OUP user menu

Respiratory viral infection predisposing for bacterial disease: a concise review

Jeanne-Marie Hament, Jan L.L. Kimpen, Andre Fleer, Tom F.W. Wolfs
DOI: http://dx.doi.org/10.1111/j.1574-695X.1999.tb01389.x 189-195 First published online: 1 December 1999


Although bacterial superinfection in viral respiratory disease is a clinically well documented phenomenon, the pathogenic mechanisms are still poorly understood. Recent studies have revealed some of the mechanisms involved. Physical damage to respiratory cells as a result of viral infection may lead to opportunistic adherence of bacteria. Enhanced bacterial adherence by specific mechanisms has been documented for respiratory cells infected with influenza A virus, respiratory syncytial virus and adenovirus in both in vitro and in vivo models. To date, results of various experimental studies indicate that different mechanisms for increased bacterial adherence induced by viruses are operating for specific viral-bacterial combinations. In the present review, a number of key findings obtained during the past two decades is presented and discussed.

  • Bacterial adherence
  • Bacterial adhesin
  • Bacterial superinfection
  • Virus-induced receptor
  • Viral glycoprotein

1 Introduction

Nasopharyngeal carriage of bacteria like Streptococcus pneumoniae and Haemophilus influenzae is a common finding in healthy individuals. How the transition from asymptomatic carriership to invasive disease occurs remains poorly understood. It has long been recognized that a preceding local viral infection appears to play an important role in this process, particularly in infections of the upper respiratory tract, like otitis media and pharyngitis.

There is also epidemiological evidence of bacterial superinfection. Excess mortality during influenza pandemics has been attributed to bacterial pneumonia with Staphylococcus aureus and S. pneumoniae and meningitis with Neisseria meningitidis. In addition, it was found that influenza vaccination offers protection against influenza-related bacterial otitis media in children.

Moreover, mixed viral-bacterial infections are associated with antibiotic treatment failure. These clinical and epidemiologic observations indicate that a better understanding of the underlying mechanisms will enable us to design more effective prevention and treatment strategies.

One possible mechanism is that viral infections facilitate bacterial colonization, adherence and translocation through the epithelial barrier, paving the way for bacterial disease. Therefore, studies have focused on changes in colonization and adherence capacity of bacteria during viral infection. Increased adherence of certain bacteria to virus-infected epithelial cells was indeed documented in various experimental models. These studies also presented evidence that a number of mechanisms are involved in this process.

In the present review we have elected to present a selection of key findings reported in the literature with regard to epidemiological and experimental data that show the important role of preceding viral infections in a variety of bacterial diseases. In addition, the latest insights into the mechanisms of increased bacterial adherence during viral infections are presented and discussed.

2 Epidemiological and clinical evidence

During the influenza pandemics of 1918 and 1957, it was recognized that the incidence of bacterial pneumonia was also increased and contributed substantially to mortality rates [13]. Especially pneumonia due to S. aureus was found to be increased. During the influenza pandemic of 1968–69, this relationship was studied in more detail in the USA. Compared to a non-epidemic period, a three-fold increase in the incidence of pneumonia due to S. aureus was noted [4].

The first epidemiological studies that focused on bacterial-viral interactions during respiratory infection were conducted in the late 1920s [5, 6]. They had the common goal of detecting changes in the bacterial throat flora during viral respiratory disease. During illness of the upper respiratory tract, H. influenzae and S. pneumoniae could be isolated from the throats of patients in greater quantities than from the same subjects when they were healthy.

Later, Nichol and Cherry compared bacteriological, virological and serological data from children hospitalized for respiratory disease with those from children without respiratory symptoms [7]. In 17% of the hospitalized patients a viral-bacterial co-infection was detected compared with only 4% in healthy controls. The incidence of viral-bacterial combinations increased with the severity of the respiratory illness, with a maximum of 31% in the most severely ill children. Another study focused on the etiology of community-acquired pneumonia in children as assessed by serology [8]. Of the children with viral infection, 39% had a bacterial co-infection. Conversely, 20% of the children with a bacterial infection showed evidence of a prior viral infection. The most frequent combination for children under the age of 5 years was S. pneumoniae with respiratory syncytial virus (RSV). In older children dual infections of S. pneumoniae with Mycoplasma pneumoniae or Chlamydia pneumoniae were seen most frequently.

Apart from pneumonia, the incidence of otitis media also increased during epidemics with influenza and RSV. Heikkinen and others found that in children with RSV respiratory tract infection and otitis media, 74% had middle ear effusion containing the same virus [9]. Other respiratory viruses were detected in the middle ear fluid less frequently. Furthermore, when virus was isolated from the middle ear fluid, co-infection with one or more bacteria was detected in 65% of cases. Influenza virus was isolated most frequently in combination with various bacteria. In patients with upper and lower respiratory tract infection with non-typeable H. influenzae (NTHI), serological evidence of a viral infection was found in 60% of the cases [10]. In one study, serological evidence of a concurrent RSV infection was found in half of the patients with respiratory infections due to Moraxella catarrhalis [11]. In otitis media, the occurrence of mixed infection is associated with antibiotic treatment failure [1113]. In 51% of patients with a dual infection, otitis media persisted for up to 12 days after institution of antibiotic treatment, compared to 35% of those with only bacterial otitis and to 19% in patients with viral otitis [12]. In addition, half of the patients with mixed infections failed to clear bacteria from the middle ear fluids within 2–4 days after starting antibiotic treatment, whereas 87% of the patients with bacterial otitis cleared bacteria successfully during the same treatment course [13].

The fact that vaccination against influenza effectively prevented bacterial otitis contributes to the concept that viral infections predispose to bacterial disease. This was proved when children were vaccinated against influenza A during an epidemic [14]. They were not only protected against influenza-induced illness, but also had fewer episodes of bacterial otitis media than non-vaccinated children. A similar protective effect was observed after passive immunization against RSV infection with immunoglobulins enriched with RSV-neutralizing antibodies (RSV-Ig). Children who received high doses of RSV-Ig developed fewer episodes of acute otitis media than controls that did not receive RSV-Ig.[15].

Finally, Kim et al. found a seasonal association between the incidence of respiratory viruses (influenza, RSV and all viruses except influenza) and pneumococcal invasive disease [16]. The incidence of both viral and pneumococcal disease peaked in the winter months. A similar association was found for influenza A infections and meningococcal infections [17]. During an influenza outbreak in November and December 1989 in England and Wales, patients with meningococcal disease showed serologic evidence of a prior influenza infection more frequently than healthy controls. Exceptionally, during the summer Hong Kong influenza epidemic in 1958 in England and Wales, a peak incidence of meningococcal meningitis was recorded. This emphasizes the role of influenza in meningococcal disease, since the occurrence of meningococcal disease is often restricted to the winter months.

3 Experimental evidence

Changes in bacterial colonization on the respiratory epithelium during viral infection were studied under various in vitro and in vivo experimental conditions (Table 1).

View this table:
Table 1

Studies on enhanced bacterial adherence during viral respiratory infection

StudyVirusBacteria (adherence effect)Cell type
In vitro studies
1978, Sanford [18]Influenza AS. aureus (enhanced)Madin-Darby canine kidney cells
1981, Selinger [21]Influenza A, rhinovirus, adenovirus, measlesS. aureus. S. pyogenes, S. pneumoniaeHeLa, Detroit
1988, Bakaletz [20]Influenza AH. influenzae (equal to controls)Tracheal cells
1989, Sanford [35,41,42]Influenza AS. aureus (enhanced)Madin-Darby canine kidney cells
1993, Saadi [24]RSVS. aureus (enhanced)HEp-2 cells
1993, Raza [43]RSVN. meningitidis, H. influenzae (enhanced)HEp-2 cells
1994, Hakansson [22]AdenovirusS. pneumoniae (enhanced)A549 cells
1996, Saadi [23]RSVB. pertussis (enhanced)HEp-2 cells
In vivo studies
1986, Plotkowski [44]Influenza AS. pneumoniae (enhanced)Tracheal cells
1992, Patel [25]RSVNon-typeable H. influenzae (equal to controls)Nasopharyngeal cells
1987, Sanford [27]Influenza AS. aureus, Streptococcus group B (enhanced)Nasopharyngeal cells
1980, Scott Giebink [26]Influenza AS. pneumoniaeOtitis media model
  • Influenza and rhinovirus enhanced adherence of all bacteria. Measles decreased adherence and adenovirus induced no difference.

  • Otitis media was outcome of experiment, see text.

Most of these studies make use of adherence assays, measuring bacterial adherence subsequent to viral infection. Here we discuss a selection of the studies that focused on quantitative differences in bacterial adherence to virus-infected as compared to uninfected cells in vitro or uninfected animals or humans in vivo.

3.1 In vitro studies

In 1978, Sanford et al. were the first to introduce an adherence assay to verify increased susceptibility of mammalian cells to bacterial adherence as a result of viral infection [18]. They exposed monolayers of Madin-Darby canine kidney (MDCK) cells to various streptococcal strains. Indeed, adherence of group B Streptococcus and various streptococcal species was found only to the cells infected with influenza virus and not to membranes of uninfected cells. Yet, these bacteria were not recognized to be human pathogens associated with increased morbidity during influenza epidemics. In subsequent adherence studies with S. aureus, however, preinfection of MDCK cells with influenza A virus significantly enhanced adherence [19]. This effect varied depending on the virus strain and bacterial strain tested.

In an attempt to create human-like conditions, Bakaletz et al. developed a tubular organ model, in which ciliated tracheal cells of chinchillas were cultured [20]. Adherence of NTHI was found to be restricted to the ciliated cells, however, adherence was not increased after a prior influenza A virus infection.

With human nasopharyngeal carcinomatous cells (Detroit 562) cells, the enhancing effect of an influenza A virus infection on staphylococcal adherence could be reproduced [21]. Amongst other viruses that were tested, rhinovirus preinfection enhanced staphylococcal and streptococcal adherence, whereas measles and adenovirus decreased adherence of both bacterial species. The authors suggested a virus-specific induced change on the host cell membrane to which all bacteria adhered in the same way.

Increased pneumococcal adherence could be found after prior adenovirus infection of human pneumocyte-like (A549) cells [22]. Moreover, this effect was restricted to adenovirus serotypes that cause respiratory infections in humans.

After infection with RSV, both S. aureus and Bordetella pertussis have been found to adhere in greater quantities than to non-infected HEp-2 cells [23,24].

3.2 In vivo studies

In a cotton-rat model increased colonization of NTHI in experimentally RSV-infected animals was found [25], as there was an increase in NTHI concentrations in nasal washings.

The synergy of a viral and bacterial infection compared to infection with a sole agent was studied in the chinchilla otitis media model [26]. Otitis media developed in 67% of the animals inoculated with pneumococci and influenza A virus, versus 21% when the animals were inoculated with S. pneumoniae alone. If the chinchillas were infected with influenza A virus only 4% developed otitis.

In another model, influenza virus-infected ferrets were used to study bacterial adherence [27]. Ferrets were experimentally inoculated with influenza A virus to study the adherence of radiolabeled S. aureus and group B Streptococcus type 1a to the nasopharynx. It was found that virus infection enhanced S. aureus adherence more extensively than streptococcal adherence. In contrast to the results of Selinger et al. [4], these results suggested that bacteria-specific mechanisms were operative in mediating adherence to virus-infected cells.

Finally, in humans, bacterial adherence to pharyngeal cells was studied in patients with naturally acquired acute respiratory illness and in volunteers experimentally infected with influenza virus [28]. Increased adherence of S. aureus to pharyngeal cells was found in both patient groups. Curiously, H. influenzae and S. pneumoniae showed enhanced adherence only in the experimentally infected individuals.

4 Pathogenic mechanisms

It is likely that many factors are involved in the phenomenon of bacterial superinfection. In Table 2 we have summarized potential pathogenic mechanisms that may contribute to bacterial adherence during viral respiratory disease, which we will briefly discuss in the following paragraphs.

View this table:
Table 2

Pathogenic mechanisms of bacterial adherence during viral infection

(1) Physical damage to local respiratory tract epithelium
impaired local defence mechanisms
– loss of cilia
– impaired function of Eustachian tube
basement membrane exposure after damage
(2) Bacterial-cellular interaction
viral glycoproteins expressed on host cell
– glycoprotein G
– neuraminidase
– hemagglutinin
virus-induced changes on host cell membrane
– CD14, CD15 and CD18
– fimbriae-associated receptors
– PAF receptors
proteins absorbed from extracellular matrix
– fibrinogen
bacterial adhesins

4.1 Physical damage to respiratory tract epithelium

The intact respiratory epithelium efficiently removes bacteria that have been aspirated into the airways, mainly by ciliary cleaning mechanisms. In primary ciliary dyskinesia, bacterial pulmonary infections are common [29]. Viral infections are thought to induce secondary ciliary impairment. RSV infection induces ciliary injury resulting in ciliostasis, clumping and loss of cilia from live cells in vitro [30]. Influenza virus infection is accompanied by cytological changes in the ciliated columnar epithelium leading to necrosis of the bronchial epithelial lining and in addition to cellular lesions in the alveoli [31]

It was shown that Pseudomonas aeruginosa adheres to injured tracheal cells of mice during systemic influenza virus infection [32]. In this study no adherence could be found to the intact virus-infected mucosa, basal membrane or to cells recovered from bronchial alveolar lavages. These observations would suggest that apart from specific adherence mechanisms to viral infected cells, there is also an aspecific or ‘opportunistic’ adherence phenomenon which occurs to injured or altered tissues.

In otitis media, impairment of the Eustachian tube function is thought to play an important role in secondary infection. Tube obstruction due to local inflammation leads to an impaired clearance and aeration mechanism. In the chinchilla model, mentioned previously, desufflation of the middle ear in animals inoculated with S. pneumoniae led to an increased incidence of otitis media compared to animals that had not undergone desufflation (45% vs. 21%) [33]. Thus, swelling of the mucosa and functional impairment of the Eustachian tube induced by viral infection may be another mechanism by which viruses promote bacterial superinfection.

4.2 Bacterial cellular interaction

Already in 1978 Bartelt et al. showed that trypsin treatment of uninfected cells reduces bacterial adherence, suggesting that a protein receptor is needed for bacterial adherence [34]. Subsequent studies revealed that antiviral antibodies blocked streptococcal adherence to influenza A virus-infected cells [35]. This suggested that viral glycoproteins expressed on the infected host cell membrane could function as receptors for bacteria. During replication of influenza virus, neuraminidase and hemagglutinin are inserted into the host cell membrane. These viral glycoproteins are potential receptors for bacteria. Indeed it was found that streptococcal adherence to influenza A-infected MDCK cells could be blocked by neuraminidase treatment [19]. In addition, hemagglutinin also acted as a receptor by which streptococci adhere during an influenza A virus infection. However, both pretreatments did not inhibit staphylococcal adherence: thus other cellular receptors might be involved in virus-induced increased staphylococcal adherence that are still undefined.

During RSV infection, glycoproteins F and G are inserted into the host cell membrane. In a recent study, glycoprotein G was found to be involved in increased binding of N. meningitidis to RSV-infected epithelial cells [36].

Apart from glycoproteins, virus infections may induce other changes of the host cell membrane that may contribute to bacterial adherence. Upregulation of pre-existing receptors like CD14, CD15 and CD18 was documented for RSV-infected HEp-2 cells [37]. Both CD14 and CD15 were found to be associated with enhanced adherence of non-piliated N. meningitidis to RSV-infected cells.

In the attachment of NTHI to epithelial cells, fimbriae, known as outer membrane protein P5 homologous fimbriae (P5 fimbriae), play a major role. When pneumocyte type II cells (A549) were preinfected with RSV, P5 fimbria-mediated binding to yet undefined receptors was enhanced, suggesting that RSV infection of these pneumocytes increased the attachment sites for NTHI [38].

Another upregulated receptor that has been associated with increased bacterial adherence is the platelet activating factor (PAF) receptor. This receptor is found on endothelium and is a receptor for pneumococci only when the cells are activated by cytokines like tumor necrosis factor a and interleukin 1α [39, 40]. Both cytokines are produced during RSV infection, but no studies have been conducted yet to address the possible role of the PAF receptor in enhanced pneumococcal adherence as a result of a prior viral infection.

Molecules present in the extracellular matrix provide another mechanism for bacterial binding. They serve as a cross-link between the bacteria and virus-induced changes on host-cell membrane. For example, fibrinogen enhances adherence of group A streptococci (GAS) to influenza A-infected cells [41]. It seems likely that either neuraminidase or hemagglutinin (or both) is (are) responsible for the direct binding of fibrinogen, because fibrinogen-mediated adherence was variable with the subtype of influenza virus used in the assay. Moreover, the authors in this study also showed that tunicamycin, an inhibitor of glycosylation of viral glycoproteins like neuraminidase and hemagglutinin, dramatically decreased fibrinogen-mediated binding of GAS to influenza virus-infected cells.

Finally, Sanford et al. studied the molecular properties of adhesins involved in staphylococcal adherence to influenza A-infected MDCK cells [42]. After heat treatment, it was shown that adhesins mediating binding to virus-infected cells were more heat-labile than adhesins involved in binding to non-infected cells.

Autoclaving and heat treatment, however, were unsuccessful in completely aborting adherence. As adhesins are proteins, it was hypothesized that denaturing surface-exposed adhesins may expose a deeper layer of adhesins, probably of another chemical structure.

In an earlier study of the same group it was shown that two other proteins, clumping factor and protein A, that are present in supernatants of heat-treated staphylococci act as adhesins to normal cells but not to virus-infected cells [33]. Their conclusion was that different subsets of adhesins might play a role in the mechanism of virus-induced bacterial superinfection compared to the adhesins involved in a primary bacterial infection.

5 Conclusions

During the last two decades both epidemiological and experimental evidence has been accumulated for a crucial role of preceding viral infection in the etiology of bacterial respiratory infections. This relationship has been documented most convincingly for otitis media in children. These epidemiological observations are now supported by experimental data, from studies with both human cell lines and in animal models. They present unequivocal evidence for enhancement of bacterial adherence by prior viral infection of respiratory cells. Moreover, these experimental studies have shed some light on the pathogenic mechanisms involved although insight is still far from complete.

Finally, understanding these mechanisms is crucial for improving the prevention and management of respiratory infections, which will be particularly helpful in this era of increasing antibiotic resistance.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
  43. [43].
  44. [44].
View Abstract