«Etiology of the common cold: Modulating factors William J. Doyle1 and Sheldon Cohen2 1Universityof Pittsburgh, 3000 Mt Royal Blvd, Glenshaw PA 15116, ...»
ed. by R. Eccles and O. Weber
© 2009 Birkhäuser Verlag Basel/Switzerland
Etiology of the common cold: Modulating factors
William J. Doyle1 and Sheldon Cohen2
1Universityof Pittsburgh, 3000 Mt Royal Blvd, Glenshaw PA 15116, USA; 2Department of
Psychology, Carnegie Mellon University, Pittsburgh, PA 15213, USA
The development of a “cold-like illness” (CLI) usually requires infection with an upper
respiratory virus such as rhinovirus, influenza virus, respiratory syncytial virus, parainflu- enza virus, coronavirus or adenovirus, among others, and the development of sufficient signs, symptoms and pathophysiologies to qualify as being ill based on personal and cul- tural definitions. A viral upper respiratory tract infection (vURTI) in the absence of overt illness (subclinical vURTI) will not be made manifest to the individual or to observers and, therefore, will not be diagnosed as a CLI. The degree of illness occurring during a vURTI is directly related to the extent of provoked inflammation, which in turn depends on the engagement of antiviral defense systems. Thus, risk factors for CLI can modulate either the vURTI risk by affecting virus exposure and/or susceptibility to infection, or the CLI risk given a vURTI by affecting immunocompetence, the provoked inflamma- tion and/or the interpretation of illness as a CLI. In this chapter, we review published studies for evidence of CLI risk-modulating factors and report that climate, crowding and perhaps female gender can affect the probability of exposure to vURTI viruses, that extant immunological factors and age can affect the probability of virus infection given exposure, that stress levels (moderated by social environment), health practices (exercise, tobacco and alcohol consumption, sleep efficiency) and genetics contribute to CLI risk most probably by modulating the immune-inflammatory response to infection, and that other factors such as pollution, home environment and certain personality traits affect CLI risk by biasing illness interpretation for a given set of symptoms and signs.
Introduction This chapter reviews those factors that are suspected or proven to influence an individual’s susceptibility to the ‘common cold’. Because the common cold is an illness attributable to a viral upper respiratory infection (vURTI), we need to consider factors that moderate an individual’s risk for infection with a common cold virus as well as those that moderate illness expression in infected individuals. Before reviewing the results of specific studies that address these issues, it is necessary to present a general background for 150 William J. Doyle and Sheldon Cohen purposes of establishing definitions and introducing certain concepts that lay the foundation for that discussion.
Definition of the ‘common cold’ The first reported use of ‘a cold’ as an illness descriptor was in 1537 and reflected the noted similarities between the symptoms and signs of the ‘disease condition’ and the physiological responses to cold temperature exposure . Indeed, a belief that cold air exposure caused the common cold was widespread during the time of Benjamin Franklin (1706–1790), who countered that developing the illness depended on contact with ill persons . Much later it was shown that most illnesses recognized as a common cold were caused by viruses that infect the upper respiratory tract . Recent definitions for the common cold note its infectious etiology, but still focus on a listing of the signs and symptoms characteristic of the illness. For example, the Merriam-Webster Medical dictionary defines the common cold as: “an acute contagious disease of the upper respiratory tract that is marked by inflammation of the mucous membranes of the nose, throat, eyes and Eustachian tubes with a watery then purulent discharge and is caused by any of several viruses” . Thus, in discussing the common cold, we are referring to a culturally accepted constellation of upper respiratory symptoms (if perceptible only by the affected person) and signs (if perceptible by both affected persons and observers)  that signals the presence of a vURTI caused by rhinovirus (RV), respiratory syncytial virus (RSV), adenovirus, influenza virus, parainfluenza virus, coronavirus and metapneumovirus, among others [6–10]. While usually self-limiting and of short duration, vURTIs can be associated with a variety of complications [8, 11, 12] that include otitis media [10, 13, 14], sinusitis , bronchiolitis , asthma exacerbations [16, 17] and pneumonia . Because the use of ‘common cold’ as an illness descriptor often carries the implicit connotation of RV infection, here we use the more inclusive term, cold-like illness (CLI) in referring to upper respiratory illness during a vURTI.
Definition of the viral symptom/sign complex
The viral symptom/sign complex (vSSC) is a summary measure of illness during a suspected vURTI and can be defined by the magnitudes and durations for a set of commonly expressed symptom and/or sign elements .
Most simply, the vSSC is measured as the area under the curve (AUC) relating the sum of vSSC element magnitudes to time for a specified period.
In research, a commonly used vSSC element set is that originally defined by Jackson to include sneezing, runny nose, nasal congestion, sore-throat, cough, malaise, chills and headache [20, 21]. In other vSSC constructions, Etiology of the common cold: Modulating factors 151 these elements are supplemented with additional symptoms/signs of an uncomplicated vURTI (e.g., confusion, insomnia, anorexia, fever, muscle ache and joint pain, among others) and/or the symptoms/signs associated with vURTI complications such as earache (otitis media), sinus pain/fullness (sinusitis), wheezing (bronchiolitis, asthma exacerbation) and chest congestion/difficulty breathing (pneumonia) . While the viruses that cause vURTIs are diverse, the vSSC for all viruses is similar with few consistently expressed elements that would allow for assignment of an illness episode to a particular virus or group of viruses in the absence of additional information such as seasonality [5, 23–25].
CLI, vSSC and vURTI relationships
The vSSC is not equivalent to a CLI, but rather the vSSC is used by an assessor to define the presence of a CLI based on past experiences and cultural context. The symptom vSSC is used by affected individuals in making judgments as to whether or not they have a CLI, while the sign vSSC is used by others to mark an individual as ‘ill’ for possible contact avoidance . Thus, persons assign themselves (and others) as to whether or not they ‘have’ a CLI based on selected aspects of vSSC and not on the presence/ absence of vURTI. Importantly, a vSSC and the derived CLI assignment are not prerequisite expressions of a vURTI [27–31]. For example, experimental exposure of susceptible adults to usual vURTI viruses (influenza A virus, RV, RSV) causes a CLI in only about 60% of those with documented infection [32–34] and nasal/nasopharyngeal detection of vURTI viruses in children is associated with a parent-identified CLI for the child in only about 60–85% of the detections . Moreover, the frequencies of vURTI complications are only partially conditioned by the vSSC or presence of a CLI [10, 14, 36]. These relationships are illustrated in the Venn diagram presented as Figure 1 where the CLI set is a subset of the vSSC set, which, in turn, is a subset of the vURTI set, but the intersection of the complication set with that for either the vSSC or CLI is not unity.
Figure 2a shows an idealized vSSC (sum of element magnitudes versus time) for a vURTI. There, the onset of increased vSSC magnitude occurs at a variable time after virus infection and the vSSC magnitude shows a curvilinear increase to a plateau and then a decrease to baseline . This type of curve embeds a number of signals that can be abstracted by an individual for purposes of assigning the presence of a CLI. These include, the AUC, the rate of change in vSSC magnitude between days or over a period of days after illness onset (slope of the vSSC rise), the maximum vSSC magnitude and the time 152 William J. Doyle and Sheldon Cohen Figure 1. Modified Venn diagram depicting the nested relationships among a vURTI, the vSSC, a CLI and complications.
during which the vSSC magnitude exceeds a certain value (width of the timewindow bounding the vSSC at a specified vSSC magnitude). Studies of CLI assignments made by adult subjects with experimental vURTIs and by parents for their children with natural vURTIs show that different persons may use different signals to make their CLI assignments [19, 22] and document a relative weighting of the vSSC elements (e.g., greater weight for rhinorrhea and nasal congestion when compared to other elements) used in vSSC construction that is not uniform across the population. For these reasons, there is not a 1:1 correspondence between an objectively measured vSSC and an individual’s subjectively constructed vSSC or between either type of vSSC and an individual’s CLI assignment. This is made explicit in the Jackson definition of a clinical ‘cold’, which requires either an individual’s assignment of a CLI and the concurrent presence of specific symptom vSSC elements or a symptom vSSC that conforms to certain criteria [20, 21]. Recognizing that perceived symptoms need not scale linearly with objectively measured signs [37–39], some investigators differentiate subjective and objective CLIs (clinical colds). For example, Cohen and colleagues defined a subjective CLI using the Jackson criteria as modified by Gwaltney and colleagues  and an objective CLI using measurable vSSC sign elements .
Figure 2b shows a more realistic vSSC that includes a discrepancy between the subjective (dashed curve) and objective vSSCs (solid curve), a pre-exposure, basal SSC (bSSC) and two (T1 and T2) subjective vSSC magnitude thresholds for assigning a CLI. From this vSSC representation, it Etiology of the common cold: Modulating factors 153
Figure 2. (a) Simple depiction of a vSSC represented by the function relating the sum of symptom or sign element magnitudes to time.
Signals for extracting the presence or absence of a CLI include the slope of the vSSC rise (Slope), the maximum vSSC magnitude (Amp), the width of the time-window bounding the vSSC at a specified vSSC magnitude (D2-D1) and the AUC (shaded area). (b) A more realistic vSSC that includes a discrepancy between the subjective (svSSC, dashed curve) and objective vSSCs (ovSSC, solid curve), a pre-exposure, basal SSC (bSSC) and two (T1 and T2) subjective vSSC magnitude thresholds for assigning a CLI.
is clear that CLI assignment can be affected by the chosen threshold criterion (either T1 or T2), by the subjective vSSC bias (difference between the subjective and objective vSSC) and by the magnitude of the bSSC. Also, the objective vSSC can be modified by changes in the inflammatory reactivity of the upper respiratory tract as measured by the magnitude of provoked inflammation for a given stimulus intensity. For example, reactivity can be dramatically increased by pre-exposure to certain inflammatory stimuli (e.g., allergy, pollutants, cold air), a phenomenon termed ‘priming’ . Priming by a non-vURTI stimulus would be manifest as an increased objective vSSC and a higher probability of CLI assignment during a vURTI. Alternatively, a vURTI can ‘prime’ the upper respiratory tract to other inflammatory stimuli such as cold air , which would increase the objective vSSC and possibly transform a subclinical vURTI into a CLI. These effects may partly explain modulation of CLI risk by certain personality traits (e.g., positive subjective vSSC bias attributable to neuroticism ), adverse environments (e.g., increased bSSC attributable to air pollution ), allergy (e.g., priming of the vSSC by household mold ), cold weather (e.g., priming of the inflammatory response to cold air by a vURTI ) and by other factors described in the Results section.
trolled by the synthesis of potent signaling chemicals, including the cytokines, and by the synthesis and/or release of effector chemicals, including the more traditional inflammatory mediators such as histamine, bradykinin and arachidonic acid metabolites . The interactions of these signaling chemicals are complex, self-amplifying and feedback modulated, and allow for detecting the nature of the threat (e.g., pollutant, allergen, virus exposure), adaptively tailoring the evolution of a threat-appropriate immuneinflammatory host response to eliminate the source of the threat, and downregulating those responses once the threat has been eliminated.
Figure 3 presents a simple interpretive model for understanding signal processing from initial virus exposure through the development of symptoms, signs and complications during the course of a vURTI and for defining the various nodes at which modulating factors can act. Briefly, virus exposure is processed by a set of biological filters that may or may not prevent infection and/or limit viral replication and viral spread to adjacent cells. These filters are tuned by environmental factors (e.g., air pollution, cigarette smoke exposure), genetic factors (e.g., HLA haplotypes), the functionality of existing physical barriers to infection (e.g., mucociliary clearance system) and the immune status of the host (e.g., extant presence of non-specific antiviral chemicals, homotypic sIgA antiviral antibody titers). If infection is established, viral sensors detect the event and trigger the activation and/or up-regulation of the innate immune system and of both the humoral and cellular components of the adaptive immune systems.
In turn, these systems up-regulate the activities of the biological filters with the teleological goals of progressively decreasing viral load, limiting viral spread to adjacent anatomical compartments, eliminating infected cells, preventing secondary bacterial infections, establishing immune memory to prevent re-infection with the same virus and healing the damaged mucosa . Failure to achieve threat-appropriate responses or to adequately coordinate the up- and down-regulation of these responses can lead to the development of an exaggerated inflammatory response as well as to vURTI complications.