Purpose To assess whether crofelemer would prevent chemotherapy-induced diarrhea (CID) diarrhea in patients with HER2-positive, any-stage breast cancer receiving trastuzumab (H), pertuzumab (P), and a taxane (T; docetaxel or paclitaxel), with/without carboplatin (C; always combined with docetaxel rather than paclitaxel). Methods Patients scheduled to receive ≥ 3 consecutive TCHP/THP cycles were randomized to crofelemer 125 mg orally twice daily during chemotherapy cycles 1 and 2 or no scheduled prophylactic medication (control). All received standard breakthrough antidiarrheal medication (BTAD) as needed. The primary endpoint was incidence of any-grade CID for ≥ 2 consecutive days. Secondary endpoints were incidence of all-grade and grade 3/4 CID by cycle/stratum; time to onset and duration of CID; stool consistency; use of BTAD; and quality of life (Functional Assessment of Chronic Illness Therapy for Patients With Diarrhea FACIT-D score). Results Fifty-one patients were randomized to crofelemer ( n  = 26) or control ( n  = 25). There was no statistically significant difference between arms for the primary endpoint; however, incidence of grade ≥ 2 CID was reduced with crofelemer vs control (19.2% vs 24.0% in cycle 1; 8.0% vs 39.1%, in cycle 2). Patients receiving crofelemer were 1.8 times more likely to see their diarrhea resolved and had less frequent watery diarrhea. Conclusion Despite the choice of primary endpoint being insensitive, crofelemer reduced the incidence and severity of CID in patients with HER2-positive breast cancer receiving P-based therapy. These data are supportive of further testing of crofelemer in CID. Trial registration Clinicaltrials.gov, NCT02910219, prospectively registered September 21, 2016.