Monday, 26 August 2013

Editor's rant: Why testing the few may not benefit the many...

Prospective screening without regard for whether the person is sick. That's what I think we need more of, in order to truly understand respiratory viruses and acute respiratory infections (ARIs).

And I
 don't just mean the scary ones like MERS-CoV or influenza A(H5N1) virus or H7N9 or H7N7 (zoonotic flu). 

I also mean the rhinoviruses, influenza A(H3N2) virus, H1N1 (seasonal flu), endemic coronaviruses (CoV; 229E, OC43, NL63, HKU1), metapneumoviruses (MPV; I use the plural because there are 4 genotypes and who knows how many immunogenetically distinct clades), respiratory syncytial viruses (RSV; same plural), adenoviruses, enteroviruses, Saffold viruses, parechoviruses, polyomaviruses, bocaviruses...etc.

Sure, there have been studies in birth cohorts and in the community among "normal"'healthy people. But they sometimes have limitations that may blur our view of what is really happening in the community. For example, such studies may:

  • Exclude certain diseases that viruses are involved in.
  • Only sample when there are signs and symptoms of disease. 
  • Only call "disease" when a certain number or combination of symptoms are present (this one irks me no end - pedantically, if your body deviates from its physiological norms, you are diseased).
  • Sample too infrequently to catch whats going on between sampling points. We don't get one virus, recover, then get another - we're a virus's favourite hang out - but because of our awesome immune system, only some of those infections make us ill enough to stop and groan.
  • Employ insensitive detection methods (cell, tissue or organ culture). In fact, if a study used culture you might as well ignore those data - they will have missed many fastidious viruses (those that don't grow easily or in the cell lines used) rendering any conclusions associating detection of a virus and disease weak or wrong.
  • Sample for too short a period or just focus on a particular season etc. 
  • Only include a pet virus or a few viruses or just those viruses known about at the time. 
Much of our understanding of each virus comes from hospital-based studies. People in this environment, whether admitted (inpatients) or presenting but being allowed back home (outpatients), represent the "tip of the iceberg" of the disease spectrum. The pointy end. The most severe cases. We may make the assumption that the viruses circulating in the community are represented by what's happening in a hospital environment - or vice versa. But how often have we tested that? Do we know if there is a lag or lead time? Does it differ by climate? Could we go further and perhaps use those numbers to predict what the burden of disease in hospital will be this "season"?

And then there's the ongoing testing issue. Research dollars generally do not fund epidemiology. Certainly not ongoing epidemiology. Even big hospitals and private testing labs cannot afford the personnel and cost of testing all respiratory samples for all "likely" viral pathogens, all the time. "Likely" having been defined with the caveats above. 

And so our epidemiology data have holes. Big ones. We read of complaints about some countries not being able to identify a viral/bacterial cause (not that a POS lab test does prove cause) of pneumonia or encephalitis...but many patients in more "developed"countries also leave hospital without ever being attached to a lab-confirmed positive result. We could reduce that, even if we could not specifically treat them. And therein lies another issue. We test for some viruses based on historical precedent - do those precedents accurately stand up today? Do we even have the data to answer that? If you are a health professional, have a look at what your local testing lab offers - does it cater for the most likely causes of ARI or just what's been used before? 
Click image to enlarge. Respiratory virus infections among the community
and in hospital-based populations. Generally more males than females
present to hospital  with clinically-defined acute respiratory infections
(ARIs). Infections in the hospital setting are shown in red, those in the
community in orange.  Most ongoing virus testing is from
hospital-based populations as is our contemporary
understanding of viral season.

Notifiable viral diseases are kept track of, and if they occurred by themselves without interaction with, or interference from, other viruses that might be enough. But they don't. 

The "One World" concept of infectious disease study - looking at animals and humans and the environment together - is great; what about the concept of "One Virus"? The days of a study looking at just one virus and from that, without testing for any other respiratory virus that may cause the same signs and symptoms, concluding what that virus is capable of, it's severity and how many cases of disease are lessened by a drug for it, in a human should be far behind us. But they are not. 

So, do we really know the viruses that call us home? And if the answer is no, how can we possibly hope to be prepared for the next virus that emerges, the next local viral outbreak of ARI or encephalitis or gastroenteritis, or the next pandemic? How can we protect our population from viral threats if we're always on the back foot?

If I ruled the world, we would do more testing we'd try not to bias our attentions toward any 1 virus, we'd screen everything for everything, we'd sample the community, we'd make the data publicly available in real time, we'd understand ARI epidemiology better and we'd use all those data to prioritize some antiviral drug development or other viral interventions. At the very least, we'd re-jig our testing panels and create a new paradigm or 2.

But I don't rule the world - nor do I have input into these sorts of decisions - perhaps you do?

Feel free to weigh in below.