Factors such as age, ethnicity, immunosuppression including HIV-1 infection, socioeconomic factors like crowding, and temporal association to respiratory viral diseases widely affect the risk of IPD [4]. In particular, a high carriage prevalence has been associated with high incidences of pneumococcal disease, including IPD [5]. Attendance to day care centres (DCC) has been recognized as an important risk factor both with regard to acquisition of pneumococcal carriage and infections in pre-school children [6, 7]. In Denmark, it is estimated that between 55–85 % of children between 6 months and 5 years attend a DCC [6, 8].

The heptavalent-pneumococcal conjugate vaccine (PCV7, Pfizer Vaccines) has been shown to be highly effective in reducing IPD caused by vaccine serotypes in both vaccinated and non-vaccinated cohorts [9-12]. PCV7 was introduced in the Danish childhood immunization programme in a 2+1 schedule in October 2007; before that, PCV7 was only recommended for children younger than 2 years of age with high-risk conditions [13, 14]. In the pre-PCV7 period, the incidence of IPD among children younger than 5 years in Denmark was approximately 25 per 100.000 population, with the highest incidence rates observed in children younger than 2 years of age [15]. The incidence of IPD nearly halved shortly after the introduction of the vaccine [16]. Pneumococcal conjugate vaccines of higher valence, including the 10-valent pneumococcal non-capsulate H. influenzae protein-D conjugate vaccine (PhiD-CV; PCV10; GlaxoSmithKline Biologicals) and the 13-valent pneumococcal protein conjugate vaccine (PCV13; Pfizer Vaccines) provide broader serotype coverage [11, 15, 17, 18]. PCV13 was started to be used in Denmark in April 2011 [28].

We found that both the carrier rates and the serotype distribution of carried isolates varied markedly with age. The high prevalence of carriage among the youngest children coincides with the observation that the incidence of IPD is highest in this age group [8]. Paediatric serotypes/groups (6, 14, 19 and 23) were found frequently both in carriers and as a cause of IPD in the youngest children in the pre-PCV7 period. These serotypes/groups are all represented in PCV7, explaining the higher coverage of the PCV7 among IPD cases in children <23 months, for whom vaccination is attempted (Fig. ​22). The predominance of serotypes included in PCV7 declined significantly with increasing age, both among carriers and IPD cases, in accordance with findings reported from other populations [2, 9]. Serotype 5 is not a frequent cause of paediatric IPD in Denmark, while serotypes 1 and 7F become more important in IPD in older children and adults [8, 16]; all these factors may determine a relatively low coverage of PCV10 among carried isolates in our study population. The higher serotype coverage of PCV13 was mainly related to the presence of serotype 6A in both carriage and in IPD (Fig. ​22). These findings are comparable to those reported from a number of carriage studies from different geographic locations [7, 11, 22, 23]. Also, the analysis of risk factors for the carriage of vaccine serotypes showed that age, gender, size of DCC, type of DCC, and month of sampling were not found to be significantly associated with an increased risk of carrying an isolate belonging to a serotype included in PCV7, PCV10 or PCV13 (data not shown).

The pneumococcal capsular polysaccharide has been described as a major determinant of the duration of carriage, relative invasiveness, disease presentation and mortality related to IPD [13, 20]. It also appears that the biochemical structure of the capsule determines the success of a given serotype in nasopharyngeal carriage [24]. In spite of differences in study design and population, a number of studies have identified a few serotypes as having the greatest invasive potential. Similar to our findings, a study from Gambia reported that serotypes/groups 1, 12 and 14 were more frequently found in patients with severe pneumococcal disease than in carriers, while serogroups 19 and 23 were often isolated from healthy controls [5]. In a study from India, serotypes/groups 1, 5, 7, 12 were prevalent among invasive isolates, but not among carriers [25]. Also, a study from Papua New Guinea [26] reported that serogroups 11, 15 and 22 were of low invasiveness, similar to what we found in our study. A study from Oxford reported that PCV7 serotypes showed to be associated with high invasiveness [27]. The consistency between these reports supports the fact that the invasiveness of capsular serotypes is a serotype-specific characteristic. However, even if capsular serotype is a factor that determines the invasiveness of a particular strain, the study population may determine the prevalence of host-related factors (such as genetic predisposition, the prevalence of co-morbid conditions and socioeconomic factors), which influence the invasiveness of serotypes.