Recently, the GAVI Alliance (formerly the Global Alliance for Vaccines and Immunization), a partnership of public and private sector organizations dedicated to “immunization for all” (http://www.gavialliance.org), started the introduction of pneumococcal vaccines in developing countries, especially in Africa , and in May 2012 PCV-13 was introduced in Ghana. Studies on the carriage rates of S. pneumoniae in healthy subjects should be conducted before the introduction of conjugate pneumococcal vaccination campaigns to provide baseline information on vaccine coverage and to allow identification of changes to the serotype distributions resulting from pneumococcal serotype replacement .
Our study provides the first large data set on S. pneumoniae carriage and serotype distribution in healthy Ghanaian children. Carriage prevalence was 31%-34% in our study compared with the 15% previously reported in Ghana . This difference might be due to the age differences between our study subjects (≤6 years of age) and those of the previous study, which included children ≤13 years of age . The overall carriage in our study was relatively low when compared with prevalence rates reported in other African countries; 22%-60% in Kenya [22, 23], 62% in Uganda , 90% in Gambia  and 35% in Tanzania . This may be due to the low numbers of children below the age of two years included in this study (Table 1). As our data on pneumococcal carriage were collected before the introduction of PCV-13, they can serve as a baseline to measure possible future serotype replacement associated with the introduction of PCV-13 .
Penicillin has been the drug of choice worldwide for the treatment of pneumococcal infections. However, penicillin resistant strains have emerged, resulting in a shift to the use of other drugs, e.g. cefotaxime, chloramphenicol and erythromycin . In an earlier carriage study from the Ashanti region of Ghana, 49% of children were found to carry S. pneumoniae and 39% of isolates showed intermediate penicillin resistance . Holliman et al.  described the history of S. pneumoniae penicillin resistance in Ghana and found that overall around 12% of isolates tested were intermediate resistant with the exception of a study in 1996 that reported 31% of isolates to be resistant . However, this high rate of resistance was suspected to be due to the methodology used in the specific study [27, 28]. In a recent study conducted in the Ashanti region from 2008 to 2010, 99% of 91 invasive S. pneumoniae isolates were penicillin sensitive . In our study, two isolates were found to be fully penicillin resistant and 45% (n = 130) showed intermediate penicillin resistance. These data suggest that intermediate resistance to penicillin has increased among S. pneumoniae in Ghana during the last 5-10 years and that the prevalence of penicillin resistance is in line with the situation in other African countries. Despite a possible increase in intermediate penicillin resistance, and depending on the site of infection, penicillin can still be used for the treatment of pneumococcal infections in Ghana.
The S. pneumoniae serotypes found show that PPV-23 covers 55% of serotypes while PCV-10 and PCV-13 cover 40% and 50% of serotypes, respectively. As PCV-13 is currently used nationwide in Ghana to vaccinate children ≤5 years with a vaccination schedule at 6, 10 and 14 weeks, there might be a high risk of replacement with serotypes not covered by PCV-13. In addition, a high proportion of non-vaccine serotypes exhibited intermediate resistance to penicillin. It is therefore imperative that nasopharyngeal carriage and penicillin resistance of S. pneumoniae be monitored regularly in children . Prior to our study, limited data were available on serotype distribution among S. pneumoniae in Ghana [18, 19]. Thus, contrary to recommendations [7, 8] recent surveillance data on prevalent serotypes were not available in Ghana when PCV-13 was introduced in May 2012. Our finding that only 50% of the serotypes found are included in PCV-13 highlights the importance of conducting such carriage studies before the introduction of any vaccine.
In this study we focused on children from nurseries and kindergarten, and as a result, very few children <11 months of age were included. This is a limitation of the study, as other studies in Africa have shown a very high carriage rates in this age group . We chose to study children attending nurseries and kindergartens as the length of time and close physical proximity to other children in such locations constitutes an optimal environment for horizontal spread of pneumococci . These sites act like reservoirs of different serotypes, which can then easily spread to the surrounding community. It is therefore of interest to see which non-vaccine serotypes might take over after vaccination . Although we only recruited children from two cities in Ghana, we are of the opinion that the data on carriage rates are representative of other regions of Ghana as we did not observe large differences between the carriage rates of the two sites, Accra and Tamale, even though they are geographically apart and have different climates. In addition, we did not observe any marked differences in serotype distribution between the two study sites (Figure 1, Tables 1 and 2).