Increased Antimicrobial Resistance Among Nonvaccine Serotypes of Streptococcus pneumoniae in the Pediatric Population After the Introduction of 7-Valent Pneumococcal Vaccine in the United States

Abstract

Background: 

The 7-valent pneumococcal conjugate vaccine (PCV7) was introduced in the United States in February 2000. The PROTEKT US study evaluated serotype distribution, PCV7 coverage and antimicrobial susceptibility among Streptococcus pneumoniae isolates collected from children aged 0 to 14 years in 2000 through 2001 (year 1; n = 2033), 2002 through 2003 (year 3; n = 1740) and 2003 through 2004 (year 4; n = 1591).

Methods: 

Serotyping was performed by Neufeld Quellung reaction. Antimicrobial susceptibilities were determined centrally according to Clinical Laboratory Standards Institute methodology and interpretive breakpoints.

Results: 

The proportion of isolates covered by PCV7 (vaccine serotypes) decreased from 65.5% (year 1) to 34.7% (year 3) and to 27.0% (year 4) (P < 0.0001) with similar changes seen at regional and state levels. The most common serotypes in year 4 were nonvaccine serotypes (NVS) 19A (19.0% of all isolates), 6A (7.8%), 3 (7.6%), 15 (6.3%) and 35B (5.8%) and vaccine serotype 19F (12.7%). NVS 19A increased relative to vaccine serotype 19F among isolates expressing the erm(B) + mef(A) macrolide-resistant genotype (P < 0.0001) between year 1 (7.8% [19A] versus 86.7% [19F]) and year 4 (45.5% [19A] versus 51.7% [19F]). Antimicrobial resistance rates (year 1 versus year 4) among NVS from nonblood (respiratory tract) sources increased for penicillin (resistant: 12.7–16.1% [P = 0.0857]; intermediate susceptibility: 20.1–31.5% [P < 0.0001]), erythromycin (21.2–31.6% [P < 0.0001]), amoxicillin–clavulanate (1.4–5.8% [P < 0.0001]) and multidrug resistance (resistance to ≥2 antimicrobial classes) (24.6–31.6% [P = 0.0034]).

Conclusions: 

The proportion of S. pneumoniae isolates from U.S. pediatric patients covered by PCV7 decreased substantially in the 4 years after vaccine introduction. However, resistance to commonly used antimicrobials, including β-lactams and macrolides, as well as multidrug-resistant strains increased significantly among respiratory tract isolates of NVS.

The management of pediatric respiratory tract infections (RTIs) has become complicated by the emergence and spread of antimicrobial resistance, in particular in Streptococcus pneumoniae, the most frequently implicated bacterial pathogen.¹ According to surveillance data covering the last 5 years, resistance rates for both β-lactams and macrolides in S. pneumoniae have surpassed 30% in the pediatric U.S. population.^2–7

One approach to combat the dangers posed by antimicrobial resistance in S. pneumoniae is vaccination. The 7-valent pneumococcal conjugate vaccine (PCV7; Prevnar; Wyeth Lederle Vaccines, Pearl River, New York, NY) was introduced in the United States in February 2000. The vaccine is licensed for children up to 5 years of age and there is a general recommendation for its use in all children <2 years of age. PCV7 provides coverage against S. pneumoniae serotypes 4, 6B, 9V, 14, 18C, 19F and 23F. Crosscoverage may extend to other serotypes belonging to serogroups represented in PCV7, in particular, protection from disease caused by type 6A as a result of immunization with type 6B.^8,9 Antimicrobial resistance in S. pneumoniae is closely associated with serotype, with resistance being higher in serotypes represented in the PCV7 vaccine than in those outside PCV7 coverage.¹⁰

The decline in invasive disease that has occurred after the introduction of PCV7 has been accompanied by reductions in penicillin and macrolide nonsusceptibility among invasive pneumococcal disease isolates.^8,11,12 However, little information is available about how antimicrobial susceptibility patterns in nonblood (respiratory tract) isolates from children have been affected since PCV7 was introduced.

The PROTEKT US (Prospective Resistant Organism Tracking and Epidemiology for the Ketolide Telithromycin in the United States) study is a longitudinal surveillance study initiated in 2000 to monitor the spread of antimicrobial resistance among respiratory tract pathogens in the United States. The program includes serotyping analysis. In the current study, a comparison was made among the serotype distribution, PCV7 coverage and antimicrobial susceptibility of S. pneumoniae pediatric isolates collected in 2000 through 2001 (year 1 of the PROTEKT US study; around the time of the introduction of PCV7) and those collected in 2002 through 2003 and 2003 through 2004.

PATIENTS AND METHODS

S. pneumoniae isolates were collected from pediatric patients (≤14 years of age with community-acquired RTIs) during the 2000 to 2001 (PROTEKT US year 1), 2002 to 2003 (year 3) and 2003 to 2004 (year 4) respiratory seasons. One hundred twelve centers in 38 states participated. Patients presented to either primary care or a hospital with one of the following conditions acquired in the community: acute otitis media, community-acquired pneumonia, acute maxillary sinusitis or acute rhinosinusitis. Sources of clinical isolates included blood, sputum, bronchoalveolar lavage fluid, middle ear fluid, nasopharyngeal swabs or aspirates, and sinus aspirates. Patients with nosocomial RTIs and those with cystic fibrosis were excluded from the study as were duplicate strains or strains originating from existing collections. The study design and methodology have been described previously.¹³

Serotyping was performed at G.R. Micro Ltd (London, U.K.) using the Neufeld Quellung reaction and Statens Serum Institute antisera (Copenhagen, Denmark). Statens Serum Institute was used as the reference laboratory for quality assurance and identification of rare serotypes.

Susceptibility to a range of antimicrobials was determined at the Clinical Microbiology Institute (Wilsonville, OR) using Clinical Laboratory Standards Institute broth microdilution methodology and Clinical Laboratory Standards Institute breakpoints when available.^14,15

Genotyping of erythromycin-resistant isolates was conducted at G. R. Micro Ltd. using methods described previously.^16,17 Statistical analysis was performed using the χ² and Fisher exact test tests and InStat software (GraphPad Software, Inc., San Diego, CA).

RESULTS

Serotype Distribution.

During the 3 years studied, 7813 pediatric S. pneumoniae isolates were serotyped. To avoid bias, the analysis was conducted only on the isolates collected from the 112 centers that participated in all years of the study (year 1, 2033; year 3, 1740; year 4, 1591).

Across the United States as a whole, serotypes covered by the PCV7 vaccine decreased in prevalence from 65.5% in 2000 through 2001 to 34.7% in 2002 through 2003 and to 27.0% in 2003 through 2004 (P < 0.0001; Table 1). Decreases were also observed in this period in all states and regions surveyed (Table 2) with the largest reduction seen in Florida. There was considerable between-state variability in vaccine serotype (VS) prevalence in the year 4 isolates ranging from 18.3% in Ohio to 38.7% in California (Table 2).

TABLE 1: Proportion of Streptococcus pneumoniae Isolates With Serotypes Covered by PCV7 According to Source and Age Group Collected in Year 1 (2000–2001), Year 3 (2002–2003) and Year 4 (2003–2004) of the PROTEKT US Study

TABLE 2: Proportion of Streptococcus pneumoniae Isolates Covered by the PCV7 Vaccine According to U.S. State and Region in Year 1 (2000–2001), Year 3 (2002–2003) and Year 4 (2003–2004) of the PROTEKT US Study

In year 1, the proportion of S. pneumoniae isolates with serotypes represented in PCV7 was higher in isolates cultured from blood than in those cultured from respiratory tract sources (Table 1). Similarly, a higher proportion of isolates collected from patients ≤2 years of age had serotypes covered by the vaccine than those from patients 3 to 14 years of age (Table 1). By year 4, these patterns were not evident and VS prevalence was <30% regardless of patient age or isolate source. The decrease in VS among isolates collected from blood is shown in Figure 1.

FIGURE 1.: Change in serotype prevalence between year 1 (2000–2001; n = 577) and year 4 (2003–2004; n = 277) for S. pneumoniae isolates collected in the PROTEKT US study and cultured from blood samples. (NT, nontypeable.)

Of the nonvaccine serotypes (NVS), the most common among year 4 isolates were 19A (19.0% of all isolates) followed by 6A (7.8%), 3 (7.6%), 15 (6.3%) and then 35B (5.8%) (Fig. 2).

FIGURE 2.: Prevalence of the most common nonvaccine S. pneumoniae serotypes collected from pediatric isolates in year 1 (2000–2001), year 3 (2002–2003) and year 4 (2003–2004) of the PROTEKT US study.

Antimicrobial Resistance in Nonvaccine Serotypes.

Rates of resistance to various antimicrobials among NVS isolates are shown in Table 3. Among isolates collected from blood, there were no significant changes in the susceptibility of NVS to the drug classes tested. By contrast, intermediate susceptibility to penicillin, and resistance to amoxicillin–clavulanate and erythromycin, increased significantly among isolates from respiratory tract sources. One NVS isolate, with a minimum inhibitory concentration of 4 μg/mL, exhibited resistance to telithromycin during the study period.

TABLE 3: Antimicrobial Resistance Among Streptococcus pneumoniae Isolates With Nonvaccine Serotypes Collected in Year 1 (2000–2001), Year 3 (2002–2003) and Year 4 (2003–2004) of the PROTEKT US Study

Multidrug resistance (resistance to ≥2 antimicrobial classes) decreased, but not significantly, in blood isolates (P = 0.356) and increased among respiratory tract isolates (P = 0.0034) during the study period.

Antimicrobial Resistance in Vaccine Serotypes.

Table 4 shows the rates of antimicrobial resistance observed for VS isolates. Among VS blood isolates, there were no significant changes in resistance rates to all drug classes tested. By contrast, significant increases in the rates of resistance to erythromycin and amoxicillin–clavulanate occurred among VS respiratory tract isolates.

TABLE 4: Antimicrobial Resistance Among Streptococcus pneumoniae Isolates With Vaccine Serotypes Collected in Year 1 (2000–2001), Year 3 (2002–2003) and Year 4 (2003–2004) of the PROTEKT US Study

Antimicrobial Resistance Among All Isolates.

Rates of resistance among all isolates are shown in Table 5. With the exception of amoxicillin–clavulanate, resistance rates to all agents tested decreased significantly among the isolates collected from blood. Among the respiratory tract isolates, penicillin resistance decreased significantly, whereas intermediate penicillin susceptibility rose significantly; the rates of amoxicillin–clavulanate and erythromycin resistance did not change.

TABLE 5: Antimicrobial Resistance Among All Serotyped Streptococcus pneumoniae Isolates Collected in Year 1 (2000–2001) and Year 4 (2003–2004) of the PROTEKT US Study

Genotyping.

Genotyping analysis of erythromycin-resistant S. pneumoniae isolates revealed a sharp increase in the proportion of strains positive for mef(A) and erm(B) between year 1 (10.9%) and year 4 (28.9%) with a corresponding decrease in the prevalence of isolates positive for either mef(A) alone (72.0% to 58.0%) or erm(B) alone (15.8% to 11.2%). Almost all mef(A) + erm(B) isolates analyzed were serotype 19F or 19A. The proportion of erm(B) + mef(A) isolates exhibiting NVS 19A increased relative to those exhibiting VS 19F between year 1 (7.8% [19A] versus 86.7% [19F]) and year 4 (45.5% [19A] versus 51.7% [19F]) (P < 0.0001).

DISCUSSION

This study demonstrates that pneumococcal serotype distribution has changed significantly in the 4 years after the introduction of PCV7. As a consequence of the reduction in VS pneumococci, the predominant pneumococcal serotypes isolated from children in 2003 through 2004 were NVS. The observed decreased prevalence of PCV7 serotypes occurred in both patient subgroups analyzed, indicating that the vaccine has been equally effective in protecting infants (0–2 years of age) and older children (3–14 years of age). The degree to which serotype distribution changed over the study period varied from state to state and it will be important to determine if this variability reflects differential rates of immunization in different parts of the country.

A limitation of the surveillance program, which collects a specified number of isolates from each site in each year, is that burden of disease cannot be measured. However, data collected since the introduction of PCV7 indicate that routine vaccination has led to a reduction in the overall incidence of invasive pneumococcal disease in young children.^{8,11,12,18–20}

Recent surveillance data suggested that pneumococcal penicillin and erythromycin resistance had declined in the pediatric population after the introduction of PCV7.^8,11,12 However, concern has been expressed that resistance may emerge in NVS pneumococci.²¹ Increased macrolide resistance in NVS among blood isolates has been noted previously, but this was offset by a much larger decrease in macrolide resistance among VS.¹² A major finding of the current analysis was the increase in antimicrobial resistance among NVS; resistance to penicillin, macrolides, cotrimoxazole, amoxicillin–clavulanate and multidrug resistance all increased between 2000 and 2004. Although these trends give cause for concern, the clinical significance of antimicrobial resistance in pneumococci cultured from the respiratory tract of children is difficult to interpret. An encouraging finding was the low rate of resistance to telithromycin, which is approved for the treatment of community-acquired pneumonia in adults.

A factor contributing to the increased pneumococcal antimicrobial resistance among NVS is almost certainly the 3-fold increase between 2000 and 2004 in the prevalence of erm(B) + mef(A) isolates. Most of these strains are resistant to most antimicrobial classes.^16,17 In addition, a high proportion of erm(B) + mef(A) isolates are international clonal strains expressing serogroup 19.^16,17,22 In the current analysis, >90% of erm(B) + mef(A) isolates were either serotype 19A or 19F. However, the relative distribution of these 2 serotypes changed significantly over the study period; whereas VS 19F predominated in year 1, by year 4, the distribution between NVS 19A and 19F was roughly equal. Although serotype switching from 19F to 19A predates vaccine introduction,²³ it is likely that immunologic selection by the vaccine explains the emergence of 19A. Lower vaccine efficacy against respiratory isolates of serotype 19F²⁴ is the probable reason that a residual 12.7% of all isolates analyzed from year 4 were of VS 19F, making it the most common remaining VS, and the second most common serotype, overall (after 19A).

Along with selective pressure of the vaccine it is likely that a principle driver of increasing resistance among NVS is antibiotic consumption.²⁵ A recent study showed that the variation in pneumococcal resistance across the United States was best explained by probable geographic variability in antibiotic selection pressure.²⁶

In conclusion, serotype distribution among pediatric S. pneumoniae isolates has changed significantly since PCV7 was introduced. Unfortunately, this trend has been accompanied by significant increases in the rates of in vitro antimicrobial resistance among NVS. Our data, which are consistent with recent observations from Alaska,²⁷ Boston,²⁸ Kentucky²⁹ and Pittsburgh,³⁰ suggest that the impact of the vaccine on antimicrobial resistance among noninvasive pneumococci has been transient. Although expanded vaccine formulations with efficacy against a wider range of pneumococcal serotypes may be useful, reducing the level of indiscriminate or inappropriate antibiotic prescribing is also essential if antimicrobial resistance is to be brought under control.