The aim of this study was to quantify the effect of a single vaccination of broilers against a HPAI H5N1 on virus transmission, and to which extent virus could spread among broilers and DOC. Vaccination with an inactivated homologous strain did not reduce transmission of H5N1 virus significantly. In both vaccine experiments, in which birds were vaccinated when they were either one or ten days old, the reproduction number was above one indicating that virus could still spread extensively. Moreover vaccination did not prevent the occurrence of clinical signs, although it seemed to reduce mortality slightly. This implies that unnoticed virus spread is unlikely, even in a vaccinated flock, which is often feared by farmers and policy makers as mentioned before [14, 15, 33].
Transmission among day-old chickens, which still had MDA, tended to be reduced compared to chickens of four weeks of age without MDA, as fewer contact birds became infected, but the reproduction number was not significantly below one. Of course the real control group, MDA-free DOC, was lacking, as this was not our research question. In various trials, the clinical signs seemed to be less severe, suggesting that these DOCs may pose a risk in the spread of the infection as they may spread the virus unnoticed. These findings indicate that broilers, including DOC, could play a role in the epidemiology of AI, as virus could spread extensively. It cannot be determined from these experiments, however, to which extent this may occur in the field, as this also depends on the number of virus incursions, and the subsequent implemented control measures.
Vaccination only induced very low titers of HI antibodies in few birds and a single vaccination was not effective in reducing transmission. The most likely explanation for the reduced efficacy in comparison to layers was the presence of maternally derived antibodies at time of vaccination. MDA in general persist in broilers for approximately 14-21 days after hatch , and for AI it has been demonstrated that MDA titers were low at 7 days after hatching . It has been demonstrated for other viral infections that MDA may interfere with an effective immune response [36, 37], like for example for Newcastle disease  and infectious bronchitis . Whether there was interference between MDA and vaccination could not be demonstrated in this study, as no group of broilers without MDA was included, because this type of broilers is not present in endemically infected areas in which broiler breeders are vaccinated regularly, and because it was not the research question of this study.
A study on AI vaccination in MDA-positive broilers  demonstrated that HI titers remained high until five weeks post vaccination, indicating that a decrease in immunity was not to be expected within the interval vaccination-challenge applied in our trials. Studies have also shown that HI levels after a single vaccination in birds with MDA reached a peak at six weeks post vaccination [15, 35], suggesting our interval being too short. We challenged at 28 days of age, however, as this was assumed to be a reasonable interval for field conditions and also considering the duration of the experiments and applied before .
Another explanation for the failing immune response is that broilers have an immature immune system, as broilers have been bred for growth characteristic, which may have an effect on both the humoral and cellular immune responses . Although the birds in our experiments were not protected, other studies have shown that AI vaccination of broilers at 10 days of age gave satisfactory antibody titers and clinical protection after challenge [35, 42, 43]. A possible explanation for the difference between their observations and ours is the use of different vaccines or adjuvantia.
In our experiment, some birds did survive the infection, although the HI titers at challenge were below 32. In the unvaccinated experiments with 4-week-old birds, for example, all inoculated birds shed virus and showed clinical sign, but some contact-infected chickens survived for more than 10 days. This was also observed in the vaccine experiments and DOC experiment. However, no association could be found between HI titer at moment of challenge and protection against contact-infection and also not between HI titer and infection after inoculation. This phenomenon has also been observed in studies on for example H7N7 in turkeys , and on Newcastle disease . In the latter experiment, vaccinated birds with low or undetectable antibody titres were protected against disease and mortality, but infection and transmission still occurred. One explanation for surviving of contact birds is that these birds had become infected with a low virus dose. As the birds were housed in pairs, the exposure dose could have been low, as some of the inoculated chickens in this experiment died before having shed a large amount of virus. However, Spekreijse et al.  showed that the case fatality rate of chickens did not differ between dose groups, and in their experiments all birds that were infected died eventually. Another explanation is that other immune responses than antibodies were induced, such as a cellular immunity, although it is generally assumed that inactivated vaccines usually induce a B cell response only . Another possibility is that the low HI titers may have been sufficiently high to induce some protection.
Of course extrapolation of results from experiments to the field should always be done with caution. Nevertheless, our results indicate that vaccination is not effective in broilers, as early vaccination does not induce a good immune response and if vaccination is applied later, the birds may be protected the moment they are slaughtered. Henning et al.  demonstrated that the risk of infection was higher in flocks vaccinated once, in comparison to two vaccinations, also suggesting that vaccination of broilers will not be very effective, but it does not seem to be feasible vaccinating broilers even twice. Although discrepancies between laboratory and field results have been observed more often , the efficacy of vaccination is usually higher under experimental conditions than under field conditions. Therefore, a proper vaccination scheme with killed vaccines seems useless, although use of other types of vaccines [17, 33] could be more successful. Adequate biosecurity measures should therefore be implemented in endemically infected countries to control AI [2, 14, 15, 17, 50].