The effects of romiplostim on bone marrow morphology were evaluated in adults with immune thrombocytopenia (ITP). Patients with platelet counts <50 × 10 9 /L, ≥1 prior ITP therapies, and no collagen at baseline received weekly subcutaneous romiplostim starting at 1 μg/kg, adjusted to maintain platelet counts between 50 and 200 × 10 9 /L. Biopsies were scheduled after 1, 2, or 3 years of romiplostim (cohorts 1, 2, and 3, respectively). Irrespective of scheduled time, biopsies were performed earlier if patients discontinued or failed to achieve/maintain a response to romiplostim. Reticulin (silver stain) and collagen (trichrome stain) were graded by two hematopathologists using the modified Bauermeister scale (0–4). Of 169 patients, 131 had evaluable biopsies; 9/131 (6.9 %) had increases of ≥2 grades on the modified Bauermeister scale (cohort 1: 0/34; cohort 2: 2/39; cohort 3: 7/58), including two with collagen. Three of the nine patients had follow-up biopsies, including one patient with collagen; changes were reversible after romiplostim discontinuation. Of the nine patients, one had neutropenia detected by laboratory test and two had adverse events of anemia, both non-serious and not treatment-related. By actual exposure (as some biopsies did not occur as scheduled), the number of patients with grade increases ≥2 were year 1: 3/41, year 2: 1/38, year 3: 5/52. Twenty-four patients sustained platelet counts ≥50 × 10 9 /L for ≥6 months with no ITP medications after discontinuing romiplostim, i.e., they entered clinical remission of their ITP. In conclusion, in patients with ITP receiving romiplostim, bone marrow changes were observed in a small proportion of patients. ClinicalTrials.gov identifier: NCT#00907478