Objectives: To describe, in a cohort of dogs with presumed primary immune-mediated neutropenia, the presenting clinical characteristics, haematology results, bone marrow characteristics, therapies used (drugs and doses), clinical response to treatment, relapse and outcome at six months and one year. Methods: Multi-institutional recruited retrospective descriptive case series with voluntary submissions. Presumed immune-mediated neutropenia was diagnosed based on a neutrophil concentration <1·5×109 cells/L on a minimum of two complete blood counts, exclusion of other causes of neutropenia based on a diagnostic bone marrow aspirate or biopsy, and exclusion of secondary immune-mediated neutropenia. Dogs meeting these diagnostic criteria between 2006 and 2013, and that had a haematocrit of ≥29% and minimum of two complete blood clounts performed after initiation of therapy, were included. Results: Information on 35 dogs was included. Neutropenia was less than 0·5×109 cells/L in most cases (21 dogs), 0·5 to ·99×109 cells/L in 11, and 1.0 to 1·49×109 cells/L in three. Eight dogs had thrombocytopenia, which was severe (<49·9×109 cells/L) in three. [Correction added on 23 May 2017, after first online publication: the cell numbers were incorrect due to errors in the conversion of cell measurements to international units. The numbers have been corrected throughout the article and Table 2.] Twenty-three dogs had myeloid hyperplasia, 10 dogs had myeloid hypoplasia and two dogs had normal myelopoiesis. Neutropenia resolved in 32 of 33 dogs within two weeks of starting corticosteroid therapy and in all dogs within one month. Relapse of neutropenia occurred in 12 cases within one year. Clinical Significance: Initial response of presumed primary immune-mediated neutropenia cases to corticosteroid therapy can be excellent. Long-term monitoring for relapse is warranted because 34% of cases relapsed during or after taper of immunosuppressive medications.