BACKGROUND: Retinal angiography has become a widely used diagnostic tool. It requires the intravenous administration of the fluorescent dyes fluorescein and indocyanin green. We recently received blood taken 8 h after retinal angiography, without our knowing it. We describe the failure of an automated flow cytometry system in the enumeration of lymphocyte subpopulations in this sample. METHODS: Cell enumeration was achieved by the use of the lyse-no wash MultiTEST procedure (Becton-Dickinson) together with the FACSCalibur cytometer. Absolute cell counts were obtained using TruCount beads. Data were analyzed automatically by the MultiSET and manually with the CellQuest softwares. RESULTS: The dot plots obtained with this sample looked quite abnormal. All monuclear cells stained brightly in the FITC channel irrespective of anti-CD3-FITC conjugate binding. This resulted in a major undercompensation for the increased spillover of the fluorescein emission into the PE-channel. PE-labeled cell and TruCount bead events coalesced. The MultiSET software failed to draw proper gatings and proved useless. Alternative manual gatings could partially rescue the analysis. CONCLUSIONS: Clinicians and cytometrists should be aware that, because of dye entry or binding, blood mononuclear cells collected shortly after retinal angiography are not suitable even for common cytometry applications.