key: cord-0046186-pwe7yc83
authors: Peiris, J. S. Malik; Madeley, Charles R.
title: Respiratory Viruses
date: 2020-06-22
journal: Manson's Tropical Diseases
DOI: 10.1016/b978-1-4160-4470-3.50050-1
sha: b3c42b13ef03abefe4819cde4008560c1522c239
doc_id: 46186
cord_uid: pwe7yc83

nan

The respiratory tract can be divided into upper and lower parts, with the boundary at the lower end of the larynx. Viral infections confi ned to the upper part (upper respiratory tract infection, URTI) are rarely life-threatening, with the exception of croup. They can be uncomfortable but do not usually call the individual's future into question. These infections do not automatically spread to the lower respiratory tract, but where the lower respiratory tract is involved the process is extensive and rarely confi ned to one lobe or even one lung. This contrasts with pneumococcal pneumonia, which is typically confi ned to one lobe of one lung (see also Chapter 11) .

Although widespread, the clinical consequences of viral infection are usually less severe than those seen in bacterial pneumonia; otherwise, such infections would be much more lethal. Severe Acute Respiratory Syndrome (SARS) and disease due to avian infl uenza subtype H5N1 are notable exceptions, where disease severity and mortality from a virus infection is particularly high. The most common manifestations of a lower respiratory tract infection (LRTI) are bronchiolitis (in infants) or an atypical pneumonia. Even when an LRTI occurs, the upper tract is also usually involved and the causative virus can usually be recovered from it.

There are no clear-cut differences between the clinical presentation(s) of any viruses in the respiratory tract. For example, although respiratory syncytial virus (RSV) is the most common cause of bronchiolitis worldwide, this clinical condition may also be caused by parainfl uenza viruses, infl uenza viruses, adenoviruses or rhinoviruses. Consequently, it must not be assumed that two patients with similar clinical illnesses will have been infected by the same virus. This is particularly so in babies and young children. Conversely, the same virus may cause a range of clinical manifestations in different patients. 3 Table 46 .1 lists those viruses generally associated with the respiratory tract. Nevertheless, other viruses may be present as part of a generalized process in which the respiratory component is only a (small) part. Table 46 .1 is divided into two sections: section A lists those viruses usually associated with respiratory tract disease. Confi rming their presence will usually identify the cause of the illness, although dual and even triple infections can occur, particularly in the compromised host. The viruses are listed in approximately descending order of importance in terms of numbers of cases annually and their potential severity. By almost any criterion RSV would head the list but the others could be ranked in a different order, depending on the age, time of year and geographical location of the population. Section B lists another three viruses which may be found in the respiratory tract of clinically normal individuals, especially children. Nonetheless, enteroviruses are increasingly recognized as aetiological agents of respiratory disease. Herpes simplex virus, too, may cause no overt lesions in the respiratory tract, although its presence indicates a potential to cause damage if the opportunity occurs -particularly in compromised patients. Reoviruses are not proven pathogens in the respiratory tract, although they are frequently isolated from the throats of children. Important features of each virus, and infection with it, are discussed below.

This virus is distributed worldwide and is found wherever it has been sought. It is frequently associated with bronchiolitis in babies -with a peak incidence at about 6 months -and is the most common virus detected, especially in children under 1 year of age who are hospitalized with respiratory infections. Large epidemics occur annually at the same season, but the seasonality of RSV epidemics may vary in different geographical regions (see Epidemiology, below). The starting date and extent of the epidemic may vary but the annual epidemic occurs reliably. For diagnostic purposes there is only one serotype, but two subtypes (A and B) have been described and they may co-circulate, with one usually predominating in any given year. No obvious differences in disease severity or pathogenesis have been documented. 3 RSV causes a substantial but variable LRTI disease burden in tropical countries. 4 In a population-based study of infants in Kenya it was found that RSV was common, approximately 36% of infections led to LTRI, 23% were severe and 3% of infected children were hospitalized. 5 More recently, it is becoming clear that RSV causes signifi cant morbidity in the elderly as well as in infants. 6

Antigenically, these are the most variable of the respiratory viruses. Both exhibit antigenic 'drift', in which the surface antigens of the virus change gradually in the face of immunological pressure from the host species, with one or two variants predominating at a given time. In showing this progressive and 'directional' antigenic change, they are unique among respiratory viruses. In addition, infl uenza A, but not infl uenza B, shows occasional major antigenic changes in the surface antigenic structures (haemagglutinin and/ or neuraminidase), and called 'antigenic shift', which may lead to a pandemic. Such pandemic infl uenza viruses are derived from avian infl uenza viruses through genetic re-assortment with animal or human strains. This results in the incorporation of new viral surface antigens to which the human population is immunologically naïve. The timing, extent and direction of either 'drift' or 'shift' have so far been completely unpredictable. However Smaller-scale infl uenza epidemics associated with antigenic drift contribute to mortality in the elderly and in those with preexisting conditions such as chronic cardiopulmonary or renal disease, diabetes, immunosuppression or severe anaemia. The risk of Reye's syndrome is increased following infl uenza in children on long-term aspirin therapy. While the morbidity and excess mortality associated with infl uenza in temperate regions is well documented, 6 the more diffuse seasonality (see Epidemiology, below) obscures the disease burden due to infl uenza in the tropics. Nevertheless, recent studies in Hong Kong and Singapore have revealed that infl uenza-associated mortality and morbidity in tropical settings is as signifi cant as in temperate climates. 7 Interestingly, infl uenza-associated mortality is not restricted solely to respiratory complications. A small proportion of cardiovascular mortality also appears to be triggered by infl uenza. 7 Avian infl uenza virus (H5N1) is currently endemic in poultry in a number of countries in Asia and Africa and has repeatedly been transmitted zoonotically to humans, often with fatal consequences. 8, 9 The associated disease was unusual in that previously healthy young adults and children are among those most severely ill. The disease presents as a rapidly progressive viral pneumonia with severe leucopenia and lymphopenia, progressing to Acute Respiratory Distress Syndrome and multi-organ dysfunction that fail to respond to standard antibiotic therapy for the pathogens causing community acquired pneumonia. Some patients also manifest a watery diarrhoea and moderate liver dysfunction. Most, though not all, patients have a history of recent exposure to sick poultry. Transmission remains zoonotic although occasional instances of limited and, so far, unsustained human-to-human transmission following close family-type contact have been reported. Early diagnosis and treatment with oseltamivir is lifesaving (see Treatment, below).

Other avian infl uenza A viruses (e.g. subtypes H9N2, H7N7) have also caused human infection. H9N2 and H7N7 have caused mild fl u-like illness or conjunctivitis (H7N7). But one fatal respiratory illness caused by H7N7 virus is documented.

The continued zoonotic transmission of these avian viruses does not imply that they are inevitably likely to lead to another pandemic. However, the unusual severity of H5N1 disease in humans gives cause for concern, because one cannot assume that the acquisition of human-to-human transmissibility (if it ever occurs) will always be associated with a signifi cant loss of virulence. Irrespective of whether or not a putative pandemic threat becomes reality, it is clear that H5N1 viruses have already had a signifi cant impact on the poultry industry, on human economic and social well-being and consequently on human health.

This virus, which resembles respiratory syncytial virus (RSV), was discovered in 2001 by van den Hoogen and colleagues in The Netherlands. 10 It is now recognized to be a separate virus in its own right, although the disease it causes, its world wide distribution and seasonality are similar to those of RSV. 11, 12 It, too, may cause infections in the elderly as well as in babies under 1 year old, and its discovery has accounted for some of the diseases in these age-groups for which no cause had been found hitherto. Retrospective serology, though, has shown that this is not a new pathogen, even for man, but has been around for a long time.

There are four serotypes of parainfl uenza, with type 4 possessing two subtypes: 4a and 4b. Types 1 and 2 typically cause croup, a high-pitched barking cough in children which is profoundly irritating to their parents. Type 3 can cause bronchiolitis or pneumonia and, less often, croup. In temperate countries, types 1 and 2 (together with RSV) are more prevalent in the winter months, whereas type 3 is unusual (among respiratory viruses) in occurring more often in spring and early summer. This dissociation between the peaks of activity of parainfl uenza type 3 and RSV has also been observed in tropical regions. 13

There are 51 different serotypes but the majority of respiratory infections involve types 1-7. Types 1, 2, 5 and 6 are usually associated with endemic disease in temperate regions, and types 3, 4 and 7 with epidemics. The higher-numbered serotypes appear in the respiratory tract from time to time but the majority of them have been found only in the gut (see Chapter 46).

Adenoviruses are unusual in that prolonged carriage (up to 2 years in some cases) may occur in the tonsils of children, often with no continuing illness. The clinical signifi cance of adenoviruses isolated from the throats of children must therefore be interpreted cautiously, especially if the strain has not been typed. However, they may cause a primary and severe pneumonia in debilitated children, in whom it may be rapidly fatal, and in some immunocompromised patients.

These are frequent causes of the 'common cold', itself a frequent winter and summer illness in temperate countries but they have a year-round seasonality in the tropics. 14 They can be diffi cult to grow in culture and are very under-reported, mainly because diagnosis is often not attempted. With over 100 serotypes, serological diagnosis is impracticable. Molecular diagnosis based on conserved parts of the viral genome has revealed that rhinoviruses are detected in a substantial number of children hospitalized with acute respiratory disease in both temperate and tropical regions. 14 However, in a proportion of cases, a rhinovirus is detected together with other respiratory pathogens and the relative contribution of rhinovirus to the illness is unclear. A better understanding of the epidemiology of rhinoviruses in apparently asymptomatic children (and adults) is needed. Rhinoviruses are now also recognized to be a signifi cant precipitating factor in exacerbations of asthma and chronic obstructive airways disease in both children and adults. 15 They have also occasionally been the sole pathogens present in the lungs of immunocompromised patients dying with respiratory signs and symptoms.

Human coronavirus (HCoV) strains 229E and OC43 have been long recognized as the second main cause of the common cold. More recently, three other coronaviruses have been detected in humans, SARS CoV (see below), HCoV-NL63 and HCoV-HKU1. The HCoV 229E, OC43, NL63 and HKU1 viruses are ubiquitous and are regularly detected in respiratory specimens of a small proportion (1-10%) of children hospitalized with acute respiratory disease and in many parts of the world. [16] [17] [18] Infection with these human coronaviruses presents as an upper respiratory tract infection, asthma exacerbation, acute bronchiolitis, pneumonia, febrile seizures and also as croup (especially NL63). HKU1 can be associated with URTI, LRTI (especially in those with underlying diseases of the respiratory tract) and with febrile seizures in children. 17 HCoV are not readily cultivable and require molecular methods (such as reverse transcription polymerase chain reaction, RT-PCR) for detection (see Diagnosis, below).

In 2003, a coronavirus causing a severe and often fatal pneumonia emerged in southern China. Within weeks of its spread to Hong Kong, the disease had also spread worldwide to affect over 30 countries across fi ve continents; a dramatic illustration of how rapidly a newly emerging respiratory disease can spread. 19 It was unusual in that it caused severe disease, which was also readily transmitted to those caring for the patients. Unlike many other respiratory viral infections, viral load in the upper respiratory tract did not peak until the second week of illness and, consequently, transmission was rare within the fi rst 5 days from onset of illness. This allowed public health measures of early case recognition and isolation to interrupt transmission within the community. SARS was a disseminated infection and not one confi ned to the respiratory tract. 19 Virus was detectable in the faeces and urine and these may also contribute to transmission under some circumstances.

The virus originated as a zoonosis. The precursor virus is present in bats (Rhinolophus spp). 20 Civet cats and other small mammals within live game-animal markets in southern China provided a reservoir and amplifi er of the virus and probably provided the opportunity for adaptation to humans. 21 While the transmission of the human-adapted virus that caused the global outbreak in 2003 has been interrupted, it is possible that the disease may reappear, either through the escape of the human-adapted SARS CoV from a laboratory or by the re-adaptation of the animal virus to effi cient human transmission.

Measles (see also Chapters 43 and 47) is often not recognized as a major cause of LRTI morbidity or mortality, and there are a number of factors that may account for this underassessment. 22 Children with measles may not always be admitted to a general paediatric ward, the aetiology may be attributed to a super infecting pathogen rather than to measles, and some patients with measles (especially when immunocompromised as a result of malnutrition, cytotoxic drug treatment or for other reasons) will fail to develop the typical rash. In patients who do not manifest typical clinical features, both clinical and laboratory diagnosis of measles is diffi cult, even in the developed world. Where the diagnosis has been actively sought in developing countries, measles is found to be a major cause of LRTI, accounting for 6-21% of morbidity and 8-50% of the mortality attributed to LRTI. The effects of the virus on the respiratory tract can be direct (giant cell pneumonitis) or indirect. The latter includes the depressive effects of the virus on the host immune system, stores of vitamin A and overall nutritional status. All of these can lead to an increased risk of super-infection with other viral or bacterial pathogens.

Human Boca viruses belong to the family Parvoviridae are associated with a proportion of lower respiratory tract disease and wheezing in children, especially those aged 6 months-2 years. 23, 24 The viral DNA is also detectable in the serum but it is not clear whether this represents infectious virus. Enteroviruses have been known for many years as causes of a range of clinical manifestations. Their role in respiratory infections is now being increasingly investigated. 25 Hantavirus pulmonary syndrome (HPS) is a rare but important cause of severe respiratory illness in the North and South American continents. Their role in respiratory disease was fi rst recognized in May 1993, when an outbreak of a severe, and frequently fatal, respiratory disease occurred in the area in the USA where the four states Arizona, Colorado, New Mexico and Utah abut. The causative agent was found to be a hantavirus, later called Sin Nombre virus. The natural host was found to be the deer mouse, Peromyscus maniculatus, the local population of which had recently increased rapidly, bringing them and their excreta more into contact with humans, and allowing the virus to cross the species gap. Related viruses causing a similar disease syndrome have since been isolated in North (e.g. New York, Bayou, Black Creek Canal viruses) and South (e.g. Andes virus) Americas, but with different species of natural rodent hosts. 26 These viruses all belong to the same hantavirus genus as those causing haemorrhagic fever with renal syndrome (HFRS) in the Old World: Hantaan, Seoul and Puumala viruses. Both HFRS and HPS have a similar febrile prodrome with thrombocytopenia and leucocytosis. In HPS, the key differences are that the capillary leakage which follows is localized to the lungs and that, with Sin Nombre virus, renal dysfunction is minimal. There was no evidence of human-to-human transmission in this outbreak, but there is evidence that some of the South American hantaviruses causing HPS may be transmitted between humans in a nosocomial setting.

A detailed analysis of the respiratory complications of the immunocompromised patient (oncology, leukaemia, transplantation) is outside the compass of this book but some mention is necessary of opportunistic infections in patients who have been immunodepressed by the human immunodefi ciency virus (HIV) or who have the acquired immune defi ciency syndrome (AIDS) (see also Chapter 20) . They are likely to contract any of the viruses already mentioned and may have diffi culty in eradicating them due to the lack of functioning cellular immunity. However, viral respiratory infections are not in themselves necessarily a life-threatening problem in patients with AIDS, with three exceptions: cytomegalovirus, measles and varicella-zoster virus.

Cytomegalovirus is an opportunist pathogen in immunocompromised patients in whom it can cause serious or even fatal respiratory complications. It is more important as an opportunist pathogen of transplant recipients (especially bone marrow transplants) than those immunocompromised through AIDS. Perinatal cytomegalovirus infection may occasionally present as pneumonitis in the newborn and (together with Chlamydiae) must be considered in the differential diagnosis. Apart from such occasional illnesses, most cytomegalovirus infections are clinically silent, although serological surveys have shown positivity rates approaching 100% in some overcrowded populations. It is also a cause of congenital malformations, especially sensorineural deafness, following maternal infection in pregnancy (see also Chapter 47).

In the immunocompromised, giant cell pneumonitis due to measles can be fatal, and may occur even in patients who have past immunity (naturally derived or vaccine induced). Chickenpox is usually trivial in school-age children but may be severe and include respiratory complications in adults and in the immunocompromised.

The diagnosis of several other agents has been undertaken in virus laboratories because these agents cause respiratory infections which overlap clinically with those due to viruses, and they are diagnosed serologically (see below). They include psittacosis, Q fever and mycoplasmosis, where isolation of the causative organism is either diffi cult or dangerous. They also include Chlamydia pneumoniae (TWAR), which is recognized as a cause of community-acquired pneumonia although diagnostic tests are not yet widely available.

The activities of these agents are under-recorded in most parts of the world. Since they are amenable to antibiotic therapy, it is important that they are diagnosed.

The aetiology and epidemiology of acute respiratory infections have been intensively studied in the temperate areas of the world. Information from tropical regions is more scanty, but what evidence there is suggests that the viruses responsible for respiratory disease in the tropics are no different from those found in temperate zones. 5, 13, [27] [28] [29] [30] However, the severity of illness and its sequelae, as well as their seasonality, may be markedly different from those in the developing world. 31 The data on respiratory infections obtained by Jacob John and his colleagues 13 in Vellore, India, and shown in Tables 46.2 and 46.3, confi rm a range of aetiological agents familiar to workers in temperate zones though with different seasonality. In temperate regions, respiratory infections have generally been shown to increase in the autumn and winter, although the exact mechanisms are still not fully understood. A similar periodicity is shown in tropical regions but this may be related to fl uctuations in rainfall or humidity rather than temperature. 31 In contrast to temperate regions, infl uenza in the tropics may occur in the summer months or all year round, and RSV in subtropical Hong Kong is a summer disease. 

The activities of infl uenza A and B remain impossible to predict and can fl uctuate greatly from year to year. The appearance of a 'new' strain of either A or B can be associated with an epidemic the size of which is likely to be greater as the size of the antigenic change increases, although other, so far unidentifi ed, virulence factors may be even more infl uential. Recently, with detection of H5N1 infl uenza A strains in migratory birds, and fatal infections in domestic poultry and a high mortality where the virus has been transmitted to humans, 8 there has been widespread anxiety that a human pandemic will follow. So far this has not happened and predicting whether, when and where this might happen is impossible. With no shift changes in infl uenza B, major epidemics are less common.

Even where high-quality, competent diagnostic services are available, not every clinical respiratory disease yields unequivocal evidence of infection by a virus or other microorganism. The proportion in which a positive diagnosis is made varies from a quarter to a half, depending on laboratory, area, population and time of year. The recent discoveries of a number of newly recognized respiratory viruses (e.g. human metapneumovirus, bocavirus, novel coronaviruses and novel hantaviruses) has highlighted the fact that there are probably still more viruses to be uncovered.

With the development and application of new methods to detect pathogens, novel respiratory viruses are likely to be increasingly recognized in the respiratory tract. It is essential, however, to differentiate asymptomatic viral carriage from infections of aetiological signifi cance, a task that requires careful epidemiological studies including the relevant controls. Berman 2 has summarized the data from developing countries and found that the percentage of hospitalized patients who were virus positive was about twice the fi gure found in those attending as outpatients. This difference is not surprising and probably refl ects both the greater opportunity to make a specifi c diagnosis in the hospitalized patient and the greater severity of their disease. The majority of trivial episodes (head colds and increased nasal secretions) are not subjected to virus diagnosis and the causative viruses are unconfi rmed.

As with most other diseases, respiratory infections are made worse by other components of the patient's environment. Poverty, malnutrition, pollution and overcrowding (common in urban environments everywhere in the world) are well recognized to contribute to the frequency and severity of respiratory illness. The effects may be direct or indirect through the presence of other disease, poor sanitation and poor personal hygiene ( Figure  46.1) .

Nevertheless, although a poor, malnourished child in a densely populated and economically deprived urban area will have many respiratory illnesses, viruses are no respecters of persons and his or her better-off cousin in a wealthy environment may also have a considerable number of infections. Where the difference lies is that the latter will be able to cope better and will have fewer longer-term sequelae, which include chronic respiratory impairment, wheezing, asthma, bronchitis and bronchiectasis.

There are three main reasons for providing a laboratory diagnosis of viral respiratory infections: for individual patient diagnosis to aid clinical management (specifi c therapy, stopping antibiotic therapy, infection control); to monitor routine virus activity in the community (epidemiology, e.g. vaccine strain selection for infl uenza); or for research investigations.

Rapid diagnosis of viral respiratory infections (i.e. in less than 3 h) has been shown to reduce antibiotic use and to be costeffective. 32 In addition, such confi rmation of the cause is useful in hospital infection control (e.g. in cohorting similar cases) and, occasionally, in deciding whether to use antiviral drugs in selected high-risk patients (see Treatment, below). The new antineuraminidase drugs for treating infl uenza provide an additional incentive (13), measles virus (7), mumps virus (1), unidentifi ed virus (7). b Acute exacerbation of bronchial asthma (8), tropical pulmonary eosinophilia (2), tuberculosis (2), foreign body aspiration (2) and membranous tracheitis (1). Reproduced with permission from Jacob John et al. 1991. 13 for making rapid diagnoses in, for example, an outbreak situation, although their cost may yet deter their widespread use. It is self-evident that individual diagnosis must be quick if it is to infl uence clinical management. Rapidity of diagnosis is also important in epidemiological studies because the clinician will lose interest in sending specimens if there is no equally rapid feedback on the cause of the patient's illness. It is not surprising that epidemiological data are patchy, but they can refl ect year-byyear variations if the population on which the studies are performed remains approximately constant.

Diagnosis of respiratory infection is achievable within 2-3 h using techniques such as antigen detection (see below). [33] [34] [35] However, these techniques are not universally available, even in hospitals in the developed world, mainly because they are labourand expertise-intensive. In the developing world this is compounded by a shortage of staff experienced in the use of such techniques, but these objections are surmountable. Enzyme immunoassays in 'kit format' which are relatively simple to perform are available for the diagnosis of infl uenza A and B and for respiratory syncytial virus. They are, however, expensive. They have adequate positive and negative predictive value of infection during infl uenza epidemics but have poor predictive values during periods of low infl uenza activity. 36 They also have poor sensitivity for the diagnosis of avian infl uenza (H5N1) disease in humans. 9 The increasing need to provide diagnosis for avian infl uenza H5N1 which is best done by sensitive molecular (e.g. RT-PCR) methods (because other options are less sensitive and virus culture necessitates biosafety level 3 facilities) is increasing the need for establishing this technology in reference laboratories investigating respiratory infections, even in a developing country setting. This may in time permit more utilization of multiplex molecular tests for investigation of a wider range of respiratory pathogens (Multiplex PCR). 37 However, these methods remain resource and expertise intensive and need to be well controlled with regular quality control exercises. Microarray methods with the potential to detect a range of pathogens in a single test are in development on a research basis but it is unclear if they have adequate sensitivity for virus detection in clinical specimens (in contrast with virus isolates where high titres of nucleic acid are present). These methods however have potential for the detection of novel pathogens and also for greater recognition of co-infection by multiple pathogens.

Laboratory diagnosis of respiratory virus infections depends on the demonstration of either virus or viral components in the patient at the acute stage of the illness, or subsequently an immune (serological) response to the virus.

There are several approaches to this. They include demonstration of: (1) viral antigens by immunofl uorescence 33, 34 or enzyme immunoassays, 35 (2) viral infectivity by growth in cell culture; or (3) viral nucleic acid by various techniques. Details of the techniques are not given here, but the advantages and disadvantages of each are indicated in Table 46 .4. Before setting up a diagnostic laboratory, the aims of the operation should be clearly thought out. If the catchment population is very large, the number of specimens may also be large and the advantages of automation (e.g. in machine-based nucleic acid amplifi cation or enzyme immunoassays) may be decisive. However, this level of abundance is rare and the number of available specimens may be too few. Automation may then be less advantageous and is often minimal except for serology (see below). Except for special studies, most of the specimens will come from hospitalized patients because of the practical diffi culties of collecting and delivering specimens from the community. Virology specimens are perishable and must be delivered to the laboratory without delay.

Such methods of diagnosis depend on good-quality specimens being taken from the patient. It is easier to take a bad specimen than a good one, and close cooperation with the laboratory will help to raise the positivity rate.

This, at present, means demonstrating an antibody response in the serum to the stimulus provided by the virus. Seeking responses in cellular immunity or antibody in other body fl uids remain research techniques only.

For a valid diagnosis, a convalescent specimen of serum (taken after enough time for a response has elapsed) is needed but may be diffi cult to collect 2 weeks after the onset from patients who may by then be totally recovered and unwilling to oblige the investigator's interest. This is particularly true with children. Nevertheless, unless an antibody response can be demonstrated 

(seroconversion or a rising titre) some uncertainty over the validity of the result will remain. The alternative is to demonstrate an IgM-class response but this suffers from the twin disadvantages that such tests are not available for all viruses and the sample (to be reliably positive) may have to be taken after the acute illness is over, with the problem(s) already mentioned.

Serology remains the routine choice for some respiratory agents which, although not viral in nature, are traditionally diagnosed by virus laboratories. These include: psittacosis, Q fever and Mycoplasma pneumoniae infection. All cause an illness with an insidious onset and are diffi cult and/or dangerous to isolate. Since all, therefore, are susceptible to antibiotics, a diagnosis is important and can be life-saving. The role of Chlamydia pneumoniae is poorly documented at present in the absence of an easily used test.

The diagnosis of chickenpox is usually clinically obvious, but measles may present problems because the skin rash is often absent in the immunocompromised patient. Immunofl uorescent examination of nasopharyngeal secretions for measles-infected cells provides a rapid diagnosis, but this is unlikely to be widely available. Cytomegalovirus can be cultured from the sputum (voluntary or induced) or detected in bronchoalveolar lavage/lung biopsy specimens, if available. Adenoviruses in the immunocompromised may be detected by culture or molecular methods in the respiratory tract, blood, urine and faeces. Detection in multiple sites is evidence of disseminated disease and is an indicator of poor prognosis and for urgent antiviral therapy.

RSV and infl uenza viruses are particularly infectious, and are notorious causes of cross-infection in hospitals. This may pose par-ticular hazards to patients at higher risk, such as those with underlying heart or lung disease (e.g. congenital heart damage or bronchopulmonary dysplasia). Transmission of RSV (as with most other respiratory viruses) is by direct contact or via infected surfaces or fomites. Infl uenza A, on the other hand, is effi ciently spread by large droplets and occasionally via small droplets.

Precautions that may help reduce the risk of cross-infection include the isolation and/or cohort nursing of infected patients and scrupulous care in hand-washing between patients. It is essential to remember that viruses can also infect medical and other hospital staff (RSV may be asymptomatic or cause a 'common cold' in adults) and be transmitted by and through them. SARS was indeed a frightening reminder that hospitals may serve as a venue for amplifi cation and dissemination of virus infections.

Amantadine, and its alternative rimantadine, were options for the prevention (in outbreaks within closed communities of high risk individuals) and less convincingly for the treatment of infl uenza A. 38 However, since 2003, increasing resistance of both H3N2 and H1N1 subtypes of infl uenza A virus has now led to its withdrawal as an option for the treatment of seasonal infl uenza. Neuraminidase inhibitors such as zanamivir and oseltamivir, which inhibit the viral enzyme neuraminidase from both infl uenza A and B (concerned with release of the virus from infected cells) are effective for the treatment and prophylaxis of infl uenza but have to be given within 48 h of onset for apparent clinical benefi t. They are expensive and are best used on those most at risk of serious illness -those at the extremes of life. They are also effective against other infl uenza A virus subtypes including avian infl uenza A H5N1. While either zanamivir (given by inhaler) or oseltamivir (given orally) can be used for prophylaxis of H5N1 infl uenza disease, given the potential for dissemination of this virus beyond the respiratory tract, the systemically active oseltamivir is the preferred option for treatment. However, experience with human cases of H5N1 avian infl uenza has shown that resistance develops rapidly and may be a major problem in widespread prophylactic or therapeutic use. These drugs are not active on other respiratory viruses, even those with viral neuraminidases. Generally, the management of viral respiratory infections is essentially symptomatic and is dealt with elsewhere (Chapter 11). Antibiotics are not routinely indicated for viral respiratory infections unless secondary bacterial superinfection occurs. The 'atypical' bacterial infections mentioned above (Q fever, mycoplasmosis and chlamydiosis) are amenable to antibiotic therapy. (T)ribavirin given as an aerosol inhalation is claimed to reduce the severity of RSV infection in infants, but this remains controversial. It is a very expensive drug, but may be life-saving in those with congenital heart and/or lung damage for whom RSV infection may be the fi nal insult which pushes them into heart or lung failure. (T)ribavirin may have some effect in infl uenza but the evidence is minimal. It has also been used in hantavirus pulmonary syndrome but, again, the evidence of effi cacy is minimal.

Aciclovir (given i.v.) is effective in the treatment of varicella or herpes simplex infections of the respiratory tract in the immunocompromised patient. It should also be used in an immunocompetent patient (usually an adult) with varicella pneumonia. Ganciclovir and foscarnet are useful in cytomegalovirus infection in the immunosuppressed, but a detailed discussion of this problem is beyond the scope of this chapter.

With the cells of the target organ immediately accessible to viruses, it is proving diffi cult to produce effective vaccines to respiratory tract viruses. 39 Other than in measles, which has a systemic phase, vaccines have had only limited success. In the tropics, even the measles vaccine has limitations because much of the impact of this virus on morbidity and mortality is during infancy, and existing measles vaccines are not effective at inducing immunity in the presence of passive maternal antibody. Newer measles vaccines, including one using canarypox virus as a vector, have been explored but not yet adopted. A second dose of conventional vaccine has also been suggested but cost makes this impractical in many countries.

Infl uenza vaccine is used for persons at high risk (e.g. patients with underlying heart, respiratory or immunocompromising diseases, patients on dialysis, the elderly) and contains antigens from two current infl uenza A subtypes (H3N2 and H1N1) and from infl uenza B. The constituents are modifi ed as the prevalent strains vary. The conventional infl uenza vaccine is formalin-killed egggrown virus and has provided useful protection, particularly in the elderly and those with pre-existing lung damage in whom even minimal protection may be enough to prevent death. An alternative approach of a live attenuated vaccine containing cold-adapted infl uenza strains has shown some effi cacy and such vaccines are now available.

The possible emergence of an H5N1 avian strain adapted to man has stimulated research into new ways to produce vaccines (e.g. using reverse genetics or a disabled adenovirus as a vector for infl uenza antigens) and for new antiviral drugs. Phase 2/3 clinical trials with such H5N1 candidate vaccines are in progress and have shown that the conventional approaches used in seasonal vaccines are poorly immunogenic with the avian H5 haemagglutinin. Therefore, novel adjuvants and whole virus vaccines are now being tried. In mathematical simulations of a new pandemic, vaccines have been shown to be the intervention with the greatest public health impact but until the virus does adapt, we do not know whether any of these novel approaches will be effective in controlling a pandemic. An experimental enteric coated vaccine to adenovirus 14 was developed for use in the US Army to combat epidemics in recruit camps but has found no application elsewhere. Prevention of RSV, severe measles and varicella in susceptible (immunocompromised or severely malnourished) contacts may also be achieved by passive immunization. There is evidence that humanized mouse antibodies or hyperimmune gamma-globulin may give some protection from, or reduce the severity or duration of, RSV infections in the more vulnerable (e.g. premature) babies, but these preparations are very expensive and their use should be confi ned to those in whom infection will be life-threatening on standard management. 40 Normal human gamma globulin is effective in preventing/attenuating measles if administered within 3 days of contact. For the prophylaxis of varicella, high-titre varicella-zoster human immune globulin (ZIG) must be used. Maximum protection (from severe disease, but not from infection) follows administration within 48 h of contact, but some benefi t may accrue if given within 10 days of exposure.

Respiratory infections are very common throughout the world and are worse where social conditions are inadequate. Much childhood respiratory tract disease is either totally due to viruses or is virus-initiated, and the same viruses appear to be involved in all regions, tropical or temperate. Epidemiological data are incomplete everywhere (but more so for the poorer parts of the world) and come mostly from hospitalized patients.

Nevertheless, RSV is a universal childhood pathogen, found everywhere it has been sought. The numbers of virologically confi rmed diagnoses each year (most them in patients in hospital) in the Newcastle and Tyneside area in the UK (population about 1 million) and from Hong Kong island (population about 0.7 million) are remarkably similar: 500-600 and 500-700 cases, respectively. There are likely to be many more in the crowded cities of India, China, the Philippines, Brazil and elsewhere. The effects of RSV (and other viruses) are exacerbated by overcrowding, malnutrition, air pollution, poor sanitation, minimal medical care, etc.

A pandemic of infl uenza A (similar to the one that swept the world in 1918/9) will be a major health problem. Whether and when it may happen is unpredictable but current preparative measures may help to reduce its impact. Respiratory disease, like diarrhoea, results in signifi cant morbidity and mortality in the developing world, has signifi cant economic consequences and will require an enormous commitment of resources to abate. Viruses and bacteria are both involved and there are few effective vaccines at present.

Estimates of world-wide distribution of child deaths from acute respiratory infections

Epidemiology of acute respiratory infections in children of developing countries

Respiratory infections

Respiratory syncytial virus infection: denominator-based studies in Indonesia

Respiratory syncytial virus epidemiology in a birth cohort from Kilifi district, Kenya: infection during the fi rst year of life

Viral pneumonia in older adults

Infl uenza-associated mortality in Hong Kong

Avian infl uenza A H5N1: A threat to human health

Update on avian infl uenza A (H5N1) virus infection in humans

Prevalence and clinical symptoms of human metapneumovirus infection in hospitalized patients

Newer respiratory virus infections: human metapneumovirus, avian infl uenza virus, and human coronaviruses

Human metapneumovirus infections among children with acute respiratory infections seen in a large referral hospital in India

Etiology of acute respiratory infections in children in tropical Southern India

New Vaccine Surveillance Network. Rhinovirus-associated hospitalizations in young children

Association of rhinovirus infections with asthma

Human coronavirus NL63 infection is associated with croup

Coronavirus HKU1 and other coronavirus infections in Hong Kong

Human coronavirus NL63 infection and other coronavirus infections in children hospitalized with acute respiratory disease in Hong Kong, China

The severe acute respiratory syndrome

Severe acute respiratory syndrome coronaviruslike virus in Chinese horseshoe bats

Isolation and characterization of viruses related to the SARS coronavirus from animals in southern China

The burden of acute respiratory infection due to measles in developing countries and the potential impact of measles vaccine

Human bocavirus and acute wheezing in children

Human bocavirus: a new viral pathogen

Microbiology of acute otitis media in children with tympanostomy tubes: prevalences of bacteria and viruses

Hantaviruses: a global disease problem

Symposium on etiology and epidemiology of acute respiratory tract infection in children in developing countries

A seven year study of WHO virus laboratory reports on respiratory viruses

Etiology of acute lower respiratory tract infections in children in a rural community in The Gambia

Report of a workshop on respiratory viral infections: epidemiology, diagnosis, treatment and preventions

Epidemiology and seasonality of respiratory tract virus infections in the tropics

Cost-effectiveness of rapid virus diagnosis of viral respiratory tract infections in pediatric patients

Rapid Virus Diagnosis: Application of Immunofl uorescence

Immunofl uorescence Antigen Detection Techniques in Diagnostic Microbiology

Direct antigen detection

Accuracy and interpretation of rapid infl uenza tests in children

Comparison of multiplex PCR assays and conventional techniques for the diagnostic of respiratory virus infections in children admitted to hospital with an acute respiratory illness

Antivirals for infl uenza: historical perspectives and lessons learned

Emerging respiratory viruses: challenges and vaccine strategies

Prevention of respiratory syncytial virus infections: indications for use of palivizumab and update on the use of RSV-IGIV