Pneumococcal serotypes in children, clinical presentation and antimicrobial susceptibility in the PCV13 era

The aim was to analyse invasive pneumococcal disease (IPD) serotypes in children aged ⩽17 years according to clinical presentation and antimicrobial susceptibility. We conducted a prospective study (January 2012–June 2016). IPD cases were diagnosed by culture and/or real-time polymerase chain reaction (PCR). Demographic, microbiological and clinical data were analysed. Associations were assessed using the odds ratio (OR) and 95% confidence intervals (CI). Of the 253 cases, 34.4% were aged <2 years, 38.7% 2–4 years and 26.9% 5–17 years. Over 64% were 13-valent pneumococcal conjugate vaccine (PCV13) serotypes. 48% of the cases were diagnosed only by real-time PCR. Serotypes 3 and 1 were associated with complicated pneumonia (P < 0.05) and non-PCV13 serotypes with meningitis (OR 7.32, 95% CI 2.33–22.99) and occult bacteraemia (OR 3.6, 95% CI 1.56–8.76). Serotype 19A was more frequent in children aged <2 years and serotypes 3 and 1 in children aged 2–4 years and 5–17 years, respectively. 36.1% of cases were not susceptible to penicillin and 16.4% were also non-susceptible to cefotaxime. Serotypes 14, 24F and 23B were associated with non-susceptibility to penicillin (P < 0.05) and serotypes 11, 14 and 19A to cefotaxime (P < 0.05). Serotype 19A showed resistance to penicillin (P = 0.002). In conclusion, PCV13 serotypes were most frequent in children aged ⩽17 years, mainly serotypes 3, 1 and 19A. Non-PCV13 serotypes were associated with meningitis and occult bacteraemia and PCV13 serotypes with pneumonia. Non-susceptibility to antibiotics of non-PCV13 serotypes should be monitored.

The aim of this study was to analyse the distribution of S. pneumoniae serotypes according to the clinical presentation of IPD and antimicrobial susceptibility in children aged ⩽17 years in a community with intermediate PCV13 coverage before its introduction in the official paediatric vaccination calendar.

Study design
We conducted a prospective observational study between January 2012 and June 2016 in children aged ⩽17 years with IPD attended in three paediatric hospitals: Hospital Vall d'Hebron of Barcelona, Hospital Sant Joan de Déu of Esplugues, Barcelona and Hospital de Nens of Barcelona. The estimated reference population aged ⩽17 years of the hospitals was 442 761, representing 31.9% of this age group in Catalonia.
IPD was diagnosed through the isolation of S. pneumoniae by culture or detection of bacterial DNA by real-time polymerase chain reaction (PCR) in any normally sterile site [15], such as blood, cerebrospinal fluid, pleural fluid or articular fluid, among others. We established the presence of S. pneumoniae DNA by amplification of the autolysin (lytA) gene and the wzg (cpsA) gene, according to published assays [16][17][18]. We included only samples that were positive for the lytA and wzg genes in real-time PCR.
All clinical samples were analysed by culture and real-time PCR.
The clinical evaluation was made by the paediatricians participating in the study according to established diagnostic criteria [19].

Data collected
Epidemiological and clinical characteristics, including age, gender, underlying medical condition, clinical presentation, intensive care unit (ICU) admission, clinical outcome and PCV13 vaccination status were collected. The method of diagnosis, the serotype and antimicrobial susceptibility were analysed.
A subject was considered vaccinated if they had received the last dose of PCV13 ⩾15 days before symptom onset.

Serotype identification and antimicrobial susceptibility
Strains isolated by culture were serotyped using the Quellung reaction or dot blot at the National Centre of Microbiology (Majadahonda, Spain) [20]. When the diagnosis was made only by real-time PCR, serotypes were identified at the Molecular Microbiology Department, Hospital Sant Joan de Déu, in accordance with previously-validated methods [15].
PCR-positive samples that were negative for the serotypes included in the sequential multiplex PCR (including all vaccine serotypes) were classified as other non-vaccine serotypes (ONV).
Susceptibility to penicillin and cefotaxime was determined by the microdilution method according to the Clinical and Laboratory Standards Institute (CLSI) criteria. Current EUCAST breakpoints were used to interpret susceptibility [21]. A minimum inhibitory concentration (MIC) study was carried out at the National Microbiology Center. The penicillin breakpoints used were: MIC >0.06 mg/l, non-susceptible and MIC >2 mg/l, resistant. For cefotaxime, the breakpoint was: MIC >0.5 mg/l, non-susceptible.

Statistical analysis
Proportions were compared using the χ 2 test or Fisher's exact test, as appropriate. All statistical tests were two-tailed and statistical significance was established as P < 0.05. If the number of cases was zero, this was corrected by adding 0.5, and P values were calculated according to Sheskin's method [22] Associations between variables were assessed using the odds ratio (OR) and 95% confidence intervals (CI).
The analyses were conducted using the Statistical Package for Social Sciences (SPSS 19.0 for Windows) and EPIDAT (program for the epidemiological analysis of tabulated data; version 3.1).

Data confidentiality and ethical aspects
Informed consent was obtained from all individual participants included in the study. The study complies with the principles of the Declaration of Helsinki and the legal structure with respect to international human rights and biomedicine and protection of personal data laws. The Ethics Committee of Hospital Sant Joan de Déu approved the study.
There was a seasonal variation, with 72% of cases detected during the cool months (October to March; P < 0.001).
Two patients died: an 18-month-old child with serotype-35F meningitis who had congenital immunodeficiency, and a 23-month-old child with septic shock caused by a non-vaccine serotype which was not identified.
A total of 92.5% of IPD cases caused by serotype 3 were overall pneumonia and 83% of IPD cases caused by serotype 3 were complicated pneumonia. Serotype 3 caused overall pneumonia and complicated pneumonia more frequently than the other clinical forms (OR 5.5, 95% CI 1.9-15.92 and OR 4.3, 95% CI 2.01-9.35, respectively).

Serotypes and method of diagnosis
The diagnosis was made by culture in 62 cases (24.5%), culture and real-time PCR in 69 (27.3%) and real-time PCR in 122 (48.2%).
PCV13 and non-PCV13 serotypes showed no significant difference in the percentage of cases diagnosed only by PCR (52.1% vs. 41.1%).
There were significant differences between serotypes in the method of diagnosis: 86.8% of cases due to serotype 3 were diagnosed only by real-time PCR vs. 38% of cases due to other serotypes (P < 0.001). By contrast, other serotypes, such as serotype 14 and 24F, were diagnosed less frequently by real-time PCR (13.3%, 2/15 and 0%, 0/6) than the other serotypes (P = 0.005 and P = 0.030, respectively).

Epidemiology and Infection
Twenty cases (16.4%) were also non-susceptible to cefotaxime (Table 4). With respect to the clinical form, 5 out of 44 cases with MIC > 0.06 were considered resistant due to the meningeal location (serotypes 14, 19A, 19F, 24A and 24F). One case (serotype 14) was also non-susceptible to cefotaxime.
There were no significant differences in the percentage of nonsusceptible penicillin strains between PCV13 (36.1%) and non-PCV13 serotypes (36.0%). PCV13 serotypes were associated with strains not susceptible to cefotaxime (P = 0.010) and with isolates not susceptible to either penicillin or cefotaxime (P = 0.010).
Serotype 11A showed non-susceptibility to cefotaxime in 60% of strains compared with 14.5% for other serotypes (P = 0.031). These strains were also non-susceptible to penicillin (P = 0.031).

Discussion
PCV13 serotypes caused 64.4% of IPD cases in children aged ⩽17 years during the study period, and the most frequent serotypes were 3, 1, 19A and 14. Other studies in Catalonia have shown a higher frequency of PCV13 serotypes, despite the decreased incidence in all age groups in this period [23]. However, other authors have found a higher percentage of non-PCV13 serotypes. Makwana et al. observed that 77.5% of cases of IPD in children aged <5 years in England and Wales between 2010 and 2016 were caused by non-PCV13 serotypes although there was a high immunisation rate with 97.3% having received ⩾1PCV13 dose [24]. A study conducted in Madrid (Spain) between 2012 and 2015, with vaccination coverages between 67% and 95%, found that non-PCV13 serotypes accounted for 68.2% of cases in children aged <15 years [25], somewhat higher than the 55.2% described by Janoir et al. [14] in France in 2011-2012 in the same age group, with a vaccination coverage of at least 90%. The differences between these results might be explained, at least in part, by a lower PCV13 coverage (about 64% in children aged <2 years) in Catalonia than in other areas because during the study period, PCV13 was not included in the Catalan immunisation schedule [12]. In fact, a systematic review by Balsells et al. in children in the post-PCV era found large differences between countries. Non-PCV13 serotypes accounted for 42.2% of childhood IPD cases, but marked regional differences were observed, from 9.2% in one Eastern Mediterranean country to 71.9% in Europe [26].
With respect to the distribution of the main serotypes, Ceyhan et al. [27] found that, in Turkish children aged <18 years with 25.5% of IPD cases having received any PCV dose, serotype 19F was the most frequent, followed by 14, 3 and 6B (in equal amounts), while Camilli et al. [28], in Italian children aged <5 years in 2012-2014 with PCV coverage at 24 months of age of about 80% found that serotypes 1, 19A and 14 were among the Table 4. *The P value indicates differences in the relationship between a serotype and non-susceptibility to an antibiotic with respect to the other serotypes.
four most frequent serotypes, together with 24F. In addition to the differences in PCV coverage, the diagnostic techniques used in these studies might help to explain the differences observed [29]. In our study, serotype 3 was the most frequent and also the most frequently diagnosed only by real-time PCR (87% of cases due to serotype 3 were diagnosed only by real-time PCR). When real-time PCR is not available for the diagnosis, another distribution of serotypes in the same age groups may be found, because some serotypes are more antimicrobial-resistant than others and real-time PCR makes it possible to diagnose the more sensitive serotypes. Obando et al. [30] found evidence of pneumococcal infection in 84% of culture-negative pleural fluid samples on the basis of ply or wzg gene detection by PCR; these cases were significantly more likely to have received antimicrobial drug therapy before sampling. In our study 48% of IPD cases were diagnosed only by real-time PCR, which served to increase the etiologic diagnosis and serotype identification. Serotype 19A was associated with age <2 years, serotype 3 with the 2-4 years age group and serotype 1 with the 5-17 years age group. Other studies have found similar results, with 100% of cases due to serotype 19A in children aged <2 years, 63.6% of cases due to serotype 3 in the 2-4 years age group and 80% of cases due to serotype 1 in children aged >4 years [21].
Compared with the pre-PCV13 era, [23] in our region the most frequent serotype was 19A (24.6%) in children aged <2 years and serotype 1 (46.8%) in children aged 2-4 years, while serotype 3, now the most frequent, represented only 2.8% and 4.2% of cases in these age groups, respectively. This may be explained by the replacement of serotypes.
PCV13 serotypes were associated with overall pneumonia and complicated pneumonia in children aged ⩽17 years, mainly serotype 19A in children aged <2 years, serotype 3 in those aged <5 years and serotype 1 in the 5-17 years age group. A Polish study conducted between 2011 and 2013 in all age groups, found that PCV13 serotypes caused 78.4% of cases of pneumonia, and serotype 3 and serotype 1 were associated with pneumonia, as in our study [31].
In our study, non-PCV13 serotypes were associated with occult bacteraemia, especially 10A and 12F, and with meningitis in the 2-4 years age group. Similarly, other authors have reported that 90% of meningitis cases and 64% of occult bacteraemia cases during 2011-2012 were produced by non-PCV13 serotypes [32].
The percentage of strains not susceptible to penicillin (36.1%) or to cefotaxime (16.4%) was higher than that described by other authors [23,25]. Serotypes 19A and 14 were associated with strains not susceptible to penicillin or cefotaxime, whereas serotype 23B and 24F were not susceptible to penicillin. Other authors have also described non-susceptibility to penicillin in serotypes 19A and 14 [28,31], serotype 24F [14,28] and serotype 23B [33]. Skoczyńska et al. [31] found that non-susceptibility to penicillin was also associated with vaccine-serotypes 19F, 6B and 9V, which in our study could not be demonstrated, perhaps due to the small number of strains studied. They found an association between non-susceptibility to cefotaxime in serotype 19A and 14, as in our study. The non-susceptibility to antibiotics of non-PCV13 serotypes is concerning and should be monitored.
Among the strengths of the study is the high percentage of cases serotyped, since serotyping was not possible in only 3.8% of cases. The same diagnostic techniques (culture and real-time PCR) were used uniformly in all cases, improving the diagnostic sensitivity and the range of serotypes identified, which permitted a more precise estimate of the burden of IPD.
There were some limitations. First, the population studied corresponds to three paediatric hospitals, although these hospitals serve 32% of the population aged ⩽17 years requiring hospitalisation in Catalonia. In the subgroup analyses, the small number of cases observed in some situations may explain why the confidence intervals of the measure of association were very wide. Finally, antibiotic sensitivity could only be studied in 48% of cases serotyped, since a large part of the cases were diagnosed by real-time PCR.

Conclusions
PCV13 serotypes were the most frequently found IPD serotypes in children aged ⩽17 years, especially serotype 19A in children aged <2 years, serotype 3 in the 2-4 years age group and serotype 1 in the 5-17 years age group.
Non-PCV13 serotypes were the main cause of meningitis, occult bacteraemia and other clinical presentations, while PCV13 serotypes were mainly responsible for pneumonia.
PCV13 and non-PCV13 cases presented a high frequency of non-susceptibility to penicillin. Non-susceptibility to both penicillin and cefotaxime was associated with serotypes 19A and 14, and serotype 19A was associated with resistance to penicillin.
The non-susceptibility to antibiotics of non-PCV13 serotypes is concerning and should be monitored to apply the appropriate disease prevention strategies.
Financial support. This work was supported by the Plan Nacional I + D + I, Instituto de Salud Carlos III -Subdirección General de Evaluación y Fomento de la Investigación Sanitaria (Projects: PI11/02081 and PI11/02345) and cofunded by the Fondo Europeo de Desarrollo Regional (FEDER) (funding for the serotyping study and for participating in congresses) and the Agència de Gestió d' Ajuts Universitaris i Recerca (AGAUR: grant number 2017 SGR 1342) (funding of the English manuscript revision).
Conflict of interest. JJ Garcia-Garcia has received honoraria for speaking at symposia (Pfizer and GSK), and financial support for attending symposia (Pfizer). None were concerned with the present study.
C Muñoz-Almagro reports travel grants from Pfizer, research grants from BioMerieux, Stat DX and Instituto de Salud Carlos III, personal fees from GSK as consultor to an advisory board and honoraria for speaking at symposia from Roche and Biomerieux. None were concerned with the present study.
For the remaining authors, no competing interests were declared.
Ethical standards. Informed consent was obtained from all individual participants included in the study. Written informed consent was obtained from a parent or guardian for participants under 16 years of age. The study complies with the principles of the Declaration of Helsinki and the legal structure with respect to international human rights and biomedicine and protection of personal data laws. The Ethics Committee of Hospital Sant Joan de Déu approved the study.
Data. No additional data are available.