The collection of 15 mL/kg of whole blood is recommended for transfusion in large animals [8, 9]. In the present study, the collection and transfusion of that volume caused statistically significant changes in clinical, haematological and biochemical parameters.
After blood collection, no changes in HR were observed, whereas there was a significant decrease at reinfusion, showing that the increased volume resulting from administration of blood minimised HR in an attempt to keep the animal’s equilibrium. In this study, the HR was only evaluated after collection, not during, suggesting that withdrawal of 15 mL/kg of blood was not sufficient to activate the sympathetic response. Malikides et al.  has observed significant HR increases in horses during and after blood collection. In small animals and humans, decreased blood volume promotes a strong sympathetic response, leading to an increase in HR [13, 14].
A significant decrease in RT was only detected in the initial time points after collection compared with the basal time point, which was probably due to decreased blood cell metabolism, but the values remained within the reference values for the species (38.5 to 39.5) . Heinius et al.  also observed this change in a study with haemorrhagic shock in pigs.
In the sheep studied herein, there was no difference in CRT, indicating that the blood collection did not cause major damage to the animals’ circulation. According to Radostits , CRT reflects the circulatory status of the animal without being a high sensitivity assessment. In our study, the RR and SBP showed no differences (p > 0.05) between time points, despite the change in blood volume due to blood loss and reinfusion, which leads us to believe that the decline and subsequent increase in blood volume were not sufficient to cause changes that may affect these parameters.
The GV and RBC showed a decrease (p < 0.01) after collection but remained at levels considered normal for the species. According to Hauptman and Chaljdry , the GV values determined after blood loss may not be accurate, because the adaptation in the plasma/blood relationship of blood cells has not yet occurred. In spite of the significant reduction in GV and RBC from Tc to T12, there was a sharper drop at T24, suggesting that this is the time required for the stabilisation of GV and RBC. In the initial days after acute blood loss, the mobilisation of erythrocytes from storage organs such as the spleen, liver, subcutaneous tissue, great vessels and pulmonary circulation occurs as an urgent compensatory measure [17, 18].
From T48, it was observed an increase in the GV and RBC, indicating that the increased production of erythrocytes is evident from 48 to 72 hours and reaches a maximum at about 7 to 8 days after the onset of bleeding . In human medicine, depending on the circumstances, it is possible to collect a unit or more at appropriate intervals (every five to seven days) before elective surgery . In this study, after 8 days, the GV and RBC values did not return to baseline (Tc0), and we observed that the withdrawal of 15 mL/kg of whole blood decreased the GV by 6.6%. If this is a fixed percentage, a new collection of blood after eight days would lead the animals to become anaemic, suggesting that eight days is not a sufficient interval for a new collection in sheep. Experiments with horses showed that blood withdrawal of 15 mL/kg caused GV to decline by 5% 24 hours after collection .
The increase in mean GV and RBC post-Tr0 are indicative of the increased supply of circulating blood constituents . The reinfusion of 15 mL/kg increased GV by approximately 3%, which is similar to the increase for cows and horses undergoing homologous transfusion [8, 9]. This demonstrates that autologous transfusion presents similar results to those of homologous transfusion, although it is more advantageous because it reduces the risks of adverse reactions.
The WBC initially increased and then decreased, showing that the proposed blood loss caused a slight leukocytic response, with values that are within the reference. These data suggest the autologous transfusion caused minimal complications. The early response of the organism to haemorrhagic injury is characterised by activation of the immune system and by an overwhelming inflammatory reaction .
Studies in human and veterinary medicine have described inflammatory responses associated with blood transfusion. In this study, reinfusion of blood caused an increase in the WBC. In an autologous transfusion model in dogs, McMichael et al.  used leukoreduced and nonleukoreduced erythrocyte concentrates and demonstrated that the inflammatory response was greater in animals that received the nonleukoreduced concentrate. Moreover, an in vitro study has shown a significant increase in seven markers of inflammation during storage comparing leukoreduced and nonleukoreduced erythrocyte concentrates .
In humans, there is an immediate 60% increase in the WBC 12 hours after transfusion, which returns to baseline within 24 hours . In the sheep studied here, the WBC increased starting three hours after transfusion and decreased from the fourth day. The leukocyte response observed in this study is not linked to the presence of antibodies against the surface of erythrocytes or other cellular antigens because the animals received autologous blood. Instead, it is possibly due to storage, which promotes lysis of leukocytes along with release of cytokines and inflammatory immunomodulators such as histamine, myeloperoxidase and eosinophil cationic protein , which triggers an inflammatory response and consequent increase in the WBC. The reduction in WBC has been shown to lessen or eliminate the inflammatory response to blood transfusions in humans and dogs .
After blood collection, there was a decrease in TP and albumin, whereas there was an increase in these parameters after reinfusion. This is most likely because the average total protein is about 7% of total plasma, whereas albumin represents about 50% of plasma total protein [10, 27]. Thus, the decrease in blood volume due to collection caused a decrease in TP and albumin. In a report by Kerr , the reduction in TP was shown to be a consequence of blood loss, and there is a tendency to lose albumin quicker than other proteins due to its small size.
Twenty-four hours after blood collection, TP and albumin values returned to similar one observed at baseline. However, Malikides et al.  observed the return of albumin to baseline 8 days after blood loss in horses. The rapid return to baseline levels of both TP and albumin is due to the existence of a secondary circulation of proteins (especially albumin) from the capillaries to the tissue fluids, which return to the bloodstream via the lymph [27, 28].
There was an increase in TP and albumin during reinfusion (p < 0.05), suggesting a preservation of proteins in stored blood plasma. At the time of reinfusion, these were returned to the animal, thus increasing the values of both TP and albumin.
In this study, we evaluated the activity of a variety of enzymes in order to demonstrate changes in liver function resulting from the loss and reinfusion of whole blood. At Tc, there was a discrete decrease in GGT activity when compared to Tc0, but the values remained within the reference values for the species, suggesting that there were no hepatic injuries resulting from blood loss. After blood reinfusion, there was increased GGT activity. However, at two time points (Tr3 and Tr48), the values exceeded the reference; these were not indicative of a lesion because the values remained within normal limits at subsequent time points. Variation in AST activity was only minor, implying no major changes as it remained within the reference values for sheep (60–280 units/L) .
There was an increase in CK activity above the reference values from Tc until T6, which was later decreased (data not showed). The increase in activity of this enzyme may be a result of haemolysis, skeletal muscle injury or contamination of the blood sample by muscle fluid during a difficult venipuncture . In this study, the need for recumbency during collection may have caused muscle damage, which increased serum levels of this enzyme. According to Kaneko et al. , the short half-life of CK allows for high blood levels to rapidly return to normal.
During collection and reinfusion there were no changes in creatinine and urea serum levels, indicating no kidney alteration during the both process.
Ca concentrations decreased after blood collection, but there was an increase after reinfusion. After reinfusion, the Ca level tended to increase (p < 0.05) markedly from Tr This illustrates that in addition to blood cells, blood reinfusion restores proteins and elements associated with them, such as Ca.
In natural cases of hypocalcaemia in bovines, there is a decrease in serum inorganic P, which can be more intense than that shown experimentally . After the collection of blood from the sheep studied here, there was a reduction in P concentration, suggesting that the decrease in Ca may have contributed to this reduction. Three hours after reinfusion, P levels increased similar to the calcium levels.