A meta-analysis was performed to evaluate the effects of supplementary n-3 polyunsaturated fatty acids (PUFA) sources in the diet on the formation of some important n-3 PUFA contents in eggs and to assess factors contributing to the conversion efficiency of omega-3 in laying hens. A dataset was constructed from 34 studies examining the impact of dietary inclusion with ingredients rich in n-3 PUFA on fatty acids profile and production performance of laying hens. The eligibility criteria were developed to obtain studies reporting required information with sufficient quality. The mixed model methodology was employed where the “study” was set as random effects and fatty acid (FA) supplements as fixed effects. Several factors were included in the models as covariates. Discrete analysis for sources of FA was also performed to compare their effects on FA formation in eggs. Significant linear positive associations were observed between the concentration of α-linolenic acid (ALA), total n-3 PUFA, and the ratio of linoleic acid (LA) to ALA (LA/ALA) in diets with the formation of eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA), total n-3 PUFA, and n6/n3 ratio in egg (P < 0.05) with different magnitudes. ALA and total n-3 PUFAs concentration had no relationship with cholesterol concentration, feed intake, and egg weight. Prediction models for DHA formation was higher for ALA as predictor variables (slope = 0.482; R2 = 0.684) than n-3 PUFAs (slopes = 0.998, R2 = 0.628). Significant interactions were found on the level of ALA × FA sources and n-3 PUFA × FA sources. Fish oil (P = 0.0148, R2 = 0.732) improved the prediction equation to estimate DHA formation. To conclude, levels of ALA, n-3 PUFA, and the ratio of LA/ALA can be used as predictor variables to estimate the formation of n-3 fatty acids in eggs. It was confirmed that although all n-3 FA sources had a positive correlation on DHA and n-3 PUFA deposition, however, fish oil showed the highest prediction model for DHA formation across all FA sources included in the dataset.
Read full abstract