A meta-analysis on the relationship between rumen fermentation parameters and protozoa counts in in vitro batch experiments

Present meta-analysis examines recent in vitro rumen batch experiments where a significant change in the protozoa population was measured after the additions of different substances thought to modify in vitro rumen metabolism. The aim is to study how protozoa number is relating to methane yield and rumen fermentation parameters. Approximately 80 % of the 46 trials from the selected 27 papers (for a total of 201 dietary treatments) tested plant-derived substances, either alone or in mixtures, and 24 trials used rumen fluid from cattle and 22 from sheep. In vitro fermentations with sheep fluid used slightly larger bottles (113 vs 94 mL, P < 0.05), but the inoculum volume, substrate amount, and fermentation duration (38.2 mL, 328.9 mg, and 32.3 h on average, respectively) did not differ (P > 0.05). Only ammonia concentration in cattle fluids was higher compared to sheep (274.2 vs 137.3 mg/L, P < 0.01). Within each trial, we calculated the percentage variation of protozoa compared to control bottles and the majority of the experimental treatments reduced the number of protozoa, by an average of-27.5 %. The relationship between fermentation parameters and variations of pro-tozoa was studied by linear regressions, adjusted for trial effect. In general, protozoa variation in fermentation liquids did not affect total gas production or the sum of volatile fatty acids con-centration. Contrarily, a decrease in protozoa numbers was associated with a significant decrease in methane production, both in absolute terms (R2 = 0.604) and as a proportion of total gas (R2 = 0.528). In terms of individual volatile fatty acids, there was a slight decrease in acetate (R2 = 0.298), an increase in propionate (R2 = 0.388), resulting in a decrease in the acetate: propionate ratio (R2 = 0.379), and no effect of the protozoa on butyrate. Finally, there appeared to be a positive relationship (R2 = 0.580) between protozoa counts and NH3 concentration. In conclusion, rumen protozoa counts of in vitro batch systems are positively associated with methane and ammonia yields, as found in vivo, and this regression study predicts that the com-plete removal of protozoa results in a reduction of about 25 % and 36 % of total ammonia and methane yields, respectively. Batch fermentation systems appear to be appropriate for testing substances that can modulate protozoa counts and related fermentative characteristics.