Sodium sulphite is used in an optional way to remove insoluble proteins from neutral detergent fibre (NDF) residue. To determine whether the recovery of both NDF and insoluble nitrogen (N) in NDF solution (NDIN) are altered by its use, both parameters were measured in a set of 12 feeds, including cereal grains: maize (MG), rye (RG) and wheat (WG); cereal co-products: maize gluten feed (MGF), distilled dried grains from barley (DDGB) and wheat (DDGW) and wheat bran (WB); protein concentrates: rapeseed meal (RSM) and expeller palm kernel (EPK); dehydrated sugar beet pulp (DBP) and oat (OH) and ryegrass (RGH) hays. Associated effects on the in situ effective degradability (ED) of both NDF and NDIN were also studied in DDGW, WB, RSM, EPK, DBP, OH and RGH. Also, ED of acid detergent fibre (ADF) and its N (ADIN) were studied in hays. Errors due to microbial contamination in the rumen on the ED of NDF, ADF, NDIN and ADIN were also established in these last seven samples using 15N infusion methods. Three rumen and duodenum cannulated wethers were used in the study. The sulphite use in NDF solution led to reductions (DDGB, DDGW, RSM and OH) and increases (RG, WG, WB and DBP) of the NDIN proportion, as well as the contribution of crude protein to NDF. These variations were associated with irregular effects on NDF residues and on ED of both NDIN and NDF. As a consequence, sulphite use does not assure the reduction of the insoluble protein contamination and it may even increase it. This methodology may also alter the degradability estimates of NDIN or NDF. Mean ruminal microbial contamination in NDF was 7·0, 10·8, 13·3, 5·4, 12·0, 35·3 and 20·0 g/kg in WB, DDGW, RSM, EPK, DBP, OH and RGH, respectively. The associated contents of microbial N in NDIN were: 59·3, 29·9, 26·2, 19·8, 37·3, 441 and 150 g/kg, respectively. Microbial contamination in ADF and ADIN (g/kg) was 3·6 and 94·5 in OH and 1·7 and 41·2 in RGH. Not correcting this contamination led to consistent undervaluations of ED of NDIN and NDF in all tested feeds, although errors only reached significance for NDIN in hays and DBP. Microbial-corrected ED of NDIN was 0·685, 0·826, 0·481, 0·389, 0·166, 0·718 and 0·425 in WB, DDGW, RSM, EPK, DBP, OH and RGH, respectively, whereas values for ADIN were 0·504 (OH) and 0·469 (RGH).