The aim of this study was to examine whether yeast derivate (YD) based on whole brewery yeast added to the creep feed of suckling and newly weaned piglets or to the creep feed of the piglets and the sow's diet prevented post weaning diarrhoea (PWD) or affected performance. Thirty sows and their litters were randomly allocated to three treatment groups: PSP (1.5 g/kg of YD to the sows’ feed from 1 wk before expected farrowing to weaning; 3 g/kg or 2 g/kg of YD added to the piglets’ creep feed from 2 wk of age until 2 wk post weaning (PW) and from wk 2 to 5 PW, respectively); PP (YD added to the piglets’ creep feed as in PSP); or C (control, no YD added). At weaning (4 wk of age) 2 individually housed piglets from all litters were subjected to either experimental Escherichia coli (E. coli) challenge or placebo treatment on d 1 to 3 PW, whereas performance was measured on 3 group-housed piglets from each litter. In individually housed piglets the faecal consistency score (FCS) was affected by an interaction between days PW, treatment group, and challenge group (P=0.005). In general, FCS was lower in placebo than in E. coli-challenged piglets and in PSP and PP piglets than in C piglets. The PSP and PP piglets had lower risk of PWD, defined as FCS > 3, on d 2 to 6 PW compared to C piglets (P=0.014 and P=0.001, respectively). This effect was evident in both placebo and E. coli-challenged PP piglets (P=0.010 and P=0.038, respectively), whereas PSP piglets only differed from C in E. coli-challenged piglets (P=0.030). In E. coli-challenged piglets faecal shedding of haemolytic E. coli was lower in PP than in C piglets (P=0.026). In placebo piglets the latency time to first observation of PWD was longer in PP than in PSP and C piglets (P=0.048 and P=0.017, respectively). The specific antibody titre in piglets or sows was not affected by YD. In group-housed piglets the medical treatment against PWD tended to occur in fewer PP than PSP and C pens within the first 3 wk PW (P=0.078). The average daily gain did not differ between treatments, but PSP piglets had an improved gain to feed ratio (G:F) in wk 0 to 5 PW (P<0.01). In conclusion, YD may prevent PWD at weaning at 4 wk of age if added to the creep feed 2 wk before weaning and PW. Adding YD to the sow as well may, however, antagonize the effect on PWD at a low pathogenic E. coli load, but may improve the G:F compared to no YD supplementation.