Deplatform Disease

View Original

Did SARS-CoV-2 cause the decline of influenza through viral interference? It seems unlikely.

The gist: Some have taken note of the disappearance of influenza with this pandemic, and argued that this is because of viral interference from SARS-CoV-2: in essence, people are becoming infected with SARS-CoV-2 and therefore rendered incapable of catching influenza. Some are taking it so far as to suggest that this explains the decline of flu rather than non-pharmaceutical interventions (NPIs) like masking, distancing, ventilation, humidity, handwashing, etc. The problem is: there’s really no evidence to support that. Experimentally, influenza seems to enhance infection of cells by SARS-CoV-2, and worsens outcomes in mice. Additionally, the rapid decline of influenza occurred before SARS-CoV-2 was widespread, weakening this argument. In epidemiologic studies from early in the pandemic, superinfection of both viruses seemed to be a very common occurrence in one site, arguing against significant interference. And this also doesn’t explain why other respiratory viruses also underwent significant decline during the COVID-19 pandemic (unless SARS-CoV-2 interferes with them too). The simplest explanation is that the NPIs crushed it. It’s possible that there is a role for interference between influenza and SARS-CoV-2 to explain the decline of the former, but in all likelihood it’s very minor if it’s there at all. Still, virus interference is an interesting subject to consider in the context of public health emergencies and should be explored further.


You can see an inverse relationship between the incidence of rhinovirus infections and the incidence of influenza infections recorded from July 2016 to June 2019 at Yale New Haven Hospital. Influenza seems to peak around the time that rhinovirus declines; for rhinovirus a similar trend is noted but it seems less marked (seasonality seems to be weaker for rhinovirus than influenza). Note however that these data are retrospective and correlational so caution is needed in their interpretation. Wu, A., Mihaylova, V. T., Landry, M. L. & Foxman, E. F. Interference between rhinovirus and influenza A virus: a clinical data analysis and experimental infection study. Lancet Microbe 1, e254–e262 (2020).

There seems to be a popular claim going around that the decline in influenza during the 2020 season is the result of viral interference, which is fascinating. Unfortunately it is being used to argue against the necessity of non-pharmaceutical interventions as means to control the spread of respiratory pathogens, which is much less fascinating and dishonest. I think that the source of this claim is likely this article from STAT which discusses the concept generally. In short, the concept refers to the propensity of having an infection with one virus disrupting your ability to become infected with another one. It’s a fairly well documented phenomenon at the level of cell culture, and it does make some intuitive sense. The immune system has a general network of pattern recognition receptors which overlap greatly for viruses, and viruses rely very strongly on their host cell for resources to replicate (they are parasites after all), which is why so many of these innate defenses converge on interferons. In particular RLRs seem to be very important sensors of viral nucleic acids as they are capable of recognizing the nucleic acids of viruses from every group in the Baltimore classification except for single-stranded DNA viruses (SARS-CoV-2 does indeed interact with RLRs and has machinery to disrupt their downstream interferon induction; a more global view of the innate immune response to SARS-CoV-2 can be found here). Interferons do a great many things, but among them, they rewire the metabolism of the cell to disrupt protein synthesis and as a result, the viruses can’t replicate their own proteins. I suppose that there’s a point of caution there about the terminology- it’s not so much that the first virus interferes with the second virus, but rather that it triggers a response from the host that makes them much more difficult to infect. Thus one could imagine a relatively simple mental model: virus 1 infects the cell, replicates some, triggers the intrinsic antiviral sensing mechanisms that result in production of interferon, and suddenly virus 2 has a really hard time establishing a productive infection in the same cell. A more detailed (though dated) discussion of interference can be found here.

Co-infection (where both pathogens are acquired at the same time) and superinfection (where hepatitis B is acquired first) have different trajectories in hepatitis D virus infection. This is considered the most severe form of viral hepatitis. Negro, F. Hepatitis D virus coinfection and superinfection. Cold Spring Harb. Perspect. Med. 4, a021550 (2014).

Beyond the cell culture level, the concept has therapeutic potential: one could imagine that if infected with a particularly dangerous virus, one could be inoculated with a harmless one (or relatively so) that would in turn disrupt the ability of the other virus to replicate and cause disease. This is known as superinfection therapy (SIT). There does appear to be some evidence of it occurring beyond the cell culture level as an elegant study from Wu et al showed, and it even pointed to interferons as the mechanism of interference. In their study, they performed a retrospective epidemiological analysis showing that there’s an inverse relationship between influenza A infections and rhinovirus infections over several years, and on statistical analysis there is indeed a significant divergence between the superinfection of the two (which is to say, if there were no effect and the phenomenon were solely due to random chance, there would be many more superinfections expected to observed than there actually were). They additionally show that infection of primary human airway epithelial cells with rhinovirus is really disruptive to future infection by influenza A virus, and upon treatment with a drug that blocks the induction of interferon, the effect disappears and co-infection readily occurs. This is really great news because rhinoviruses are relatively harmless, so one could imagine affording people a temporary protection in the setting of a pandemic influenza through deliberate infection with rhinoviruses. Of course, the issue is that you can only keep up production of interferon for so long before its effects become toxic- so that does raise concerns about the potential of superinfection as a prophylaxis. Additionally, the cell culture model is a bit reductive; there could be other components of the immune system mediating nonspecific protection (i.e. trained immunity- though I will note that I have some significant skepticism about this concept’s therapeutic potential beyond what has been established for BCG immunotherapy in bladder cancer; but the intellectually honest thing to do would be to explore it nonetheless). I think an animal model could have been valuable in assessing that- maybe even a human challenge study for the future. Finally, I think it would have been very valuable to see experimentally what happened when you inoculated the cells with influenza before rhinovirus, as, if the hope is a therapy, that would be the typical scenario you would be dealing with- a patient who has severe influenza rather than one in the throes of a brutal rhinovirus infection. That said, my read of the study was not that it was immediately striving for clinical application of these results but rather as an explanation of some key aspects of host-pathogen interaction and seasonality of respiratory viruses, and I think it accomplishes that quite well given the limitations of its retrospective design. However, given that the data are retrospective, and limited to a single site, and given that the supporting experiments are in vitro, we need to be very cautious with our interpretations of the study here. One could imagine these data used to justify a prospective cluster randomized trial where before flu season a large cohort of individuals is recruited to be inoculated with a rhinovirus and at another site they receive a sham control and comparing the incidence of influenza and its associated complications.

Some examples of how viruses avoid host cell induction of interferon. García-Sastre, A. Ten strategies of interferon evasion by viruses. Cell Host Microbe 22, 176–184 (2017).

I think you all know me well enough by now though to know there’s a “but” coming. The thing is, the picture presented until now might lead you to forget that superinfection and co-infection (when infection with both the pathogens occurs at the same time as opposed to one and then the other as in superinfection) do happen, and it’s typically not a good thing. Influenza for instance is known for being complicated by bacterial pneumonia (one of my favorite reviews- yes, I have favorite reviews. I’m a giant nerd)- though you could reasonably demur that this is quite different from two viruses superinfecting the same host, given that interferon can have more complex roles in the course of bacterial infection. Superinfection can take many complex forms though. For instance, patients who have hepatitis B virus infection can develop a much more severe form of the condition if they also acquire the hepatitis D (delta) satellite; hepatitis D cannot replicate unless hepatitis B is also present within the cell, and it leads to a more severe disease (though the form it takes depends on whether superinfection or co-infection occur, emphasizing the importance of timing). The thing is, the existence of viral superinfections that make things worse is kind of an important challenge to SIT. For one thing, given how broadly interferon acts to prevent viral infections, based on what we have thus far established, how could it even happen? Well, therein is the mess: interferon is such a powerful and ubiquitous mechanism of host defense, that for any viral infection, there needs to be some way of getting around it if productive replication is to occur. One of the most obvious ways is simply to hide. Coronaviruses for example can avoid detection of their nucleic acids (which would then trigger interferon signaling) by using double-membraned vesicles. Alternatively, the virus could take an active role in suppressing the induction of interferon-stimulated genes (ISGs) or interferon itself. Orf9b of coronaviruses disrupts key signaling at the mitochondria that can induce interferon production. We could go on- the point is viruses have to be really good at getting around interferon to function as viruses. A recent preprint even suggests that this is why B.1.1.7 variants are both more lethal and more transmissible than other variants of SARS-CoV-2: they are better at getting around the innate antiviral immunity we have. The fact that viruses can suppress interferon creates a problem because this would indicate some cells are at less than their baseline level of vulnerability in the course of a viral infection because there are processes at work actively preventing them from inducing their antiviral protective mechanisms. The thing is that the suppression of interferon, realistically, cannot be permanent (in fact the preprint about B.1.1.7 I just linked suggests that the delay in its production and then subsequently turning it on contributes to spread of the virus by activating an inflammatory cascade that prompts the host to cough, thereby allowing it to spread), so this suggests that there is actually a finite window of increased susceptibility to superinfection, reinforcing that whole theme of “timing is everything.”

Preinfection with influenza results in worse outcomes in mice and enhanced SARS-CoV-2 replication in the respiratory tract. Bai, L. et al. Coinfection with influenza A virus enhances SARS-CoV-2 infectivity. Cell Res. 31, 395–403 (2021).

Early on in the pandemic I remember some of my friends who treated patients describing those who had both influenza and COVID-19 at the same time (sometimes they even had influenza A, B, and COVID-19 at the same time), which sounds… just really terrible. There was also concern for the possibility of a twin-demic: influenza and SARS-CoV-2/COVID-19. That thankfully did not pan out. There was however a center in Wuhan which noted shockingly high rates of superinfection of influenza and SARS-CoV-2 (57.3 %), and in particular superinfection with influenza B and SARS-CoV-2 was associated with worse outcomes (studies are not entirely consistent on the effect this has on patient outcomes). Of course, that result needs to be interpreted with a lot of caution: it’s a single-center study, and it’s retrospective, but it did comprise the largest sample of superinfected patients that I could find in the literature. You might also expect the analysis of these results to be subject to some important confounders, e.g. individuals who are superinfected with both viruses might have poorer health at baseline or immunological deficiencies that would make the superinfection more serious. It’s hard to assess the meaning of these results without matching patient to singly infected patients. The study does also allude to the fact that COVID-19 mortality was noted to be lower in regions where uptake of influenza vaccines was higher, which is interesting but also likely not a reliable basis for making firm conclusions. A helpful starting point would be to examine how the two viruses interact in a cell culture system, which fortunately Bai et al did. Influenza A virus very clearly enhanced the ability of SARS-CoV-2 to infect multiple cell lines. They further validated this in a mouse model where some mice received both SARS-CoV-2 and influenza (influenza first 2 days earlier) and others received only SARS-CoV-2 and sacrificed 2 days after SARS-CoV-2 infection:

The lung histological data in Fig. 2e further illustrate that IAV and SARS-CoV2 [superinfection] induced more severe lung pathologic changes, with massive cell infiltration and obvious alveolar necrosis, compared to SARS-CoV-2 single infection or mock infection.

They called it coinfection in the paper but it didn’t occur at the same time, so it’s technically superinfection. They then demonstrate that at least part of this explainable by the fact that influenza infection enhances expression of ACE2, which is the receptor for SARS-CoV-2, and an interferon-stimulated gene (ISG). Bai et al also showed that using A549 cells, influenza virus, but not HPIV, HRSV, nor HRV3 enhanced infection by SARS-CoV-2.

The principal issue I have with this study is it examined what happens if you have influenza and then SARS-CoV-2 but not the reverse.

Still, that’s a lot of evidence that influenza + SARS-CoV-2 = not good. Admittedly, it’s mouse evidence and cell culture evidence and we all know the adages, but at a minimum it is clearly not consistent with a viral interference between influenza and SARS-CoV-2, and epidemiologic evidence suggests significant potential for superinfection. There’s also another piece (emphasis mine):

The findings in this report are subject to at least four limitations. First, an ecologic analysis cannot demonstrate causality, although the consistency of findings across multiple countries is compelling. Second, other factors, such as the sharp reductions in global travel or increased vaccine use, might have played a role in decreasing influenza spread; however, these were not assessed. Third, viral interference might help explain the lack of influenza during a pandemic caused by another respiratory virus that might outcompete influenza in the respiratory tract (10). This possibility is less likely in the United States because influenza activity was already decreasing before SARS-CoV-2 community transmission was widespread in most parts of the nation. Finally, it is possible that the declines observed in the United States were just the natural end to the influenza season. However, the change in the decrease percent positivity after March 1 was dramatic, suggesting other factors were at play.

Young et al. also contributed a letter:

Petersen, E. et al. Comparing SARS-CoV-2 with SARS-CoV and influenza pandemics. Lancet Infect. Dis. 20, e238–e244 (2020)

The precise mechanisms explaining the rapid decline in influenza cases this year may be multifactorial, but we hypothesise that the implementation of public health measures such as social distancing, the use of protective face masks and frequent hand washing following the outbreak of COVID-19 might have contributed to our observation. It is possible that fewer influenza tests were performed this year due to the focus on COVID-19. However, in China, influenza A and B were typically tested in suspected COVID-19 cases with flu-like symptoms [2]. In addition, in the USA, the WHO reported high levels of influenza testing after the outbreak of COVID-19, with a similar number of influenza tests at the peak and 7 weeks after the peak [3]. We also cannot exclude the influence of viral interference (an immune response to one viral infection prevents a second infection) in limiting influenza spread, although co-infection with influenza and SARS-CoV-2 has been reported [4].

In short, it seems doubtful from my vantage point that viral interference explains the decline of influenza to any significant extent. Certainly, the existence of interference does not necessarily mean that no coinfections would occur, but given all the non-pharmaceutical interventions that were implemented it’s not hard to see why influenza disappeared. This part is where people usually counter, accusatorily, with why it didn’t make COVID-19 disappear too then. Well:

  1. Seasonal influenza is less transmissible. It has an R0 of 1.3 most years.

  2. It also has a much shorter incubation period.

  3. Presymptomatic/asymtomatic spread of influenza is much more limited.

  4. We also know that SARS-CoV-2 spread is driven by overdispersion which isn’t true of influenza.

So in short, it’s really not a stretch to suggest that NPIs really did crush influenza. After all, these were observed with multiple endemic respiratory viruses, like RSV, as well; while one could argue for a universally interfering effect of SARS-CoV-2 across all these respiratory viruses, that requires this pesky thing called evidence. The role of viral interference between SARS-CoV-2 and influenza remains to be validated, but I see no compelling evidence for it, and some lower quality evidence against it.

References

1.Wu, A., Mihaylova, V. T., Landry, M. L. & Foxman, E. F. Interference between rhinovirus and influenza A virus: a clinical data analysis and experimental infection study. Lancet Microbe 1, e254–e262 (2020).

2.Greenwood, V. et al. A viral mystery: Can one infection prevent another? Statnews.com https://www.statnews.com/2021/01/31/a-viral-mystery-can-one-infection-prevent-another/ (2021).

3.Whitaker-Dowling, P. & Youngner, J. S. Viral interference-dominance of mutant viruses over wild-type virus in mixed infections. Microbiol. Rev. 51, 179–191 (1987).

4.Bowie, A. G. & Unterholzner, L. Viral evasion and subversion of pattern-recognition receptor signalling. Nat. Rev. Immunol. 8, 911–922 (2008).

5.Sadler, A. J. & Williams, B. R. G. Interferon-inducible antiviral effectors. Nat. Rev. Immunol. 8, 559–568 (2008).

6.Sparrer, K. M. J. & Gack, M. U. Intracellular detection of viral nucleic acids. Curr. Opin. Microbiol. 26, 1–9 (2015).

7.McNab, F., Mayer-Barber, K., Sher, A., Wack, A. & O’Garra, A. Type I interferons in infectious disease. Nat. Rev. Immunol. 15, 87–103 (2015).

8.van Puffelen, J. H. et al. Trained immunity as a molecular mechanism for BCG immunotherapy in bladder cancer. Nat. Rev. Urol. 17, 513–525 (2020).

9.Lowery, S. A., Sariol, A. & Perlman, S. Innate immune and inflammatory responses to SARS-CoV-2: Implications for COVID-19. Cell Host Microbe (2021) doi:10.1016/j.chom.2021.05.004.

10.Hepatitis D. Who.int https://www.who.int/news-room/fact-sheets/detail/hepatitis-d.

11.Negro, F. Hepatitis D virus coinfection and superinfection. Cold Spring Harb. Perspect. Med. 4, a021550 (2014).

12.DiNardo, A. R., Netea, M. G. & Musher, D. M. Postinfectious epigenetic immune modifications - A double-edged sword. N. Engl. J. Med. 384, 261–270 (2021).

13.García-Sastre, A. Ten strategies of interferon evasion by viruses. Cell Host Microbe 22, 176–184 (2017).

14.McCullers, J. A. The co-pathogenesis of influenza viruses with bacteria in the lung. Nat. Rev. Microbiol. 12, 252–262 (2014).

15.Olsen, S. J. et al. Decreased influenza activity during the COVID-19 pandemic - United States, Australia, Chile, and South Africa, 2020. MMWR Morb. Mortal. Wkly. Rep. 69, 1305–1309 (2020).

16.Thorne, L. G. et al. Evolution of enhanced innate immune evasion by the SARS-CoV-2 B.1.1.7 UK variant. doi:10.1101/2021.06.06.446826.

17.Bai, L. et al. Coinfection with influenza A virus enhances SARS-CoV-2 infectivity. Cell Res. 31, 395–403 (2021).

18.Ziegler, C. G. K. et al. SARS-CoV-2 receptor ACE2 is an interferon-stimulated gene in human airway epithelial cells and is detected in specific cell subsets across tissues. Cell 181, 1016-1035.e19 (2020).

19.Young, G. et al. Rapid decline of seasonal influenza during the outbreak of COVID-19. ERJ Open Res. 6, 00296–02020 (2020).

20.Biggerstaff, M., Cauchemez, S., Reed, C., Gambhir, M. & Finelli, L. Estimates of the reproduction number for seasonal, pandemic, and zoonotic influenza: a systematic review of the literature. BMC Infect. Dis. 14, 480 (2014).

21.Standl, F., Jöckel, K.-H., Brune, B., Schmidt, B. & Stang, A. Comparing SARS-CoV-2 with SARS-CoV and influenza pandemics. Lancet Infect. Dis. 21, e77 (2021).

22.Petersen, E. et al. Comparing SARS-CoV-2 with SARS-CoV and influenza pandemics. Lancet Infect. Dis. 20, e238–e244 (2020).

23.Sneppen, K., Nielsen, B. F., Taylor, R. J. & Simonsen, L. Overdispersion in COVID-19 increases the effectiveness of limiting nonrepetitive contacts for transmission control. Proc. Natl. Acad. Sci. U. S. A. 118, e2016623118 (2021).

24.Feng, L. et al. Impact of COVID-19 outbreaks and interventions on influenza in China and the United States. Nat. Commun. 12, 1–8 (2021).

25.Baker, R. E. et al. The impact of COVID-19 nonpharmaceutical interventions on the future dynamics of endemic infections. Proc. Natl. Acad. Sci. U. S. A. 117, 30547–30553 (2020).