A long feature in a Chinese weekend newspaper in May 2026 laid out the case for CAR-T cell therapy in autoimmune disease. It was unusually careful for a general-interest piece: it named the trials, counted the patients and did not call anything a cure. Since then, more data have appeared. What follows is a map of the four clinical readouts that matter most to a patient considering China, together with one regulatory fact that has not moved at all.
That last point is the one most often lost in translation. Autoimmune disease is the fastest-moving area of CAR-T research in the world, and Chinese centres are now publishing in the New England Journal of Medicine, Nature Medicine and Med. But no CAR-T product is approved in China for any autoimmune indication. Access is through registered clinical trials and, at a small number of centres, physician-initiated programmes. A clinic advertising approved CAR-T for lupus in China is making a claim that needs checking.
Why a cancer therapy works on autoimmunity
Autoimmune disease begins when the immune system loses tolerance and attacks the body's own tissue. There are more than a hundred such conditions. Taken together they affect roughly 5 to 10 percent of the population in many surveys, occur two to three times as often in women as in men, and are often described as the third largest category of chronic illness after cancer and cardiovascular disease.
B cells sit near the centre of many of them. Their normal job is to make antibodies against foreign threats. When they malfunction, they produce autoantibodies, present the body's own antigens to T cells and secrete inflammatory signals that keep the attack going. Most conventional treatment suppresses this activity broadly with steroids and immunosuppressants. It helps many patients, but a subgroup remains refractory, and the drugs generally have to be continued indefinitely.
CAR-T offers a different proposition. A patient's T cells are engineered to recognise a surface marker such as CD19, then returned to the body to remove the B cells carrying it. The observation that has excited researchers is what happens next: when B cells eventually return, they appear to come back naive, without the autoreactive memory that drove the disease. That is the immune reset hypothesis, and it is why the phrase drug-free remission appears so often in this literature. It is a hypothesis with growing support, not an established mechanism.
The short version
The target: B-cell markers, most often CD19, because B cells produce the autoantibodies and sustain the immune memory.
The idea: deep depletion followed by repopulation with naive B cells, so the disease-driving memory is not restored.
The evidence: a German phase 1/2a trial in 24 patients, a Chinese phase 1 trial in 11 patients with autoimmune haemolytic anaemia, and smaller Chinese lupus series.
The rule: no CAR-T product is approved in China for autoimmune disease. Access means a trial or a physician-initiated programme.
Where the story started
In August 2021 a team at Friedrich-Alexander-Universität Erlangen-Nürnberg led by Georg Schett reported in the New England Journal of Medicine the first use of CAR-T in a patient with autoimmune disease. The patient was a 20-year-old woman with severe refractory systemic lupus erythematosus who had failed seven conventional treatments, including glucocorticoids, immunosuppressants and monoclonal antibodies. After an infusion of autologous CD19 CAR-T cells, her B cells were cleared, her lupus symptoms improved rapidly, and no cytokine release syndrome, neurotoxicity or prolonged cytopenia was observed.
A single case proves very little. But it established that the approach could be attempted outside cancer, and it set off the programme of trials that produced the 2026 readouts.
One patient, three diseases, one infusion
The most striking published case came in April 2026 in Med, from the same Erlangen group. A 47-year-old woman had been diagnosed in 2014 with severe refractory autoimmune haemolytic anaemia, in which the immune system destroys its own red cells, leaving her dependent on regular transfusions. Nine conventional approaches had failed. She was subsequently diagnosed with antiphospholipid syndrome in 2015 and immune thrombocytopenia in 2019. Three separate autoimmune diseases, all traced to malfunctioning B cells.
After a life-threatening relapse, her physicians used CD19 CAR-T on a compassionate basis. The reported sequence is unusually detailed. CAR-T cell levels peaked on day 9 while circulating B cells disappeared. The patient stopped needing transfusions on day 7. Haemoglobin began rising from day 9 and reached the normal range after day 25. Platelets recovered slowly across eleven months to about half of normal. Antiphospholipid antibodies became undetectable, no venous thrombosis was observed, and the patient was discharged on day 10. CAR-T cells became undetectable by day 322, at which point new B cells reappeared. No cytokine release syndrome occurred.
This was reported as the first time a single CAR-T infusion produced remission of three B-cell-driven autoimmune diseases without further transfusion or treatment. It is also a single compassionate-use case, in one patient, at one centre, with no control. Its value is as a demonstration of what immune reset might look like, not as an estimate of how often it works.
The four readouts that changed the conversation
By 2026 the field had moved from case reports to cohorts. The table below collects the four results most often cited, including the two from Chinese institutions that matter most to international patients considering treatment here.
| Group and journal | Patients and diseases | Reported result |
|---|---|---|
| Erlangen, Nature Medicine, January 2026 | 24 patients: 10 lupus, 9 systemic sclerosis, 5 idiopathic inflammatory myositis; phase 1/2a | 22 of 24 met the six-month endpoint; 90% of lupus patients improved, all sclerosis patients had no further progression, 80% of myositis patients had moderate or major improvement |
| Chinese Academy of Medical Sciences, NEJM, January 2026 | 11 patients with relapsed or refractory autoimmune haemolytic anaemia; 5 compassionate use, 6 in a phase 1 trial | All 11 in complete remission at three months, with normalised blood counts and no immunosuppressants; median drug-free remission 11.5 months, longest 21 months |
| Peking Union Medical College Hospital, Med, April 2026 | 6 patients with lupus complicated by immune thrombocytopenia | All 6 improved, 3 reached complete remission and stopped immunosuppressants; median follow-up 12 months; no severe toxicity |
| Six German centres and Hangzhou, Lancet Rheumatology, September 2026 | 108 patients with refractory autoimmune disease, treated between March 2021 and July 2025 | Largest safety cohort to date; cytokine release syndrome in 80%, but two thirds of the whole cohort had grade 1 only, with one grade 3 case |
The Erlangen cohort is the most informative on durability. At a mean follow-up of 13 months, with the longest beyond 23 months, none of the 24 patients had relapsed or progressed, and none remained on other immunosuppressants or glucocorticoids. The Chinese Academy of Medical Sciences result is the most informative on a disease that is rarely discussed in the autoimmune CAR-T conversation: refractory autoimmune haemolytic anaemia, where all eleven patients had failed at least three prior treatments.
Safety deserves the same care as efficacy. In the Chinese phase 1 trial, one patient developed grade 3 immune effector cell-associated haematotoxicity, and no other severe toxicity was reported. That is a small number with limited follow-up, not a clean bill of health. The 108-patient safety cohort gives a fuller distribution of short-term events, and longer-term risks such as prolonged cytopenia, hypogammaglobulinaemia and infection remain under study.
The rule that has not changed
China has approved nine CAR-T products, all of them for blood cancers or, since June 2026, one solid tumour. Not one carries an autoimmune indication. The Chinese trials above were conducted under research protocols or compassionate-use provisions, and the same is true of physician-initiated programmes at a small number of centres. Enrolling in a trial is not the same as buying an approved therapy, and trial places are limited by eligibility criteria that most applicants will not meet.
This distinction has practical consequences for overseas patients. A trial may not accept international participants, may require long residence in China for monitoring, and will not come with a commercial price quotation. A programme that offers CAR-T for lupus as a paid, routinely available service with a fixed price is describing something different from the published evidence, and that difference should be clarified in writing before any commitment.
A published complete remission in a Chinese hospital is a research result. Until a regulator approves an autoimmune indication, the honest description of access in China is clinical trial or physician-initiated programme, never approved therapy.
What to ask before you travel
- Is the proposed treatment part of a registered clinical trial, and if so what is the registration number and who is eligible?
- Does the centre accept international participants, and what duration of stay and follow-up monitoring does the protocol require?
- If it is not a trial, under what regulatory pathway is the treatment being offered, and is that documented?
- Has the centre published its own outcomes, or is it quoting results from another institution or another product?
- Which CAR construct is used, and what is the plan for cytokine release syndrome, neurological toxicity and prolonged cytopenia?
- What does the written quotation include, and what happens if manufacturing fails or the patient becomes ineligible after apheresis?
- Has your own specialist at home reviewed the proposal and the alternatives to it?
What this means if you are considering treatment in China
The science here is worth taking seriously. Chinese groups are contributing original trial data in a field that barely existed five years ago, and drug-free remission in refractory autoimmune disease is a meaningful result. That is exactly why the boundary matters: enthusiasm is what makes an unapproved indication easy to oversell.
Our role is to help you establish which category a proposal falls into before money or travel is involved. We can help you obtain the trial registration, clarify whether international patients are accepted, and confirm what a centre's written quotation covers. We do not recommend treatment, and we do not make claims about outcomes.
The indications page sets out where CAR-T is approved today and where it remains experimental.
This article is educational and is not medical advice. Whether any therapy is appropriate can only be decided by a treating specialist with access to the full medical record.