Purpose: A meta-analysis was conducted to summarize and estimate the effect of roughage-based diet on shedding of generic E. coli and colonic pH in cattle compared to grain-based diets.
Methods: A literature review was conducted using PubMed and articles were screened for relevance. References from selected publications were reviewed to identify additional work. Studies of the effects of roughage versus grain diets in cattle on pre-slaughter fecal colonic pH and total generic E. coli counts were included in the meta-analysis. Studies using composite fecal samples or inoculated cattle were excluded. Standardized mean differences in generic E. coli counts and colonic pH were analyzed using Likelihood and Bayesian approaches. Due to significant between-study variation, a random effects model was fitted for each outcome assuming alternate underlying distributions. Sensitivity analyses were conducted by eliminating one study at a time from the analysis.
Results: Five studies involving 181 presumably uninfected cattle were included in the meta-analysis. Based on the best estimate from the aggregate studies, diets consisting of roughage alone, roughage and molasses, and modified grain diets reduced generic E. coli counts by -3.08, -2.79, and -1.09 log/CFU, respectively and increased colonic pH by 4.67, 3.64 and 2.62 units, respectively.
Significance: Meta-analysis can be a useful tool in summarizing the effectiveness of existing food safety interventions. Given the conflicting research on the effect of cattle diet on fecal shedding of STEC and the small sample sizes in existing studies, a meta-analysis is warranted.