Amônia respiratória na avaliação do status nutricional proteico em bovinos

Abstract

This work was carried out with the objective of assessing the use of breath ammonia as a bioindicator of protein nutrition in cattle and to assess the adjustment of predictive models of urinary nitrogen (UN) excretion with breath ammonia (BA), serum urea nitrogen (SUN), and rumen ammonia nitrogen (RAN). The assessed treatments were diets with 10%, 13%, and 16% of crude protein (CP) in the dry matter. Six non-castrated Holstein calves were used, with a mean age of 9 months and a mean initial body mass of 279.91 kg, in a 3x3 double Latin square design. Diets were provided at 7:00 am and 7:00 pm. The assessment was carried out at 12:00 am, 06:00 am, 12:00 pm, and 06:00 pm. The relationship of dietary CP content on breath ammonia flow, and ammonia concentration in expired air and rumen was assessed. Additionally, breath NH3 method was analyzed in comparison with SUN and RAN. To choose the most appropriate model, the models were tested by means of multiple linear regression, observing a significance level of 5% and based on the coefficient of determination. The multiple models tested involved dry matter intake, dietary protein content, protein intake, BA, SUN, and RAN. A linear increase (P <0.05) was observed in the intake and apparent digestibility of CP, as well as nitrogen excretion in urine (NEU) and feces, and positive linear effect (P <0.05) of flow, NH3 concentration, SUN, and RAN with increased dietary protein (P <0.05). Expired NH3 and SUN showed a significant difference (P <0.05) between the sampling times, in which the highest averages occurred at 12:00 pm. There was also a positive correlation (r >0, 5) of flow, ammonia concentration, SUN and RAN with CP intake, dietary CP levels, and NEU. The highest correlation found was between RAN and NEU (r = 0.85). The linear regression model of nitrogen excretion in urine due to protein content in the diet (P <0.05) showed a determination coefficient of 37.9. The best model with BA was adjusted (P <0.05) including dietary protein content and crude protein intake (R2 = 42.5) with a variance inflation factor (VIF) lower than 2.01. The best model with SUN was adjusted (P <0.05) including dietary protein content (R2 = 31.8; VIF = 1.49). The best model with the RAN was adjusted (P < 0.05) including dietary protein content and dry matter intake (R2 = 67.8; VIF < 1.6). It is concluded that BA presents a lower adjustment in multiple regression than RAN, and greater adjustment than models with SUN and only with dietary protein content. What’s more, breath ammonia excretion can be used as a bioindicator of protein nutrition and BA can be used to predict urinary nitrogen excretion in cattle.