Norovirus, gastroenteritis, and indoor environmental quality

Recently, I began reviewing the literature on indoor bioaerosol dynamics. Using ISI’s Web of Knowledge, I compiled a bibliography of about 400 articles. I printed the abstracts, assembled them alphabetically, and started reading. However, I only got as far as ‘B’ before becoming sidetracked in reaction to a comment by Beggs and Kerr (2000) that surprised me: ‘… it is now evident that the airborne route is a significant mode of spread in outbreaks of acute viral gastroenteritis.’ Whoa! I’m aware that airborne transmission is key for tuberculosis, measles, and chickenpox and also strongly suspected to play a significant role for influenza, rhinovirus, and severe acute respiratory syndrome (SARS) (Li, 2011). But viral gastroenteritis? Isn’t that issue entirely in the domains of water quality, sanitation, personal hygiene, and food safety? How could airborne transmission play an important role? And, if airborne transmission was indeed significant (combined with the reasonable expectation that such transmission would mainly occur indoors), then would this topic not lie well within the scope of Indoor Air? I returned to the Web of Knowledge to identify literature related to airborne transmission of viral agents that cause gastrointestinal diseases. A substantial body of published work mainly focuses on norovirus (NoV) as a causative agent. My reading of about 50 abstracts plus the full text of about 20 articles persuades me (i) that airborne transmission is indeed an important exposure pathway for acute gastroenteritis caused by NoV; (ii) that the indoor air research community possesses knowledge and tools that could usefully contribute new insight into modes of transmission and means of prevention; and (iii) that, thus far, there has been little overlap between the research communities concerned with NoV-induced illness and those addressing indoor environmental quality. As a noteworthy exception to this last point, an Indoor Air article by Morawska (2006) reviewed the state of knowledge on indoor transmission of viral infections by droplets, including some discussion of Norwalk-like viruses. Morawska mentions gastrointestinal illness as one outcome of concern. It is also noteworthy that while her article has 55 references, only two were published in Indoor Air, and both of these concerned the SARS epidemic. In this editorial, I summarize the evidence concerning airborne transmission of NoV as a cause of acute viral gastroenteritis, the significance of this issue for those concerned with indoor environmental quality, and pose some unanswered questions where I think regular readers of Indoor Air could play a useful role in future research. The illness ‘hyperemesis hiemis,’ or ‘winter vomiting disease,’ was first described by Zahorsky in 1929 (Caul, 1994; Patel et al., 2009). A viral cause was identified about 40 years later through application of immune electron microscopy to stool samples obtained from volunteers who became ill after ingesting bacteria-free fecal filtrates derived from an outbreak at an elementary school (Kapikian et al., 1972). That Norwalk virus, named for the town in Ohio where the school was located, represented the first discovery of what is understood to be a group of related viruses, now classified as the genus Norovirus. In historical literature, these viral agents are collectively referred to as Norwalk-like viruses, Norwalk agents, or small round-structured viruses (SRSVs). The viral particles are approximately spherical with a diameter of 25–30 nm. The genome has been sequenced (Lambden et al., 1993). Very sensitive detection of the virus is now commonly practiced, based on reverse transcription polymerase chain reaction (RT-PCR). That development has greatly advanced investigations of environmental contamination and modes of transmission (Kageyama et al., 2003). Common symptoms of NoV-induced illness include vomiting, diarrhea, and stomach pain. Experimental ingestion by volunteers and outbreak investigations indicate that illness typically occurs within a day or two following exposure. For otherwise healthy individuals, the symptoms generally resolve within a few days after onset. Acute gastrointestinal illness caused by the NoV is a major cause of both epidemic and sporadic disease. ‘Noroviruses are now recognized as being the leading cause of epidemics of gastroenteritis and an important cause of sporadic gastroenteritis in both children and adults’ (Glass et al., 2009). For the United States alone, Mead et al. (1999) have reported that 23 million cases annually might be caused by Norwalk-like viruses, with only 40% being food associated. Musher and Musher (2004) suggest caliciviruses, of which ‘Norwalk-like viruses are the prototype’ might be responsible for 74 million cases of acute gastrointestinal illness annually in the United States. The associated economic burden has not been carefully appraised, but a crude estimate for the United States indicates a minimum impact in the range $3–9 billion per year, simply based on the scale of the economy (gross domestic product = $14 trillion/year), the size of the US population (307 million) and the assumption that each of the 23–74 million illnesses per year causes an average loss of one average person-day’s contribution to the GDP. An important feature of NoV-induced illness is the occurrence of intense outbreaks in which a substantial proportion of occupants of a particular indoor environment become ill during a relatively short period (days to weeks). Studies of these outbreaks provide important information about transmission pathways and risk factors. Published studies present and discuss evidence of airborne transmission and the role of indoor environmental contamination for NoV-caused outbreaks across a broad range of indoor environments (see Table 1). Although most research on viral gastroenteritis has focused on conditions in developed countries, the illness is widespread. According to Patel et al. (2009), ‘Human caliciviruses (i.e., viruses belonging to the genera Norovirus… and Sapovirus) are now recognized as a leading cause of diarrhoea worldwide among persons of all ages.’ Diarrheal diseases are a major factor in the global burden of disease, causing an estimated 2.5 million deaths annually for children under 5 (Kosek et al., 2003). The underlying causes of diarrheal diseases in developing countries are generally regarded to be lack of access to safe-drinking water, inadequate sanitation facilities for managing human waste, and poor hygiene. Might viral agents that are transmitted indoors play a significant contributory role to this global public health problem? Although the modes of infection are not fully understood, airborne transmission has been inferred to occur in many outbreaks. Patel et al. (2009) write, ‘Several characteristics of NoVs facilitate their spread in epidemics including (i) the low infectious dose of NoVs (<10 viral particles); (ii) prolonged duration of viral shedding, even after symptoms resolve, increases the risk of secondary spread; (iii) the stability of the virus in relatively high concentrations of chlorine and at a wide range of temperatures (from freezing to 60°C); and (4) repeated infections can occur throughout life with re-exposure, likely because lack of complete cross-protection against the diverse NoV strains and inadequate long-term immunity.’ Another interesting characteristic of NoV is that the strains that cause human illness have thus far resisted all efforts to grow them in cell culture. It seems astonishing that viral replication that evidently occurs so readily and robustly in vivo cannot be replicated in vitro. Practically, this limitation impairs our ability to study the environmental persistence and influencing factors of NoV infectivity. Vomiting is a common symptom of NoV illness. It sometimes occurs in public spaces, no doubt because symptom onset can be precipitous. Vomitus has been shown to be a carrier of NoV (Greenberg et al., 1979; Wu et al., 2005). Furthermore, many outbreak studies show that being exposed to a vomiting event is a significant risk factor for becoming infected (e.g., Chadwick and McCann, 1994; Marks et al., 2000; Wikswo et al., 2011). It is suggested that vomiting produces aerosol droplets containing viral particles. One transmission pathway entails such droplets being inhaled by exposed individuals, depositing in the upper respiratory tract, and subsequently being swallowed along with respiratory mucus. Alternatively, aerosol droplets produced during vomiting could settle onto indoor surfaces that might then be transferred to hands of exposed individuals through physical contact with the surfaces. Use of those hands for food preparation or eating without first being effectively washed could lead to ingestion of the viral particles, again inducing disease. In a variation of these pathways, inadequately cleaned vomitus from a heavily trafficked flooring surface, such as carpet, might become a secondary source of infectious particles that are liberated into the indoor air mechanically through resuspension. Those particles could either be directly inhaled (pathway 1) or settle onto other surfaces for subsequent hand-to-mouth transmission (pathway 2). This latter chain of events has not been explicitly discussed in the outbreak literature that I reviewed, but does follow from several lines of evidence, including the substantial resuspension of particles by walking on carpet (Cheng et al., 2010), vacuuming (Corsi et al., 2008), and the apparent transmission of an SRSV infection to carpet installers 12 days after the end of an outbreak (Cheesbrough et al., 1997). Not significantly discussed in the NoV outbreak literature, but also potentially important, is the generation of aerosol droplets from toilet flushing (Barker and Jones, 2005). At the peak of illness, the level of excreted NoV can be up to a trillion genomic copies per gram of feces (Atmar et al., 2008). When combined with frequent and watery stools, droplet generation from toilets would seem to pose significant risks of viral dissemination both directly (especially in public toilet rooms with many stalls) and indirectly via surface contamination. The literature I reviewed to prepare this editorial is rich and informative. Reading and reflecting on it taught me much about an important subject that lies within our field of indoor environmental quality and health and yet has not been significantly addressed by members of our research community. Here are some issues where I saw large gaps in knowledge or understanding that we could help fill. First, while the important role of aerosol transmission of NoV is often cited, the supporting evidence is as yet circumstantial. I found no studies that reported efforts to detect NoV in indoor air. Given the small infectious dose combined with episodic release events, it might be difficult to detect NoV by direct filter sampling. However, one interesting possibility to explore would be the use of ventilation system filters as sampling media. Could RT-PCR be applied to detect NoV in the dust collected on used heating, ventilating, and air-conditioning (HVAC) filters operated in a recirculating airstream? This sampling concept was recently described for characterizing bacterial and fungal aerosols indoors (Noris et al., 2011). Second, what are the characteristics of aerosol droplets generated by vomiting that is caused by viral gastroenteritis? The number and size of droplets as well as the extent to which they contain NoV particles are crucially important factors that could affect person-to-person transmission risk. Large droplets remain airborne only briefly, limiting the time for exposure by direct inhalation. Such droplets would settle onto surfaces close to the point of release where they could cause localized fomite contamination that then might be transmitted through direct hand contact or via resuspension. For example, a droplet of 100 μm in diameter with the density of water settles at a speed of 25 cm/s and so would only remain airborne in a room for a period of roughly 10 s. During this time, with typical indoor air speeds of 20 cm/s, the droplet could travel only about 2 m from its source. On the other hand, a unit-density droplet of 10 μm in diameter has a much smaller settling velocity, about 0.3 cm/s. It would remain airborne for a few minutes following release and consequently have the potential to travel distances of 10 m or more from its release location. Deposition for a 10-μm particle would remain a stronger fate than removal by ventilation, and so, contamination of upward-facing surfaces by droplet settling would be important over a broader area than for 100-μm droplets. Finally, if droplets containing viral particles were produced with a diameter of 1 μm or less, they would settle slowly and mostly remain airborne until the room air was filtered or exchanged by ventilation with outdoor air. These small particles might maximize the potential for direct transmission by inhalation but be associated with greatly reduced transmission by means of droplet contamination of fomites. A third area with important research needs is in viral particle release from contaminated fomites. Here, a particular concern would be fibrous surfaces that have become contaminated with feces or vomitus from an infectious individual and that are subject to mechanical disturbance. Specific examples include soiled bedding and clothing, and also carpeting, especially in areas with high traffic. Can infectious virus-containing particles be released from these materials and become airborne? What are the particle sizes and release rates? Note that although the viral particles are themselves tiny (25–30 nm), the size of airborne particles or droplets containing these viruses could be much larger, containing materials other than just the virus. Better understanding of mechanisms of transmission is important for many purposes. To what extent can concerned individuals limit their exposure risk by wearing respiratory masks in public spaces in addition to practicing careful hand hygiene? To what extent do building factors such as ventilation rate and HVAC filter efficiency influence the transmission and outbreak risk? Can practical measures be developed and adopted to reduce the frequency or intensity of epidemic outbreaks? Can the overall rate of sporadic illness caused by NoV be moderated by better indoor environmental hygiene? Is gastroenteritis caused by NoV an appropriate topic for Indoor Air? I think so. Why? There is no doubt that the illness is itself of major concern and also that it is transmitted most commonly in indoor environments. Those two criteria might not by themselves be sufficient without being able to answer ‘yes’ to one of these questions. (i) Do indoor environmental conditions influence the spread of the disease? (ii) Is it possible to reduce the rate of spread of this disease through changes in the design, construction, operation, and maintenance of buildings (and of vehicles that transport humans)? Although not yet conclusive, I think that there is sufficient evidence to make a case for affirmative answers to both of these questions. To achieve a firmer basis for the answers, we must develop a better understanding than we have today of the transmission mechanisms for NoV and of the influence of indoor environmental conditions on those mechanisms.