A preliminary but important SJS/TEN signal
A new 2026 pharmacovigilance study raises an unusual question for Stevens-Johnson syndrome and toxic epidermal necrolysis cases involving cancer patients: can immune checkpoint inhibitor therapy change the apparent toxicity of another medication? The study, currently posted as a medRxiv preprint and not yet peer reviewed, analyzed more than 13.7 million deduplicated reports submitted to the FDA Adverse Event Reporting System (FAERS) from 2016 through 2025. The investigators examined whether exposure to immune checkpoint inhibitors - drugs that alter immune signaling through pathways such as PD-1, PD-L1, and CTLA-4 - changed reported associations between other medications and serious adverse events.
The most striking SJS/TEN finding involved the fluoroquinolone antibiotic moxifloxacin. In the latest preprint version, the investigators reported a moxifloxacin-by-immune-checkpoint-inhibitor interaction odds ratio of 100.51, with a 95% confidence interval of 38.58 to 261.84. Only nine SJS/TEN reports involved joint exposure to moxifloxacin and an immune checkpoint inhibitor, so the size of the estimate should not be mistaken for a measured incidence rate. The finding is a signal that warrants further study, not proof that the combination causes SJS/TEN in an individual patient.
Why the interaction concept is different from a conventional drug-safety signal
Most medication-causation analysis asks whether a particular drug is known to cause SJS/TEN and whether the patient's timing and clinical course fit that association. The new study asks an additional question: whether the immune environment created by checkpoint blockade may modify the relationship between a second drug and a severe reaction.
That concept is biologically plausible. Immune checkpoint inhibitors are designed to remove inhibitory signals from T cells so the immune system can attack cancer more effectively. The authors note experimental and clinical observations suggesting that PD-1/PD-L1 blockade can lower the threshold for drug-specific T-cell responses and that hypersensitivity to previously tolerated medications can emerge after checkpoint therapy. If that mechanism is confirmed, culprit-drug analysis in oncology patients could become more complex than simply selecting one medication from a list.
What the study does not establish
The limitations are substantial. FAERS is a spontaneous-reporting database. Reports can be incomplete, duplicated despite deduplication efforts, influenced by publicity, and affected by the fact that seriously ill oncology patients commonly receive many drugs at the same time. FAERS does not provide a reliable denominator for calculating incidence, and an interaction odds ratio is not the same thing as a patient's probability of developing SJS/TEN.
The study is also a preprint. It has not yet gone through journal peer review. The nine jointly exposed moxifloxacin-SJS/TEN reports are a small absolute number, and the analysis cannot eliminate confounding from infection, cancer severity, other medications, or the checkpoint inhibitor itself. For those reasons, the finding should be described as an emerging pharmacovigilance signal rather than a proven drug interaction.
Why moxifloxacin already presents a difficult legal issue
Current U.S. moxifloxacin labeling already identifies Stevens-Johnson syndrome and toxic epidermal necrolysis among severe and sometimes fatal reactions. That means this preprint does not create a simple new failure-to-warn theory. Moxifloxacin is also widely available as a generic drug, making exact product identification important because federal preemption can sharply limit traditional warning claims against generic manufacturers.
For a real SJS/TEN case, the analysis would need to identify the exact moxifloxacin product and manufacturer, the immune checkpoint inhibitor and its manufacturer, the timing of both exposures, whether the patient had previously tolerated moxifloxacin or related fluoroquinolones, all competing medications, the cancer indication, the latency to the first symptoms, and the clinical reasoning used to identify the culprit drug. Depending on those facts, the scientific interaction question may be more important to causation than to a standalone product-liability theory.
What this means for SJS/TEN case screening now
The practical takeaway is not that patients should avoid moxifloxacin or immune checkpoint therapy without medical direction. Instead, oncology-related SJS/TEN intake should capture both categories of exposure. If a patient developed SJS/TEN after moxifloxacin, it is now worth asking specifically whether the patient was receiving or had recently received an immune checkpoint inhibitor such as a PD-1, PD-L1, or CTLA-4 agent.
Pharmacy records, infusion records, oncology treatment calendars, hospital medication-administration records, and the exact onset chronology may be particularly important. A history of prior tolerance to the antibiotic could also become relevant to the proposed immune-modification hypothesis, although it would not by itself prove causation.
What would strengthen this signal?
This signal would become substantially more persuasive if the work is peer reviewed; if independent datasets reproduce the moxifloxacin-specific interaction; if mechanistic studies demonstrate checkpoint-dependent amplification of moxifloxacin-specific T-cell responses; if regulators or manufacturers recognize the interaction; or if well-documented clinical case series show a consistent exposure pattern. Until then, the study is best viewed as a credible early warning for researchers and case evaluators rather than evidence of a new established tort.
Sources
Immune checkpoint blockade reshapes drug-associated toxicity: a pharmacovigilance atlas of drug-ICI interactions, medRxiv v3 DOI: 10.64898/2026.08.31.26361880. Direct source
DailyMed - Moxifloxacin tablet labeling Direct source
Earlier ICI/SJS-TEN FAERS analysis, medRxiv DOI: 10.1101/2025.06.21.25330030. Direct source