In the present study, blood was the biological material most often used for diagnosis. The isolation of pneumococcus in culture is considered the gold standard for the definition of pneumococcal disease, although the positivity of blood culture is low, especially in children . It is worth highlighting the importance of routine blood culture for patients suspected of having bacteremia, particularly in relation to the meningitis diagnosis, in which blood culture is often neglected when just the culture of cerebrospinal fluid is targeted .
Besides the low positivity of blood culture, another aggravating factor for proper diagnosis, especially for pneumonia, is cases that are cared for in clinics and treated empirically, without collection of biological samples for identification of microorganisms. Diagnosis is also hampered by previous use of antibiotics before the collection of biological material . It is estimated that around 50% of cases of pneumonia remain undiagnosed . Therefore, the extent of pneumococcal disease is underestimated and difficult to assess.
In the present study, one possible reason for the higher frequency of meningitis is its compulsory notification and a more active surveillance system when compared to pneumonia.
After the introduction of PCV7 in the United States, beyond the direct effect on the reduction of cases of pneumococcal disease in the vaccinated population, the herd effect was also observed, which acted indirectly to prevent disease in non-vaccinated individuals as a consequence of reduced circulation of vaccine serotypes . The data presented in the current study were through 2013, so a further observation period in the unvaccinated population would be necessary to detect the full effect of herd immunity.
In Brazil, PCV7 was introduced in 2002, and was provided free of charge only to people at high risk of acquiring IPD and in private clinics, reaching a small portion of the population, verifying the absence of response to this intervention.
In 2010, Brazil introduced PCV10 into its routine National Immunization Program, aiming to minimize the impact of pneumococcal disease. Research on the effectiveness of PCV10 implementation for public health services in Brazil has indicated a reduction in children’s hospitalizations for pneumonia ; protection of nasopharyngeal pneumococcal carriers against vaccine serotypes ; and reduction of IPDs cases in all age groups, especially in children under 2 years old .
In 2011 and 2012, vaccination coverage of PCV10 in Brazil was 82.1% and 88.4%, respectively . In the cities of Ribeirão Preto, Araraquara, Franca and Barretos, the coverage of PCV10 in children under 5 years old was around 86% , not reaching the ideal rate of 95% recommended by the Ministry of Health for efficient vaccination coverage for PCV10. In addition, completion of 3 doses of the vaccine was necessary in order to ensure the reduction of IPD cases , which explains the persistence after 2010 of some serotypes contained in the vaccine.
Classically, pneumococcal diseases mainly affect children and the elderly . However, an understanding of pneumococcal epidemiology should involve all age groups. In mid-2012 in Argentina and Bolivia, most cases of IPDs occurred in patients between 2 and 5 years old. In Brazil, there was a higher incidence in people between 30 and 49 years old , coinciding with our data.
Meningitis is commonly associated with children under 2 years old, and mostly under 1 year old . Contradicting our data, in Uberlândia, the Federal District  and Rio Grande do Norte , pneumococcal meningitis prevailed in children under 5 years old. In the region studied, meningitis occurred mainly in adults.
From 2001 to 2011 in Brazil, 282,593 cases of meningitis were identified; 100,559 of those were bacterial meningitis, of which 13,469 were related to S. pneumoniae . From 1999 to 2010, it was observed that 32.5% of bacterial meningitis in Rio Grande do Sul was caused by pneumococcus . In the present study, the percentage of patients with meningitis was higher than that.
Considering that in Brazil, PCV10 is given free only to children under 2 years old, the occurrence of IPDs in patients over 5 years old is relevant information, especially taking into account that in recent decades, increases in the elderly Brazilian population have accelerated . This highlights the importance of constant surveillance, aiming to detect changes in the circulation of serotypes in this age group.
With the use of conjugate vaccines, a decrease in vaccine serotypes, and an increase in those not included in the vaccines, a phenomenon known as serotype replacement has been observed in several countries [28,29,30,31,32,33]. Therefore, when evaluating the effect of pneumococcal vaccines, the overall rate of disease occurrence may not decrease significantly, because of an increase in other non-vaccine serotypes .
The increase in non-vaccine serotypes detected in the present study suggests serotype replacement. This conflicts with a study done in São Paulo that showed no significant increase in non-vaccine serotypes in IPDs after the introduction of PCV10 . Fluctuations in the occurrence of serotypes can happen naturally or by selective pressure caused by vaccines, so increases in non vaccine serotypes should be carefully evaluated. Complexes and varied factors are involved in the replacement phenomenon, making it difficult to analyze .
Currently, the most advanced technology for pneumococcal vaccine is polysaccharides containing capsular conjugate vaccines linked to a carrier protein. However, there have been studies of the implementation of a new generation of vaccines based on common protein components of S. pneumoniae that provide serotype independent immunity . According to the World Health Organization, adoption of this new technology will need to provide equivalent or greater benefit than that obtained with conjugate vaccines to prevent pneumococcal disease .
Thus, changes in the incidence of pneumococcal serotypes associated with the disease after the use of conjugate vaccines must be distinguished from normal serotype temporal changes .
In this scenario, serotype replacement after vaccination may raise questions regarding the effect of this intervention. However, some non-vaccine serotypes seem to be less invasive than those present in the vaccine, so that reduction of IPDs becomes a more positive effect than eventual serotype replacement .
The target for prevention of pneumococcal diseases with vaccines is the bacterial capsule. The effect of vaccination can be reduced by the classical phenomenon of capsular exchange, in which a vaccine serotype changes its capsular locus and begins to express other non-vaccine serotypes .
Noteworthy is an increase in serotypes 3, 12F, 11A and 8, reinforcing the need for continued vigilance. With the exception of Serotype 3 included in PCV13, no other serotypes are included in the available vaccine formulations (PCV10 and PCV13).
An example of serotype replacement is what happened with 19A, which is cited as one of the emerging serotypes in the United States since the introduction of PCV7 . However, it stabilized after an increase in the use of PCV13, in which it is included . During the present study, prevalence of serotype 19A did not change, which is similar to what was reported in São Paulo . There was a predominance of 19A in children under 5 years old. This is similar to results found in the United States, where the frequency of this serotype in children under 6 years old was substantial .
In Brazil, as in the Caribbean and some other Latin American countries, there is low frequency of serotype 19A . Although a case-control study on the impact of PCV10 in Brazil showed cross-protection between serotypes 19F and 19A , an increase in the circulation of Serotype 19A was reported in 2012 .
In the Central Laboratory of Paraná, in 2001–2002, pneumococcus was the most common etiologic agent in acute bacterial meningitis, with higher incidence of serotypes 14, 23F and 3 . In our study, serotypes 14, 3 and 19F were more involved in meningitis. The potential for invasion expressed by Serotype 1 has been linked to outbreaks and fatal cases of meningitis in Africa  and in some countries in Europe , but in the present study, Serotype 1 was not detected among the main causes of meningitis. In Latin America, this serotype has been considered a major cause of IPDs in all ages .
Serotype 3 has a specific virulence factor for pneumonia associated with severe disease cases , especially in the elderly . This corroborates the present study, in which this serotype was important in pneumonia.
In RRAS 13, serotypes 14, 6B, 6A, 19A and 3 were among the leading causes of IPDs in children younger than 2 years old. Serotypes 14 and 6B are included in PCV10; the remaining serotypes are included only in PCV13. In 2010, countries like Italy  and the U.S.  replaced PCV7, in use since 2000, with PCV13 because of its higher serotype coverage.
In the age group of 60 years old or older, serotypes 3, 9 V, 19F and 23F were the most common. One study found that in Argentina, most IPDs were caused by serotypes 14, 1 and 5, and in Bolivia the most frequent serotypes were 14, 6B and 18F .
Among the specific serotypes of PCV13, serotypes 6A and 19A did not present a significant increase during the course of the present study. Serotype 3 presented a progressive increase (Tables 3 and 4), especially in cases of meningitis and pneumonia (Table 5), with emphasis on patients over 20 years old (Table 6). This increase should be monitored to evaluate the circulation of this serotype in the community.
This research has limitations. Because it is a retrospective study, results and conclusions are based on social and microbiological pre-existing information, thus are subject to information bias. In addtion, there is to consider the use of sampling related to convenience criterion adopted for the study.