Following our recent interview with Prof Pablo Murcia, Professor of Integrative Virology at the University of Glasgow (Click here to watch). We were contacted by a colleague of Dr David Marlin from his Animal Health Trust (AHT) days, Dr Richard Newton (BVSc MSc PhD FRCVS), who is now the Director of Epidemiology & Disease Surveillance at Cambridge University Vet School. Richard is an expert in equine infectious diseases and he and David worked together on a number of studies at the AHT. We feel you may find his insight into bird flu and horses interesting!
Bird flu virus infections in horses – heeding some scientific caution with recent findings of antibodies in a small number of horses in Mongolia
Dear David
We followed with interest the recent thread and interview with Professor Pablo Murcia referring to his group’s recently published work on detecting apparent immune responses against the bird flu virus known as H5N1 in a few horses in Mongolia. We strongly agree with Pablo’s urging for ongoing vigilance and indeed our group, Equine Infectious Disease Surveillance (EIDS), based at the University of Cambridge and generously funded by the British horseracing and breeding industry, plays a vital and ongoing role in identifying and raising awareness about different emerging equine disease risks. We fully acknowledge the critical importance of the possibility of bird flu viruses emerging among horses, as history attests it having occurred previously. For example, this phenomenon readily explains the sudden emergence in North America in 1872 of an equine epidemic known as the ‘great epizootic’. Although at that time influenza viruses had not yet been discovered, the ‘great epizootic’ behaved just like equine flu – in particular, the signs were severe but clinically typical and the disease spread rapidly through horses travelling on the railroad network between major cities across the American continent, equivalent to today’s long-distance movement of horses and their infections by road and air transport.
However, the reason for contacting you is in order that your members might also understand that some in the scientific community are exercising some caution regarding the evidence from Pablo’s work, albeit we recognise the significance of the threat posed by cross-species transmission of influenza viruses between birds and other species, including horses. A letter to this effect has recently been published online by the Equine Veterinary Journal (EVJ). So, whilst recognising the threat we feel that it is important to consider the following points in order that the most balanced and informed perspective might be taken on where horses presently sit on the spectrum of risk of cross-species transmission of H5N1 bird flu virus.
- The limited seropositivity findings actually provide some reassurance
There were 2160 blood samples included in the study and having applied several tests for detecting antibodies against influenza viruses, it was inferred that 99.6% of the samples were negative by one test (nine positives) and 99.9% were negative by a different test (two positives). As the two positive samples by the second test were from different herds, in different Mongolian provinces, in different years, this argues against there having been horse-to horse transmission. - Acknowledging the potential for false positive results in serological testing
It is generally accepted that no diagnostic test is 100% accurate all of the time and that should always be be considered when evaluating results from studies applying diagnostic assays, as there may be the possibility of false positive and false negative results. What this means is that some samples will return positive test results when they are in fact negative (i.e. false positives) and vice versa, false negative test results returned on samples that are in fact positive. The extent to which accurate results (and by simple subtraction false results) are likely is usually known for a diagnostic test and these measures are commonly referred to as the ‘sensitivity’and ‘specificity’ of the test. In simple terms, because the number of positive H5 antibody test results was so low it is possible that even with high test specificity, that the positive results reported may not truly confirm that these horses were exposed and mounted an immune response to an H5 influenza virus. - Only low-level antibody titres were reported on single samples
One of the antibody tests that was used provided a figure (known as a ‘titre’, based on how much the sample could be diluted whilst still detecting presence of antibody) for how much antibody was detected in two positive samples. This reported‘titre’was though quite low compared to the amount of antibody that would usually be expected in an animal that had mounted an immune response against an influenza viral infection. It was not possible in this sort of cross-sectional study to provide the best evidence based on measuring antibody levels as a firm indicator of response to a viral infection – this would involve taking two samples from the same animal, the
first taken before or at the time of the infection (known as an acute sample) and the second after recovery or clearance of the infection (convalescent sample). With this approach a marked increase in antibody levels seen between acute and convalescent samples in the same animal is a firm indication of the animal responding immunologically to the infection. - Lack of epidemiological analysis to contextualise the positive antibody findings
The horses with positive antibody findings had no history of observed clinical signs and although additional information on the animals was collected, no results were provided to put the antibody findings in any sort of context. Although relating antibody findings backwards to prior events can be very difficult, sometimes this approach can provide interesting insights into possible previous events. One example, nearly 18 years ago now, was our investigation of a respiratory disease outbreak in a pack of English foxhounds in which evidence of antibodies to H3N8 equine influenza was discovered (37.5% of 72 hounds were antibody positive), consistent with prior viral exposure. Influenza as the cause of the investigated outbreak in 2007 was soon ruled out, but simply comparing the dates of birth/age of the antibody positive and antibody negative hounds, indicated that the time of exposure must have been several years before, most likely in spring 2003, when equine flu was known to be actively circulating in horses in the UK.
So, we caution that a finding of a very low proportion of horses positive for antibodies to H5 influenza virus may not represent the optimal evidence for confirming cross-species transmission and heightened additional transmission risk from horses. False conclusions as to the extent of this having already occurred may have unwelcome unintended consequences for the equine industry. Further work is required to clarify the current risks, as well as the need for heightened surveillance and testing of cases of acute respiratory disease in horses, in particular including well vaccinated animals.
Richard Newton
Equine Infectious Disease Surveillance (EIDS), Department of Veterinary Medicine,
University of Cambridge, UK
Watch the original video interview or listen to the podcast with Prof Pablo Murcia one of the researchers behind the study that found horses that have been infected with H5N1 – avian flu. HERE
