Certain probiotics may prevent the development of antibiotic-associated diarrhoea (AAD) and Clostridium difficile-associated diarrhoea (CDAD), but their effectiveness depends on both strain and dose. There are few data on nutritional interventions to control AAD/CDAD in the spinal cord injury (SCI) population. The present study aimed to assess (1) the efficacy of consuming a commercially produced probiotic containing at least 6·5 × 10⁹ live Lactobacilluscasei Shirota (LcS) in reducing the incidence of AAD/CDAD, and (2) whether undernutrition and proton pump inhibitors (PPI) are risk factors for AAD/CDAD. A total of 164 SCI patients (50·1 (sd 17·8) years) with a requirement for antibiotics (median 21 d, range 5–366) were randomly allocated to receive LcS (n 76) or no probiotic (n 82). LcS was given once daily for the duration of the antibiotic course and continued for 7 days thereafter. Nutritional risk was assessed by the Spinal Nutrition Screening Tool. The LcS group had a significantly lower incidence of AAD (17·1 v. 54·9 %, P< 0·001). At baseline, 65 % of patients were at undernutrition risk. Undernutrition (64·1 v. 33·3 %, P< 0·01) and the use of PPI (38·4 v. 12·1 %, P= 0·022) were found to be associated with AAD. However, no significant difference was observed in nutrient intake between the groups. The multivariate logistic regression analysis identified poor appetite ( < 1/2 meals eaten) (OR 5·04, 95 % CI 1·28, 19·84) and no probiotic (OR 8·46, 95 % CI 3·22, 22·20) as the independent risk factors for AAD. The present study indicated that LcS could reduce the incidence of AAD in hospitalised SCI patients. A randomised, placebo-controlled study is needed to confirm this apparent therapeutic success in order to translate into improved clinical outcomes.