**4. Discussion**

Patients diagnosed with chronic pulmonary disease were enrolled to analyze the association between influenza vaccination and the prevention of pneumonia. Our study indicated that patients who received vaccination one year prior to the study had a significantly decreased risk of developing pneumonia. Additionally, patients who received vaccination consecutively for two and three years showed a continuous reduction of their risk of pneumonia. Although the patients who received influenza vaccinations only in the second year did not show any decrease in their risk of developing pneumonia, the data was not statistically significant. Repeated influenza vaccination was demonstrated to be effective for preventing severe and fatal influenza infection in elderly individuals [17]. Our study showed that influenza vaccination can lower the risk of pneumonia in individuals aged ≥ 65. A previous study used health administrative databases to investigate the association between influenza vaccination and the all-cause death of elderly people >65 years. They also concluded that influenza vaccination was associated with reductions in the total hospitalizations for pneumonia and influenza and all-cause mortality during the influenza season [21]. Moreover, some studies, such as Li et al., confirmed that previous pneumococcal and influenza vaccination in elderly patients reduced the length of hospital stay and reduced the risk of bacteremia [22], which was similar to our study. However, Demirdogen Cetinoglu et al. showed that influenza vaccination did not affect the clinical outcome of hospitalized adult CAP patients. This result may be related to the low influenza vaccination rate in the elderly [23]. Therefore, further research is needed to analyze this in the future.

Several studies have shown that influenza vaccination not only prevents influenza infection but also reduces the risk of several diseases in certain groups of patients. According to a retrospective cohort study, the administration of the influenza vaccination in elderly patients with diabetes reduced risks of hospitalization, lung failure, and 12-month mortality [15]. Influenza vaccination in elderly individuals reduced the risk of acute kidney injury in a nested-control study. Although the actual mechanism is unclear, influenza vaccination is proposed to be associated with the reduction of the inflammation cascade [19]. A study showed that influenza vaccination reduced dementia risk in patients with chronic kidney disease [24]. Another study also showed that influenza vaccines might prevent cardiovascular disease [16]. A high-dose influenza vaccine (60 μg of hemagglutinin from each of the three viral strains against 15 μg of hemagglutinin from each of the four viral strains) might be more effective in reducing poor clinical outcomes in patients who have a heart failure or myocardial infarction history [20]. In another study, it was found that influenza infection increased the risk of atrial fibrillation, and people who received influenza vaccination showed a lower atrial fibrillation risk [18]. Influenza vaccination could decrease respiratory failure risk in patients with chronic obstructive pulmonary disease [13]. Our study found that influenza vaccination could significantly decrease the risk of pneumonia.

One of the most serious complications of influenza infection is bacterial pneumonia, which increases morbidity and mortality [25]. Evidence shows that older patients, or those with severe illness with pneumonia, had higher 30-day mortality rates [5]. In general, influenza viruses cause only the desquamation of the epithelial cells in the respiratory tract, but this disrupts the outermost part of the mucosal defense and promotes secondary bacterial pneumonia. The most common pathogens associated with secondary infection are *S. pneumoniae*, *Staphylococcus aureus*, and *Haemophilus influenzae* [26]. In Taiwan, to prevent *S. pneumoniae* infection in elderly individuals, the governmen<sup>t</sup> has provided free vaccination of the 23-valent pneumococcal polysaccharide vaccine for people aged > 75 since 2008 [27]. We also considered that one of the factors for lowering pneumonia risk in our enrolled patients might be *S. pneumoniae* vaccination. However, only 257 (1.7%) patients in our study population had received *S. pneumoniae* vaccine. Therefore, the observation that influenza vaccination could reduce the risk of pneumonia in elderly individuals may not be overestimated.

The prevalent strains of influenza viruses in Taiwan are A/H1N1, A/H3N2, and B viruses. For the prevention of an influenza outbreak, the health authority of the Taiwanese governmen<sup>t</sup> recommends trivalent influenza vaccine use. However, because of genetic divergence, influenza B viruses are divided into two lineages: Victoria and Yamagata lineages. Simultaneous cocirculation of these lineages has been observed, and a mismatch of the vaccine lineage and circulating strains has been reported [28–30]. This might decrease the effectiveness of influenza vaccines. However, there was no mismatch of the vaccine strains and the circulating strains in our study during the observation period (2010–2013).

Our study has several limitations. First, in Taiwan, influenza vaccines are purchased from different pharmaceutical companies. The manufacturing process of influenza vaccines varies between pharmaceutical companies. For instance, the inactivated influenza vaccine (i.e., Optaflu ©) produced by Novartis contains whole virus, whereas the vaccine produced by Sanofi Pasteur (i.e., Fluzone ©) contains split virus. We did not examine whether the protective effect of both these influenza vaccines is the same. In addition, the protective effect may be different for different age groups. Second, we were unable to obtain potentially relevant personal behavioral information, such as alcohol consumption, smoking habits, and body mass index. These confounding factors may have affected the outcome. Third, Taiwan's National Health Insurance system involves the Taiwanese population. Our data accurately reflects the situation in Taiwan; however, our results may not be applicable to other regions. Fourth, the National Health Insurance Research Database (NHIRD) consists of claim data. The database used does not contain information about clinical parameters, such as pneumonia infection severity. Therefore, we could not further distinguish the severities of respiratory diseases. Fifth, the laboratory data and microbial culture data that may affect the occurrence of pneumonia infection are not included in the database. Moreover, information about from sputum cultures or viral swabs and the causative agents of pneumonia was not available. Thus, the cause of pneumonia identified could not be divided into viral or bacterial.
