Disruptions of the intestinal microbiome are involved in the pathophysiology of acute gastrointestinal graft versus host disease in patients after allogeneic stem cell transplantation. Systemic broad-spectrum antibiotics display a major risk factor for the loss of protective commensal gut bacteria. Abstract Background Maintaining gastrointestinal (GI) microbiome diversity plays a key role during allogeneic stem cell transplantation (ASCT), and loss of diversity correlates with acute GI graft versus host disease (GvHD) and poor outcomes. Methods In this retrospective analysis of 161 ASCT patients, we used serial analyses of urinary 3-indoxyl sulfate (3-IS) levels and GI microbiome parameters within the first 10 days after ASCT to identify potential commensal microbiota–sparing antibiotics. Based on antibiotic activity, we formed 3 subgroups (Rifaximin without systemic antibiotics, Rifaximin with systemic antibiotics, and Ciprofloxacin/Metronidazole with/without systemic antibiotics). Results Mono-antibiosis with Rifaximin revealed higher 3-IS levels (P < .001), higher Clostridium cluster XIVa (CCXIVa) abundance (P = .004), and higher Shannon indices (P = .01) compared to Ciprofloxacin/Metronidazole with/without systemic antibiotics. Rifaximin followed by systemic antibiotics maintained microbiome diversity compared to Ciprofloxacin/Metronidazole with/without systemic antibiotics, as these patients showed still higher 3-IS levels (P = .04), higher CCXIVa copy numbers (P = .01), and higher Shannon indexes (P = .01). Even for this larger cohort of patients, the outcome was superior with regard to GI GvHD (P = .05) and lower transplant-related mortality (P < .001) for patients receiving Rifaximin plus systemic antibiotics compared to other types of systemic antibiotic treatment. Antibiosis with Ciprofloxacin/Metronidazole (n = 12, P = .01), Piperacillin/Tazobactam (n = 52, P = .01), Meropenem/Vancomycin (n = 16, P = .003), Ceftazidime (n = 10, P = .03), or multiple systemic antibiotics (n = 53, P = .001) showed significantly lower 3-IS levels compared to mono-antibiosis with Rifaximin (n = 14) or intravenous Vancomycin (n = 4, not statistically significant). Conclusions Different types of antibiotic treatments show different impacts on markers of microbiome diversity. The identification of antibiotics sparing commensal bacteria remains an ongoing challenge. However, Rifaximin allowed a higher intestinal microbiome diversity, even in the presence of systemic broad-spectrum antibiotics.