Incidence of BNP at the herd level
One of the main objectives of this study was to identify the incidence of BNP at the herd level. For the calculation of the incidence the average number of calves born on the farm was available. This could have introduced a bias, as stillborn calves would not have been included in the denominator of the calculation. Using the average number of calves born on the farm also could have introduced a bias, which we cannot exclude. However, all farmers included in the study had reported that their herd size had been stable over the considered time period, therefore we are confident that the number of calves born over several years did not vary considerably. BNP affects calves during their first 4 weeks of life. Male Holstein-Frisian and Red Holstein calves are usually sold at an age of two weeks, thus it is possible that some cases of BNP were missed when BNP occurred after the calves were sold to another farm. However, most of the participating farms (77 %) had German Fleckvieh, crossbred or Brown Swiss. Male calves from those farms are typically not sold before they have reached a certain age or weight and do not leave the farm during the time when BNP may occur.
The highest incidence reported in a herd from our study (19 %) was substantially higher than the incidence rates reported in other studies, which was up to only 10 % . Although the BNP incidence was low on most of the affected farms, some herds were obviously more severely affected. The incidence of BNP varies markedly among different countries and also among the federal states of Germany [4, 6]. The farms enrolled in our study were from nine federal states of Germany. However, most of the farms with cases of BNP were localized in Bavaria. Kasonta et al.  investigated the influence of different patterns of PregSure BVD vaccinations on the incidence of BNP. Their data showed that different vaccination regimens resulted in differences in alloreactive antibody concentrations. Therefore, the differences in incidence rates among federal states can potentially be explained by the different vaccination protocols that exist. Apart from the varying incidence among the federal states of Germany, we found considerable differences among the participating farms within the federal states. Presumably, the use of different vaccination regimens at the herd level is the reason for that observation, but other factors (e.g. colostrum management) could have had an influence on the incidence as well.
The present study shows a remarkable increase of BNP cases in 2008, reaching a peak in 2009. Although PregSure BVD was licensed in 2004, an increased BNP incidence was not noted until 2007. The same phenomenon was observed in New Zealand: the first cases of BNP occurred there in 2011, which was 3 years after the introduction of PregSure BVD . Kasonta et al.  interpreted the temporal delay as being associated with the repeated immunizations that are required for creating a titre of alloreactive antibodies high enough to induce clinical BNP in the calves.
Various reasons exist for the decline of BNP cases on farms in our study since 2010. Owing to increased knowledge concerning aetiopathogenesis, farmers adapted their management practices, especially colostrum and vaccination management . The safest way to prevent further BNP cases is to dispose of colostrum coming from known ‘BNP dams’ and to feed calves colostrum from cows that have not been vaccinated with PregSure BVD. Before PregSure BVD was withdrawn from the German market in April 2010, many farmers had already switched to a different BVD vaccine or discontinued vaccination against the BVDV. Nevertheless, calves with BNP are still observed on some farms, but in much smaller numbers . Therefore, the antibody titre seems to remain high for a long time, even though PregSure BVD is no longer used. It is uncertain how much longer further cases of BNP will occur, but they will most likely stop when the last PregSure BVD-vaccinated cows are removed.
Data matching between the two databases
Another objective of our study was to investigate whether the number of calves with BNP registered at the NPhVS reflects the real number of cases in Germany. More than one-half of the cases and farms had not been listed in the NPhVS database of the Paul-Ehrlich-Institut. Three-hundred cases of BNP were reported in our study that occurred before 2009, which was the year when BNP was recognized as a potential serious adverse effect of PregSure BVD. A retrospective report of these cases was technically possible, but seems unlikely to happen, so most of these cases are most likely not in the database. Although veterinarians are requested to report adverse events in accordance with the veterinary medical association’s professional code of conduct, they and the farmers do not directly benefit from reporting cases to the NPhVS. Thus, they are increasingly reluctant to fill in reporting questionnaires. Although we discovered a remarkable extent of underreporting, considering the population of 4.3 million dairy cows in Germany , the overall incidence is still very low.
The severity of economic losses in affected herds due to medical treatments of BNP calves, costs for the veterinarian, and especially case fatalities, is difficult to estimate. However, an appreciable pecuniary loss can be assumed for farms with high incidence. In addition, less obvious losses could be caused by subclinical BNP cases in which calves could be more susceptible to other diseases, especially infections due to immunosuppression. Thus, in an epidemiological study, 2.5 % of control calves had thrombocytopenia and leucocytopenia without clinical signs of BNP , and another research group induced subclinical BNP in 20 % of calves after feeding with colostrum from ‘BNP dams’ .
Confirmation of BNP via blood analysis or necropsy
To reduce the bias of misclassification and, thus, maximize the probability of including farms with actual cases, only farms with at least one confirmed case of BNP were included in this study. Therefore, we excluded some farms from the study because even though the farmers had observed calves with signs consistent with BNP within their herds, they did not initiate an examination to verify the aetiology.
As reflected by the high proportion (58 %) of cases that were confirmed by haematology and or necropsy, veterinary practitioners demonstrated a great commitment to elucidate the previously unknown aetiology of this disease.
The spectacular clinical signs that are most often seen in affected calves, and the severe process of BNP are very similar among cases. Therefore, it can be assumed that affected farmers and their veterinarians are aware of the disease pattern and that most of the calves included in this study without disease confirmation likely did have BNP. Some cases of BNP were probably missed at the very beginning of the occurrence when farmers and veterinarians were not aware of the disease or in case of sudden death. Furthermore it is not possible to get any information about subclinical cases of BNP within a herd without blood sampling the calves, thus subclinical BNP could not be taken into account.
Because there are no advantages (e.g. compensation or specific treatments and better prognoses for the calves after disease diagnosis) for farmers and veterinarians in confirming additional cases, the proportion of investigated cases will decline and reports on the occurrence of BNP to official agencies will become less common over the years.