A meta-analysis of variability in continuous-culture ruminal fermentation and digestibility data

Alexander Nikolov Hristov, C. Lee, R. Hristova, P. Huhtanen, J. L. Firkins

Research output: Contribution to journalArticle

37 Scopus citations

Abstract

A meta-analysis was conducted to compare ruminal fermentation and digestibility data and variability between continuous-culture (CC) experiments and in vivo data. One hundred eighty CC studies representing 1,074 individual treatments, published in refereed journals between 1980 and 2010 were used in this analysis. Studies were classified into 2 groups based on the type of CC used: CC systems specified as rumen simulation techniques (RUSITEC) and non-RUSITEC CC systems (non-RUSITEC). The latter was a diverse group of systems, all of which were termed CC by the investigators. The CC data were compared with a data set of in vivo trials with ruminally cannulated lactating dairy cows (data from a total of 366 individual cows). The reported neutral detergent fiber (NDF) concentration of the diets fed in the 3 data sets was, on average (dry matter basis), 44, 34, and 32%, respectively. The average total volatile fatty acid (VFA) concentration for the RUSITEC and non-RUSITEC data sets was 67 and 80% (respectively) of the total VFA concentration in vivo. The average concentration of acetate was also lower for the CC data sets compared with in vivo and that of propionate was considerably lower for RUSITEC compared with in vivo, but butyrate concentrations were similar between the CC and in vivo data sets. Variability in the VFA data was generally the highest (higher coefficients of variation and variance) for the non-RUSITEC data set, followed by RUSITEC, and was the lowest for in vivo. Digestibilities of NDF and particularly organic matter were lower in the CC data sets compared with in vivo; the average NDF digestibility was 34.2, 45.5, and 53.0% for RUSITEC, non-RUSITEC, and in vivo, respectively. Variability in nutrient digestibility data followed the pattern of variability of the VFA data: highest variability for the non-RUSITEC data set, followed by RUSITEC, and the lowest for in vivo. This analysis showed that CC systems are generally characterized by lower total VFA and acetate concentrations, extremely low counts or lack of ruminal protozoa, and lower organic matter and NDF digestibilities than in vivo. Overall, variability was much greater for CC than for in vivo experimental data.

Original languageEnglish (US)
Pages (from-to)5299-5307
Number of pages9
JournalJournal of Dairy Science
Volume95
Issue number9
DOIs
Publication statusPublished - Sep 1 2012

    Fingerprint

All Science Journal Classification (ASJC) codes

  • Food Science
  • Animal Science and Zoology
  • Genetics

Cite this