The most important risk factor for CAP is age. So much so, that the incidence triples or quadruples when age is >65 years.2 The reason is unclear, but several factors may be involved, related to immune senescence due to weakness (innate and adaptive) of the immune system, increased asymptomatic aspirations in this age group,6 or the convergence of other risk factors (RF) that do not acquire statistical significance in younger people, but that take center stage in older age groups, possibly due to an additive effect. One of these factors may be tobacco smoking, the main RF for COPD, which in turn is one of the main RFs for CAP. Thus, it is often difficult to determine the specific weight of smoking in the development of CAP, especially considering that smoking is associated with lower socioeconomic status, poor diet, increased alcohol consumption, and reduced physical activity, that are also RFs for respiratory infection.7 Nevertheless, enough studies using statistical analysis with logistic regression models are now available to allow the differentiation of any confounding factors, showing that tobacco smoking in itself increases the risk of pneumonia. In this regard, a population-based study2,3 was published that demonstrated an increased risk of CAP in smokers that was also directly related to the duration of tobacco consumption and the number of cigarettes consumed. Concurrently, it has been shown that this risk disappears when smokers quit. In fact, after 5 years, the risk decreases by 50%.3 Moreover, if we consider that one of the criteria for causality between two factors is dose-response, the increase in risk associated with an increase in intensity of consumption and the overall amount of tobacco consumed further reinforces the notion that tobacco smoking is a direct causal agent of CAP. On the other hand, we know that tobacco smoking causes morphological changes in the epithelium of the bronchial mucosa, with loss of cilia, mucous gland hypertrophy and increased goblet cells that may favor the presence and spread of microbes in the bronchial tree. An inflammatory reaction is produced in the airway causing macrophage and neutrophil activation that releases proteases. Oxidative stress and cytokine release are triggered, leading to both innate and adaptive immune response.8–14 This may in turn make the bronchial mucosa epithelium more sensitive to the inflammatory aggression of the infection itself.15–17 Moreover, it is now known that tobacco smoking inhibits some of the key functions of the innate and adaptive response, including the response of two Toll-like receptors (TLR2),18 nuclear factor kappaB (NF-κB),8 CD4-lymphocyte proliferation (LTCD4),12,19 maturation of dendritic cells,10 and opsonization and phagocytosis capacities.20 Thus, tobacco smoking can alter immunity against infection, affecting the modulation of intra- and intercellular signaling of epithelium and immune cells, and suppressing the activation of important elements of the innate and adaptive immune response.17,21 The fact that tobacco smoking increases susceptibility to bacterial infection is well established.22,23

Passive smoking is considered a major epidemiological problem. In 2004, an international study including 192 countries found that 40% of children, 33% of male non-smokers and 35% of female non-smokers, were passive smokers.24 There were wide geographical variations that could be explained by the different stages of the tobacco smoking epidemic in the country surveyed, because passive smoking is directly related to active smoking rates. In the year 2005, prevalence data in Spain estimated that about 50% of adults were exposed to second-hand smoke.25 Later, as in most developed countries, legislation was introduced to regulate tobacco smoking in public places. This has reduced exposure to tobacco smoke by approximately 20%–25%, due to falling consumption in public places and in the workplace, but not at home. Accordingly, in recent years there has been growing interest in understanding the effect of passive tobacco smoking and some studies suggest that it could also involve a higher risk of respiratory infections in both the children of smoking parents,26,27 and in adults.28,29 Allowing smoking in the home has also been shown to be an important predictor of health loss in elderly people with CAP.30 In fact, lesions in the respiratory epithelium, connective tissue and vascular endothelium of the lung caused by tobacco smoke can occur even at low smoke concentrations,31 supporting the notion that exposure to tobacco smoke could be an important risk factor for CAP development in passive smokers.

Two separate case-controlled,32 population-based studies have also demonstrated that passive smoking is a risk factor for developing pneumococcal bacteremia in immunocompetent adults, resulting in an odds ratio (OR) of 2.6 and attributable risk (AR) of 31% in subjects aged <65 years,28,33 and OR 2.2 and AR 13% in those over 65 years of age.28,32

The most common causative organism of CAP, regardless of care level and severity, is Streptococcus pneumoniae. This has also been shown to be the agent most frequently linked to smoking, especially in patients with COPD. An in vitro study has shown increased adherence of S. pneumoniae to the epithelial cells in the oral cavity of smokers,34 persisting for up to three years after smoking cessation. This may produce greater oropharyngeal colonization and lead to greater chance of developing CAP.

Smoking has also been related with invasive pneumococcal disease (IPD), which in approximately 80% of cases is due to pneumonia.35 This association was reported in a study by Pastor et al. in 1995, with an OR of 2.6 in smokers between 24 and 64 years of age and OR of 2.2 in smokers older than 65 years. Attributable risk (AR) was 31% in the first group and 13% in the second group.3 Another population-based, case-controlled study28 also highlighted that smoking is the largest independent RF for IPD in immunocompetent adults, with an OR=4.1 in current smokers (AR 51%). Furthermore, there is a dose-response effect, and smoking cessation dramatically reduces the risk after 10 years, when it becomes the same as that of nonsmokers.

The special relationship of active smoking with pneumococcus has been demonstrated by some studies that have shown tobacco-induced changes in the clearance and phagocytosis of S. pneumoniae in the lungs,36 and inhibition of the anti-pneumococcal activity of some innate immunity antimicrobial peptides.37,38 Thus, tobacco smoking impairs immunity to infection, especially against some microorganisms, including S. pneumoniae, and it seems well established that active smoking increases the risk of pneumonia.

However, the information as to whether active smokers have a worse CAP prognosis is not so clear. Three longitudinal studies39–41 involving large patient cohorts found higher mortality in smokers with pneumonia, while two metaanalyses42,43 found that tobacco smoking did not affect prognosis. Two large retrospective series that studied CAP mortality in relation to tobacco smoking have shown contradictory results.44,45 Finally, a recent study showed that tobacco smoking was an independent risk factor for septic shock in pneumococcal pneumonia,46 and another epidemiological study concluded that smoking was an independent determinant of mortality in bacteremic CAP caused by S. pneumoniae.47 Considering all these data, we can conclude that there is evidence to suggest worse prognosis of pneumococcal pneumonia in active smokers, while the picture is not so clear if we consider overall CAP, independent of the etiology.

Less frequent, but not less important, is CAP due to Legionella pneumophila. Its incidence varies according to published series: some report sporadic cases while in others it is the second or third most common etiologic agent. This depends on epidemiological factors, the completeness of the diagnostic method used, and the severity of CAP, among other factors. In population-based studies, L. pneumophila represents 1%–3% of diagnosed microorganisms.48 This type of CAP usually affects young individuals, often without underlying disease.49 Smoking is the most important risk factor in these subjects, as the risk may be increased by 121% for each pack of cigarettes consumed daily, with OR of 3.48,50 and as much as 7.49 if analyzed in people without comorbidities. This increased risk is attributed to the difficulty in eradicating the microorganism from the bronchial tree, due to the deterioration of the respiratory mucosa and impaired cilia caused by tobacco smoking that, depending on the individual's immune status, facilitates entry into and subsequent invasion of the alveolar macrophages.

Tobacco smoking may also favor the presence of diseases which are in turn known risk factors of CAP, including:

• -

  Periodontal diseases directly related to the development of CAP.51 Poorer oral hygiene has been shown to potentially lead to increased respiratory infection.52,53 Additionally, smokers show higher subgingival bacterial colonization than non-smokers, which is directly related to the number of cigarettes smoked by day. Accordingly, smokers of fewer than 10 cigarettes/day present periodontitis between 2.5 and 6 times more often than non-smokers (OR 2.79), while OR is 5.88 for those smoking >30 cigarettes/day; OR decreases dramatically if the habit is quit and can reach OR 1.15 after 11 years.54

• -

  Viral upper respiratory tract infections. Upper respiratory tract infections (URTI) are a risk factor of CAP, independent of other related factors,2,7 and smoking independently influences the presence of prior viral infection. Blake et al.55 observed in a cohort of soldiers diagnosed with URTI that 22.7% were smokers, vs 16% non-smokers (RR: 1.5). Increased susceptibility to experimental viral infection has also been observed in smokers.56

At present there is enough scientific evidence to show that active tobacco smoking is an important risk factor for CAP26,28,29,33; it has a direct and independent effect on the risk of CAP, but it may also act indirectly causing chronic bronchitis or COPD which, in turn, are well recognized risk factors for CAP.7,33 Therefore, one of the main ways of preventing CAP is to intervene in the smoking habit, advising its cessation, which may reduce the risk of CAP by 50% after five years.3 Furthermore, a reduction of 14% in IPD has been observed every year28 and, after 10 years, the risk of IPD can reach the same level as that of non-smokers.

On the other hand, pneumococcal vaccine is concomitantly recommended. In fact, the American scientific societies propose tobacco cessation combined with pneumococcal vaccine in patients hospitalized for CAP,57 and further recommend the pneumococcal vaccine for active smokers in the population aged between 19 and 64 years, regardless of the presence of comorbidities.58 This recommendation is even more relevant for active smokers, in whom lower rates of influenza and pneumococcal immunization have been observed compared to former smokers and non-smokers.59

At present, two types of vaccine are available: the 23-valent polysaccharide vaccine and 13-valent conjugated vaccine. The former has been used in Spain since 1999, and protective effect has been found in studies not designed to assess its effectiveness,7,60 while clinical trials to test its efficacy against IPD do not establish such an effect, and effectiveness on CAP in general cannot be estimated.61 Therefore, pneumococcal immunization with the 13-valente conjugate vaccine, authorized for adults in Europe since October 2011 and in Spain since July 2012, has been recently recommended.62 The main advantage over the former is that, being conjugated, it is capable of inducing a T-dependent immune response that provides a better immune response and generates immune memory.63 Additionally, it has been described to generate IgA responses in mucosa, reducing nasopharyngeal colonization by serotypes represented in the vaccine, thus breaking the infectious cycle and producing group immunity.64 Furthermore, the serotype coverage of the 13-valent conjugate vaccine reaches up to 82% of CAP-producing agents in healthy adults.65 Therefore, it seems clear that the smoking population should be considered as a target for pneumococcal vaccination strategies.66

The authors declare no conflicts of interest.