From Aortitis to Sweet's: The Immune Spectrum of G-CSF Adverse Events

Granulocyte colony-stimulating factor (G-CSF) and its PEGylated formulation (pegfilgrastim) are indispensable for preventing chemotherapy-induced neutropenia and for stem-cell mobilization. Beyond hematopoiesis, G-CSF modulates innate/adaptive immunity and may precipitate immune-mediated events in susceptible hosts. Clinical signals span large-vessel vasculitis/aortitis and non-vascular neutrophil-dominant phenotypes, yet timing, spectrum, and management remain variably characterized.

To identify and synthesize primary clinical reports of autoimmune/rheumatic disease triggered or exacerbated by exogenous G-CSF (filgrastim or pegfilgrastim), describing phenotypes, latency, management, outcomes, and implications for practice.

We performed a narrative review of primary clinical studies, including observational cohorts and case reports or short case series, identified through targeted searches of PubMed/MEDLINE, Scopus, and Embase, complemented by investigator-curated references. Eligible reports described autoimmune, inflammatory, or rheumatic manifestations occurring after exposure to exogenous G-CSF. Predefined phenotypes included large-vessel vasculitis/aortitis; neutrophilic dermatoses (e.g., Sweet's syndrome, neutrophilic dermatosis of the dorsal hands); cutaneous small-vessel vasculitis; inflammatory or crystal arthritis (e.g., calcium pyrophosphate deposition disease); pulmonary inflammatory injury (e.g., diffuse alveolar hemorrhage); and clear flares of established autoimmune disease. Narrative or mechanistic reviews and trials of GM-CSF pathway blockade were excluded from qualitative synthesis.

Twenty-four studies met inclusion criteria: two observational cohorts and twenty-two single-patient reports/short series. Cohorts quantified a low but reproducible burden of pegfilgrastim-associated aortitis, with stereotyped involvement of the aortic arch/proximal branches and typical latency of 7-15 days. Case-level data confirmed marked elevation of acute phase reactants, usually negative autoantibodies, favorable outcomes after G-CSF withdrawal with or without short glucocorticoid courses, and recurrence/migration on re-exposure. Non-vascular events clustered earlier (2-7 days) and included biopsy-proven Sweet's/NDDH, leukocytoclastic vasculitis, CPPD flares, and rare diffuse alveolar hemorrhage; granulomatous dermatitis and perioperative pyoderma gangrenosum were also observed. Re-exposure information suggested phenotype-specific risk: recurrent CPPD with pegfilgrastim was mitigated by switching to short-acting filgrastim; selected limited aortitis resolved without steroids; refractory aortitis responded to IL-6 blockade, enabling uninterrupted chemotherapy.

Exogenous G-CSF can precipitate a coherent spectrum of immune-mediated toxicity with distinct, clinically actionable timing windows: vascular events peak 7-15 days after pegfilgrastim, whereas cutaneous, articular, and pulmonary manifestations arise within 2-7 days. Recognition hinges on temporal linkage, imaging/histology, and exclusion of mimics. Management is phenotype-aware-drug withdrawal ± short glucocorticoids, formulation switching for CPPD, and targeted IL-6 blockade in selected refractory aortitis-allowing preservation of hematologic benefits while minimizing immune toxicity. Prospective surveillance and structured re-exposure studies (including formulation tailoring) are priorities to move from signal detection to prevention.