Pneumococci remain the main cause of complicated pediatric pneumonia in the post-pandemic era despite extensive pneumococcal vaccine use

Abstract

Nucleic acid amplification tests (NAATs) greatly enhance the capacity to identify the etiology of pediatric complicated pneumonia. However, the use of pneumococcal conjugate vaccines could reduce the importance of Streptococcus pneumoniae in pediatric complicated pneumonia with the potential emergence of other bacterial agents. Using an expanded NAAT in culture negative pleural fluid or empyema samples collected in 2010–2024 (n = 554) in Portugal, we show that S. pneumoniae remains the most frequent agent despite decades of pneumococcal conjugate vaccine use and the COVID-19 pandemic. A rebound in pediatric complicated pneumonia occurred post-pandemic, including a rise in cases by Streptococcus pyogenes and Haemophilus influenzae. Empiric therapy of pediatric complicated pneumonia should still consider S. pneumoniae as the most likely cause, even in countries where the pneumococcal conjugate vaccine is in the national immunization program with a high uptake.

Introduction

The advent of pneumococcal conjugate vaccines (PCVs), although including only a fraction of the > 100 serotypes known, raised the possibility of effectively preventing a significant part of pneumonias, including complicated cases. The initial 7-valent PCV was mostly targeting serotypes identified in blood, irrespective of disease presentation, but the 13-valent PCV (PCV13) already included the most frequent serotypes identified in pediatric complicated pneumonias (pneumonias occurring with parapneumonic effusion or empyema, PCP).

The etiologic diagnosis of PCP remains challenging, with culture of empyema, pleural fluid or blood being frequently negative and highlighting the importance of nucleic acid amplification tests (NAATs) [1,2,3]. Several bacteria are reported frequently as causing PCP, including: Streptococcus pneumoniae, Staphylococcus aureus (methicillin resistant and susceptible), Streptococcus pyogenes (or Lancefield group A streptococcus – GAS), Haemophilus influenzae, Mycoplasma pneumoniae, Mycobacterium tuberculosis and other species of the Streptococcus genus [4, 5]. Notwithstanding the dominance of S. pneumoniae as a causative agent of PCP in Portugal, the number of samples testing negative for S. pneumoniae increased from 43/135 in 2010–2015 to 76/163 in 2016–2019 (p = 0.013) [1, 2], raising the possibility of increases of other bacterial species as causes of PCP. This, together with a high PCV13 uptake in Portugal (> 95%), led us to expand our NAAT to identify other possible bacterial causes of these infections.

Methods

We included culture-negative empyema and pleural fluid samples from children and adolescents (pediatric patients) (< 18 years) recovered in 52 hospital laboratories and pediatric departments located throughout Portugal. This is a service offered by the central laboratory to participating hospitals and samples are collected prospectively.

For all the samples received, we performed four real-time PCR (rPCR) multiplex reactions targeting different bacterial species (Table 1) and simultaneously detecting human DNA [1].

Table 1 Primers and probes used in the nucleic acid amplification test

Results

Between July 2010 and June 2024, we analyzed 544 samples. Up to 2020, samples were tested prospectively solely for S. pneumoniae and the results were reported previously [1, 2]. These samples were in storage and were re-tested retrospectively with the reactions targeting the additional six pathogens. From 2021 onwards samples were tested prospectively. The age of the patients ranged from 2 months to 17 years (data missing for 10 patients). 197 samples were negative for all bacteria tested (36.2%). In eight samples, we were not able to perform all PCR reactions due to limited sample availability, with the majority being positive for S. pneumoniae (6/8). The number of samples analyzed each year and positive identifications in the NAAT are presented in additional Table 1 [see Additional file 1].

The distribution of bacterial species identified in our samples is presented in the Fig. 1. Most samples were positive for S. pneumoniae (n = 289, 53.1%), followed by 48 samples positive for S. pyogenes (8.8%). H. influenzae and S. aureus were found in smaller numbers (n = 12, 2.2% and n = 6, 1.1% each, respectively), three samples were positive for M. tuberculosis (0.6%) and two for M. pneumoniae (0.4%). S. agalactiae was not detected in any sample. In additional Table 2 we present the distribution into age groups of the pathogens detected in the samples. S. pyogenes was more frequent among samples collected in children up to 23-months of age, whereas S. pneumoniae was detected more frequently in patients 2–17 years of age (Fisher exact test, p < 0.001).

Fig. 1

Number of samples from pediatric patients (< 18 years) indicating any identified pathogens (Portugal, 2009/10 – 2023/24)

In 13 samples, two bacterial species were detected: S. pneumoniae and S. pyogenes (n = 7), S. pneumoniae and H. influenzae (n = 4), S. pneumoniae and S. aureus (n = 1), and S. pneumoniae and M. tuberculosis (n = 1).

Discussion

Despite the high uptake of PCV13 in Portugal, most PCPs in our study were still caused by S. pneumoniae. S. pyogenes ranked second, with approximately half of the positive cases (26/48) being detected in the two last epidemiological years (Fig. 1), concurrent with an outbreak of pediatric invasive GAS infections in Portugal [6]. However, PCPs due to S. pyogenes were also detected in almost all epidemiological years, reinforcing the importance of this pathogen even outside outbreak contexts. In contrast, in a single center in Australia, the increase of S. pyogenes in PCP in 2022–2023 was much more marked and contributed significantly to a rise of PCP incidence [7].

Another difference in our study from previous reports is the proportion of infections caused by S. aureus, since it is frequently described as a major agent, often reported more frequently than H. influenzae [5, 8], in contrast to the situation in Portugal where there were twice as many infections by H. influenzae (n = 12, 2.2%). As with S. pyogenes infections, approximately half of the H. influenzae cases were reported in the last two epidemiological years (4 cases in 2022–2023 and 3 cases in 2023–2024). In these two seasons, an increase in the number of submitted samples was seen, after a marked decline in 2020–2021. We do not believe this is due to any change in our surveillance but to be consistent with the resurgence of respiratory diseases in the post-pandemic years noted in several countries [9]. From 2022–2023 to 2023–2024, a decrease in samples in which no bacterial agent was detected and an increase in the proportion of samples positive for S. pneumoniae were also found (Fig. 1). Similarly, the single center study from Australia found that the post-pandemic increase in incidence of PCP was accompanied by a decrease of the number of samples in which no pathogen was identified [7]. As in other countries, in Portugal there was a rebound in overall invasive pneumococcal disease in children in 2022–2023 [10], with this trend persisting through at least 2023–2024. The increase in PCP samples seen in 2022–2024 was driven mostly by serotype 3, which was also the dominant serotype pre-COVID-19 pandemic [1, 2, 10].

During the autumn of 2023 there was an increase of M. pneumoniae infections in several countries of the northern hemisphere. However, the number of infections reported in two Portuguese hospitals between April 2022 and September 2023 remained low [11] and in our study no increase of PCP cases due to M. pneumoniae was seen.

We found cases where the DNA of two bacterial species was detected, suggesting possible co-infections in these cases, as already reported [8]. Since we have no data regarding the immune status or other comorbidities of the patients, we cannot say if these occurred in a particularly susceptible population.

Our work has limitations. Since we did not collect any clinical information, we cannot explore potential differences in severity between cases caused by different species nor if cases where no pathogen was identified correspond to patients under antimicrobial treatment for longer before sample collection. Additionally, although the study involved both pediatric and microbiology departments, it was not designed for the estimation of PCP incidence, which may have changed during the study period.

More than two decades after extensive use of PCVs and after the major perturbation induced by the COVID-19 pandemic, PCPs are still caused mainly by S. pneumoniae. Diagnostic tests and empiric therapy of PCPs should continue to focus on S. pneumoniae, even in countries with a high PCV uptake. However, the etiology of more than a third of cases in our study remained unknown and other microbial agents can emerge as important causes of infection, reinforcing the need to expand NAATs to identify the etiology of these complicated infections.