Choice of colostrum
As not all batches of colostrum supplied by various farmers had a volume exceeding three litres, a choice had to be made between feeding a homogenous mixture of different batches and feeding different colostrum batches to different calves. The obvious advantage of the first option would have been standardisation of the colostrum offered to the calves. This approach would have required thawing and thorough mixture of a number of batches of colostrum, followed by re-freezing and re-thawing for use. This might have damaged the (unknown) aetiological principle. Another possible disadvantage would have been the dilution of the aetiological agent below a threshold of effect. Because of this, it was decided to feed different batches of colostrum to the individual calves, where possible. In the case of the twins (calves 5 and 6), there were two batches of colostrum of three and one litre left. It was decided to feed them equal volumes; hence these two batches were mixed. The fact that calves 1 and 2 received a mixture of two colostrum batches was due to a misunderstanding but in retrospect proved to be very valuable (see the section on aetiology and pathogenesis).
In the reviewed literature only very few papers deal with normal haematology of very young calves, and either report no precolostral samples nor the exact times post colostrum  or only indicate an interval during which samples were taken [18–20]. However, none of the articles report values below the reference ranges used here. As demonstrated by the values obtained from the control calves, it is unlikely that the observations reported in the experimental calves are normal occurrences in neonatal calves.
Except in calf 1, the courses of thrombocyte counts show a similar pattern in all experimental calves: there was an initial drop within a few hours after colostrum intake, followed by a temporary rise for a few days, and a final drop to extremely low levels in calves 2, 3, 4, and 6. Calf 5 made a remarkable recovery. On the basis of an average lifespan of thrombocytes of 5 to 9 days , about 10 to 20% of thrombocytes can be expected to be removed and replaced per day, or roughly 0.5 to 1% per hour. Thus, even sudden and complete cessation of thrombocyte release from bone marrow could not explain the drop in peripheral thrombocyte counts recorded in the experimental calves in this trial. Hence peripheral destruction must be the main cause.
The drop in thrombocytes cannot be explained by differences in handling following blood sampling. Furthermore, the time between sampling and testing did not affect the number of thrombocytes notably. During a pre-trial blood was sampled and examined at different points up to 72 hours post sampling. There was no notable change in the values obtained. Also the repeatability was evaluated by testing the same blood samples 10 times in a row. There again, no notable differences were found. Therefore it is concluded that the values obtained for the thrombocytes and leukocytes in this study were not influenced by the methods applied. It is also highly unlikely that this drop was caused by transportation of the calves, as the drop was already seen in the first hours after ingestion of colostrum, which was before transportation in calves 1 - 4.
The undulating course in leukocyte counts (Figure 2) may reflect initial peripheral destruction and transendothelial migration, subsequent recruitment, and final depletion due to bone marrow damage. The extremely low levels in affected calves clearly show that both mononuclear cells and granulocytes are affected. Again, no such pattern was seen in the leukocyte counts of the control calves (Figure 2).
Most of the experimental calves were somewhat subdued during the first few days. It is assumed that this was due to the transport to the barn on a hot summer day, and the change in environment.
A remarkable finding was that despite the evidence of generalised bacterial infections in five of the six calves, there were only three days on which one of two calves (calves 5 and 6) had a body temperature above 40°C. Therefore, fever does not seem to be a consistent characteristic of BNP. This is in contrast to findings described earlier [1, 22].
The first clinical signs in calf 2 were blood in the faeces (day 4) and petechiae on day 6. Calves 3 and 4 showed similar clinical courses as calf 2. They had distinct amounts of blood in the faeces on days 8 and 11, and petechiae on days 9 and 10, respectively. Compared with data given in the literature (average age of 16 days  or 17 days ) the clinical signs of the bleeding disorder in the experimental calves of this study occurred much earlier. This may be explained by the fact that the calves in this study were examined daily, with special attention being paid to any signs of bleeding disorder, whereas the cases described in the literature may have been detected at later stages of the disease.
Calves 1 and 5 did not develop unequivocal clinical signs of BNP, but had small amounts of blood in the faeces on several days. The postcolostral rise in globulin levels indicates satisfactory resorption of colostrum in all calves. Therefore, failure of resorption cannot be the reason for the differences in the clinical and haematological outcomes.
The high incidence of processes associated with local or systemic bacterial infections in the experimental calves, requiring intensive treatment, despite isolated housing needs an explanation. Impairment of both local and systemic resistance due to the leukopenia observed seems to be the most plausible reason. Whether calves exposed to the aetiological agent develop either BNP or systemic infections cannot be decided on the basis of available data.
Gross pathology and histopathology
The histological findings in the bone marrow of the six experimental calves indicate that the disease is not an all-or-nothing phenomenon; instead, there can be gradual differences. As Figure 5 shows, there are distinct differences in the bone marrow histology of the six experimental calves. Bone marrow histology of calves 1 and 5 was judged normal, while the bone marrow of calves 3 and 4 showed distinct features of BNP, with no megakaryocytes and only few precursor cells being left. Findings in calves 2 and 6 were not as clear-cut. Although the bone marrow of calf 2 was also judged as being indicative of BNP, the bone marrow of both femur and sternum was not as severely affected as in calves 3 and 4; there were few megakaryocytes left, and other precursor cells were only moderately reduced. Calf 6 showed the least changes, with both megakaryocytes and precursor cells present, but in reduced numbers. Therefore, it cannot be excluded that this calf was affected by BNP, but either at an early stage or in a less severe form. It is not known if such a less severe form of the disease exists, but there is anecdotal evidence that some affected calves do survive the disease.
The control calves were not euthanized due to ethical and economic considerations, therefore no post-mortem examinations or bone marrow examinations are available for these calves. However, their blood values never showed any indication of disturbed production of blood cells. Therefore, it is unlikely that these calves would have shown any bone marrow changes.
In summary, the clinical, haematological and post mortem findings in three of the six experimental calves clearly met the defining characteristics of BNP [1–3, 5–9, 13]. Whether the thrombocytopenia and blood admixture in the faeces on calf 6 on the day of euthanasia heralded BNP or were consequences of septicaemia remains open to speculation. The findings of bone marrow histology in this calf can be explained by both processes.
Speculations on aetiology
The fact that calves of various breeds have been simultaneously affected in Europe [13–15, 23], as well as among the experimental calves of this study all but rules out an exclusively genetic cause of the syndrome.
Most other possible causes of bone marrow depletion in young calves have been excluded . In view of the very short interval between colostrum intake and changes in blood values, any infectious agent transmitted via colostrum would have to have very unusual properties. Also, a cell-mediated mechanism seems unlikely, as the colostrum batches had been stored at -20°C for at least two months. This mainly leaves toxins and antibodies as candidates for aetiologic principles. (Low molecular) toxins would be expected to cross the placental barrier during pregnancy. Therefore, haematological changes would be expected in precolostral calves. All references in the literature, except for one  describe no changes in precolostral calves. One possible hypothesis currently being discussed is an immunopathological reaction due to antibodies - which could be transmitted via colostrum. Many research groups are working on identifying these specific antibodies in the colostrum of cows that have had affected calves. To our knowledge, no specific antibodies have been identified that are linked to the disease. However, all cows whose colostrum was used had been vaccinated against BVD with the vaccine PregSure-BVD®, while none of the dams of the control calves had been vaccinated against BVD. This vaccine was taken off the market in Germany in April, 2010, and in all of Europe in June, 2010, by a voluntary decision of the company. It might be possible that vaccination with this vaccine induces the production of specific antibodies which are transmitted via colostrum to the calves and get absorbed and than bind to both peripheral blood cells and stem cells in the bone marrow of the calves. However, these are speculations, and no evidence has yet been published in the literature.
As no precipitous drops in the erythrocyte counts nor signs of haemolysis were recorded, peripheral destruction of erythrocytes does not seem to be a component of the syndrome. The long lifespan of erythrocytes (100 days)  explains the relatively slow decline in haematocrit despite bone marrow depletion of all precursor cell lines.
According to the history given by the owners of the dams of the experimental calves and the owners of the dams of the control calves, none of the cows had been vaccinated against BVD; yet all the former and five out of six of the latter group had antibodies against BVD, indicating previous natural infection. In view of the high prevalence of BVD-seropositive animals among the cattle population, this is not surprising, and offers no explanation for the development of BNP in experimental calves.
The aetiological role of colostrum seems to be confirmed by the results of this study. The data presented in Table 1 indicate that all batches of colostrum used had the potential to cause BNP. The colostrum donor cows were selected specifically because they had already had at least one calf affected by BNP. This may explain the high proportion of calves that reacted to the ingestion of colostrum seen in this study. Even with under-detection and under-reporting of cases, the incidence of BNP in the field is relatively low. Therefore, it must be assumed that only few cows have the aetiological principle in their colostrum. Some individual predisposition can also be speculated to be present in the calves. This is evidenced by calves 1 and 2: each received three litres of the same mixture of colostrum batches from two cows. Since the calves were born within about half an hour of each other, the intervals between birth and ingestion of colostrum were comparable. Also, postcolostral globulin concentrations in serum were similar (Table 3). Thrombocyte counts in calf 2 dropped below the reference range within two hours after colostrum intake and subsequently never increased. This calf was the first one to develop signs of BNP and had to be euthanized on day 9. By contrast, calf 1 showed neither a drop in thrombocyte counts nor unequivocal signs suggestive of BNP. Therefore, it seems that the specific colostrum is not pathogenic for all calves, but possibly only to calves with an appropriate individual predisposition.
Some results of calves 5 and 6 also suggest that BNP may not be an all-or-nothing phenomenon. Furthermore, personal observation of patients in our clinic shows that some calves recover from overt disease, and some calves are affected subclinically, showing reduced blood values of thrombocytes and leucocytes, but no clinical signs of BNP (like calf 5). Whether the severe infectious processes that are a plausible explanation for the episodes of moderate fever, and necessitated euthanasia in calves 1, 5 and 6 were due to the lack of immune cells remains unclear. An alternative explanation for the fever would be massive cell destruction. The fact that the calves did not receive colostrum from their own dams may also have contributed to this course of events.
Since the haematological changes occurred well before any treatments were administered, any influence of treatments on the course of events seems very unlikely.