BACKGROUND: Anti GM-CSF autoantibodies (aAb) have been related to acquired pulmonary alveolar proteinosis (PAP) and described in cases of severe infections such as cryptococcosis and nocardiosis in previously healthy subjects. Whether there are different anti-GM-CSF autoantibodies corresponding to these phenotypes is unclear. Therefore, we examined anti-GM-CSF autoantibodies to determine whether amount or neutralizing activity could distinguish between groups. METHODS: Plasma samples gathered in the National Institute of Health from patients with anti GM-CSF aAb and either PAP (n = 15), cryptococcal meningitis (n = 15), severe nocardiosis (n = 5) or overlapping phenotypes (n = 6) were compared. The relative amount of aAb was assessed using a particle-based approach, reported as a mouse monoclonal anti-human GM-CSF as standard curve and expressed in an arbitrary Mouse Monoclonal Antibody Unit (MMAU). The neutralizing activity of the plasma was assessed by inhibition of GM-CSF-induced intracellular phospho-STAT5 (pSTAT5) in monocytes. RESULTS: Anti-GM-CSF aAb relative amounts were higher in PAP patients compared to those with cryptococcosis (mean 495 ± 464 MMAU vs 197 ± 159 MMAU, p = 0.02); there was no difference with patients with nocardiosis (430 ± 493 MMAU) nor between the two types of infections. The dilution of plasma resulting in 50% inhibition of GM-CSF-induced pSTAT5 (approximate IC50) did not vary appreciably across groups of patients (1.6 ± 3.1%, 3.9 ± 6% and 1.8 ± 2.2% in PAP patients, cryptococcosis and nocardiosis patients, respectively). Nor was the concentration of GM-CSF necessary to induce 50% of maximal GM-CSF-induced pSTAT5 in the presence of 10 MMAU of anti-GM-CSF aAb (EC50). When studying longitudinal samples from patients with PAP or disseminated nocardiosis, the neutralizing effect of anti-GM-CSF aAb was relatively constant over time despite targeted treatments and variations in aAb levels. CONCLUSIONS: Despite different clinical manifestations, anti-GM-CSF antibodies were similar across PAP, cryptococcosis and nocardiosis. Underlying host genetics and functional analyses may help further differentiate the biology of these conditions.