inflammatory cascades including IEC-HS. If the hypothesis is confirmed, it would create a structural safety distinction between rapid-manufacturing programs (zola-cel, rap-cel) and slower conventional programs (Kyverna, Cabaletta Bio, Autolus, CRISPR Therapeutics) — one that could reshape the competitive landscape and influence how regulators approach approval pathways for each type.

Is it safe to enroll in CAR T trials for autoimmune disease right now?

BMS’s zola-cel program is currently paused; enrollment in Breakfree-2 is not open as of publication. Novartis’s rap-cel autoimmune trials are also paused following three IEC-HS deaths. Patients and neurologists considering participation in any autoimmune CAR T trial should be aware that both of the most advanced CD19-directed programs in this space have experienced serious safety events and are under voluntary suspension pending regulatory review. CAR T trials in oncology — a distinct risk-benefit context — are not affected. Physicians with questions about specific trial eligibility and current enrollment status should contact the relevant trial sponsors directly or check ClinicalTrials.gov current status for updates.

When might BMS restart enrollment in Breakfree-2, and what has to happen first?

BMS is targeting an early 2027 restart, pending completion of its FDA interaction process. Before restarting, the company must present its full case review of the patient death to the FDA, demonstrate that the manufacturing reagent change identified as a contributor to ICANS events in its lupus trial has been corrected, and satisfy the FDA that the data monitoring committee’s clearance of the MS cohorts is supported by adequate safety data. The timeline assumes productive regulatory engagement and no additional safety findings during the pause period.