A Population-Based Assessment of the Impact of 7- and 13-Valent Pneumococcal Conjugate Vaccines on Macrolide-Resistant Invasive Pneumococcal Disease: Emergence and Decline of Streptococcus pneumoniae Serotype 19A (CC320) With Dual Macrolide Resistance Mechanisms

Summary

Macrolide-resistant Streptococcus pneumoniae containing both mef(E)/mel and erm(B) rapidly expanded (2005–2009) but decreased following pneumococcal conjugate vaccine 13 introduction. Macrolide-resistant serotypes not represented in vaccines increased modestly. Selective pressures of macrolide use and vaccine introductions were associated with these changes.

Abstract

Background

Macrolide efflux encoded by mef(E)/mel and ribosomal methylation encoded by erm(B) confer most macrolide resistance in Streptococcus pneumoniae. Introduction of the heptavalent pneumococcal conjugate vaccine (PCV7) in 2000 reduced macrolide-resistant invasive pneumococcal disease (MR-IPD) due to PCV7 serotypes (6B, 9V, 14, 19F, and 23F).

Methods

In this study, the impact of PCV7 and PCV13 on MR-IPD was prospectively assessed. A 20-year study of IPD performed in metropolitan Atlanta, Georgia, using active, population-based surveillance formed the basis for this study. Genetic determinants of macrolide resistance were evaluated using established techniques.

Results

During the decade of PCV7 use (2000–2009), MR-IPD decreased rapidly until 2002 and subsequently stabilized until the introduction of PCV13 in 2010 when MR-IPD incidence decreased further from 3.71 to 2.45/100000 population. In 2003, serotype 19A CC320 isolates containing both mef(E)/mel and erm(B) were observed and rapidly expanded in 2005–2009, peaking in 2010 (incidence 1.38/100000 population), accounting for 36.1% of MR-IPD and 11.7% of all IPD isolates. Following PCV13 introduction, dual macrolide-resistant IPD decreased 74.1% (incidence 0.32/100000 in 2013). However, other macrolide-resistant serotypes (eg, 15A and 35B) not currently represented in PCV formulations increased modestly.

Conclusions

The selective pressures of widespread macrolide use and PCV7 and PCV13 introductions on S. pneumoniae were associated with changes in macrolide resistance and the molecular basis over time in our population. Durable surveillance and programs that emphasize the judicious use of antibiotics need to continue to be a focus of public health strategies directed at S. pneumoniae.

Streptococcus pneumoniae is a commensal of the human nasopharynx and an opportunistic pathogen. The pneumococcus causes numerous diseases ranging from localized infections (eg, otitis media and pneumonia) to severe invasive disease (eg, sepsis and meningitis), with the highest burden of pneumococcal disease occurring in young children [1]. In the 1980s and 1990s, β-lactam resistance emerged in the pneumococcus, complicating the choice of treatment regimens. Macrolide antibiotics became a major alternative to the use of β-lactams for the treatment of suspected upper respiratory tract infections and community-acquired pneumonia caused by pneumococci. However, macrolide resistance rapidly emerged in S. pneumoniae following introduction and widespread use of new semisynthetic macrolides (eg, azithromycin, clarithromycin) [2, 3].

Two major macrolide resistance phenotypes are observed in pneumococci [4]. An M phenotype and an MLS_B phenotype, mediated through macrolide efflux and ribosomal target site modification, respectively [5]. The M phenotype results from the expression of the macrolide resistance efflux pump encoded by mef(E)/mel, 2 cotranscribed genes of the macrolide efflux genetic assembly (Mega), that confers moderate-level resistance to 14- and 15-membered macrolides [6]. The MLS_B phenotype is due to the presence of the ribosomal methylase encoded by erm(B), which results in high-level macrolide resistance as well as resistance to the chemically distinct lincosamides (clindamycin) and streptogramin B that target overlapping ribosomal sites [4]. Throughout the 1990s, the expansion of macrolide-resistant invasive pneumococcal disease (MR-IPD) in the United States was largely due to mef(E)/mel-mediated macrolide efflux [5, 7, 8]. Both mef(E)/mel and erm(B) are contained on conjugative transposable elements or smaller remnants of such elements. Horizontal gene transfer and clonal expansion are major contributors to the dissemination of macrolide resistance determinants in pneumococci [9, 10].

In 2000, the first pneumococcal conjugate vaccine containing 7 capsular polysaccharides of serotypes 4, 6B, 9V, 14, 18C, 19F, and 23F conjugated to the CRM197 diphtheria protein (PCV7) was licensed for use in children aged <5 years in the United States [11, 12]. However, serotype replacement was observed [13, 14], and in 2010, an additional 6 serotypes (1, 3, 5, 6A, 7F, and 19A conjugated to CRM197) were added to the PCV7 serotypes to create a 13-valent conjugate vaccine (PCV13), which replaced PCV7 in the United States [12, 15]. PCV7 and PCV13 vaccination was highly effective against vaccine-serotype IPD and also was associated with decreased pneumococcal upper respiratory carriage of vaccine serotypes [16], thus providing [17] herd protection to unvaccinated older children and adults.

Using a well-established population-based surveillance network and molecular typing, we followed MR-IPD in Atlanta, Georgia, for more than 20 years. We previously documented the emergence of MR-IPD cases in Georgia Health District-3 (HD-3; 1994–1999) and the initial impact of PCV7 (2000–2002) on the incidence of macrolide-resistant S. pneumoniae using population-based assessment [2]. Our aim in the current study was to investigate the trends in incidence and molecular basis of MR-IPD in the Atlanta metropolitan area over time during the PCV7 and PCV13 eras. The emergence and decline of MR-IPD due to serotype 19A, which belongs to clonal complex 320 (CC320) was defined. The clone contained dual macrolide resistance mechanisms (both mef(E)/mel and erm(B)). The doubling of MR-IPD due to non-PCV serotypes was also observed.

METHODS

Surveillance

IPD has been tracked in Atlanta since 1994 using active, population-based surveillance [7] as part of the Centers for Disease Control and Prevention (CDC) Active Bacterial Core Surveillance of the Georgia Emerging Infections Program [2, 18]. For this study (2003–2013), S. pneumoniae isolates from normally sterile sites were collected from all hospitals and laboratories within the Georgia HD-3, which consists of the core metropolitan Atlanta counties (Clayton, Cobb, DeKalb, Douglas, Fulton, Gwinnett, Newton, and Rockdale), with a 2010 population of 3682873. Population census data (2000 and 2010) and annual post-census estimates were obtained from the US Census Bureau.

Bacterial Strains

Streptococcus pneumoniae isolates were tested for antibiotic susceptibility by the CDC [6, 19] and serotyped using standard Quellung reactions [18] and polymerase chain reaction (PCR) [20]. Antibiotic minimum inhibitory concentrations (MICs) were determined using broth microdilution assays [19]. Isolates with an erythromycin MIC of ≥1 µg/mL were resistant, with 0.5 µg/mL were intermediate resistant, and with <0.5 µg/mL were susceptible. All nonsusceptible pneumococcal isolates (resistant and intermediate resistant) were analyzed using PCR for the presence of macrolide resistance genetic determinants [2]. Of 1170 macrolide nonsusceptible isolates from 2003–2013, 1131 (96.6%) were available for the molecular determination of the macrolide resistance genotype.

Macrolide Resistance Gene Detection

Genomic DNA was isolated from erythromycin nonsusceptible strains using crude lysis [7] or InstaGene Matrix (BioRad). OneTaq DNA polymerase and deoxynucleotide solution mix (New England Biolabs) were used. PCR amplification of erm(B) of a 551 bp product was performed using primers KG1F (5ʹ-TTGGAACAGGTAAAGGGCATT) and KG1R2 (5ʹ-TTTGGCGTGTTTCATTGCTTG) [7]. Detection of mef(A/E)/mel was determined by the presence of a 555 or 456 bp product from primers KG8 (5ʹ-GTATCATGTCACTTGCTATGCC) [6] and KG10 (5ʹ-ACACCTAGCTTGCCTACAAGTG). As mef(E) is more common than mef(A) in the United States[4], mef(E) is reported. The multilocus sequence type was determined for 35 dual macrolide-resistant genotype isolates using traditional methods, whole genome sequences [9, 21] and the pubmlst.org/spneumoniae database.

Statistical Analysis

We used χ² analyses Prism 5 (GraphPad) to compare proportions of invasive cases of S. pneumoniae and noncases as previously described [22]. Population data were used to calculate incidence rates, reported as cases per 100000 population [2].

RESULTS

Changes in Macrolide-Resistant Invasive Pneumococcal Disease: 2003–2013

The emergence of MR-IPD in Atlanta (1994–1999) and the initial impact of PCV7 introduction (2000–2002) have been previously reported [2]. From 2003 to 2009, the incidence of MR-IPD ranged from 3.36 to 4.34/100000 population (26.2%–32.3% of all IPD isolates; Figure 1, Table 1). After introduction of PCV13 in 2010, the incidence of MR-IPD decreased to 2.45/100000 by 2013 (P < .0001). From 1999 to 2013 the MR-IPD incidence decreased 73.7% (from 9.3 to 2.45/100000; Figure 1).

Figure 1.

The incidence (1994–2013) of macrolide-resistant invasive pneumococcal disease in metropolitan Atlanta, Georgia (black line), as cases per 100000 population. The incidence by macrolide resistance genotype: mef(E)/mel (white squares), erm(B) (white diamond), or the dual macrolide resistance genotype (mef(E)/mel and erm(B)) (white triangles) is also shown. Abbreviation: MR-IPD, macrolide-resistant invasive pneumococcal disease.

Table 1.

Incidence of Macrolide-Resistant Invasive Pneumococcal Disease 2003–2013 in Atlanta, Georgia, by Age Group, Macrolide Resistance, and Macrolide Resistance Genotype

Incidence	2003	2004	2005	2006	2007	2008	2009	2010	2011	2012	2013
All ages
Overall invasive pneumococcal disease	14.02	12.29	12.89	13.50	13.23	12.02	11.99	11.51	9.46	8.03	8.90
Macrolide resistance	4.09	3.36	3.48	4.34	3.46	3.88	3.71	3.82	2.62	2.72	2.45
mef(E)/mel	3.49	2.72	2.67	2.59	1.83	2.22	1.64	1.60	1.46	1.61	1.50
erm(B)	0.22	0.34	0.35	0.52	0.75	0.91	0.87	0.74	0.34	0.54	0.63
Dual resistance	0.19	0.08	0.39	1.12	0.68	0.65	1.11	1.35	0.79	0.57	0.32
Aged <2 years
Macrolide resistance	23.82	14.74	16.21	18.79	18.88	25.48	20.36	18.57	16.80	5.28	7.16
mef(E)/mel	20.42	11.06	13.51	12.16	10.49	13.80	10.67	6.55	8.96	5.28	6.14
erm(B)	0.00	0.00	0.00	1.11	2.10	3.19	1.94	1.09	2.24	0.00	1.02
Dual resistance	2.27	2.46	2.70	5.53	4.19	7.43	7.76	10.92	4.48	0.00	0.00
Aged 2–4 years
Macrolide resistance	7.89	4.73	3.70	4.84	6.53	6.19	8.22	7.77	2.08	2.78	2.03
mef(E)/mel	7.89	3.94	1.48	3.45	4.35	1.55	5.06	2.12	0.69	1.39	2.03
erm(B)	0.00	0.00	0.74	0.69	0.73	1.55	1.27	0.71	0.69	0.00	0.00
Dual resistance	0.00	0.00	1.48	0.69	1.45	3.09	1.27	4.95	0.69	1.39	0.00
Aged 5–17 years
Macrolide resistance	0.36	0.36	0.39	0.61	0.48	0.60	0.48	0.67	0.00	0.48	0.31
mef(E)/mel	0.18	0.36	0.39	0.20	0.32	0.45	0.16	0.17	0.00	0.48	0.15
erm(B)	0.18	0.00	0.00	0.00	0.16	0.00	0.32	0.00	0.00	0.00	0.15
Dual resistance	0.00	0.00	0.00	0.20	0.00	0.15	0.00	0.50	0.00	0.00	0.00
Aged 18–39 years
Macrolide resistance	1.73	1.06	1.45	2.29	1.61	1.00	1.80	0.85	1.20	0.73	0.98
mef(E)/mel	1.62	0.82	1.32	1.72	0.80	0.78	0.54	0.53	0.65	0.27	0.45
erm(B)	0.11	0.23	0.00	0.11	0.57	0.11	0.45	0.21	0.09	0.09	0.45
Dual resistance	0.00	0.00	0.00	0.46	0.23	0.00	0.81	0.00	0.46	0.36	0.09
Aged 40–64 years
Macrolide resistance	4.17	4.34	4.54	5.08	3.63	4.13	2.80	3.85	3.30	4.03	3.70
mef(E)/mel	3.10	3.47	3.50	2.13	1.71	2.16	1.12	1.19	1.74	2.57	1.98
erm(B)	0.24	0.43	0.70	1.02	0.81	1.41	0.84	1.19	0.55	0.83	0.95
Dual resistance	0.36	0.00	0.35	1.83	0.81	0.47	0.75	1.19	0.92	0.64	0.77
Aged ≥65 years
Macrolide resistance	14.77	11.72	10.05	13.32	9.75	12.23	13.63	15.33	7.31	8.74	5.94
mef(E)/mel	13.76	9.96	7.18	10.36	5.32	8.15	6.26	8.82	4.52	4.21	4.45
erm(B)	1.02	1.76	0.96	0.99	2.22	2.85	3.31	2.68	0.35	2.59	1.19
Dual resistance	0.00	0.00	1.44	1.48	1.77	1.22	3.68	3.83	2.09	1.94	0.30

Incidence reported as cases per 100000 population. Macrolide-resistant invasive pneumococcal disease strains that were not polymerase chain reaction positive for erm(B) or mef(E)/mel may be ribosomal mutations or an unknown mechanism.

Emergence of MR-IPD Due to erm(B) and the Dual Macrolide Resistance Genotype

In the PCV7 era, from 2003 to 2009, mef(E)/mel-mediated MR-IPD declined, from 3.49 to 1.64/100000 (P < .0001; Figure 1, Table 1). This decline was offset by an emergence of isolates with erm(B)-mediated resistance after 2003. The incidence of isolates containing erm(B) alone significantly increased from 2003 to 2009, from 0.22 to 0.87/100000 (P = .0002; Table 1) due to increases in serotypes 19A, 15A, and 23A (Supplementary Table 1). Also noted was the emergence of isolates with a dual macrolide resistance genotype (mef(E)/mel and erm(B); Table 1). This genotype was not observed prior to 2000 in our population but steadily increased from 2003 to 2010, 0.19 to 1.35/100000 (P < .0001; Figure 1, Table 1). After introduction of PCV13 in infants in 2010, MR-IPD declined from 2009 to 2013, 3.71 to 2.45/100000 (P = .0103). The dual macrolide resistance genotype declined from 2009 to 2013, 1.11 to 0.32/100000 (P < .0001; Figure 1, Table 1). The overall incidence of MR-IPD with mef(E)/mel or erm(B) alone stabilized from 2009 to 2013 (Figure 1, Table 1; Supplementary Tables 1 and 2).

Age-Specific Changes in Macrolide-Resistant Invasive Pneumococcal Disease

In those aged <2 years, the mean incidence of erm(B) alone or dual macrolide resistance (mef(E)/mel and erm(B)) increased from 0.00 and 2.48 to 2.13 and 5.72/100000 (mean incidence 2003–2005 compared to 2006–2009, P = .0077 and P = .0106), respectively (Table 1, Figure 2A). Following PCV13 introduction, MR-IPD rapidly declined 65.7% in those aged <2 years by 2013 (2006–2009 mean incidence of 20.88–7.16/100000 in 2013; P = .0033; Figure 2A, Table 1). Reductions in MR-IPD were observed for the 3 genotypes: mef(E)/mel dropped 47.9% (11.78–6.14/100000), erm(B) dropped 51.2% (2.09–1.02/100000), and, after reaching 10.92/100000 in 2010, the dual macrolide resistance genotype disappeared by 2012 (Figure 2A, Table 1).

Figure 2.

The incidence (2003–2013) in metropolitan Atlanta, Georgia, of macrolide-resistant invasive pneumococcal disease in individuals (A) aged <2 years and (B) aged ≥65 years. Incidence is shown as cases per 100000 population, with white bars for mef(E)/mel incidence, vertical striped bars for erm(B) incidence, and black bars for the dual macrolide resistance genotype (mef(E)/mel and erm(B)) incidence.

In those aged ≥65 years, MR-IPD incidence caused by isolates containing erm(B) increased from 1.02 to 3.31/100000 (2003–2009) and the dual macrolide resistance genotype increased from 0.00 to 3.68/100000 (P = .0024). Conversely, the incidence of isolates containing mef(E)/mel declined from 13.76 to 6.26/100000 (Table 1, Figure 2B). After the introduction of PCV13, MR-IPD decreased 51.4% (2006–2009 mean incidence of 12.23 to 5.94/100000 in 2013, P = .0013; Figure 2B, Table 1). This reduction was due to decreases in all resistance genotypes: 40.8% reduction of mef(E)/mel (7.52 to 4.45/100000, P = .0486), 49.1% reduction of erm(B) (2.34 to 1.19/100000, P = .0546), and 85.3% reduction of the dual resistance genotype (2.04 to 0.30/100000, P = .0169; Figure 2B, Table 1).

MR-IPD in those aged 2–4 years fluctuated from 2003 to 2009, and a 68.5% reduction following PCV13 introduction was experienced (2006–2009 mean incidence of 6.45–2.03/100000 in 2013, P = .0254; Table 1). In low-risk age groups (those aged 4–17, 18–39, and 40–64 years), MR-IPD did not change significantly (Table 1).

Macrolide-Resistant Serotypes

PCV7 or PCV13 Serotypes

MR-IPD caused by PCV7 serotypes (4, 6B, 9V, 14, 18C, 19F, and 23F) continued to decline from 2003 to 2009, from 1.43 to 0.09/100000 (P < .0001). Remarkably, by 2013 no cases of MR-IPD caused by PCV7 serotypes were detected in the population (Figure 3, Table 2). Of the additional 6 serotypes contained in PCV13 (1, 3, 5, 6A, 7F, and 19A), serotypes 1 and 5 did not cause MR-IPD and serotype 7F caused only three cases of MR-IPD in our population from 2003–2013 (Figure 3, Table 2). MR-IPD caused by serotype 3 remained low from 2003 to 2013 at a mean incidence of 0.07/100000 (Table 2). The incidence of MR-IPD caused by serotype 19A significantly increased from 2003 to 2009, 0.93 to 2.15/100000 (P = .0001), which was due to increases in erm(B) (Supplementary Table 1) and dual macrolide resistance genotypes (Figure 4). Following PCV13 introduction in 2010, MR-IPD due to serotype 19A sharply declined to 0.43/100000 by 2013 (P < .0001; Figure 3, Table 2). MR-IPD caused by serotype 6A declined from 2003 to 2009, 0.57 to 0.06/100000 (P = .0003), and no cases occurred from 2010 to 2013 (Figure 3, Table 2).

Figure 3.

The serotype distribution of Streptococcus pneumoniae macrolide-resistant invasive disease isolates in metropolitan Atlanta, Georgia, as percent of isolates (A) 2002 (2 years after PCV7 introduction), (B) 2003–2005, (C) 2006–2009, and (D) 2013 (3 years after PCV13 introduction). Sizes of pie charts are scaled (based on 2002) to represent cases per year. The following are illustrated: (A) 100% (74 cases), (B) 135% (301 cases, mean 100/year), (C) 162% (480 cases, mean 120/year), and (D) 116% (86 cases). Other pneumococcal serotypes represent nonvaccine serotypes. Serotypes contained in pneumococcal conjugate vaccine 13 are underlined. Abbreviation: PVC, pneumococcal conjugate vaccine.

Table 2.

Incidence of Macrolide-Resistant Invasive Pneumococcal Disease 2003–2013 in Atlanta, Georgia, by Serotype

Incidence	2003	2004	2005	2006	2007	2008	2009	2010	2011	2012	2013
Overall invasive pneumococcal disease	14.02	12.29	12.89	13.50	13.23	12.02	11.99	11.51	9.46	8.03	8.90
Macrolide-resistant invasive pneumococcal disease	4.09	3.36	3.48	4.34	3.46	3.88	3.71	3.82	2.68	2.72	2.45
PCV7 serotype	1.43	0.79	0.60	0.36	0.10	0.13	0.09	0.19	0.15	0.03	0.00
4	0.11	0.08	0.00	0.00	0.00	0.00	0.00	0.00	0.03	0.00	0.00
14	0.57	0.11	0.14	0.13	0.00	0.03	0.00	0.06	0.03	0.00	0.00
18C	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00
19F	0.14	0.04	0.18	0.16	0.03	0.00	0.06	0.03	0.00	0.03	0.00
23F	0.25	0.15	0.14	0.03	0.00	0.00	0.03	0.03	0.00	0.00	0.00
6B	0.22	0.19	0.14	0.03	0.00	0.06	0.00	0.00	0.00	0.00	0.00
9V	0.14	0.23	0.00	0.00	0.07	0.03	0.00	0.06	0.09	0.00	0.00
PCV13 serotype	1.61	1.73	1.88	2.51	1.83	1.92	2.30	1.88	1.08	0.77	0.51
1	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00
3	0.11	0.08	0.04	0.13	0.03	0.00	0.09	0.16	0.00	0.15	0.09
5	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00
19A	0.93	1.43	1.56	2.12	1.66	1.80	2.15	1.72	1.08	0.59	0.43
6A	0.57	0.23	0.28	0.23	0.10	0.13	0.06	0.00	0.00	0.00	0.00
7F	0.00	0.00	0.00	0.03	0.03	0.00	0.00	0.00	0.00	0.03	0.00
Nonvaccine serotype	1.04	0.83	0.99	1.47	1.53	1.83	1.32	1.75	1.39	1.92	1.93
12F	0.00	0.15	0.04	0.00	0.00	0.00	0.00	0.00	0.06	0.00	0.00
15A	0.04	0.04	0.11	0.26	0.51	0.42	0.36	0.28	0.33	0.35	0.37
15B	0.00	0.00	0.07	0.03	0.07	0.03	0.00	0.00	0.06	0.06	0.11
15C	0.00	0.00	0.00	0.03	0.00	0.03	0.00	0.00	0.00	0.09	0.11
22F	0.00	0.00	0.00	0.07	0.00	0.13	0.09	0.13	0.03	0.15	0.17
23A	0.00	0.00	0.00	0.00	0.14	0.26	0.06	0.09	0.06	0.15	0.14
33F	0.68	0.34	0.60	0.55	0.41	0.45	0.33	0.19	0.27	0.44	0.23
35B	0.04	0.04	0.04	0.13	0.07	0.10	0.09	0.28	0.18	0.32	0.43
11A	0.07	0.08	0.00	0.10	0.03	0.10	0.03	0.19	0.09	0.12	0.17
6C	0.11	0.11	0.11	0.20	0.24	0.19	0.18	0.28	0.24	0.21	0.09
Others	0.11	0.08	0.04	0.10	0.07	0.13	0.18	0.31	0.06	0.03	0.11

Incidence reported as cases per 100000 population.

Abbreviation: PCV, pneumococcal conjugate vaccine.

Figure 4.

Streptococcus pneumoniae macrolide-resistant invasive pneumococcal disease isolates with dual resistance genes (mef(E)/mel and erm(B)) were first observed in the Atlanta, Georgia, surveillance area in 2000: 19F (black bars), 19A (white bars). Other serotypes (striped bars) with the dual macrolide resistance genotype were first observed in 2006 (11A), 2008 (23A), 2009 (15A, 33F, and two nontypeable strains), 2010 (two 22F and one nontypeable), 2012 (6C and 35B), and 2013 (serotype 3).

Non-PCV7 or PCV13 Serotypes

Although the overall incidence remained <2/100000, MR-IPD caused by nonvaccine serotypes significantly increased from 2003 to 2013 (Figure 3, Table 2, Supplementary Tables 1 and 2), essentially doubling. The incidence of serotype 15A (primarily erm(B)-mediated) increased approximately 10 fold (from 2003 to 2013, 0.04 to 0.37/100000; P = .0033), as did serotype 35B (exclusively mef(E)/mel-mediated, 0.04 to 0.43/100000; P < .0001). Serotypes 22F (P = .0230), 15B (P = .0367), and 15C (P = .0047) also increased due to mef(E)/mel, and 23A (P = .0379) increased due to erm(B) (Table 2, Supplementary Tables 1 and 2). The incidence of MR-IPD cases due to serotype 33F, exclusively mef(E)/mel-mediated, remained stable, with a mean incidence of 0.41/100000 from 2003 to 2013, causing 9%–15% of MR-IPD (Figure 3, Table 2, Supplementary Tables 1 and 2).

Dual Macrolide Resistance Serotypes

In 2003, the first MR-IPD serotype 19A, ST320 isolate with the dual macrolide resistance genotype (mef(E)/mel and erm(B)) was identified (Figure 4). The dual-resistance genotype significantly increased from the 2003 to 2009 (P < .0001, Figure 1 and 4, Table 1). Dual macrolide-resistant isolates from 2003 to 2005 belonged to CC320 (three 19F included ST271 and two ST3039, while the 19A isolates were all ST320). All dual macrolide-resistant isolates from 2007 to 2009 were also found to belong to CC320, two 19F isolates (ST236 and ST3039), twenty-three 19A isolates (ST320), and one 19A isolate (ST1339). These isolates contain the newly recognized mobile genetic element Tn2010 [10]. Following PCV13 introduction in 2010, 19A isolates with the dual macrolide resistance genotype rapidly declined (Figures 1 and 4). The dual macrolide resistance genotype was also identified in serotypes 3, 6C, 11A, 15A, 22F, 23A, 33F, 35B, and nontypeable S. pneumoniae (Figure 4).

DISCUSSION

PCVs have been introduced into populations worldwide, and each time major reductions in the burden of pneumococcal disease have been noted [12]. PCVs provide individual protection and reduce transmission and asymptomatic nasopharyngeal carriage of the serotypes contained in PCVs (eg, herd protection) [23, 24]. PCVs also reduce the incidence of antibiotic-resistant S. pneumoniae by decreasing serotypes that carry multiple, antibiotic-resistant determinants [12]. In a previous study, we found MR-IPD incidence decreased dramatically in the Atlanta population after the introduction of PCV7, primarily due to the reduction of disease due to PCV7 serotypes that contain the mef(E)/mel efflux macrolide resistance determinant [2]. The reductions of IPD and MR-IPD observed in metropolitan Atlanta following PCV7 introduction reflected the rest of the United States [25].

However, as we report here, the incidence of MR-IPD stabilized in Atlanta from 2003 to 2009. We found a continued decrease in MR-IPD due to mef(E)/mel-containing PCV7 serotype isolates from 2003 to 2009 with use of PCV7. Also, MR-IPD caused by serotype 6A declined throughout the PCV7 era despite not being included among PCV7 serotypes. Protection against serotype 6A by PCV7 occurred through immunological cross-protection by serotype 6B, a serotype included in PCV7 [26]. The effect of PCV7 on 6A was initially unclear until the discovery of serotype 6C, which was historically not distinguished from serotype 6A [26] by serologic capsule typing and for which serotype 6B did not provide cross-protection.

In contrast, MR-IPD due to erm(B)-only isolates and predominantly isolates that contain both determinants emerged and increased in the PCV7 era (Figure 1). Much of this was due to an increase in MR-IPD serotype 19A isolates with erm(B) or dual macrolide resistance. The incidence of MR-IPD caused by erm(B)-resistant serotype 15A and 23A isolates and mef(E)/mel-resistant serotype 35B isolates also increased following PCV7 introduction. We recently defined the mobile genetic elements that disseminate macrolide resistance in these emerging strains [9]. Although the routine use of PCV7 in children reduced MR-IPD caused by PCV7 serotypes and serotype 6A, serotype replacement was observed [2, 13, 24, 27] and was likely driven by the selective pressures of PCV7 and continued high-level macrolide usage [3, 28].

Streptococcus pneumoniae that contain the dual macrolide resistance determinants mef(E)/mel and erm(B) were first noted in the late 1990s from the United States, Japan, and South Africa [29–32]. Many of these strains were noted to belong to a single 19F multidrug-resistant clone and subsequently identified as Taiwan^19F-14 or PMEN14 [32, 33]. MR-IPD caused by dual macrolide-resistant serotype 19A S. pneumoniae belong to CC320 (formerly CC271) [9, 10, 34], a PMEN14 lineage. This genotype was first identified in the Atlanta surveillance area in 2000 in three serotype 19F isolates (ST236, ST271, and ST3039), all clonal complex CC320 [9]. From 2006 to 2010 in the Atlanta population, in the United States, and worldwide, the incidence of dual macrolide-resistant serotype 19A CC320 cases steadily increased [8, 10, 14, 34] (Figure 4). Dual macrolide-resistant isolates were found to contain the mobile genetic element Tn2010 and were multidrug resistant [9]. Tn2010 is a large 26.4-Kb element with Mega (mef(E)/mel) and Tn917 (erm(B)) inserted at 2 distinct sites into a tet(M)-carrying Tn916-like conjugative transposon [4, 10]. The 19A isolates of the clonal lineage of Taiwan^19F-14 arose by capsular switching and clonal expansion under vaccine selection. Other 19A isolates were capsule switching clones of other PCV7 serotypes or an expansion of isolates in the 19A clonal complex [35]. “Soft selective sweeps” appear important in the evolution of pneumococcal multidrug resistance phenotypes and the emergence of vaccine escape clones [33].

Introduction of PCV13 in infants (2010) decreased the incidence of MR-IPD caused by dual resistance isolates of serotype 19A. By 2013, no invasive disease caused by dual macrolide resistance genotype isolates was detected in individuals aged <18 years in Atlanta. The greatest reductions in MR-IPD after PCV13 were observed in those aged <2 years (74%) and those aged 2–4 years (66%) and in largely unvaccinated populations (those aged ≥65 years; 51.4%). As of March 2012, only 56% of US children aged <5 years were vaccinated with PCV13 [36], so the rapid decline in all age groups from 2010 to 2013 may be attributed to both individual and herd protection. These data are consistent with overall IPD reductions seen throughout the United States [17]. Until recently, only the 23-valent pneumococcal polysaccharide vaccine (PPV23) was recommended for those aged ≥65 years who are at high risk of pneumococcal disease. PPV23 provides some protection against IPD but does not impact pneumococcal carriage or transmission [12, 24]. In August 2014, the US Advisory Committee on Immunization Practices released a time-limited PCV13 recommendation for routine use in the those aged ≥65 years [37, 38]. This introduction may further reduce the incidence of IPD and MR-IPD in this population.

Use of macrolides remains a powerful selective pressure for macrolide resistance in S. pneumoniae [3, 28]. While PCVs have reduced the burden of MR-IPD [2, 8], macrolide resistance in pneumococci has continued to increase worldwide. In Atlanta, <27.5% of IPD isolates in 2013 were macrolide resistant, while in China >90% of S. pneumoniae isolates were macrolide resistant [39, 40]. Although the current burden of disease remains low, the emergence of MR-IPD due to nonvaccine serotypes such as 15A and 35B in our population is a concern. The data indicate a need for continued monitoring of IPD and MR-IPD serotypes for future pneumococcal conjugate vaccine formulations [12].

Macrolide resistance in S. pneumoniae remains an important clinical problem. Macrolides are commonly prescribed for treating suspected bacterial upper respiratory infections but should be used judiciously, especially in settings of >25% high-level pneumococcal macrolide resistance, such as in Atlanta HD-3. Though macrolide treatment is not recommended for known IPD, macrolides are recommended empiric therapy for community-acquired pneumonia in otherwise healthy adults [41]. Amoxicillin is the recommended first-line therapy for community-acquired pneumonia caused by bacteria, and macrolide treatment is only indicated for children with a history of anaphylaxis to β-lactam antibiotics or evidence of atypical pathogens [42]. Data regarding specific macrolide use during the study in Atlanta are not available, but US national data for 2010 indicate azithromycin was the most frequently prescribed at 174 prescriptions per 1000 persons [43]. Point-of-care diagnostics for community-acquired pneumonia and upper respiratory infections can lead to initial pathogen-directed therapy and decrease empiric therapy. As noted here, the introduction of pneumococcal conjugate vaccines has significantly lowered community levels of pneumococcal macrolide resistance.

In summary, serotype 19A, especially CC320 with dual macrolide-resistant mechanisms, rapidly emerged in Atlanta following PCV7 introduction as a dominant serotype replacement macrolide and multiantibiotic-resistant clone. The incidence of MR-IPD, due largely to the impact on serotype 19A, decreased dramatically among those aged <2 years as well as in those aged 2–4 years and ≥65 years following the introduction of PCV13 in 2010. Expanded pneumococcal conjugate vaccines continue to reduce the burden of IPD and are important tools in combating antibiotic resistance. However, macrolide resistance in some nonvaccine serotypes (albeit at low overall incidence) has increased. Our study emphasizes the importance of ongoing surveillance for invasive S. pneumoniae disease and the need for continuous programs that emphasize the judicious use of antibiotics.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. We thank Yih-Ling Tzeng for critical review of the manuscript; Lillian Morgan, Wendy Baughman, Amy Tunali, and others of the Georgia Emerging Infections Program; Bernard Beall and the Centers for Disease Control and Prevention (CDC) Streptococcus Lab; and Matthew Moore of the CDC Respiratory Disease Branch.

Financial support. This work was supported by the National Institutes of Health (RO1 A1070829 to D. S. S.) and the CDC-sponsored Emerging Infections Program.

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.