The subject of protein quality assessment of foods and diets was addressed at the Codex Committee on Vegetable Proteins (1982–1989), FAO/WHO (1989, 2001) and WHO/FAO (2002) expert reviews. These international developments are summarized in this manuscript. In 1989, a Joint FAO/WHO Expert Consultation on Protein Quality Evaluation reviewed knowledge of protein quality assessment of foods, and specifically evaluated amino acid score corrected for protein digestibility, the method recommended by the Codex Committee on Vegetable Proteins. The report of the Consultation published in 1991 concluded that the Protein Digestibility-corrected Amino Acid Score (PDCAAS) method was the most suitable approach for routine evaluation of protein quality for humans. The Consultation recognized that the amino acid scoring pattern proposed by FAO/WHO/UNU (1985) for preschool children was at that time the most suitable pattern for calculating PDCAAS for all ages except infants in which case the amino acid composition of human milk was recommended to be the basis of the scoring pattern. The rat balance method was considered as the most suitable practical method for predicting protein digestibility by humans. Since its adoption by FAO/WHO (1991), the PDCAAS method has been criticised for a number of reasons. The FAO/WHO (2001) Working Group on analytical issues related to protein quality assessed the validity of criticisms of the PDCAAS method. While recognizing a distinct regulatory use of protein quality data, the Working Group recommended that the PDCAAS method may be inappropriate for the routine prediction of protein quality of novel and sole source foods which contain high levels of anti nutritional factors; and that for regulatory purposes, the method should be revised to permit values of >100 for high quality proteins. In evaluating the recommendations of the Working Group, the WHO/FAO (2002) Expert Consultation on Protein and Amino Acid Requirements endorsed the PDCAAS method with minor modifications to the calculation method but also raised several issues. These included the calculation of scoring patterns; prediction of amino acid digestibility by faecal and ileal methods; reduced bioavailability of lysine in processed proteins; truncation of the amino acid score and consequent PDCAAS value; protein digestibility as a first limiting factor in determining the overall available dietary nitrogen; and the calculation of amino acid score for a dietary protein mixture. These concerns were considered particularly important in relation to the regulatory aspects of protein quality of foods, and their resolution was urgently recommended through a new separate expert review.

Three screening tests for protein quality, modified limiting amino acid score (MLAAS), net weight gain (NWG) and net protein ratio (NPR), were compared. Two experiments using young broiler chickens were conducted in a temperature-controlled room at 28·5 ± 0·5° with no adaptation to cages and diets, or at 31 ± 0·5° with 2 d adaptation to cages and diets. Nine isoenergetic diets containing nominally 100 g crude protein/kg supplied by legume meals and one isoenergetic N-free diet were randomly allocated to chicks in single cages in each side of a four-tier battery brooder. Each dietary treatment had eight replicates. The chickens had access ad lib. to diet and drinking water throughout a 14 d observation period. Body weight and feed were measured at the start on day 7 and at the end on day 21. The results indicated that keeping the chickens at 31 ± 0·5° and giving them a 2 d adaptation period decreased the variability of chickens' responses to each treatment. MLAAS, NWG and NPR methods could distinguish legume proteins of high, medium and low feed values. MLAAS correlated well with NWG (r 0·90; P < 0·001) and NPR (r 0·78; P < 0·01) in evaluating the protein quality of grain legumes used as sole sources of protein for meat chickens. However, MLAAS did not predict the exact order of NWG and NPR. Growth was limited because dietary methionine, the first limiting amino acid, provided only 27·6–55·2% of the recommended proportion in the protein. Although the results should be interpreted cautiously since a small sample size was used, it was concluded that the MLAAS calculation could be used as a reasonable estimate of the relative protein quality of most grain legumes, but that NWG and NPR were better methods as they detected limiting factors other than limiting amino acids in raw and processed legumes.