Determination of digestible energy values and fermentabilities of dietary fibre supplements: a European interlaboratory study in vivo.

The performance of methods to determine energy conversion factors for dietary fibre (DF) supplements and fermentability (D) values of their non-starch polysaccharides (NSP) was investigated. Heats of combustion, digestible energy (DE) and D values were determined on five DF supplements in five European laboratories on five separate occasions. In each instance the DF supplements were fed to juvenile male Wistar rats at two doses, 50 and 100 g/kg basal diet, for 3 weeks with food and faeces collected in the 3rd week. Among-laboratory variations in heats of combustion (delta Hc) were < 2%. DE values (kJ/g dry weight) at the upper and lower doses respectively were: 10.4 and 9.9 for a high-methoxyl apple pectin, 9.5 and 9.4 for a sugar-beet DF supplement, 12.2 and 12.7 for soyabean DF supplement, 3.8 and 4.0 for maize bran, and 0.3 and 0.3 for Solka-floc cellulose. Variations among laboratories, among occasions and among animals were < 1, < 2 and < 2.5 kJ/g respectively. The among-occasion: among-laboratory variance ratio for DE was 0.5, suggesting the method performed equally well in all laboratories. There was no evidence of learning of fatigue or fatigue in the performance of the method. D values were also independent of dose and at the high and lower doses were: pectin 0.92 and 0.95, sugar-beet NSP 0.68 and 0.68, soyabean NSP 0.86 and 0.88, maize bran 0.17 and 0.18, cellulose 0.07 and 0.06. Among-laboratory variance tended to increase with decreasing fermentability and ranged from 0.03 to 0.18. The DE and D data were not significantly different from a previously proposed relationship DE = 0.7 x delta Hc x D, where delta Hc is the heat of combustion of the supplement. We conclude that while the among-laboratory variation in the D of difficult-to-ferment NSP is too large for the reliable prediction of energy value the method for the direction determination of DE is both reproducible and repeatable, that DE is independent of dosage of DF supplement up to 100 g/kg diet, and that it is safe to discriminate between energy values with a precision of 3 kJ/g. The conversion of both DE and D to net metabolizable energy for the purpose of food labelling, tables and databases is described.