This is the first assessment of pneumococcal carriage in Aboriginal people in WA in the PCV era. The PCV program aimed to reduce the burden of invasive pneumococcal disease (IPD) but we also anticipated that it would alter the serotypes that are carried in the nasopharynx. While pneumococcal carriage rates were high, 7vPCV serotypes were rarely carried. Despite a lack of data on carriage serotypes in WA prior to introduction of 7vPCV, it is likely that the vaccination program has contributed to reducing carriage of 7vPCV serotypes. IPD due to 7vPCV serotypes accounted for 40% of IPD in 1997-2001 compared with 12% in 2005-2007 [2
]. The high carriage rates of non-7vPCV serotypes may counterbalance the benefits of the vaccine, assuming that some non-7vPCV serotypes may display similar or higher disease potential [14
The high pneumococcal carriage rates found in this study correspond well with studies in the Northern Territory of Australia where prevalence of carriage surpassed 80% in children aged 2-4 years in a cross-sectional study carried out in 2002 and 2004 [15
]. Pneumococcal carriage rates were lower in older adults in WA than Mackenzie et al. found in the Northern Territory but too few samples were collected from people aged ≥ 65 years in our study to enable accurate comparison. Elsewhere in the world, studies in indigenous populations and in developing countries have found comparably high rates of pneumococcal carriage [16
]. Risk factors for high carriage rates in Australian Aboriginal people including crowding and indoor smoking [4
] were widespread among our study participants.
Four of the 12 most common carriage serotypes (6C, 16F, 23B and 34) are not covered by any currently licensed vaccine (). Serotype 6C was the most frequently carried serotype in our study (). In Alaska and the UK, carriage of serotype 6A declined following 7vPCV programs, while the proportion of 6C carriage isolates increased [7
]. While data on carriage of 6A in WA are not available for the pre-7vPCV era, it is possible that a similar “replacement” took place following the 7vPCV program, given that 6C was the most commonly carried serotype in our study. Vaccination with 7vPCV, which includes 6B, elicits cross-reactive antibodies to 6A but does not appear to protect against disease caused by 6C [20
]. Meanwhile Cooper et al. have reported that 13vPCV elicits cross-protective functional antibodies to 6C in addition to covering 6A and 6B [22
]. We could expect that carriage of serotype 6C may now decrease following the introduction of the 13vPCV program for all children.
Serotype 16F has been one of the most common carriage serotypes across Australia since the introduction of 7vPCV and is a predominant cause of OM and tympanic membrane perforation in Australian Aboriginal children [23
]. Serotype 19A was the third most common serotype isolated in our study, and was the most common serotype in children < 2 years of age. We expect to see a reduction in carriage of 19A over the coming years following the introduction of 13vPCV. Despite their inclusion in 7vPCV, serotypes 23F and 19F were the fourth and fifth most prevalent serotypes found in this study. This was unexpected and requires monitoring over the coming years to determine whether the 13vPCV program reduces carriage of these serotypes. The high frequency of nonserotypeable isolates circulating in the population warrants close monitoring, and improved techniques are needed to identify whether these are non-capsular strains or novel serotypes.
23vPPV serotypes made up half (49.9%) of all serotypes found in this study. 23vPPV was included on the immunization schedule for Aboriginal children at 18 months of age and for Aboriginal adults ≥ 55 years in WA. The relatively high prevalence of 23vPPV serotypes in this study suggests it has little effect on carriage. Outbreaks of IPD caused by serotypes 12F and 19A in 2010 and 2011 indicate 23vPPV may not have been effective in preventing IPD or this may reflect the limited 23vPPV coverage in WA [2
]. We rarely identified serotypes 1 and 12F in the nasopharynx (), despite the outbreaks of IPD due to these serotypes during the study period [5
]. It is not possible to determine whether we were observing replacement disease due to “vaccine pressure” or whether this was part of the natural variation in incidence of these serotypes. This could be expected due to the opportunistic nature of our carriage surveillance resulting in swabs not being collected at the time and in locations where serotype 1 or 12F was circulating. Furthermore, serotypes 1 and 12F have high invasive potential which suggests more transient carriage, so they are rarely isolated from the nasopharynx even in very large carriage studies [25
Between 1999 and 2003 carriage rates were almost 50% lower in non-Aboriginal children than in Aboriginal children in the Kalgoorlie-Boulder cohort [12
]. During that time period Aboriginal children began receiving 7vPCV while non-Aboriginal children did not routinely receive 7vPCV until 2005. More recently, between November 2007 and May 2009, a cohort of 186 non-Aboriginal children < 36 months old experiencing recurrent acute OM (rAOM) and 81 healthy controls were recruited in a carriage study in WA. S. pneumoniae
, non-typeable H. influenzae
and M. catarrhalis
were carried by 41%, 56% and 43% of the rAOM children and 26%, 19%, and 15% of controls [28
]. The carriage rates in both the rAOM and control groups are considerably lower than we found in Aboriginal children of the same age in this study. Ongoing simultaneous surveillance of carriage in the Aboriginal and non-Aboriginal populations is required to examine and compare the effect of vaccination in both groups and give a broad overview of carriage in the WA population. Surveillance of bacterial nasopharyngeal carriage in non-Aboriginal children < 5 years has recently begun in the metropolitan Perth area, and will enhance our overview of carriage in the region.
Pneumococcal antibiotic resistance remains relatively uncommon in this population apart from cotrimoxazole, which is important when considering empiric therapy for common infections such as OM and pneumonia. Reduced susceptibility to penicillin was observed in one-quarter of isolates, but was particularly common in nonserotypeable isolates that were also more likely to be resistant to cotrimoxazole and erythromycin. We compared our antibiotic susceptibility data with resistance rates for 261 isolates (Lehmann et al., unpublished data) from a cohort of 100 Aboriginal children born in Kalgoorlie-Boulder and followed for 2 years between April 1999 and January 2003 [29
]. Cotrimoxazole resistance was more common in our isolate collection (30.7% versus 22.6%, p = 0.01), but susceptibilities to the other antimicrobial agents were not significantly different. It will be important to continue to monitor antibiotic susceptibility for the emergence of multi-resistant strains.
Our study had some limitations. The opportunistic nature of swab collection may not give a representative sample of the population. Sampling in adults was biased towards women: 77% of adults aged over 20 years were female because generally mothers and grandmothers accompanied their children to medical services, where we had the opportunity to ask them if they would like to participate in the study. However, carriage rates were not significantly different in men and women aged ≥ 20 years (p = 0.79). Vaccination data from ACIR could only be accessed for 56.4% of children swabbed, so our findings may not reflect the true number of vaccinated children. The large expanse of land and diverse regions covered by our surveillance could give rise to seasonal or geographical variation in prevalence which we could not ascertain as samples were frequently collected from limited numbers of people living in many different communities visited in different years and seasons.
Since July 2011, 13vPCV has replaced 7vPCV in the WA immunization schedule. We therefore anticipate a shift in serotype distribution, with a decrease in circulation of the six additional 13vPCV serotypes now covered in the immunization schedule. Over 30% of our isolates were 13vPCV serotypes. It is essential to closely monitor changes in carriage and IPD (rates, serotype distribution and antimicrobial susceptibility patterns) over the coming years to ensure appropriate vaccine policies are in place and to achieve the best outcome for the WA Aboriginal population [30