Variação da composição nutricional nas partes anatômicas e em relação à classe taxonômica de animais silvestres: uma revisão sistemática com uso de imputação de dados com inteligência artificial

Abstract

The meat derived from hunting wild animals is essential for the food security of indigenous peoples and traditional communities. However, information about the nutritional profile of these foods is still scarce. This study conducted a systematic review to examine the influence of the anatomical part and taxonomic class of the animal (mammals, birds, reptiles) on the nutritional composition of the meat. We used the PRISMA protocol to select articles from the Web of Science, Scopus, and Medline/PubMed databases, applying eligibility criteria: (i) original articles, published in any language, without date restrictions, and (ii) presenting composition data of meat from wild animals consumed by human populations; we excluded: (i) studies presenting secondary data and (ii) not detailing the methodology used. A quality questionnaire and a concordance analysis (Fleiss' Kappa = 1.00) were employed to ensure the robustness of the included studies. We used artificial intelligence techniques (SMAPE - Symmetric Mean Absolute Percentage Error) to estimate nutritional values not reported in 21 articles, covering 26 animal species and 10 nutrients (macronutrients and micronutrients). Considering the nature of the data, we performed parametric or non-parametric tests, adopting a p-value of 0.05 to determine the statistical significance of the results and, in situations of significant differences, we proceeded with post hoc tests. Our results reveal significant variations in the nutritional composition between anatomical parts and animal classes. Reptile viscera have more than double the lipid content and triple the iron content when compared to muscles. In the case of mammals, viscera contain five times more omega-6 and selenium, four times more iron and manganese, respectively, and almost double the zinc compared to muscles. We observed differences between classes, with birds having a lipid content more than 90% higher than that found in mammal muscles and 20% higher than that in reptiles; zinc in mammals, with values more than 100% higher than in birds; and selenium in reptiles, exceeding that in birds by more than 400%. However, we did not identify significant differences in iron content between mammals and birds, which may be attributed to the method of slaughtering birds rather than physiological aspects. We conclude that game meat, in the context of populations consuming this food, should be treated as a relevant resource both in terms of access, regarding food security, and utilization, due to its nutritional quality. This study not only fills a gap in a comprehensive nutritional profile of game meat but also introduces, for the first time in the context of analyzing this resource, a methodology for handling missing data, offering an analytical solution to address data scarcity, considering the ethical and legal dilemmas in wildlife analysis. The conclusions of this work can guide nutritional aspects in food security policies, as well as wildlife management decisions in conservation policies, seeking a balance between conservation and subsistence.