Skip Navigation

Indirect Effect of Conjugate Vaccine on Adult Carriage of Streptococcus pneumoniae: An Explanation of Trends in Invasive Pneumococcal Disease

  1. Laura L. Hammitt1,2,
  2. Dana L. Bruden1,
  3. Jay C. Butler1,
  4. Henry C. Baggett1,a,
  5. Debby A. Hurlburt1,
  6. Alisa Reasonover1 and
  7. Thomas W. Hennessy1
  1. 1Arctic Investigations Program, National Center for Infectious Diseases, Centers for Disease Control and Prevention, and
  2. 2Alaska Native Tribal Health Consortium, Anchorage, Alaska
  1. Reprints or correspondence: Dr. Laura Hammitt, CDC Arctic Investigations Program, 4055 Tudor Centre Dr., Anchorage, Alaska 99508 (lhammitt{at}cdc.gov)

Abstract

BackgroundUse of heptavalent protein-polysaccharide pneumococcal conjugate vaccine (PCV7) has been associated with decreases in PCV7-type invasive pneumococcal disease and nasopharyngeal (NP) carriage in children. Vaccine use has also indirectly decreased the rate of invasive disease in adults, presumably through decreased transmission of pneumococci from vaccinated children to adults

MethodsWe conducted NP carriage surveys in 8 villages in Alaska in 1998–2004. Streptococcus pneumoniae isolates were characterized by serotype and antimicrobial susceptibility. We analyzed trends in serotype distribution, antibiotic resistance, and factors associated with adult carriage of PCV7-serotype pneumococci before and after the introduction of PCV7 in 2001

ResultsWe collected 15,598 NP swabs; overall, 52% of adults living in the villages surveyed participated in the colonization study. The proportion of adult carriers with PCV7-type pneumococcal carriage decreased from 28% of carriers in 1998–2000 to 4.5% of carriers in 2004 (P<.0001). Among adults, the proportion of colonizing isolates that were resistant to penicillin decreased from 13% in 1998–2000 to 6% in 2004 (P=.05), whereas the percentage of isolates with intermediate susceptibility to penicillin increased from 12% in 1998–2000 to 19% in 2004 (P<.01). Adults were more likely to carry PCV7-type pneumococci if they lived with a child <5 years old or if they lived with a child who had not been age-appropriately vaccinated with PCV7

ConclusionsPediatric vaccination with PCV7 has resulted in decreased PCV7-type pneumococcal carriage among adults and helps to explain recent decreases in the rate of PCV7-type invasive pneumococcal disease among adults

Streptococcus pneumoniae is a leading cause of bacterial meningitis, community-acquired pneumonia, acute otitis media, and sinusitis, and it causes significant morbidity and mortality in the United States and throughout the world. It is also frequently present in the nasopharyngeal (NP) flora of healthy persons. Virtually all children are colonized with S. pneumoniae sometime during the first 2 years of life [1, 2]. Since the addition of a heptavalent protein-polysaccharide conjugate vaccine (PCV7) to the routine childhood vaccination schedule in 2000, numerous studies have documented declining rates of colonization with PCV7 serotypes and a lower incidence of PCV7-type invasive pneumococcal disease among young children [36]

Children are more commonly colonized with S. pneumoniae than are adults, and the highest rates of carriage are among preschool-aged children [7]. Adults living with preschool-aged children are more likely to be colonized than are adults who do not live with children [8], and adults who are in close contact with young children may be at greater risk for invasive disease than are adults who do not have contact with children [9, 10]. PCV7 is currently not recommended for adults, but its widespread use in children could indirectly affect the rate of invasive disease in adults by reducing PCV7-type colonization through decreasing the transmission of PCV-type pneumococci from children. Indeed, the decrease in rates of PCV7-type invasive disease among vaccinated children has been accompanied by a decrease in the rates of PCV7-type invasive disease among adults [3, 6]

PCV7 was included in the statewide vaccination program in Alaska beginning on 1 January 2001. In 2001–2003, the rate of PCV7-type invasive pneumococcal disease in children <2 years of age decreased 86% in Alaska, compared with the rate noted for the prevaccine years 1995–2000 [11]. During the same periods, PCV7-type invasive disease in adults ⩾18 years of age decreased by 40% in Alaska (P<.001; Centers for Disease Control and Prevention [CDC], unpublished data). To better characterize the indirect effect of pediatric PCV7 use on the rate of infection in adults within communities and households, we assessed NP carriage of vaccine and nonvaccine serotypes in children and adults in rural villages in Alaska before and after the introduction of PCV7

Subjects and Methods

Setting and subjectsWe conducted annual, community-wide surveys of pneumococcal NP colonization in 8 rural villages in Alaska (total population of the villages according to the 2000 US Census, 3868) conducted in 1998–2004 during April and May of each year. The initial purpose of the surveys conducted during 1998 through 2000 was to evaluate interventions aimed at decreasing the prevalence of antimicrobial-resistant pneumococcal colonization [12, 13]. The study was continued to assess changes in pneumococcal colonization after the introduction of PCV7. Village populations ranged from 200 to 800 persons, and 95% of village residents were Alaska Native persons. Participation was open to all residents of all ages. The study was approved by the institutional review boards of the Alaska Area Native Health Service, the Indian Health Service, and the CDC. Written, informed consent was obtained from adult study participants for themselves and for their children. Only study participants <5 years of age and ⩾18 years of age were included in the analysis

Data collectionDemographic data and household information were collected during interviews, by use of a questionnaire. We reviewed the individual medical charts of the study participants, to obtain PCV7 immunization history and to document visits to the clinic and antimicrobial use that occurred during the 6 months before the NP swab was collected. We also reviewed the charts of children who were not study participants but were residing with an adult who was a study participant, to assess pneumococcal immunization history

Laboratory methodsFor each study participant, a sterile calcium alginate–tipped swab (Calgiswab type 1; Harwood Products) was inserted through the naris into the nasopharynx and then was plated directly onto a gentamicin-blood agar medium and was incubated at 37°C in a CO2-enriched environment. Specimens were transported to the CDC laboratory in Anchorage, Alaska, where the presence of S. pneumoniae was identified by α-hemolysis and optochin susceptibility testing. Serotyping was performed by Quellung reaction with the use of reagents from Statens Serum Institut. Isolates with an optochin zone of inhibition of <14 mm were considered to be S. pneumoniae if the colony morphology was suggestive of the species and if a serotype was identified. Isolates were classified as nontypeable pneumococci if they had an optochin zone of inhibition of ⩾14 mm but could not be serotyped. Antimicrobial susceptibility testing was performed using gradient strip diffusion (E-test; AB Biodisk). S. pneumoniae isolates were classified as susceptible, intermediate, or resistant to 5 antimicrobial agents, according to the 2002 definitions of the NCCLS (now the Clinical Laboratory Standards Institute) (table 1). Isolates designated as either intermediate or resistant were categorized as nonsusceptible

Statistical methodsWe defined children as having been age-appropriately vaccinated if they had received the recommended number of PCV7 doses (per recommendations of the Advisory Committee on Immunization Practices) at the time of enrollment [1]. Pneumococcal isolates that were 1 of the 7 serotypes in PCV7 (i.e., 4, 6B, 9V, 14, 18C, 19F, and 23F) were classified as PCV7-type serotypes. Although PCV7 may provide some protection against certain serotypes related to those found in the vaccine (e.g., 6A), we classified pneumococci of all other serotypes, including nontypeable pneumococci, as non–PCV7-type serotypes. Persons who had multiple serotypes isolated (78 [5%] of 1597 pneumococcal carriers ⩾18 years of age and 54 [5%] of 1062 pneumococcal carriers <5 years of age) were classified as carriers of the PCV7-type serotype if ⩾1 isolate was a PCV7-type serotype. For persons who had multiple serotypes isolated, data on the isolate with the highest MIC were used in the reporting of antibiotic resistance trends

Data were double-entered into Paradox (version 9.0; Corel), and analyses were performed using SAS software (version 9.0; SAS Institute). Data from the prevaccine years of 1998–2000 were combined and were used as “baseline” data. We used the Cochran-Armitage test of trend to examine the proportion of persons carrying any pneumococci, PCV7-type pneumococci, and antimicrobial-resistant pneumococci across the years of the study. We used an exact P value for the Cochran-Armitage test of trend, to examine carriage of individual S. pneumoniae serotypes across the years of the study. Because we examined trends for a large number of individual serotypes, a lower P value (P=.002 [i.e., .05 divided by the number of serotypes tested]) was used to denote statistical significance

We compared the proportion of adults carrying PCV7-type pneumococci by household characteristic (presence of children <5 years of age in the household or presence of children in the household who were age-appropriately vaccinated), by use of logistic regression. In the logistic regression, we adjusted for study year and adult age class. In addition, we evaluated antibiotic use, the number of household members, and the sex of the study participants as possible epidemiological confounders; variables remained in the model if they changed the primary coefficient of interest (household characteristic) by >15%. P values are exact where appropriate and were 2-sided

Results

During annual surveys conducted from 1998 through 2004, a total of 15,598 NP swabs were collected from residents of the 8 communities (mean, 2228 participants/year). The average rate of participation among villagers was 65% for children <5 years of age and 52% for adults ⩾18 years of age

Of the study participants, 309 (99%) of 311 children <5 years of age had received at least 1 dose of PCV7, and, as of 2004, 246 (79%) of these 311 children had been age-appropriately vaccinated. These data are similar to the findings for all children <5 years of age whose charts were available for review. Vaccine history was available for 384 of 409 children living in the study villages: by 2004, a total of 381 (99%) of these children had received at least 1 dose of PCV7, and 305 (79%) had been age-appropriately vaccinated

During 1998–2004, the proportion of study participants colonized with S. pneumoniae remained stable among children <5 years of age (59% at baseline and 61% in 2004; P=.91 for trend; median percentage, 59%; range, 51%–61%), but there was an upward trend among adults ⩾18 years of age (13% at baseline and 26% in 2004; P<.0001 for trend; median percentage, 18%; range, 13%–27%). This trend of increased carriage of S. pneumoniae in adults was observed among adults in all age classes (figure 1)

Figure 1

Pneumococcal colonization among study participants, Alaska, 1998–2004. S. pneumoniae, Streptococcus pneumoniae

Among children <5 years of age who were colonized with S. pneumoniae the proportion with PCV7-type pneumococcal carriage decreased from 55% at baseline to 5% in 2004 (P<.0001 for trend; median percentage, 18%; range, 5%–55%). Among adults colonized with S. pneumoniae carriage of PCV7-type pneumococci decreased from 28% to 5% over this same period (P value for trend, <.0001). This trend of decreased carriage of PCV7-type pneumococci among adults was observed for adults in all age classes (table 2). Accordingly, because PCV7-type colonization decreased but overall colonization did not, there has been a marked increase in the proportion of adults with colonization due to non–PCV7-type pneumococci. Non-PCV7 serotypes for which a significant increase in carriage has been noted among adults are 12F, 17F, 19A, 23A, and 34 (table 3). Carriage of these serotypes (with the exception of serotype 12F, for which sample size was insufficient) also significantly increased among children <5 years of age (table 3)

Table 1

Antimicrobial susceptibility breakpoints

Table 2

Heptavalent protein-polysaccharide pneumococcal conjugate vaccine (PCV7)–type colonization among persons colonized with Streptococcus pneumoniae, by age class and year, Alaska, 1998–2004

Table 3

Carriage of heptavalent protein-polysaccharide pneumococcal conjugate vaccine (PCV7)–type and non–PCV7-type pneumococci among colonized adults ⩾18 years of age, Alaska, 1998–2004

After PCV7 became widely used, there were significant decreases, among adults ⩾18 years of age, in the proportion of colonizing isolates that were resistant to penicillin (from 13% at baseline to 6% in 2004; P=.05 for trend), intermediate to erythromycin (from 13% at baseline to 8% in 2004; P=.05 for trend), and intermediate to ceftriaxone (from 9% at baseline to 4% in 2004; P=.01 for trend) (table 4). After an initial decrease from 12% during the baseline period to 9% 2 years after the introduction of PCV7, the proportion of isolates that were intermediate to penicillin increased in 2003 and 2004 (19% in 2004; P=.01 for trend). This increase was primarily accounted for by an increase in colonization by isolates of serotype 19A. If serotype 19A isolates are removed from the analysis, colonization by isolates with intermediate susceptibility to penicillin does not change over the study period (P= .24). Analysis of trends in susceptibility to penicillin and cotrimoxazole among individual non–PCV7-type serotypes (3, 6A, 10A, 11A, 16F, 17F, 19A, and 35B) revealed an increase in nonsusceptibility to cotrimoxazole in isolates of serotypes 10A and 35B, beginning in 2002–2003. The proportion of isolates of 35B that were cotrimoxazole nonsusceptible was 0% (0 of 18 isolates) at baseline, 0% (0 of 11 isolates) in 2001, 0% (0 of 21 isolates) in 2002, 21% (10 of 48 isolates) in 2003, and 25% (7 of 28 isolates) in 2004 (P=.002 for trend). The proportion of isolates of serotype 10A that were cotrimoxazole nonsusceptible was 0% (0 of 6 isolates) at baseline, 0% (0 of 23 isolates) in 2001, 4% (1 of 23 isolates) in 2002, 23% (7 of 30 isolates) in 2003, and 22% (2 of 9 isolates) in 2004 (P=.004 for trend)

Table 4

Nonsusceptibility to antimicrobial agents among colonizing Streptococcus pneumoniae isolates from adults ⩾18 years of age, Alaska, 1998–2004

To assess the association of exposure to young children with adult colonization with PCV7-type pneumococci, we analyzed adult colonization according to whether the adult was living with a child <5 years of age. The rate of carriage of PCV7-type pneumococci among colonized adults decreased significantly both among adults with a child in the household (from 32% at baseline to 7% in 2004; P<.001 for trend) and among adults with no child in the household (from 25% at baseline to 3% in 2004; P<.001 for trend). Adults living with a child had a higher likelihood of carrying PCV7-type pneumococci than did adults who were not living with a child (adjusted odds ratio [OR], 1.81 [95% confidence interval {CI}, 1.31–2.50]; the OR was adjusted for adult age class and the study year) (table 5). The association of adult carriage of PCV7-type pneumococci with presence of children in the household was not substantially influenced by the number of household members, the sex of the adult, or antibiotic use prior to culture

Table 5

Carriage of heptavalent protein-polysaccharide pneumococcal conjugate vaccine (PCV7)–type pneumococci among adults ⩾18 years of age who were carriers of Streptococcus pneumoniae and were not living with children <5 years of age, as well as among adult carriers living with children <5 years of age, Alaska, 1998–2004

To assess the effect of pediatric vaccination with PCV7 on the colonization of household members, we analyzed carriage of PCV7-type pneumococci for each adult living with a child, according to the PCV7 vaccination status of the child. Adults living in households in which at least 1 child had been age-appropriately vaccinated with PCV7 had a lower likelihood of carrying PCV7-type pneumococci than did adults living in households in which no child was age-appropriately vaccinated with PCV7 (adjusted OR, 0.49 [95% CI, 0.28–0.83]; adjusted for year) (table 6). The association of carriage of PCV7-type pneumococci with the vaccination status of children in the household who were <5 years of age was not substantially influenced by the number of household members, the sex of the adult, or antibiotic use prior to culture

Table 6

Carriage of heptavalent protein-polysaccharide pneumococcal conjugate vaccine (PCV7)–type pneumococci among adults ⩾18 years of age who were carriers of Streptococcus pneumoniae, by whether children in the household who were <5 years of age had been age-appropriately vaccinated or not, Alaska, 2001–2004

Discussion

This analysis of pneumococcal colonization among adults and children during the 3 years before and the 4 years after the introduction of pediatric PCV7 documents a significant, consistent decrease in PCV7-type pneumococcal colonization and a concomitant increase in non–PCV7-type pneumococcal colonization in both vaccinated children and unvaccinated adults. A considerable indirect effect of PCV7 is demonstrated by the significant decrease in PCV7-type pneumococcal colonization of adults, even among adults with no household exposure to young children. Our finding that vaccine use for children has an indirect effect on PCV7-type pneumococcal colonization in adults supports recent observations of a decrease in the rate of PCV7-type invasive pneumococcal disease in adults. The existence of this indirect effect is further supported by the finding that adults living with vaccinated children are less likely to be colonized with PCV7-type strains than are adults living with unvaccinated children. Reductions in adult carriage of antibiotic-resistant pneumococci have also been observed

The dynamics of pneumococcal colonization and disease have been the subject of numerous studies dating back to the early 1900s. Transmission of the pneumococcus between household members has been described in the classic pneumococcal literature and in more-recent studies as well [8, 1418]. Colonization, particularly the acquisition of a new serotype, has been shown to be a key event in the development of invasive pneumococcal disease [17, 19]. The findings of these historic studies continue to have relevance in the vaccine era and lend support to the hypothesis that the use of PCV7 for children would have an indirect effect on colonization and disease in adults. Our findings from an ongoing survey of community colonization conducted both before and after introduction of PCV7 demonstrate the continued significance of household transmission in the spread of the pneumococcus and corroborate the role of colonization in pathogenesis

Data from vaccine trials and from soon after the introduction of PCV7 have suggested that replacement of vaccine serotypes by nonvaccine serotypes in the nasopharynx could occur after widespread vaccination [4, 2022]. Among children in our study, the magnitude of replacement colonization with non–PCV7-type pneumococci was similar to the decrease in PCV7-type pneumococcal colonization, and overall pneumococcal colonization remained stable. However, in adults, this replacement phenomenon resulted in an overall increase in pneumococcal colonization, compared with the colonization noted in the baseline years of the study. The reason for this increase is unclear, but it may be related to the dynamic nature of pneumococcal transmission within communities. Alternatively, mathematical and mouse models have demonstrated that pneumococci vary in their competitive abilities, and the increase in colonization among adults may reflect a relatively greater ability of non–PCV7-type pneumococci to colonize the adult oropharynx in the absence of PCV7 types [23]. This effect would theoretically be present in children as well, but it may be attenuated because of competition from other bacteria that frequently colonize the pediatric oropharynx or because of the proposed phenomenon of steric inhibition, in which vaccine-induced antibodies bind to the polysaccharide capsule and prevent the interaction of pneumococcal surface proteins with human epithelial binding sites [24]

Non–PCV7-type pneumococci have historically been considered to be less clinically important, because they caused fewer cases of invasive pneumococcal disease and were associated with a lower likelihood of antibiotic resistance. However, Brueggemann et al. [25] found that certain non–PCV7-type pneumococci have high invasive disease potential. Furthermore, several studies have documented small increases in non–PCV7-type pediatric invasive pneumococcal disease after the introduction of PCV7, presumably because of increased transmission from persons colonized with non-PCV7 types [3, 5, 26, 27]. An increase in non–PCV7-type invasive pneumococcal disease has also been observed in HIV-infected adults and in Alaska Native adults [11, 28, 29]. This increase is consistent with our finding of increased carriage of non–PCV7-type pneumococci among adults after introduction of childhood PCV7. HIV-infected persons and Alaska Native persons are at increased risk for invasive pneumococcal disease; the occurrence of replacement disease in these sentinel populations may signify a limit to the usefulness of the currently available vaccine and emphasizes the importance of ongoing surveillance and development of extended-valency vaccines

PCV7 serotypes have been significantly associated with antibiotic resistance, and vaccine introduction has resulted in a reduction in the proportion of cases of pediatric invasive pneumococcal disease due to antibiotic-resistant isolates [3, 5, 11]. However, the coincident phenomenon of replacement colonization in both children and adults means that non–PCV7-type pneumococci are now more likely to be exposed to antibiotic pressure and to commensal NP flora that may be capable of transferring genes encoding antibiotic resistance. One large multicenter study found that penicillin nonsusceptibility has increased slightly among pediatric non–PCV7-type invasive isolates since the introduction of PCV7 [5]. Our results show an overall decrease in carriage of antibiotic-resistant pneumococci since introduction of the vaccine; however, data from 2003 and 2004 provide some evidence of increasing resistance. In the present study, the driving force behind the observed increase in antimicrobial resistance was an increase in carriage of serotypes already known to be resistant to antimicrobials (e.g., serotype 19A); however, there also was evidence that at least 2 non–PCV7 types (10A and 35B) have recently acquired resistance to cotrimoxazole. It remains to be seen whether this trend continues and whether significant increases in resistant non–PCV7-type invasive pneumococcal disease will occur in children and adults

The current analysis has limitations. Participating villages were a convenience sample and were not randomly selected. They were, however, representative of typical rural villages in that ∼95% of residents were Alaska Native people, had similar economic and cultural factors, and had similar health care opportunities. Analysis comparing participants in the present study with nonparticipants demonstrated that participants were younger, used the health care services at the clinic more frequently, and used more antibiotics than did nonparticipants [30]. However, these differences were similar across all years of the study, which allows for comparability between study years. Our methodology did not account for the natural variation in pneumococcal carriage that may occur over the course of one or many years, and our study design does not allow us to make conclusions about the direction of pneumococcal transmission among members of a household. The direction of transmission was assumed to be from the group with the highest colonization to that with the lowest, but this cannot be proven using these data. In addition, we did not collect information on receipt of the 23-valent pneumococcal polysaccharide vaccine by adult participants. One early study that used a quadravalent vaccine containing ∼30–60 μg of polysaccharide from serotypes 1, 2, 5, and 7 demonstrated reduced colonization with these serotypes [31]. However, more-recent studies that used commercially available 14- and 23-valent pneumococcal polysaccharide vaccines did not demonstrate any influence of immunization on colonization [3235]

Despite limitations in the vaccine supply, the introduction of PCV7 into the routine childhood immunization schedule has yielded a dramatic decrease in both pediatric and adult PCV7-type invasive pneumococcal disease. The present study demonstrates that the indirect effect in adults is the result of decreased carriage of PCV7-type pneumococci among adults, presumably because of decreased transmission of PCV7-type pneumococci from children to adults. Continued surveillance is important to monitor other possible indirect effects of vaccine use, such as serotype replacement and changes in rates of antibiotic-resistant invasive pneumococcal disease

Acknowledgments

We thank the study participants and the clinic staff in each of the participating villages. We thank Marcella Harker-Jones and Maria Warnke, in addition to Alisa Reasonover, for identification and serotyping of the pneumococci. We also thank Helen Peters, Jim Gove, Catherine Dentinger, Susan Seidel, Lorraine Alexander, Daniel Feikin, Lisa Chiou, Karen Miernyk, Karen Rudolph, Carolyn Zanis, Carolynn DeByle, Marilyn Getty, and Ken Petersen for recruitment of study participants

Footnotes

  • Presented in part: 4th International Symposium on Pneumococci and Pneumococcal Diseases, 9–13 May 2004, Helsinki, Finland (abstract EPI-03); 5th International Symposium on Pneumococci and Pneumococcal Diseases, 2–6 April 2006, Alice Springs, Australia (abstract PO7.26)

    Potential conflicts of interest: none reported

    Financial support: US Department of Health and Human Services through the Centers for Disease Control and Prevention (CDC)

  • Present affiliation: Division of Global Migration and Quarantine, National Center for Infectious Diseases, CDC, Atlanta, Georgia

  • The findings and conclusions in this article are those of the authors and do not necessarily represent the views of the CDC

  • Received September 16, 2005.
  • Accepted January 12, 2006.

References

| Table of Contents