Blood Transfusions May Predict Deadly Complication of Leukemia Treatment, Study Finds

Acute promyelocytic leukemia (APL) was once among the most feared diagnoses in hematology, with patients routinely dying within days of diagnosis from catastrophic bleeding. Today, thanks to two remarkable drugs—all-trans retinoic acid (ATRA) and arsenic trioxide (ATO)—the disease is curable for the vast majority of those affected. Yet a shadow still hangs over this success story: differentiation syndrome, a potentially life-threatening complication that strikes during the very treatment that saves lives. Now, a new study published in the Annals of Hematology has brought researchers closer than ever to understanding what drives this syndrome, and it points to a surprisingly simple warning sign that clinicians can already track at the bedside: the number of red blood cell transfusions a patient receives.

Differentiation syndrome occurs when ATRA and ATO, which work by forcing the immature leukemic cells of APL to mature and eventually die, trigger a violent inflammatory reaction in the process. Patients can develop fever, fluid accumulation in the lungs, low blood pressure, and kidney failure. In severe cases, the syndrome can be fatal. Despite decades of clinical experience, the biological mechanisms underlying the condition have remained poorly understood, and doctors currently have no reliable way to predict which patients will develop it. The new research, led by Giulia Falconi, Luca Guarnera, and Maria Teresa Voso of the University of Rome Tor Vergata, together with colleagues across Italy, set out to change that by examining the clinical, molecular, and cellular fingerprints of the syndrome.

The team analyzed two independent groups of APL patients. The first was a real-life cohort of 34 patients treated at multiple Italian centers, reflecting the diversity of everyday clinical practice. The second comprised 18 patients enrolled in the landmark APL0406 clinical trial, the study that established the ATRA-ATO combination as the standard of care. From each patient, the researchers collected blood samples at baseline and at sequential time points during treatment, allowing them to track molecular changes as therapy unfolded. Crucially, the analysis included both patients who developed differentiation syndrome and those who did not, providing a natural comparison group.

The most striking clinical finding concerned transfusions. Patients who went on to develop differentiation syndrome required a median of nine units of red blood cells during the induction phase of treatment, compared with just three units in patients who did not develop the syndrome—a difference that was highly statistically significant. Even after adjusting for other clinical variables in a multivariable analysis, the transfusion burden retained its independent association with the syndrome. In other words, patients whose disease demanded more blood support early in treatment were markedly more likely to experience the complication later on.

Why would transfusion requirements track so closely with differentiation syndrome? The answer, the researchers suggest, lies in the biology of erythroid maturation—the process by which red blood cell precursors develop into mature cells. When the team performed RNA sequencing on the patients’ blood samples, they identified 93 genes that were differentially expressed in patients who developed the syndrome. When these gene signatures were mapped onto known biological pathways, they showed striking enrichment for processes related to hematopoietic stem cell differentiation and erythroid maturation. This suggests that the same differentiation programs that ATRA and ATO deliberately activate in leukemic cells may also be perturbing normal blood cell development, particularly the red cell lineage, in susceptible patients.

The molecular data were complemented by measurements of inflammatory signaling molecules. Cytokine profiling of the blood samples revealed elevated levels of interleukin-6, a potent inflammatory messenger, with a trend toward even higher concentrations at the moment the syndrome declared itself. Interleukin-6 is a well-known driver of systemic inflammation and has been implicated in cytokine release syndromes seen with other cancer therapies, including CAR-T cell treatment. Its presence here reinforces the picture of differentiation syndrome as an inflammatory storm, one that unfolds in a molecular environment already primed by disordered blood cell maturation.

But the study did not stop at the leukemic cells themselves. One of its most provocative contributions concerns the endothelium—the delicate layer of cells that lines blood vessels and controls what passes between the bloodstream and surrounding tissues. The researchers grew human pulmonary microvascular endothelial cells, or HPMECs, in the laboratory and exposed them to ATRA and ATO. The drugs induced marked morphological changes in these cells and significantly increased their permeability, allowing fluid and molecules to leak across the endothelial barrier. This laboratory phenomenon closely mirrors the capillary leak that characterizes differentiation syndrome in patients, in which fluid escapes from blood vessels into the lungs and other organs, causing potentially deadly swelling and respiratory distress.

Encouragingly, the researchers found a way to blunt this endothelial damage. When dexamethasone, a corticosteroid commonly used to prevent and treat differentiation syndrome, was added to the cell cultures alongside ATRA and ATO, the increased permeability was significantly attenuated. The finding provides a mechanistic explanation for a clinical practice that has existed for years: steroids work, at least in part, by stabilizing the endothelial barrier against the leak-inducing effects of the leukemia drugs. It also suggests that the laboratory model could be used to test other protective agents in the future.

Taken together, the results support a new model of differentiation syndrome as a three-part process. First, erythroid dysregulation—the same maturation programs triggered in leukemic cells—disrupts normal red cell development, driving the anemia that necessitates transfusions. Second, ATRA and ATO directly injure the endothelial lining of blood vessels, creating a capillary leak phenotype. Third, an inflammatory environment rich in interleukin-6 amplifies the damage. Each element feeds the others, and the transfusion burden emerges as a visible downstream marker of the earliest, hidden phase of the cascade.

The clinical implications could be significant. Red blood cell transfusion counts are recorded routinely for every APL patient, making them an early, inexpensive, and universally accessible biomarker. Patients requiring frequent transfusions during induction could be flagged for closer monitoring, earlier steroid intervention, or enrollment in prospective studies of preventive strategies. The researchers caution that their findings, while compelling, come from a combined cohort of 52 patients, and validation in larger, independent populations will be essential before transfusion burden is adopted into formal risk stratification. Still, in a complication that has long defied prediction, the idea that a number already written in every patient’s chart might reveal who is at highest risk is a rare and welcome piece of good news—one that could help make the cure for acute promyelocytic leukemia not just effective, but safer.

Subject of Research: Differentiation syndrome pathogenesis in acute promyelocytic leukemia

Article Title: Transcriptomic profile and endothelial changes during differentiation syndrome in acute promyelocytic leukemia

Article References: Falconi, G., Guarnera, L., Lazzaroni, F., Galossi, E., Ottone, T., Attardi, E., Piazza, R., Lumia, E., Silvestrini, G., Travaglini, S., Divona, M., Fabiani, E., Curti, A., Martini, V., Mazzone, C., La Barbera, E. O., Sica, S., Mannelli, F., Gurnari, C., … Voso, M. T. (2026). Transcriptomic profile and endothelial changes during differentiation syndrome in acute promyelocytic leukemia. Annals of Hematology. https://doi.org/10.1007/s00277-026-07180-0

Image Credits: AI Generated

DOI: 10.1007/s00277-026-07180-0

Keywords: acute promyelocytic leukemia, differentiation syndrome, all-trans retinoic acid, arsenic trioxide, transcriptomics, endothelial permeability, interleukin-6, red blood cell transfusion, erythroid maturation, dexamethasone, biomarker, hematology