Chimeric antigen receptor T-cell therapy has delivered some of the most dramatic results in modern cancer medicine, producing durable remissions in patients with relapsed or refractory blood cancers who had exhausted every other option. Yet the very power that makes engineered T cells so effective against leukemia, lymphoma, and multiple myeloma also makes them dangerous. When billions of reprogrammed immune cells flood a tumor-laden body, they can ignite a runaway inflammatory reaction known as cytokine release syndrome, or CRS, the most prominent and potentially life-threatening toxicity of CAR T-cell treatment. A comprehensive review published in Clinical Cancer Bulletin by Yajing Zhang and Weidong Han synthesizes the current state of CRS management, from molecular mechanisms and grading systems to emerging drugs, artificial intelligence-based prediction tools, and next-generation CAR designs that aim to defuse the storm before it starts.
The biology of CRS begins the moment engineered T cells recognize their target. CAR T cells are autologous lymphocytes genetically modified to express synthetic receptors that bind tumor-associated antigens such as CD19 or BCMA. Upon antigen engagement, the cells unleash perforin and granzyme to kill malignant cells while simultaneously secreting a cocktail of pro-inflammatory signaling proteins, including interferon-gamma, tumor necrosis factor alpha, GM-CSF, and interleukin-2. These signals then recruit and activate innate immune players, chiefly monocytes and macrophages, which respond by pouring out interleukin-6 and interleukin-1 beta. This amplification loop between adaptive and innate immunity is what transforms a targeted attack into a systemic cytokine storm. Vascular endothelial cells become activated, blood vessels leak, and patients develop the classic clinical picture: fever, fatigue, hypotension, hypoxia, and in the worst cases multi-organ dysfunction. In severe instances, the syndrome can escalate into an immune effector cell-associated hemophagocytic lymphohistiocytosis-like syndrome and contribute to disruption of the blood-brain barrier, setting the stage for neurological complications.
How sick a patient becomes is not a matter of chance. Tumor burden at baseline, the specific CAR construct used, and the magnitude of T-cell expansion all shape the intensity of the inflammatory response. The intracellular costimulatory domain built into the receptor plays a particularly important role: constructs carrying the CD28 domain drive rapid T-cell expansion and are associated with a higher incidence of acute, severe adverse events, whereas 4-1BB-based constructs expand more slowly and persistently, producing a more gradual and generally less toxic immune response. Higher CAR T-cell doses enhance antitumor efficacy but also raise the risk of massive simultaneous T-cell activation. Patient factors matter as well. Elevated pre-infusion inflammatory markers such as interleukin-6 and C-reactive protein, advanced age, cardiovascular or renal comorbidities, certain lymphoma subtypes including primary mediastinal B-cell lymphoma, and pediatric or young adult patients treated for acute lymphoblastic leukemia all signal elevated risk. Bridging therapy to shrink the tumor before infusion is one practical strategy to lower the odds of explosive cytokine release.
Accurate grading is the backbone of treatment decisions. The consensus system from the American Society for Transplantation and Cellular Therapy classifies CRS into four grades based on fever, hypotension, hypoxia, and organ toxicity, deliberately moving beyond fever alone to incorporate the need for vasopressors and supplemental oxygen. Grade 1 disease, essentially fever by itself, is usually managed conservatively with antipyretics and observation. Grade 2, marked by hypotension or mild hypoxia, requires hospital monitoring and often pharmacologic intervention. Grade 3 and 4 cases, involving refractory low blood pressure, multisystem injury, shock, or respiratory failure, demand escalation to corticosteroids, vasopressors, and intensive care, with mechanical ventilation reserved for the critically ill. Because the boundary between CRS and sepsis can be blurred, clinicians must constantly weigh early immunosuppression against the danger of masking an underlying infection, a balancing act that defines much of the day-to-day management of these patients.
At the center of the pharmacologic arsenal sits tocilizumab, a monoclonal antibody that blocks the interleukin-6 receptor and remains the only drug approved by the FDA specifically for CRS. By interrupting the interleukin-6 signaling axis, it rapidly reverses fever, hypotension, and capillary leak without impairing the antitumor activity of the CAR T cells, and in most patients one or two doses suffice to control grade 2 or higher disease. Its speed has made it the emergency drug of choice, stocked in treatment centers with STAT ordering protocols. But tocilizumab is not without caveats. Blocking the receptor can cause circulating interleukin-6 levels to rebound, and there is concern that unbound cytokine may then cross the blood-brain barrier, potentially triggering or worsening immune effector cell-associated neurotoxicity syndrome, known as ICANS. This interplay between the two syndromes complicates therapy, since ICANS typically emerges days after CRS onset and, unlike CRS, does not respond to tocilizumab because the antibody penetrates the central nervous system poorly.
For patients who fail interleukin-6 blockade or who develop neurological symptoms, corticosteroids such as dexamethasone and methylprednisolone serve as the crucial second line. Historical worries that steroids might blunt CAR T-cell proliferation have been tempered by more recent evidence showing that early administration at moderate doses and limited duration does not significantly compromise antitumor efficacy. Typical regimens involve intravenous dexamethasone at 10 milligrams every six hours or methylprednisolone at 1 to 2 milligrams per kilogram per day, tapered over three to five days according to clinical response. Beyond steroids, a growing set of targeted agents is expanding the toolkit. Anakinra, an interleukin-1 receptor antagonist, crosses the blood-brain barrier readily and is increasingly valued for tocilizumab-refractory CRS and concurrent neurotoxicity. Lenzilumab, targeting GM-CSF, and emapalumab, targeting interferon-gamma, are under clinical evaluation for severe or steroid-refractory disease. In a striking example of drug repurposing, the beta-blocker metoprolol has been shown in laboratory and clinical studies to directly inhibit interleukin-6 translation in human monocytes, reducing CRS severity without harming CAR T-cell function and opening a novel therapeutic target: the protein synthesis machinery of inflammatory cells.
Monitoring is where modern CRS care increasingly meets data science. Routine protocols call for daily ferritin, C-reactive protein, and cytokine panels during the first ten days after infusion, or longer in high-risk patients, because rising interleukin-6, ferritin, and CRP levels precede clinical symptoms and correlate with severity. Soluble interleukin-2 receptor and coagulation markers flag the hemophagocytic syndrome variant and coagulopathy. Artificial intelligence is now pushing surveillance further. A model called PrCRS, built on U-Net and Transformer architectures with transfer learning, can predict severe CRS one to three days before symptom onset. Other multimodal machine learning systems analyze vital signs, laboratory results, and early cytokine profiles to identify high-risk patients within hours of infusion, and an explainable algorithm proposed by Bogatu and colleagues incorporates domain literature to detect CRS from cytokine peak levels despite limited training data. Real-time dashboards linked to electronic health records are already being piloted at academic centers, alerting care teams and enabling preemptive tocilizumab or steroid administration based on algorithmic triggers.
None of these tools work in isolation, and the review emphasizes that the hallmark of contemporary CRS management is multidisciplinary coordination. Optimal outcomes emerge when hematologists, oncologists, intensivists, neurologists, infectious disease specialists, pharmacists, and specialized nurses operate under unified protocols with rapid triage pathways and seamless escalation from ward to intensive care. Many centers have established cellular therapy toxicity boards, early warning systems, and dedicated CAR T-cell response teams equipped with predesigned protocols, immediate drug access, and intensive care beds on standby. Standardized operating procedures embed CRS and ICANS flowcharts directly into electronic medical records, while simulation training for nurses and residents sharpens recognition of early warning signs. Age also shapes strategy: pediatric patients generally tolerate higher cytokine loads, so clinicians lean on supportive care and reserve steroids for refractory cases, whereas elderly patients, vulnerable to cardiovascular and renal decompensation, warrant a lower threshold for early intervention, cardiac monitoring, and bridging therapy to reduce disease burden before infusion.
Looking ahead, the field is moving from reactive treatment to engineered prevention. Ruxolitinib, a Janus kinase 1/2 inhibitor, dampens downstream signaling from multiple cytokines and may help when interleukin-6 blockade alone is insufficient. Siltuximab, which binds interleukin-6 directly rather than its receptor, offers an alternative where receptor blockade falls short. More radically, next-generation CAR constructs are being designed with safety built in: bispecific and split-signaling receptors that require dual-antigen recognition to activate, ON/OFF-switch systems controlled by administered small molecules, suicide genes such as inducible caspase 9, and so-called armored CARs that co-express anti-cytokine payloads like interleukin-1 receptor antagonists to neutralize inflammation at its source. Researchers are also probing the upstream triggers of CRS, implicating inflammasome activation, pyroptotic cell death, and tissue-resident macrophages, and pursuing multi-omics biomarker discovery to shift management from reactive to proactive. Together, these advances promise to widen the therapeutic window of CAR T-cell therapy, allowing its remarkable curative potential to reach broader and more fragile patient populations safely.
Subject of Research: Management of cytokine release syndrome following CAR T-cell therapy
Article Title: Management of Cytokine Release Syndrome (CRS) following CAR T-cell therapy: a comprehensive review
Article References: Zhang, Y., & Han, W. (2025). Management of Cytokine Release Syndrome (CRS) following CAR T-cell therapy: a comprehensive review. Clinical Cancer Bulletin, 4(1), Article 15. https://doi.org/10.1007/s44272-025-00044-0
Image Credits: AI Generated
DOI: 10.1007/s44272-025-00044-0
Keywords: CAR T-cell therapy, cytokine release syndrome, tocilizumab, ICANS, immunotherapy toxicity, interleukin-6, corticosteroids, anakinra, biomarkers, artificial intelligence prediction, risk stratification, hematological malignancies
Cite Scienmag News
APA
MLA
Chicago
Nathaniel Bowman. (October 1, 2026). Taming the Cytokine Storm: How Doctors Are Making CAR T-Cell Therapy Safer. Scienmag. https://scienmag.com/taming-the-cytokine-storm-how-doctors-are-making-car-t-cell-therapy-safer/
Nathaniel Bowman. “Taming the Cytokine Storm: How Doctors Are Making CAR T-Cell Therapy Safer.” Scienmag, 1 October 2026, https://scienmag.com/taming-the-cytokine-storm-how-doctors-are-making-car-t-cell-therapy-safer/. Accessed 1 October 2026.
Nathaniel Bowman. “Taming the Cytokine Storm: How Doctors Are Making CAR T-Cell Therapy Safer.” Scienmag. October 1, 2026. https://scienmag.com/taming-the-cytokine-storm-how-doctors-are-making-car-t-cell-therapy-safer/
Copy citation
Download RIS
Tags: anakinraartificial intelligence in predicting CRSartificial intelligence predictionBiomarkersCAR-T Cell TherapyCAR-T-cell therapy safetycorticosteroidscytokine release syndromecytokine release syndrome managementemerging drugs for CRS mitigationengineered T cells in cancer treatmentgrading systems for CRS severityhematological malignanciesICANSimmunotherapy toxicityinflammatory responses in immunotherapyinnovations in CAR T-cell safety protocolsinterleukin-6molecular mechanisms of CRSnext-generation CAR T-cell designsrisk stratificationrisks of cytokine storm in cancer treatmentstocilizumabtumor-associated antigens in CAR therapy
