- Design
- Retrospective multicentre cohort
- Population
- 108 patients with autoimmune disease receiving autologous CAR-T (Germany CD19; China CD19-BCMA)
- Primary outcome
- Occurrence and severity of CRS and ICANS
- Effect
- CRS 80% (grade ≥2 in 14%); ICANS 3%; B cells OR 1.31 (1.06–1.67), IL-6 OR 2.28 (1.30–4.65)
CAR-T cell therapy is moving from haematology into refractory lupus, systemic sclerosis, myositis and vasculitis. This retrospective cohort pooled 108 patients with autoimmune disease treated with autologous CAR-T cells at six German centres (CD19) and one Chinese centre (CD19-BCMA), between 2021 and 2025.
Cytokine release syndrome occurred in 80%: grade 1 in 66%, grade 2 in 13% and grade 3 in one patient. Neurotoxicity (ICANS) occurred in 3%, all grade 1 or 2. Patients who developed CRS had higher baseline B-cell counts (median 43 vs 6 cells/μL), IL-6 (10 vs 3 pg/mL) and CRP. On multivariable analysis, B-cell count (OR 1.31, 95% CI 1.06 to 1.67) and IL-6 (OR 2.28, 1.30 to 4.65) remained associated.
The message is reassuring and practical: CRS in autoimmune patients is common but overwhelmingly mild, and patients with high B-cell burden and systemic inflammation at baseline were at higher risk, though with no validated cut-off every patient still needs standard monitoring. CAR-T for autoimmune disease is available in India only in trials or a few centres, but referral questions are beginning to arrive.
- Expect cytokine release syndrome in most autoimmune patients after CAR-T, usually grade 1 with fever.
- Measure baseline B-cell count, IL-6 and CRP before CAR-T; higher values predict CRS.
- High B-cell counts or raised IL-6 may flag higher CRS risk, but no validated cut-off exists, so every patient still needs standard CRS monitoring.
- Neurotoxicity was rare (3%) and mild in this autoimmune cohort.
- Refer candidates for CAR-T only to centres running structured programmes or trials.
Why it matters
As CAR-T enters rheumatology, simple baseline tests may identify who needs closest monitoring.
Don't overread it
Retrospective, mixed products and centres; the biomarkers are associations, not validated thresholds.
The statistics, in plain English
An odds ratio of 2.28 for IL-6 means higher IL-6 was associated with roughly double the odds of CRS per unit on the scale the authors used; without the unit and a threshold, it indicates direction and strength rather than a cut-off to act on.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for rheumatology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free