|AFM Surveillance United States 2014-2017 - Credit CDC|
Acute flaccid myelitis (AFM) is a rare polio-like illness that affects a person’s nervous system - specifically the spinal cord - and is characterized by sudden weakness in one or more arms or legs, along with loss of muscle tone and decreased or absent reflexes.
AFM is a subset of conditions that fall under a broader `umbrella' of syndromes called Acute Flaccid paralysis (AFP), which may include myelitis, peripheral neuropathy, myopathy, Guillain-Barré syndrome (GBS), toxic neuropathy, and other muscle disorders.
While the exact causes of Acute flaccid myelitis aren't fully understood, it has been linked to a number of viral infections, including West Nile Virus, Adenoviruses, and a number of (polio and non-polio) enteroviruses, including EV-71 and more recently, EV-D68.A nationwide outbreak of EV-D68 was linked to a concurrent spike in AFP cases across the United States in the fall of 2014, and while a pretty strong circumstantial case has been made over the past couple of years (see EID Journal Enterovirus D68 Infection in Children with Acute Flaccid Myelitis, Colorado, USA, 2014), no definitive causal link to EV-D68 has been established.
Since then we've seen a number of EV-D68 outbreaks around the world (see here, here & here) where a concurrent spike in AFM cases have also been reported.Today the ECDC's Eurosurveillance Journal has a review of the literature, and an examination of the available evidence using the Bradford Hill criteria, that attempts to establish a potential link between EVD68 infection and AFM.
The authors found that analysis based on the Bradford Hill criteria suggested good evidence for EV-D68 being a cause of AFM. Today's review is a lengthy, detailed one, which most of my readers will want to read it in its entirety.
The link and a few small excerpts follow:
The association between acute flaccid myelitis (AFM) and Enterovirus D68 (EV-D68) – what is the evidence for causation?(EXCERPT)
Amalie Dyda1, Sacha Stelzer-Braid2,3, Dillon Adam1, Abrar A Chughtai1, C Raina MacIntyre1,4
Historically, the incidence of EV-D68 has been low, with sporadic cases and small clusters of mild illness reported. Whether this represents under-ascertainment or true low incidence in the past is unclear, but active EV surveillance studies in several countries including Germany in 2013-14, Hong Kong in 2014, France in 2014 and China in 2011-15 suggest that EV-D68 was a rare cause of clinical infection in the past [18-21]. Since 2014, the number of reported infections and clusters has increased. In addition, severe complications including AFM have been reported since 2014 . Several clusters of AFM in recent years were associated with EV-D68 and a large outbreak of EV-D68 in 2014 in the US was associated with severe respiratory illness.
Our application of the Bradford Hill criteria suggested good evidence for EV-D68 being a cause of AFM. While EVs in general are neurotropic, AFM has never previously been associated with EV-D68. It could be that incidence of EV-D68 was genuinely been much lower in the past, so that rare complications of infection have not been apparent. An analogous example is the association between Zika virus and microcephaly which was only recognised during a large-scale epidemic in Brazil in 2015 [62,63]. However, retrospective analysis of a large outbreak of Zika virus in French Polynesia two years earlier showed the same association with microcephaly but was not recognised at the time . In addition, clades A, B and B1 of EV-D68 were highly neurovirulent in animal studies, with specific tropism for motor neurons . It appears that these strains which evolved after the year 2000 are capable of causing AFM, as demonstrated in a mouse model, while the original Fermon and Rhyne strains do not cause AFM .
There is a need for phylogeographic epidemiology to ascertain temporal and geographic changes in the virus and whether such changes could explain why AFM is newly associated with the virus. Genetic changes in the virus which have rendered it more neuropathic could explain the association, and several clades have been shown to be highly neurovirulent. Phylogenetic studies have demonstrated that strains isolated in recent outbreaks are very divergent from the original Fermon strain isolated in 1962 . Clades B1 and B2 caused the 2014 outbreak  and clade B3 caused an outbreak of severe EV-D68 infection in the Netherlands in 2016 . Strains in clade B1 have mutations in structural and non-structural proteins, which could play a role in the reported neurovirulence of these strains , and all EV-D68-infections in human AFM cases were attributed to clade B. However, mouse studies showed that multiple clades (A, B and B1) cause paralysis . The observation that clade B1 was associated with AFM in 2014 may be due to the much higher incidence of clade B1 infection in 2014. More research is needed to study biological gradient and to quantify measures of association between EV-D68 and AFM.
Given that the association of AFM with EV-D68 is recent, there is a strong case for systematic and enhanced EV surveillance, which will enable investigation of epidemiological data for measures of association. While past studies and EV surveillance showed that EV-D68 was a rare cause of EV infection, there has been a change in disease epidemiology since 2014, including a rise in the incidence of clade B infections. The lack of association between AFM cases and EV-D68 in the US in 2015 and 2016  does not detract from our analysis, as AFM is a clinical syndrome with multiple possible aetiologies. More recent AFM cases could be due to a different aetiology, as other EVs continue to cause AFM, or could reflect the difficulty in isolating the virus from tissue and CSF. The Bradford Hill criteria are a tested and systematic method for evaluating causality and could be applied to other EVs.
(Continue . . .)In summary, the application of the Bradford Hill criteria suggests that EV-D68 causes AFM. AFM has not previously been associated with EV-D68, and a mouse model shows that the original Fermon strain does not cause AFM, whereas the 2014 outbreak strain does . It appears that the incidence of this infection and the clade-specific epidemiology have changed. Phylogeographic epidemiology will further our understanding of the temporal and spatial spread of increasingly neurovirulent clades and improve risk analysis. Further investigation into this relationship is important because of the severity of AFM, ongoing outbreaks of AFM and because there is currently no treatment for AFM related to EV-D68, and no vaccine to prevent infection [24,65].