According to WHO estimates for 2012, around 450 million cases of pneumonia occur worldwide every year, causing 3 million deaths, and accounting for 5.5% of overall mortality worldwide.1,2 It is the fourth cause of death worldwide, and is a particularly a serious threat to children and the elderly.1,3

Bacterial infections, being more common, have been more extensively studied. In contrast, research into viral community-acquired pneumonia (CAP), despite its growing epidemiological significance in developing countries and in the pediatric population, has been limited.4 Assuming the rate of diagnosis to still be lower than real incidence, around 200 million cases of viral pneumonia occur annually throughout the world, half of which are in children.5 Viral pneumonia is of great interest due to its impact on infant mortality, its role as a facilitator of bacterial infections (co-infections), and its ease of transmission, a factor which has transformed it into a worldwide threat.

In this review, we will focus on CAP caused by respiratory viruses in immunocompetent patients.

It is important to distinguish between CAP of viral and bacterial origin. Clinical, radiological and laboratory variables that are commonly used to distinguish between these entities are listed in Table 1.

Epidemiological studies have been published by many authors. Ruiz-González et al.6 included patients with viral and intracellular bacterial pneumonia in the same group. They concluded that pneumonia caused by intracellular pathogens affected older patients, had a more insidious disease course, and often did not produce leukocytosis. Johnstone et al.7 found that patients with viral pneumonia had more cardiac comorbidities and were older; Ma et al.8 found that institutionalization led to a greater risk of viral CAP. Liu et al.9 reported that viral pneumonia caused more cough and less pleuritic pain, while Jennings et al.10 found myalgia to be the symptom most commonly associated with the viral disease. Despite the many publications, predicting the viral etiology from clinical parameters is difficult and often inaccurate, as affirmed recently by Viasus et al.11

With regard to radiological changes, focal alveolar infiltrates have been traditionally associated with the bacterial entity, and bilateral, interstitial infiltrates with the viral form.12 However, recent studies using chest computed tomography (CT) have shown that a viral etiology cannot be ruled out by the appearance of localized alveolar infiltrates, and these may even signal the onset of many viral pneumonias.13

Viral infections do not usually affect the number of leukocytes, so the use of acute phase reactants, such as procalcitonin, as biomarkers may be of great help in reaching a diagnosis.14 Procalcitonin production depends on the presence of circulating tumor necrosis factor (TNF-α); in viral infections, macrophages produce interferon-α that can inhibit TNF-α, suppressing the elevation of procalcitonin, thus suggesting a viral origin.15

Despite these assertions, there is no gold standard for differentiating the etiology of pneumonia.16 Moreover, we must not forget that CAP, whether viral or bacterial, is a dynamic entity: differences in biomarker values or the appearance of infiltrates on radiology are only snapshots of an active process that can vary widely from day to day.

The list of viruses that can cause respiratory infection is long (Table 2). In this review, we will focus particularly on seasonal respiratory viruses.

Other viruses that have recently received considerable media attention, such as H5N1 influenza virus or coronaviruses (responsible for severe acute respiratory syndrome or the Middle East respiratory syndrome, MERS-CoV) will be examined in less depth.

The relative incidence and clinical and epidemiological characteristics of the different respiratory viruses are listed in Table 3.

Overall, syncytial respiratory virus (SRV) remains the primary causative agent of CAP in children, and the main cause of severe pneumonia in this population.17,18 Since the widespread introduction of PCR techniques, rhinoviruses are more often detected as the causative agent of viral pneumonia in children.19–22 However, because it is so frequently detected in asymptomatic individuals (15%), the etiological role of rhinovirus continues to be questioned, although it could be an indication of real, subclinical infections.23 Recently identified pathogens, such as metapneumovirus and human bocavirus,24–27 are less commonly encountered. Although the prevalence of viral pneumonia caused by adenovirus is low (2%–12%), it must be identified, as it can lead to necrotizing pneumonia.28 In this case, PCR techniques are much more sensitive than antigen detection techniques.29

In the case of adults, the most commonly detected viruses are influenza virus (IV), SRV and parainfluenza,5,10,30,31 although incidence varies depending on the diagnostic techniques used. These viruses are also the most important in Spain.32 Pneumonias caused by other viruses are more rarely reported and include outbreaks of rhinovirus,33,34 adenovirus35 (particularly serotype 14 in military institutions36), coronavirus,37 metapneumovirus,38 and even bocavirus (in immunocompromised patients).39

Infections involving both respiratory bacteria and viruses or 2 different viruses are common. The most widely accepted hypothesis is that the viral infection occurs first, followed by the bacterial form. Viral-mediated activation of proinflammatory molecules, such as interleukin-10, is thought to attract large numbers of neutrophils and macrophages to the lung. The arrival of these cytokines amplifies the immune response, causing inflammatory damage and preventing the proper clearance of bacteria.40 Bacterial superinfection worsens the prognosis of the original viral infection. Indeed, research into the influenza pandemics of 1918, 1957, and 1968 shows that most deaths were caused by a secondary bacterial infection.41 During the 2009 H1N1 pandemic, 4%–24% of cases presented secondary bacterial infection.42–44 In infections caused by other viruses, however, especially H5N1 avian influenza, the associated pneumonia appears to be caused (more frequently) by direct viral action.45

In clinical practice, this type of co-infection is particularly common in children (up to 45% of cases with CAP), and mainly involves pneumococcus,46–49 thus increasing clinical severity.50Mycoplasma pneumoniae and several species of Chlamydophila51 are also common,51 while simultaneous co-infection with 2–3 viruses is not unusual.52

CAP of mixed etiology has been characterized less in adults than in children, and prevalence is estimated at less than 5%.7,13 The most common combinations reported are rhinovirus+pneumococcus and influenza A virus+pneumococcus. More serious infections have been identified with the combination of viruses with Legionella pneumophila.53

Data on morbidity and mortality in bacterial/viral co-infection are contradictory. Hong et al.54 consider that these co-infections are no more severe than purely bacterial infections, and affect older patients and those with chronic lung diseases. In contrast, Johansson et al.55 and Seki et al.56 found that mixed etiology pneumonia was associated with higher severity scale scores and poorer progress.

Some very well-known families of respiratory viruses have produced new species and some very virulent serotypes, which in recent years have caused epidemics with significant associated morbidity and mortality.57

The great H1N1 IV-A epidemic of 2009 was unusual: transmission was enhanced by the lack of prior immunity, causing a serious medical situation worldwide.42 Studies performed subsequently showed that the younger population and certain risk groups were more affected, resulting in the loss of many more life-years.58 However, from an inter-generational perspective, the burden of death and major complications was no higher than that caused by normal seasonal influenza.59,60

In 2003, an outbreak of coronavirus in the Far East caused a severe acute respiratory syndrome. In the Middle East in 2012, another outbreak of a new coronavirus occurred, called Middle East respiratory syndrome coronavirus (MERS-CoV).61 In both epidemics, the incidence was reduced to a few hundred cases, partly due to the protective measures implemented,62 but mortality was high due to high virulence.63 Clinical symptoms were not only limited to rapidly developing bilateral pneumonia, but also included acute renal failure and severe hematological disturbances.63,64

Recombinations in animals of different IV subtypes are a global concern, as they are potentially capable of passing the species barrier, and occasionally propagating among humans.65 The serious cases of H5N1 avian influenza recorded in 2005 in Southeast Asia,66 or more recently the H3N2-porcine variant in the United States (2012),67 and H7N9 avian influence in China (2013),68 are examples of this new threat. Fortunately, surveillance systems for monitoring animal reserves and activating a rapid medical response are improving.69

The detection of virus or viral antigens in upper (nasopharyngeal aspirates) and lower (bronchoalveolar lavage and induced sputum) respiratory tract samples is usually based on culture and immunofluorescence microscopy. The detection of antibodies generated during the process of viral infection has also been used for years: seroconversion in 2 samples obtained over a period of time is suggestive of a new exposure to the pathogen. The introduction of polymerase chain reaction (PCR) techniques has improved the detection of respiratory viruses, particularly with the use of real-time techniques.3

The preferred specimen for upper respiratory samples, particularly in children, is nasopharyngeal aspirate, which combines a mixture of nasal and retropharyngeal secretions.70 The use of sterile cotton wool swabs yields a similar sensitivity for most viruses, with the exception of SRV.71,72 Devices that combine a nylon fiber swab and a container of universal transport medium provide a high diagnostic yield at all ages, comparable to nasopharyngeal lavage.30,73

Samples from the lower respiratory tract, being obtained at the site of the infection, have obvious advantages when determining the cause of pneumonia. However, contamination may occur during sampling, as the specimen passes through the upper airways. Induced sputum is often obtained from children, but bronchoalveolar lavage is preferred in adults.3 In contrast, the invasive character of other forms of diagnosis, such as transthoracic puncture, has led to their disuse.70

In sample processing, PCR-based systems are 2–5 times more sensitive than conventional methods for the detection of respiratory viruses.30 This technique is particularly useful in the elderly, who often have a lower nasopharyngeal viral burden than children.74,75 Moreover, some viruses (such as bocavirus) are hard to culture, and antigen testing may be equivocal (e.g. parainfluenza or adenovirus), so they can only be satisfactorily identified using PCR.76–78 A standard identification test has been developed that uses a PCR platform combined with microarrays to simultaneously determine more than 15 different respiratory viruses.79 These platforms are the techniques most commonly used in our hospitals at the moment.80

Since the end of the last century, empirical antibiotic treatment of CAP has been initiated as soon as possible, as this has been shown to reduce morbidity and mortality.81

However, no prospective studies have examined the benefits of antibiotics in pneumonias with a high suspicion of viral origin. The possibility of bacterial superinfection or a mixed viral-bacterial etiology, progressive improvements in antibiotic tolerability, and the medical-legal requirement to “use of all available measures”, mean that physicians are unlikely to resist prescribing antibiotics in a patient presenting with pneumonia. This dilemma has only been prospectively studied in children, in the context of SRV bronchiolitis epidemics; the results, while not definitive, do not support the generalized administration of antibiotics.82

Experience in the use of antiviral drugs in respiratory viruses can be divided into 3 areas: (a) treatment for influenza virus, which is relatively effective and well documented, thanks mainly to studies performed during the 2009 pandemic83; (b) treatment for SRV, that has been mostly explored in children and immunosuppressed populations, although results remain unclear84; and (c) treatment of other respiratory viruses, which has not been studied in depth and remains largely unevaluated.85

In infectious and contagious diseases, particularly of the respiratory tract, barrier methods are essential for preventing infection. Use of masks and gloves and hand-washing have been shown to be effective in reducing transmission rates in the healthcare setting.105 Social isolation of patients during the clinical phase of the disease is also strongly recommended and reduces overall incidence.106 However, it is difficult to implement these measures properly.106

Immunization plays a very important role in prevention, but is only available for a few viruses. Anti-influenza A and B vaccines have been shown to reduce transmission during seasonal influenza epidemics in the general population,107 but their effect on the course of pneumonia or on mortality is not so clear108; nor is their efficiency in children younger than 2 years, although they continue to be administered in many countries.109 In contrast, they appear to be very effective in elderly institutionalized subjects.110 They are currently recommended in Spain for patients with respiratory comorbidities or immunosuppression, individuals over 65 years of age, and healthcare workers.111

In addition to vaccines, chemoprophylaxis with neuroaminidase inhibitors has been successfully tested during seasonal influenza epidemics.112 As yet, no effective vaccine is available for SRV, but palivizumab has been used as chemoprophylaxis. This is a humanized monoclonal antibody that has shown a reduction of up to 50% in the incidence of pneumonia and associated hospital admissions in neonates with a high risk of infection.113

The authors declare that they have no conflict of interests.