Invasive pneumococcal diseases in Chinese children: a multicentre hospital-based active surveillance from 2019 to 2021

ABSTRACT

This study aimed to provide data for the clinical features of invasive pneumococcal disease (IPD) and the molecular characteristics of Streptococcus pneumoniae isolates from paediatric patients in China. We conducted a multi-centre prospective study for IPD in 19 hospitals across China from January 2019 to December 2021. Data of demographic characteristics, risk factors for IPD, death, and disability was collected and analysed. Serotypes, antibiotic susceptibility, and multi-locus sequence typing (MLST) of pneumococcal isolates were also detected. A total of 478 IPD cases and 355 pneumococcal isolates were enrolled. Among the patients, 260 were male, and the median age was 35 months (interquartile range, 12–46 months). Septicaemia (37.7%), meningitis (32.4%), and pneumonia (27.8%) were common disease types, and 46 (9.6%) patients died from IPD. Thirty-four serotypes were detected, 19F (24.2%), 14 (17.7%), 23F (14.9%), 6B (10.4%) and 19A (9.6%) were common serotypes. Pneumococcal isolates were highly resistant to macrolides (98.3%), tetracycline (94.1%), and trimethoprim/sulfamethoxazole (70.7%). Non-sensitive rates of penicillin were 6.2% and 83.3% in non-meningitis and meningitis isolates. 19F-ST271, 19A-ST320 and 14-ST876 showed high resistance to antibiotics. This multi-centre study reports the clinical features of IPD and demonstrates serotype distribution and antibiotic resistance of pneumococcal isolates in Chinese children. There exists the potential to reduce IPD by improved uptake of pneumococcal vaccination, and continued surveillance is warranted.

KEYWORDS:

• Invasive pneumococcal disease
• serotype
• Chinese children
• multicentre
• hospital-based

Introduction

Streptococcus pneumoniae (SP) is one of the leading infectious causes of invasive infection worldwide, especially in children and the elderly. Nearly 1 million children die every year from pneumococcal diseases, and the highest mortality rates were in Africa and Asia [1]. The World Health Organization estimates that China accounts for 12% of SP infections and 3.6% of SP deaths worldwide in children under 5 years of age [1]. There is limited information about invasive pneumococcal disease (IPD) available in China to date, probably due to inadequate laboratory facilities, lack of specimen submission, difficulties in specimen culturing, and irrational use of antibiotics [2,3]. Pneumococcal conjugate vaccines (PCVs) have decreased the total number of IPD cases caused by serotypes covered by vaccines in many countries and regions, however, previous studies had reported the vaccination rates are low due to the high cost of such in China [4–6], estimated 2017 national three-dose coverage in private market was 1.3% for PCV among children aged 1−59 months [7].

Of the > 90 different pneumococcal serotypes that have been identified based on antigenic differences in their capsular polysaccharides, limited serogroups account for most paediatric IPDs worldwide, with serogroups 1, 3, 6, 8, 14, 19, and 23 being the most common ones [8–10], meanwhile, serotype shifts have been observed in countries and regions where PCVs are widely used [11–13]. By retrieving the laboratory information system in 18 hospitals from 2012 to 2017, a total number of 1138 IPD were retrospectively enrolled, with meningitis, pneumonia with bacteraemia, and bacteraemia without focus accounting for 89.4% of all cases [14]. Serogroups accounting for IPDs in Chinese children were 19F, 14, 19A, 6B, and 23F, reported in 2010 [15], while Zhang et al.’s study [16] identified 23F, 19A, 19F, 3, and 14 were common serotypes that caused IPDs in both Chinese children and adults from 2010 to 2015. Meanwhile, increasing resistance of pneumococci to conventional antibiotics has made the treatment of infection more limited than ever. The data from these studies have mainly focused on pneumococcal isolates [15,16], or some single-centre studies investigated IPD in certain hospitals or areas, which do not reflect current trends or offer regional representation of China.

In this study, we enrolled IPD cases and pneumococcal isolates from paediatric patients in China, with the purpose of providing data on the clinical features of IPD, and current trends in the serotype and antimicrobial resistance patterns of pneumococcal strains to gain more information to revise treatment strategies and establish preventive policy guidelines.

Methods

Study design and participants

This was a hospital-based and cultured-confirmed surveillance study comprising of children aged ≤18 years of age with proven IPD from a network of 19 hospitals from 19 cities in 5 geographical regions (north, east, south central, northwest and northeast) across mainland China from January 2019 to December 2021. A workshop for participating clinicians and microbiologists was held at Beijing Children’s Hospital at the beginning of the study to standardize procedures. Forums about the project were held every 6 months for quality control of participant enrolment and to share updates about project status. The study was approved by the ethics committee of each hospital.

Case definition and clinical data

IPD is defined as infection with S. pneumoniae identified in blood, cerebrospinal fluid and/or any other sterile body fluid specimens, and a clinical diagnosis suggesting invasive disease. Clinical data, including demographic details, vaccination history, the presence of risk factors, and disease outcomes were reviewed and collected using a systematic data abstraction form. Severe pneumonia was defined according to previous literature [17].

Serotyping and multi-locus sequence typing (MLST)

All pneumococcal isolates were collected and transferred to Beijing Children’s Hospital for typing and antimicrobial susceptibility test. Isolates were serotyped with the Quellung reaction using antisera obtained from Staten’s Serum Institute (Copenhagen, Denmark) as described previously [18]. MLST was determined by sequencing and comparing the polymerase chain reaction products of housekeeping genes aroE, gdh, gki, recP, spi, xpt, and ddl [19]. The BURST analysis (https:// pubmlst.org/spneumoniae/) was used to investigate relationships among the isolates and to assign strains to a clonal complex (CC) based on the stringent group definition of six of the seven shared alleles [20]. The goeBURST software (http://www.phyloviz.net/goeburst/) was used to investigate relationships among the isolates and to assign strains to a clonal complex (CC) based on the stringent group definition of six of the seven shared alleles.

Antimicrobial susceptibility testing

Antimicrobial susceptibility was tested by establishing the minimum inhibitory concentration (MIC) value using an E-test (Liofilchem, Italy) for penicillin, ceftriaxone, clindamycin, erythromycin, tetracycline, levofloxacin, trimethoprim/sulfamethoxazole, and vancomycin. The interpretation of results was based on cutoff values defined by the Clinical and Laboratory Standards Institute (CLSI), which classified the isolates as susceptible, intermediate, or resistant [21]. S. pneumoniae ATCC49619 (American Type Culture Collection, Manassas, VA, USA) was used as a standard reference strain to quality check media and susceptibility.

Statistical analysis

Descriptive statistics including frequencies and percentages for categorical variables, were calculated to describe the characteristics of the cohort. Categorical variables were analysed using a chi-squared or Fisher’s exact test. Continuous variables that did not follow a normal distribution were described with median and interquartile range (IQR) values and compared using the Mann–Whitney U test. Differences were considered statistically significant when P < 0.05. All analyses were performed using SPSS (version 25.0; IBM Corporation, Armonk, NY, USA).

Results

A total of 478 cases of IPD were enrolled during the study period, stratified as 266, 117, 95 in 2019, 2020, and 2021, respectively (Supplementary figure 1A). The clinical data of all cases were available and collected; overall, 355 isolates were recovered among the cases (Supplementary table 1). There was significant variation in IPD cases overall by month, and the peak of onset was from October to January, accounting for 49% (234/478) of all cases. (Supplementary figure 1B).

Demographic information

Among the patients enrolled, 260 (54.4%) were male and 218 were female. The median age was 35 months (range, 1 month-15 years; IQR, 12–46 months). 40. 4% (193/478) of patients were under 2 years old and 46.7% (223/478) were aged 2–5 years. The overall frequency of age distribution is illustrated in Table 1. Nine (1.9%) patients had received pneumococcal vaccinations before infection; 6 of them were vaccinated with one-dose 23-valent pneumococcal capsular polysaccharide vaccine (PPSV23), 2 were vaccinated with three-dose PCV13, and 1 was vaccinated with one-dose PCV7. Four hundred nine patients had not received a PCV or PPSV, and the vaccination history was unknown in 60 patients.

Table 1. Demographic information and risk factors of IPD patients.

Characteristic	Total n (%) n = 478	Septicaemia n (%) n = 180	Meningitis n (%) n = 155	Pneumonia n (%) n = 133	Others^a n (%) n = 10	P-value^b
Male sex	260 (54.4)	97 (53.9)	82 (52.9)	77 (57.9)	4 (40)	0.671
Age, median (IQR, month)	35 (12,46)	37 (15,43)	31 (10,53)	27 (13,43)	27 (9,41)	0.099
Age group (months)
<6	24 (5)	3 (1.7)	17 (11)	4 (3)	0 (0)	0.000
6–23	169 (35.4)	76 (42.2)	53 (34.2)	56 (42.1)	5 (50)
24–59	223 (46.7)	87 (48.3)	53 (34.2)	58 (43.6)	4 (40)	0.078
≥60	62 (13)	14 (7.8)	32 (20.6)	15 (11.3)	1 (10)	0.002
Risk factors	105 (22)	29 (16.1)	52 (33.5)	24 (18)	0 (0)	0.000
Immunocompromised	45 (9.4)	23 (12.8)	9 (5.8)	13 (9.8)	0 (0)	0.097
Primary immunodeficiency	8 (1.7)	2 (1.1)	2 (1.3)	4 (3)	0 (0)	0.445
Neoplastic diseases	30 (6.3)	17 (9.4)	2 (1.3)	11 (8.3)	0 (0)	0.002
Nephrotic syndrome^c	5 (1)	2 (1.1)	1 (0.6)	2 (1.5)	0 (0)	–
Sarcoidosis	1 (0.2)	1 (0.6)	0 (0)	0 (0)	0 (0)	–
Systemic lupus erythematosus	1 (0.2)	0 (0)	1 (0.6)	0 (0)	0 (0)	–
Congenital heart disease	12 (2.5)	1 (0.6)	5 (3.2)	6 (4.5)	0 (0)	–
Chronic respiratory disease^d	6 (1.3)	0 (0)	1 (0.6)	5 (3.8)	0 (0)	0.000
Trauma or surgery	20 (4.2)	5 (2.8)	15 (9.7)	0 (0)	0 (0)	0.000
Cochlear implant	4 (0.8)	0 (0)	4 (2.6)	0 (0)	0 (0)	–
Cerebrospinal fluid leakage	18 (3.8)	0 (0)	18 (11.6)	0 (0)	0 (0)	0.000
Source of samples
Blood	344 (72)	180 (100)	43 (27.7)	117 (88)	4 (40)	–
Cerebrospinal fluid	46 (9.6)	0 (0)	46 (29.7)	0 (0)	0 (0)	–
Pleural effusion	15 (3.1)	0 (0)	0 (0)	15 (11.3)	0 (0)	–
Blood + cerebrospinal fluid	66 (13.8)	0 (0)	66 (42.6)	0 (0)	0 (0)	–
Blood + pleural Effusion	1 (0.2)	0 (0)	0 (0)	1 (0.8)	0 (0)	–
Other specimens^e	6 (1.3)	0 (0)	0 (0)	0 (0)	6 (60)	–
Isolates collected	355 (74.3)	141 (77.9)	114 (73.5)	93 (70.5)	7 (70)	–
Clinical course
Sepsis	179 (37.4)	51 (28.3)	77 (49.7)	50 (37.6)	1 (10)	0.000
Septic shock	29 (6.1)	5 (2.8)	18 (11.6)	6 (4.5)	1 (10)	0.002
ICU admission	187 (39.1)	30 (16.7)	96 (61.9)	60 (45.1)	1 (10)	0.000
Mechanical ventilation	53 (11.1)	5 (2.8)	31 (20)	17 (12.8)	0 (0)	0.000
Hospital-stay, median (IQR, days)	14 (8,23)	9 (7,14)	22 (14,31)	13 (9,17)	26 (18,37)	0.000
Died (case fatality)	46 (9.6)	4 (2.2)	30 (19.4)	12 (9)	0 (0)	0.000

Abbreviations: IQR, interquartile range; ICU, intensive care unit.

^a Include arthritis and/or osteomyelitis (n = 8), perforative peritonitis with acute appendicitis (n = 1), and infective endocarditis (n = 1).

^b Comparation between septicaemia, meningitis, and pneumonia.

^c Include lymphomas, leukemias, solid tumour, aplastic anaemia, Langerhans cell histiocytosis, hemophagocytic syndrome, and other diseases of the blood-forming organs.

^d Include congenital laryngeal chondromalacia, bronchitis obliterans, and asthma.

^e Include joint effusion (n = 3), abdominal dropsy (n = 1), blood + hydropericardium (n = 1), and blood + joint effusion (n = 1).

Clinical syndrome

Septicaemia was the most frequent presentation, reported in 37.7% (180/478) of the cases, followed by meningitis and pneumonia, which were found in 32.4% (155/478) and 27.8% (133/478) of all cases. Other presentations were shown in Table 1. Sepsis was found in 179 (37.4%) patients, with a higher proportion present among meningitis cases (P<0.05). Severe pneumonia accounted for 46.6% (62/133) of all pneumonia cases. The age distribution was different among disease types, meningitis was more common in patients aged <6 months and >5 years of age compared to septicaemia or pneumonia. (P<0.05) (Table 1).

Risk factors and underlying diseases

The percentage of individuals who had risk factors or underlying diseases for acquiring IPDs was 22% (105/478) (Table 1). The most common condition was an immunocompromised state (45/105, 42.9%), with a high proportion of neoplastic diseases (30/45, 66.7%). Patients with risk factors or underlying clinical conditions were older than those without risk factors; the median age at IPD diagnosis among children with clinical conditions was 46 months (IQR, 27–88 months) compared to 30 months (IQR, 11–43 months) in children without any clinical conditions (P<0.05).

Clinical outcomes

Of all patients, 187 (39.1%) patients experienced ICU admission and 53 (11.1%) underwent mechanical ventilation, with a greater proportion of these events occurring among meningitis cases (P<0.05). A total of 48 patients died and 46 of them died from IPD, with a case fatality rate of 9.6% (46/478). Patients with or without comorbidities had the same case fatality rate (8.1% vs. 10.1%). Based on the clinical syndrome, the case fatality rate was noted to be higher in meningitis cases (19.4%, 30/155) compared to pneumonia (9%, 12/133) and septicaemia cases (2.2%, 4/180) (P<0.05).

Serotype distribution

A total of 34 different serotypes were detected among 355 pneumococcal isolates in the study. The top 5 detected serotypes were 19F (n = 86, 24.2%), 14 (n = 63,17.7%), 23F (n = 53,14.9%), 6B (n = 37,10.4%), 19A (n = 34, 9.6%), comprising 76.9% (273/355) of the total isolates, followed by 15A (n = 10), 9 V (n = 8), 6A (n = 8), 3 (n = 7), 23B (n = 5), 23A (n = 4), 12F (n = 4), and 15C (n = 4) (Figure 1). Supplementary table 2 illustrates the overall frequency of individual serotype distribution. The coverage rates for PCV13 and PPSV23 were 85.1% (302/355) and 86.5% (307/355), respectively. The coverage rate for PCV13 was higher in patients < 5 years old (261/310, 84.2%) than in those >5 years old (31/45, 68.9%) (P = 0.012). Non-PCV13 serotypes included 15A (n = 10); 23B (n = 5); 23A (n = 4); 12F (n = 4); 15C (n = 4); 10A (n = 3); 15B (n = 3); 7C (n = 2); 13 (n = 2); 42 (n = 2); 24F (n = 2); and 9A, 6C, 8, 11A, 21, 27, 16F, 17F, 18F, 19B, 24A, and 9L (each n = 1).

Figure 1. Serotype distribution of 355 pneumococcal isolates. *6A was not covered by PPSV23. **Others were serotypes with one or two isolates, including 7C (n = 2), 13 (n = 2), 42 (n = 2), 24F (n = 2), 6C, 9A, 21, 27, 16F, 18F, 19B, 24A, 9L, each 1.

Serotype distribution for PCV13 did not differ substantially when analysed by age. The five serotypes (19F, 14, 23F, 6B, and 19A) that were most often responsible for causing septicaemia were the same types most often responsible for causing meningitis and pneumonia, just in a different order according to the proportion of cases they caused, and these serotypes accounted for about 80.1% (113/141), 71.9% (82/114), and 78.5% (73/93) of septicaemia, meningitis, and pneumonia cases. Isolates of serotype 23F was more frequently caused septicaemia (P = 0.001, χ²= 13.636), while serotype 14 was more frequently caused pneumonia (P = 0.031, χ2 = 6.977) (Supplementary figure 2). Serotype distribution did not vary much by geographical region. 19F, 14, 23F, 19A, and 6B accounted for 77.5%, 72%, 77.6%, 68.5%, 78.1%, and 87.1% of cases in the north, middle, east, northwest, south and northeast regions of China, respectively. Serotype distribution of strains from vaccinated cases were shown in supplementary table 3.

MLST

Among the 355 isolates studied, 109 ST types were identified by MLST analysis. Of these, ST271 (52/355, 14.6%) was the most predominated, followed by ST876 (50/355, 14.1%), ST320 (25/355, 7%), ST81 (16/355, 4.5%) and ST13646 (15/355, 4.2%). A phylogenetic tree was constructed based on single-locus variants (SLV) in seven housekeeping genes, and a total of 19 clonal complexes (CC) and 30 singletons were identified. The most common clonal complex was CC271 (97/355, 27.3%,), followed by CC876 (15.8%), CC6325 (5.9%), CC81 (4.8%), and CC90 (4.2%) (supplementary table 4).

The distribution of STs among the different serotypes is shown in Figure 2. Serotypes showed diverse ST distribution. Serotype19F was dominated by ST271 (51/86, 59.3%), serotype 14 was dominated by ST876 (47/63, 74.6%), serotypes 19A by ST320 (24/34, 70.6%), serotype 23F was dominated by ST13646 (15/53, 28.3%) and ST81 (14/53, 26.4%), serotype 6B was dominated by ST90 (11/37, 29.7%) and ST902 (7/37, 18.9%). In addition, serotype 15A was dominated by ST11972 (6/10, 60%), serotype 15C was dominated by ST8589 (3/4, 75%), serotype 3 was dominated by ST4655 (3/7, 42.9%), and serotype 12F was dominated by ST6945 (3/4, 75.0%).

Figure 2. Population snapshot of S. Pneumoniae strains through goeBURST analysis.

Note: 13 isolates of new STs were not included in the figure.

Antimicrobial susceptibility testing

According to the revised CLSI breakpoints, the prevalence rates of non-sensitivity to penicillin were 6.2% and 83.3% in the non-meningitis and meningitis isolates, respectively. The proportions of isolates that were non-sensitive to ceftriaxone were 7.5% in the non-meningitis isolates and 37.7% in the meningitis isolates, respectively. There seemed increases from 2019 to 2021 in the proportion of isolates that were resistant to ceftriaxone, from 8.9% to 15.5% according to meningitis criteria, and from 0% to 5.2% according to non-meningitis criteria. Up to 97.7% and 98.3% of isolates were not sensitive to clindamycin and erythromycin. The majority of isolates were resistant to tetracycline (94.1%) and to trimethoprim/sulfamethoxazole (70.7%). All of the isolates were susceptible to vancomycin and levofloxacin. The sensitivities and MICs for different antimicrobials are shown in Table 2. Antimicrobial susceptibility testing of all isolates for penicillin and ceftriaxone using both the meningitis and non-meningitis CLSI susceptibility cutoffs. Penicillin resistant rate is much higher while using cutoffs for meningitis. Compared to non-PCV-covered serotypes, the PCV13 covered serotypes 19F, 14, and 19A, showed high resistance to penicillin and ceftriaxone (shown in Table 3), especially in 19F-ST271, 14-ST876 and 19A-ST320 (shown in supplementary table 5). Antibiotic resistance rates to penicillin, ceftriaxone, clindamycin, erythromycin, tetracycline and to co-trimoxazole were similar across different regions (shown in supplementary table 6).

Table 2. Antimicrobial susceptibility pattern of the 355 Streptococcus pneumoniae strains.

Antimicrobials	No. of isolates (n)	Susceptibility				MIC (mg/ml)
		Susceptible (n)	Intermediate (n)	Resistant (n)	Non-sensitive rate (%)	50%	90%	Range
Penicillin
Meningitis	114	19	–	95	83.3%	1.5	3	<0.016-4
Non-meningitis	241	226	14	1	6.2%	1	2	<0.016-6
Ceftriaxone
Meningitis	114	71	28	15	37.7%	0.5	1	0.008-3
Non-meningitis	241	223	15	3	7.5%	0.5	1	0.006-3
Clindamycin	355	8	1	346	97.7%	>256	>256	0.125 to >256
Erythromycin	355	6	1	348	98.3%	>256	>256	0.125 to >256
Tetracycline	355	21	9	325	94.1%	12	24	0.094 to >32
Levofloxacin	355	355	0	0	0	0.5	0.75	0.38-2
Trimethoprim/ sulfamethoxazole	355	104	32	219	70.7%	6	>32	0.094 to >32
Vancomycin	355	355	0	0	0	0.38	0.5	0.19-0.75

MIC, minimum inhibitory concentration.

Table 3. Antimicrobial resistance pattern of the main serogroup of Streptococcus pneumoniae strains.^a

Serotype	No. of isolates	MIC50 (ng/mL)	MIC90 (ng/mL)	MIC range (ng/mL)	Penicillin (n)			MIC50 (ng/mL)	MIC90 (ng/mL)	MIC range (ng/mL)	Ceftriaxone (n)
					Meningitis	Non-meningitis					Meningitis		Non-meningitis
					R	I	R				I	R	I	R
Included in PCV13 serotypes	302	1.5	3	<0.016-4	271	31	0	0.5	1	0.006-3	89	29	26	3
19F	86	2	3	0.032-4	85	19	0	1	2	0.023-3	36	28	25	3
14	63	1.5	2	<0.016-3	61	2	0	0.5	0.75	0.012-1	24	0	0	0
23F	53	1	3	<0.016-3	52	6	0	0.38	1	0.032-1	15	0	0	0
6B	37	0.75	1.5	0.023-3	31	1	0	0.38	0.5	0.008-1	3	0	0	0
19A	34	1.5	2	0.032-4	33	3	0	0.5	1	0.016-1.5	10	1	1	0
9V	8	0.023	0.032	<0.016-0.5	1	0	0	0.008	0.047	0.008-0.5	0	0	0	0
6A	8	1	1.5	0.032-1.5	7	0	0	0.25	0.75	0.064-0.75	1	0	0	0
3	7	0.032	0.125	<0.016-0.125	1	0	0	0.016	0.19	0.008-0.19	0	0	0	0
Others	6	0.023	0.047	<0.016-0.047	0	0	0	0.012	0.047	0.006-0.047	0	0	0	0
Non-PCV13 serotypes	53	0.38	1.5	<0.016-6	31	1	1	0.125	0.75	0.006-3	3	4	3	1
15A	10	1	1	0.023-2	8	0	0	0.38	1.5	0.006-2	0	2	2	0
23B	5	1	1.5	0.023-1.5	4	0	0	0.5	0.5	0.012-0.5	0	0	0	0
12F	4	0.023	0.125	<0.016-0.125	1	0	0	0.016	0.125	0.016-0.125	0	0	0	0
23A	4	0.75	1.5	0.38-1.5	4	0	0	0.125	0.5	0.125-0.5	0	0	0	0
15C	4	0.016	0.023	<0.016-0.023	0	0	0	0.012	0.016	0.008-0.016	0	0	0	0
Others	26	0.064	2	<0.016-6	14	1	1	0.032	0.75	0.006-3	3	2	1	1

^a Antimicrobial susceptibility testing of all isolates for penicillin and ceftriaxone using both the meningitis and non-meningitis CLSI susceptibility cutoffs.

MIC, minimum inhibitory concentration.

Discussion

A marked decrease in the number of IPD cases (from 266 to 95 per year) was observed from 2019 to 2021, which we believed was attributable to increased mask-wearing and social distancing during the coronavirus disease 2019 (COVID-19) pandemic. A previous study based on surveillance data from 26 countries confirmed the significant and sustained reductions of invasive bacterial infection, which also appeared to coincide with the stringency of COVID-19 restrictive measures and with changes in the movement of people [22]. Our study is in accordance with previous findings [6,15]. that, in China, IPD was more common among children <5 years old, a majority of patients (82%) were in the 6-59-month age group, which is similar to results from other low-income countries [23]. and among children in the United States, Brazil, and Germany [24–26].

Although disease types found in this study were parallels between those in our previous study [6,14] and other studies abroad [23, 24], there were some differences in proportions and risk factors of each disease. Septicaemia was the most frequent presentation, especially in patients with neoplastic diseases and other immunocompromised states. The mortality rates of children with IPD were approximately 2.5%–23.5% in previous studies [6,23,27], depending on the disease type. Case fatality rates in this study were 9.6% and patients suffering from meningitis had higher rates of ICU admission and mortality. Meningitis was more common in patients aged <6 months and >5 years compared to other disease types; this may be because trauma and cerebrospinal fluid leakage were risk factors among patients >5 years of age in this study. According to the revised WHO classification and treatment of childhood pneumonia at health facilities, severe pneumonia was detected to be as high as 46.6% of all pneumonia cases. These cases had pleural effusion, atelectasis, pneumothorax purulent and many other serious manifestations. A previous study assumed that 1.8 million pneumonia deaths occurred in children aged 1–59 months in 2000, mainly in developing regions, and our study confirmed the disease burden of pneumococcal pneumonia in China. Meanwhile, although case numbers of arthritis and/or osteomyelitis, abdominal infection, and infective endocarditis were limited, further surveillance is also needed for these rare clinical types of IPD.

PCV13 and PPSV23 have been available in China, but in our study, only 9 (1.9%) patients had received pneumococcal vaccines. This may be because PCVs are not part of the routine childhood immunization programme and are currently available only for infants whose parents actively pay for them. It was previously reported that serotypes included in conjugated vaccines were more prevalent in children before the use of pneumococcal vaccines, our results confirmed that most IPD in children in China continues to be caused by serotypes contained in available PCV13 vaccines, reinforcing the urgency of administering the pneumococcal vaccine to Chinese children.

As PCVs have substantially reduced the incidence of IPD caused by vaccine serotypes in many countries, serotype replacement is a concern in the use of vaccines that target a proportion of diverse antigenic types. Variation in serotype distribution is based on geographical differences; for example, 19A, 12A, 12F, 24F, 35B, 33F, 22F, 15A, and 15B became common IPD serotypes after the 13-valent PCV was introduced in Belgium, South Asia, and Japan [28–30], while, in our study, the most frequent serotypes were 19F, 14, 23F, 6B, and 19A, similar to our previous findings among Chinese children [6,15] during the past 10 years and Zhang et al.’s study [16] of pneumococcal isolates from both children and adults, albeit with a decrease of serotype 3 compared to the latter study. The distribution of serotypes and vaccine coverage was similar among different age groups and regions. The high coverage rate of PCV13 and PPSV23 confirmed that stains of vaccine-covered serotypes were still popular in among Chinese population.

In consistent with previous studies [13,15,16], more than 97% of strains were resistant to erythromycin and clindamycin in our study. The penicillin non-susceptibility rates in meningitis and non-meningitis isolates were 83.6% and 6.6%, both of these rates are higher compared to in previous studies [16] and abroad [23], which may be explained by variations in the population included and the wide use of antibiotics in China [31]. Serotypes 23F, 19F, 19A, and 14 exhibited higher resistance rates compared to serotypes 9 V and 3. The high levels of resistance in serotypes 19F, 19A, 14, and 23F, may be associated with the widespread international spread of ST271, ST320, ST876, and ST81. Another noticeable finding in this study was that the resistance rates of PCV13 serotypes were generally higher than those of non-PCV13 serotypes, but we should also notice that 15A, 23B, and 23A were potential serotypes that may spread after the use of PCV13, and these also showed high resistance rates to penicillin and other antibiotics. The same finding was reported in Japan, i.e. that IPD cases caused by serotypes 12F [32] and 15A [33] increased in the post-vaccine era, and both of these serotypes are highly resistant. Thus, the resistance levels of these serotypes should be carefully monitored under the dual pressure of antibiotics and vaccines.

Our study has some limitations. Our surveillance study underestimated disease rates because of failure to obtain cultures, culturing limitations, and previous antibiotic treatment. Recruitment across all sites was not uniform; several sites had few pneumococcal isolates collected and transported because of technical issues. The population denominator was not well known, and surveillance was limited to culture-confirmed IPD.

Conclusion

This is a multicentre study focused on IPD in children of China. The data of disease type, clinical features, risk factors, prognosis of IPD, and characteristics of pneumococcal isolates of the study are representative of the Chinese paediatric population. Public policies favouring IPD vaccination should be encouraged, which would decrease the morbidity and mortality rates of IPD in China.

Acknowledgements

The authors are grateful to the doctors and patients contributed to this study. Special thanks to the professor Ge-Tu Zhao Ri for the help in revising the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work is funded by Beijing Natural Science Foundation (No. L202004) and Respiratory Research Project of National Clinical Research Center for Respiratory Diseases (No. HXZX-202106).