PROGNOSIS OF DIARRHOEA IN THE NEWBORN CALF: STATISTICAL ANALYSIS OF BLOOD BIOCHEMICAL DATA J.C. Fayet, J. Overwater To cite this version: J.C. Fayet, J. Overwater. PROGNOSIS OF DIARRHOEA IN THE NEWBORN CALF: STATISTI- CAL ANALYSIS OF BLOOD BIOCHEMICAL DATA. Annales de Recherches Vétérinaires, INRA Editions, 1978, 9 (1), pp.55-61. <hal-00900979> HAL Id: hal-00900979 https://hal.archives-ouvertes.fr/hal-00900979 Submitted on 1 Jan 1978 HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
PROGNOSIS OF DIARRHOEA IN THE NEWBORN CALF: STATISTICAL ANALYSIS OF BLOOD BIOCHEMICAL DATA J.C. FAYET J. OVERWATER Laboratoire d Eco-Pathologie I.N.R.A.. C.R.Z.V. de Theix, 63110 Beaumont. Résumé. PRONOSTIC DE LA DIARRHEE CHEZ LE VEAU NOUVEAU-NE : ANALYSE STATISTIQUE DES DONNEES BIOCHIMIQUES SANGUINES. ― L analyse statistique multidimensionnelle (composantes principales, analyse discriminante) de onze paramètres sanguins mesurés chez des veaux atteints de diarrhée, lors de la première intervention a permis d aboutir aux résultats suivants : Le paramètre constamment modifié dans cette affection est évidemment l équilibre acidobasique (acidose métabolique non compensée). Compte tenu du traitement institué, ce paramètre n est pas le facteur limitant dans la réussite de la thérapeutique. En revanche, la discrimination entre animaux morts et survivants se fait sur les variables du catabolisme, l urée en particulier. De nouvelles thérapeutiques devraient prendre en compte ce facteur, jusque là très souvent négligé et pourtant capital. Introduction. During diarrhoea in the newborn calf, the use of blood plasma analysis, in order to improve the treatment and prognosis, is not possible in all cases due to the fast evolution of the disease, which often does not allow enough time to wait for laboratory results. However the question might well be asked whether by measuring two or three well chosen parameters if one could not get an idea of the likely outcome of a particular treatment. The question is particularly pertinent in the case of a calf of little value compared to the cost of treatment. Success in correlating the clinical signs and the biochemical state of the calf could also improve the treatment of this condition. Furthermore, a better understanding of the biochemical parameters when a given treatment has failed should give some new therapeutic orientations. It was with these objectives that certain biochemical parameters, judged to be important, were measured before treatment in order to correlate these results retrospectively with the survival or death of the animals. Materials and methods. ANIMALS Fifty five male and female calves with profuse diarrhoea, received by the experi-
mental treatment unit of the research station during the calving seasons of 1974 and 1975 were used in this analysis. Their age, at the time of admisson, varied from four to twenty days. The aetiology of the diarrhoea was unknown. The clinical condition of the calves at the time when the first blood analysis was carried out, was variable: some of the calves had been ill for several days before being admitted, others were brought in within 24 hours of the first signs of diarrhoea appearing. Twenty one died. The 34 surviving calves were returned to their owners after having started to regain weight. The time spent in the wait varied from 2 to 21 days. BIOCHEMICAL ANALYSIS As soon as the calves arrived, a blood sample was taken from the jugular vein in order to determine the following parameters: 1 ) ph 2) COZpartial pressure 3) OZpartial pressure 4) HCO 3concentration 5) Na! concentration 6) K concentration 7) Cl concentration 8) haematocrit 9) blood urea concentration 10) phosphorus concentration 11) total blood protein concentration. The treatment of the calves, based almost exclusively on their rehydration, was decided using these measurements. The analysis of these blood parameters was carried out twice daily. TREATMENT The basic treatment for all calves was the same although it was adapted for each individual calf. The treatment consisted of intraveinous infusions of a sodium bicarbonate solution mixed with glucose and saccharose. The quantity of this solution administered depended upon the live weight of the calf and the bicarbonate deficit as determined from the blood analysis. In the majority of cases antibiotics were administered orally based on the results of an antibiogram carried out on the faeces. Speaking generally, colimycine and polymyxine B were the antibiotics the most used.
MATHEMATICAL TREATMENT OF THE RESULTS Two programs were used: 1) analysis of the principal components, 2) the step by step analysis of the discriminant function. The principle of the analysis of the principal components is given in the work of Sauvant et al. (1973) and is found in more detail in the works of Lefebvre (1976) and Daget (1976). The analysis of the principal components sets out to resolve a system of hierarchical reference axis so that on decreasing the number of dimensions of the space in which are projected the individuals, the loss of information is at a strict minimum. This method is used on a descriptive basis. The analysis of the discriminant function is a technique used, in this case, in order to study the divergence of two populations (surviving and dead calves). It is a case of defining a function of the recorded parameters so that the maximum difference appears between the two groups. Results. 1) CHARACTERISTICS OF THE PARAMETERS STUDIED The characteristics of the 11 parameters as defined above are shown in Table 1. They confirm the results previously found (Fayet, 1968; Tennant et al., 1968). The noncompensated metabolic acidosis is always present. The dehydration results in a high haematocrit. 2) ANALYSIS OF THE PRINCIPAL COMPONENTS a) Correlation table between the parameters. The correlation table (Table 2) shows that certain parameters have a very similar behaviour (for example 1 and 4 the bicarbonate concentration varies directly with the blood ph, a phenomenon which is well known). On the other hand there is an inverse relationship between other parameters, for
example 2 and 3: the partial pressure of CO2 and 02 are inversely related. b) Principal components. Graphical representation. The first four principal components explain nearly 77% of the total inertia (Table 3). Table 4 shows the correlation between the parameters and each of the four principal components. Figures 1 and 2 represent the projection of the &dquo;parameter&dquo; and the &dquo;animal&dquo; vecteurs onto the principal plains 1 and 2 and 1 and 3. In Figures 1 and 2 the principal axis 1 (horizontal) is equivalent to the acidbase balance. The ph and the bicarbonate concentration find themselves situated around the positive portion of this axis outside of the 0.8 circle. Axis 2 is characteristized by the three parameters, the blood phosphate, the blood potassium and the blood urea being relatively grouped together. These three parameters can be considered as indicating the state of catabolism. The fact that the acid-base balance and the catabolism can be visualised by 2 orthogonal axis shows that there is no relationship between these two criteria. The position of each of the calves on the same graph demonstrates that the surviving calves are found preferentially in the lower part and the dead calves, on the other hand, in the upper zone. The X2 test confirms that this difference is highly significant (X2= 15,03). The axis 3 (Figure 2) is defined essentially
by the blood sodium and to a lesser extent by the blood potassium concentrations which are inversely related. This axe could represent the outcome, as reflected by the plasma, of the loss of electrolytes in the faeces. As the loss of electrolytes in the faeces increases, so the blood sodium concentration decreases, whilst the blood potassium concentration (a measure of catabolism) has a tendency to increase. Although the loss of potassium is proportional to the importance of the diarrhoea, the blood potassium concentration increases. This increase is due to the release of cellular potassium. 3) DISCRIMINANT FUNCTION ANALYSIS The step by step discriminant function analysis between the group of dead and the group of surviving calves gives the results shown in Table 5. Three parameters were used: blood urea concentration, blood chloride concentration and the haematocrit. Using the equation of the discriminant function the group to which each calf should have been placed is determined. Under these conditions, 78% of the calves were correctly alotted. Errors in affecting the calves existed for both groups. Discussion. On the eleven parameters considered, it is obvious that not all represent the same value in describing the condition of the calves. The first axis given by the principal component analysis represents without any doubt the acid-base balance. One might have thought that among the parameters of this first axis would be found the most suitable parameter for differentiating between the dead and the surviving calves. This was shown however not to be the case. The localisation of the individuals on the graph demonstrates this fact, and the discriminant analysis further confirms it. However, one can find an explanation for this phenomena. Given the fact that the basic treatment consisted of a rigourous correction of the acid-base balance, it is not this parameter which will allow one to differentiate, a priori, between the surviving and the dead. On the other hand, the blood urea concentration (or the blood potassium concentration) is a measure of nitrogen catabolism which was not the object of any specific treatment. Under these conditions it is understandable that the discrimination between the two groups is made by the parameters the most representative of catabolism, that is to say the blood urea concentration. The fact that there exists a certain number of errors in affecting the calves in both groups signifies that the treatment administered was not entirely appropriate. Certain calves died although the blood analysis carried out on arrival suggested that these calves should have lived. An improvement in the treatment undertaken might be in the administration of a mixture of amino-acids. The exact composition of this mixture remains to be determined, based on the exact needs of the diarrhoeic newborn calf. It is very probable that other parameters, not analysed, have an importance for the prognosis given on arrival. The correlation matrix (Table 2) shows an opposition between the partial pressure of 02in veinous blood and the parameters of the acid-base balance, particularly the partial pressure of C02. In order to counteract the metabolic acidosis, the body normally reacts by decreasing the partial pressure of CO 2, If 2 one now considers the classic dissociation curve for oxyhaemoglobine, one sees that a lowering of the partial pressure of C0 2leads to a decrease in the amount of Oz given up to the tissues (the Bohr effect). This results in an increase in the partial pressure of 02. 2 In the case of the calf, the lowering of the partial pressure of CO 2exists only rarely. One might therefore ask whether the increase in the 02 partial pressure is not due rather to a deficient use of oxygen by the tissues. This could be the consequence of the hypovolaemia and the weak irrigation of the whole organism. Conclusion. The group of results obtained confirms that a non-compensated metabolic acidosis is a phenomenon obtained currently in all diarrhoeic calves (first principal axis). Catabolism is an important phenomenon in these animals (second principal axis). This fact is not surprising considering the
small quantity of food ingested by these sick animals. There exists no relationship between the acid-base balance and catabolism. These states are found on the orthogonal axis. The acidosis is certainly due to the loss of the bases in the faeces and to the state of fasting. It is thus a direct consequence of the diarrhoea. It is possible that the catabolism which occurs with the diarrhoea is in fact the expression of badly functioning physiological mechanisms. One could thus imagine that the catabolism which occurs could be more or less pronounced depending upon the causal agent (virus, bacteria). The use of three parameters, easily measured (urea, haematocrit and chloride) allows a prognosis to be given with a probability of exactitude of 78% following a treatment based on the rehydration and the compensation of the acidosis and the loss of Na+ ions. The model so defined could be used to test the efficiency of new treatments compared with that used in this study. In fact, the basic treatment used could be maintained and other treatments aimed at counteracting the catabolism could be added. A new principle component and discriminant analysis on the results obtained would measure the efficacity of the treatment. The graphs would show, after disposing the parameters on the axis that it was in fact the same type of pathology. The discriminant analysis would indicate if the parameters of catabolism were still the most important. Accepted for publication, October 1977. Summary. Blood samples were taken from 55 diarrhoeic calves (neo-natal diarrhoea) at the time of the first therapeutic intervention. Eleven blood parameters were measured in these samples. The whole group of measurements were analysed by their principal components (Tables 3 and 4). The results are expressed in visual form by Figures 1 and 2. The first principal axis corresponds to the acid-base balance while the second axis can be considered as corresponding to the catabolism. The discriminant analysis (Table 5) shows ihas the parameter having the best prognosis value is the blood urea concentration. By adding two other easily measured parameters (the haematocrit and the blood chloride concentration) the probability of classifing correctly the calves into one or other of the groups dead or surviving is approximately 80%. References DAGET J., 1976. Les modeles math6matiques en 6cologie. Masson (Paris), 172 p. FAYET J.C., 1968. Recherches sur le m6tabolisme hydromin6ral chez le veau normal ou en 6tat de diarrh6e. 11. L ionogramme plasmatique et le ph sanguin. Rech. V6t. (1), 109-115. LEBART L. et FENELON J.P., 1971. Statistiques et informations appliqudes. Dunod (Paris). 457 p. LEFEBVRE J., 1976. Introduction aux analyses statistiques multidimensionnelles. Masson (Paris). 219 p. SAUVANT D., FEHR P.M., RODOLPHE F., TOMASSONE R. et DELAGE J., 1973. Etude des interrelations entre les critbres de production et de composition lipidique du lait de ch6vre par deux methodes d analyse factorielle. Ann. Biol. Amm. Bioch. Biophys. 13, 107-129. TENNANT B., HARROLD D. et REINA-GUERRA M., 1968. Hypoglycemia in neonatal calves associated with diarrhoea. Cornell Vet. 58, 136-146.