Influence of processing of full fat soya beans included in diets for piglets. I. Performance

The programme reported examined the effect of different processing conditions for full fat soya beans (FFSB) during micronization (two steam inlets, being full or low steam-(FS or LS) x two temperatures, being high cook and low cook (HC and LC); Trial 1) and extrusion (four exit temperatures of 70, 90, 110 and 170degreesC; Trial 2) on performance in post-weaned piglets over the live weight range 10-27 kg fed diets containing 300 g processed FFSB/kg. Trypsin inhibitor activity (TIA, mg pure trypsin inhibited/g of sample) ranged from 16.0, 5.2, 4.1 and 2.0 for LSLC, LSHC, FSLC and FSHC, respectively (Trial 1) and from 24.1, 16.8, 6.4 and 2.9 with increasing extrusion temperature (Trial 2). In Trial 1, piglets offered the FS FFSB diets had significantly higher daily live weight gain (DLWG, 652 versus 487 g/day, P < 0.001), consumed significantly more feed (feed intake, FI, 791 versus 679 g DM/day, P < 0.001) and had a significantly better feed conversion ratio (FCR, 1.211 versus 1.404, P < 0.001). There were no effects of temperature or steam x temperature interactions. Performance characteristics were highly correlated with reduced intake of TIA (dietary intake of TIA x feed intake; P < 0.001 for DLWG and P = 0.002 for FI and FCR). In Trial 2, there was a significant linear improvement in DLWG (P = 0.004, but not quadratic) a significant change in FI (P = 0.026 quadratic, but not linear) and a significant linear improvement in FCR (P < 0.001, but not quadratic) with increasing extrusion temperature. DLWG and FCR were negatively correlated to increasing daily TIA intakes (P < 0.05). Both trials allowed an assessment of the response to TIA intake; performance was optimum at low daily intakes of TIA (between 0.5 and 1.5 g/day) irrespective of processing method. Decisions on optimum processing conditions should be based on TIA levels in FFSB, the rate of inclusion and subsequent feed intake (giving, collectively, daily TIA intake) and not simply on named processes. (C) 2004 Elsevier B.V. All rights reserved.