The study was conducted as a randomized, double-blind, placebo-controlled study at Mwanamugimu Nutrition Unit, Mulago Hospital, Kampala, Uganda, between March 2014 and October 2015. The study enrolled 400 children between 6 and 59 months of age with SAM. The children either received one daily dose of probiotics consisting of 10 billion colony-forming units of two probiotic strains BB-12 and LGG (ratio 1:1) or placebo. Probiotics or placebo was administered during hospitalization and during a subsequent 8-12 week outpatient treatment period. The primary outcome was number of days with diarrhea during hospitalization. Secondary outcomes included number of days with diarrhea during outpatient treatment, diarrhea incidence, diarrhea severity according to the Vesikari scale, fever, vomit, pneumonia, weight gain and nutritional recovery. All outcomes were analyzed separately for in- and outpatient treatment. A stool diary was developed collecting data on stool frequency and consistency. The diary was filled by caregivers and used to calculate diarrhea outcomes. The validity, reliability and sensitivity of the diary were assessed. Data from the intervention study were used to identify predictors of days with diarrhea and dehydration during hospitalization. In addition the association between days with diarrhea as well as dehydration and mortality during hospitalization were assessed. Results The study children had a mean age of 17.0 months, 58% were boys, 66% had edematous malnutrition and 14% were HIV-seropositive. There was no effect of probiotics on days with diarrhea during inpatient treatment (adjusted difference +0.2 days [95% CI -0.8; 1.2], p=0.69). However a reduction in days with diarrhea was found in the probiotic group during outpatient treatment (adjusted difference -2.2 days [95% CI -3.5; -0.3], p=0.025). There was no difference between the probiotic and placebo groups with regard to diarrhea incidence and severity, vomiting, fever, pneumonia, weight gain or nutritional recovery. Fortysix patients died, with 26 patients from the probiotic group and 20 patients from the placebo group (p=0.38). Caregivers understood the content of the stool diary (stool frequency and consistency), and after three days of training, they were able to fill the diary with high reliability. The diary had a high sensitivity to measure changes in stool frequency and consistency. Stool frequency and especially stool consistency correlated strongly with dehydration assessed by medical doctors (p=0.0018 for increasing stool frequency and p<0.0001 for increasing stool consistency). Out of 592 diarrhea episodes reported during hospitalization, 355 were categorized as hospitalacquired and 237 as admission diarrhea episodes. Young age and HIV predicted days with diarrhea and dehydration during hospitalization. Both days with diarrhea and dehydration were strongly associated with increased mortality during hospitalization. The mortality risk increased 1.4 times (95% CI: 1.2; 1.6) per day of diarrhea and 2.3 times (95% CI: 1.6; 3.6) per point on the dehydration scale. Conclusion LGG and BB-12 did not reduce the number of days with diarrhea during inpatient treatment of children with SAM. However, they reduced days with diarrhea during the subsequent outpatient treatment by 2.2 days corresponding to a reduction of 26%. Probiotics may therefore have a role in community-based treatment of children with uncomplicated SAM or in follow-up of children discharged from inpatient treatment of SAM with medical complications, but more studies are needed to investigate this. The stool diary was shown to have high validity, good reliability and high sensitivity. It may be useful in other diarrhea studies among young children in low-income countries. Diarrhea and dehydration were both strongly associated with increased mortality during hospitalization of children with SAM. Improved management of diarrhea and prevention of hospital-acquired diarrhea therefore seems to be important to reduce mortality in this vulnerable patient group.