The SENTRY Antimicrobial Surveillance Program was established in 1997 to monitor pathogens and changes in antimicrobial resistance patterns of organisms collected from patients with various infections (15). Every year, participating laboratories at medical centers in all 9 US Census Divisions identify requested surveillance pathogens by routine microbiologic methods and submit a subculture to the SENTRY program, along with basic demographic information about the case patient and limited information about the clinical setting where the case patient was treated. Pathogen confirmation is conducted by the central laboratory (JMI Laboratories, North Liberty, Iowa) by various techniques including colony morphology, biochemical algorithms, MALDI-TOF MS, PCR assays, and/or sequencing, as needed. From the SENTRY program collection of stored pathogens, S. pneumoniae isolates cultured from MEF samples from children 5 years of age or younger that were submitted from 35 participant laboratories during 2011–2021 were included in this study. Because tympanocentesis is not routinely performed among children with OM in the US and collection technique was not reported, we assumed that the MEF had been collected from a tympanocentesis performed on children with complicated OM (children with recurrent OM, children with a poor response to conventional therapy or children with a clinical relapse) or with a swab from otorrhea from a spontaneously perforated OM (16–18). We restricted serotyping to S. pneumoniae isolates collected from 2011 through 2021, after introduction of PCV13 into the US pediatric immunization program.

Serotypes were identified using in silico analysis based on Pneumococcal Capsular Typing methodology (PneumoCaT). Specifically, DNA from S. pneumoniae isolates was extracted using the KingFisher Cell and Tissue DNA kit on a KingFisher™ Flex Magnetic Particle Processor (Thermo Scientific) workstation. Total genomic DNA was used as input material for library construction and sequencing on a NextSeq 1000 Sequencer (Illumina, San Diego, California, USA) using NextSeq™1000/2000 P2 Reagents (300 cycles). The DNA libraries for the NextSeq1000 Sequencer were prepared using the Illumina DNA™ library construction protocol and index kit. Rigorous quality control metrics were applied to library construction including verification that the “Quality Score” (% ≥Q30) for the run was above 75%, the percentage of clusters passing filter was ≥60%, and the loading concentration was ≥95. Each raw data set was de novo assembled using SPAdes 3.11.1. PneumoCaT (v. 1.2.1) used a two-step approach to assign capsular type to S. pneumoniae genomic data. In the first step, if the reads matched >90% to one or more of the 92 serotypes for S. pneumoniae plus 2 additional subtypes/molecular types, a capsular type was assigned. If more than one loci matched, then in the second step, a variant-based approach that utilizes the capsular type variant database to distinguish serotypes within a subgroup/genogroup was applied to make a call on the serotype. If the coverage value against the reference sequence was ≤90%, then the serotype analysis reported “Failed” as its value. Average depth of coverage across the matching region of ≥30X was considered acceptable for PneumoCaT to analyze the data (19, 20). The nucleotide sequences used to assign a serotype were submitted to the National Center for Biotechnology Information (NCBI) and were assigned accession numbers SAMN41612263 - SAMN41612461.

Minimum Inhibitory Concentrations (MICs) were determined by the broth microdilution method using Clinical and Laboratory Standards Institute (CLSI) guidelines (2018) (21). Reference broth microdilution panels were manufactured at JMI Laboratories (2015–2021) or purchased from Thermo Fisher Scientific (2011–2014) (Cleveland, Ohio, USA) using freshly prepared drug stocks and stored at −80°C until use. Testing was performed in cation-adjusted Mueller-Hinton broth supplemented with 2.5% to 5.0% lysed horse blood (MHB-LHB). Antimicrobial susceptibility testing was performed for amoxicillin-clavulanic acid, penicillin, ceftriaxone, clindamycin, erythromycin, trimethoprim/sulfamethoxazole (TMP/SMX), levofloxacin, and vancomycin. MIC results were interpreted as susceptible, intermediate, or resistant according to the CLSI recommendations (22). Non-meningitis breakpoints were used for penicillin and ceftriaxone. Non-susceptible isolates were defined as those isolates that were intermediate or resistant to a selected antimicrobial agent. Multidrug resistance (MDR) was defined as resistant to three or more classes of antimicrobials. The MIC₅₀ was defined as the MIC of a given antimicrobial drug that inhibited growth of 50% of isolates, and MIC₉₀ was defined as the MIC that inhibited growth of 90% of the isolates.

Percentages of cases due to each serotype, non-susceptible to the selected antimicrobials, and covered by PCVs were calculated. PCV coverage was estimated for PCV7 (4, 6B, 9 V, 14, 18C, 19F, and 23F), PCV13 non-PCV7 (1, 3, 5, 6A, 7F, and 19A), PCV13 (PCV7 + PCV13 non-PCV7 serotypes), PCV15 (PCV13, 22F, and 33F), PCV15 non-PCV13 (22F and 33F), PCV20 non-PCV13 (8, 10A, 11A, 12F, 15B, 22F, and 33F), PCV20 (PCV13, PCV20 non-PCV13 serotypes), and non-PCV20 (serotypes not covered by PCV20). While serotypes 6C and 15C are not included in any PCV formulations, we also present PCV coverage estimates where these were grouped with PCVs that contained conjugates 6A or 15B due to potential prevention based on immunological or epidemiological evidence of cross-protection (23–25). Percentages were also stratified into two age groups (<2 years and 2–5 years) and two periods (to reflect potential changes associated with PCV13 introduction; Period 1 [P1: 2011–2016] and Period 2 [P2: 2017–2021]). Periods and age groups were compared for individual serotypes and PCV serotype groups by Chi-square test; p-values < 0.05 were considered significant. Non-susceptibility data were stratified by PCV group (PCV13, PCV20 non-PCV13, PCV15 non-PCV13, Non-PCV20). Analyses were performed using STATA (26).

A total of 199 S. pneumoniae isolates from middle ear fluid among children 5 years of age or younger were collected by the SENTRY program from 2011 to 2021. All 199 S. pneumoniae were serotyped (by sequencing) and tested for antimicrobial susceptibility (Table 1). Of these, 113 isolates (56.8%) were from children <2 years of age. Most isolates were collected from children receiving care in the ear, nose, and throat (ENT; n = 89, 44.7%) or pediatrics (n = 61, 30.7%) departments. Approximately 60% of isolates came from male children, and about 50% were isolated between 2017 and 2021. While isolates were reported from all 9 US Census regions, most (57.3%) originated from the West North Central region.

Among the 199 S. pneumoniae isolates that were serotyped, twenty-three unique serotypes were identified. Roughly 60% of isolates were represented by 6 serotypes: 35B (16.6%), 15B (14.6%), 15A (7.5%), 19A (7.5%), 19F (7.5%), and 3 (7.0%) (Figure 1). Serotype 3 represented a higher percentage of total isolates among children 2–5 years of age than among those age <2 years (12.8% vs. 2.7%, respectively; p-value = 0.01), whereas serotype 15B represented a smaller percentage in older children (9.3% vs. 18.6%, respectively; p-value = 0.06). Comparing study periods, serotypes 35B and 15B represented the highest percentage of detected serotypes in P1 and P2, respectively (Figure 2 and Supplementary Table S1) whereas the percentage due to serotype 19A was higher in P1 than in P2 (10.8% vs. 4.1%, respectively; p-value = 0.08) and serotype 19F was higher in P2 than in P1 (11.3% vs. 3.9%, respectively; p-value = 0.048; Figure 2 and Supplementary Table S1). Serotype distribution by age group and study period are presented in Supplementary Table S1.

In children aged ≤5 years, PCV13, PCV15, and PCV20 serotypes accounted for 23.1%, 30.2%, and 54.8% of S. pneumoniae isolates, respectively and 45.2% of cases were due to Non-PCV20 serotypes (Table 2). The PCV20 non-PCV13 serotypes and PCV15 non-PCV13 serotypes contributed 31.7% and 7.5% of cases, respectively (Table 2). Considering the age group, while PCV20 coverage was similar for children in both age groups, PCV20 non-PCV13 coverage was higher for children <2 years than in children 2–5 years of age (37.2% vs. 20.9%, respectively; p-value = 0.01) and this was due to a larger percentage of serotype 15B isolates detected in the younger age group. Overall, the percentage of serotypes included in PCV20 was higher in P2 than P1 (57.7% vs. 48.0%, respectively; p-value = 0.17) as was also true for PCV20 non-PCV13 serotypes (35.1% vs. 25.5%, respectively; p-value = 0.14) (Table 2). Serotype distribution by age group and study period are presented in Supplementary Table S2.

Non-susceptibility to penicillin, erythromycin, and TMP/SMX was common overall (45.2%, 49.8%, and 37.4%, respectively), but only 9.6% for amoxicillin-clavulanic acid (Table 3). Overall, 13.6% (n = 27) of isolates were MDR; most were PCV13 serotypes (n = 15) (Table 3). The proportions of non-susceptible isolates were similar across time periods for these antimicrobials, except for PCV13, for which proportions were lower in P2 than in P1 (Supplementary Table S3). Non-susceptibility to at least one antimicrobial was observed among 17 of the 23 serotypes identified. The majority of non-susceptibility was accounted for by serotypes 19A, 15A, 15B, 15C, 23A, 33F, and 35B (Table 4). More than 50% of serotype 19A isolates were non-susceptible to all antimicrobials tested, and 87% (n = 13) were MDR.