Invasive pneumococcal disease (IPD) continues to cause significant morbidity and mortality in the paediatric population of Spain, even after the introduction in our country of the pneumococcal conjugate vaccine.1 The incidence of IPD in children under 5 years in Spain ranges between 16.8 (2012) and 12.3 (2015) per 100000 inhabitants.1,2

In some cases, it requires admission to the paediatric intensive care unit (PICU) due to the severity of the symptoms. Some forms of IPD, such as sepsis or meningitis, require intensive care from onset, with management based on the continuous monitoring and intensive treatment that these forms require once they are diagnosed. Other forms, such as complicated pneumonia, do not always require admission to the PICU to be cured and could be managed initially in a hospital without paediatric intensive care services.3

The aim of our study was to analyse the epidemiological, microbiological and clinical characteristics of cases of IPD potentially requiring admission to the PICU.

The study included 263 patients with IPD (155 male and 108 female), with a mean age of 45 months (standard deviation [SD], 37.1), out of who 50 (19%) required admission to the PICU at some point in the course of disease (Group 1). The remaining 213 patients did not require admission to the PICU (Group 2). The incidence of IPD peaked in February (n=54; 20.5%) and March (n=44; 16.7%), while admission to the PICU peaked in March (n=10; 20%) and December (n=7; 14%) (Fig. 1A).

Figure 1.

Characteristics of patients based on whether or not they required admission to the paediatric intensive care unit (PICU). (A) Month of the year invasive pneumococcal disease was diagnosed. (B) Clinical presentation. (C) Complications. (D) Sequelae.

(0.32MB).

The length of stay was significantly longer in patients that required admission to the PICU (19.8±12.5 days) compared to those that did not (9.6±5.8 days) (P<.001). We found that 46.7% of patients with underlying disease required admission to the PICU compared to 17.3 of patients that were previously healthy. The ratio of the number of patients admitted to the PICU to the number not admitted was 4.17 times greater in patients with underlying disease compared to previously healthy patients, and this association was statistically significant (OR, 4.17; 95% CI, 1.44–12.12).

The forms of IPD found in the total sample were (n; %): complicated pneumonia (152; 57.4%), uncomplicated pneumonia (43; 16.3%), occult bacteraemia (25; 9.5%), meningitis (19; 7.2%), septic shock (7; 2.7%), mastoiditis (7; 2.7%), osteoarticular infection (7; 2.7%), orbital cellulitis (2; 0.8%), peritonitis (1) and pancreatitis (1). In group 1, the patients admitted to the PICU included 100% of patients with septic shock (OR, 5.95; 95% CI, 4.53–7.82), 84.2% of patients with meningitis (OR, 32.94; 95% CI, 9.11–119.10) and 15.2% of patients with complicated pneumonia (OR, 0.57; 95% CI, 0.30–1.50), although in the latter case the association with PICU admission was not statistically significant. Group 2, composed of the patients that did not require PICU admission, included 95% of the patients with uncomplicated (OR, 0.18; 95% CI, 0.04–0.75) (Table 1 and Fig. 1B).

In both groups 1 and 2, 64.4% of patients with IPD experienced some form of complication; 57.8% were respiratory complications and 5.3% neurologic complications. Of the patients with neurologic complications during the acute phase, 85.7% required admission to the PICU (OR, 33.3; 95% CI, 7.17–154.8), as did 15.8% of the patients with respiratory complications (OR, 0.613; 95% CI, 0.33–1.14) (Fig. 1C).

Of the 260 patients with IPD for who follow-up data were available, 85.5% were free of sequelae at 6 months. Two patients died and 1 did not attend the follow-up visit. Out of the patients with immediate sequelae, 22 (8.5%) had respiratory sequelae and 8 (3.1%) neurologic sequelae (including 2 patients with hearing loss); 1 patient had high blood pressure, 1 required a radical mastoidectomy and 1 had cavernous sinus thrombosis. In group 1, 14.3% of patients developed neurologic sequelae (OR, 37.1; 95% CI, 4.4–309-9), followed in frequency by respiratory sequelae in 10.2% (OR, 1.4; 95% CI, 0.48–3.93) (Fig. 1D).

Of all patients with IPD, 84.8% (223/263) had not received antibiotherapy at the outpatient level before admission, and 20.6% of these subset of patients (46/223) belonged to group 1 (OR, 0.43; 95% CI, 0.145–1.26; P=.115).

We found that 50.4% (132/263) of patients, without differences between groups, had received at least 1 dose of pneumococcal conjugate vaccine. Only 28.1% were correctly vaccinated for their age (8.7% with PCV7; 2.3% with PVC10 and17.1% with PCV13). We did not find an association between the need for admission to the PICU (group 1) and complete PCV13 vaccination status (OR, 0.48; 95% CI, 0.18–1.3) (P=.135).

The serotypes identified most frequently in both groups were serotype 3 (53 [20.2%], of who 17% required PICU admission), serotype 1 (48 [18.3%], of who 12.5% required PICU admission), other nonvaccine serotypes (33 [12.5%], of who 21% required PICU admission), serotype 19A (17; 6.5%, of who 23% required PICU admission) and serotype 14 (15 [5.7%], of who 20% required PICU admission). The serotypes identified most frequently in patients of group 1 were (n, %): serotype 3 (9; 18%), other nonvaccine serotypes (7; 14%), serotype 1 (6; 12%) and serotype 19A (4; 8%). The most frequent nonvaccine serotypes in group 1 were serotypes 10A (2; 4%) and 11 (2; 4%) (Fig. 2). There was only 1 case of vaccine failure in this group, corresponding to a patient with infection by serotype 3. When it came to genotyping, the only salient finding in group 1 was detection of sequence type 306, corresponding to serotype 1, in 4/6 cases; the rest of sequence types and clonal complexes varied and were not associated significantly with PICU admission.

Figure 2.

Frequency distribution of serotypes in patients that required admission to the paediatric intensive care unit (PICU) and patients that did not.

(0.18MB).

Admission to the PICU is a possible development in the context of IPD, so the exploration of the epidemiological, microbiological and clinical factors may be associated with an increased likelihood of admission to the PICU is a relevant pursuit. In our study, we identified factors associated with an increased probability of admission to the PICU, such as age less than 1 year at the time of diagnosis, the presence of underlying disease that increases the risk of IPD, specific presentations of IPD such as septic shock or meningitis and the months of the year corresponding to seasonal viral disease outbreaks.

The two seasonal peaks in the incidence of IPD and the peak in the proportion of patients with IPD requiring PICU admission coincided with the annual outbreaks of bronchiolitis due to respiratory syncytial virus (RSV) and influenza, as described in previous studies.5–9 However, while the incidence of IPD peaked in February and March (and the influenza outbreak occurred between week 50 of 2013 and week 12 of 201410), the months corresponding to the highest proportion of IPD cases requiring PICU admission were November and December (overlapping with the RSV season that happens between week 40 of one year and week 4 of the following year10). This could be explained by different factors, such as the lower age of patients with bronchiolitis, the virulence of the virus involved in the RSV or flu season, and the previous health of the patient.6,7

Outpatient management with antibiotherapy previous to hospital admission does not have a protective effect against admission to the PICU, as we found in our series and in reviewing the work of other authors.11

The total length of stay was greater in patients admitted to the PICU, which was consistent with the findings of previous studies.12 When it came to the forms of IPD that required admission to the PICU, we found that the only feature present in 100% of the patients admitted to the PICU was haemodynamic instability. Although a diagnosis of meningitis was in itself a criterion for admission to the PICU, it was not always made at the outset and patients with a discharge diagnosis of meningitis that were not admitted to the PICU had favourable outcomes. This was because 39.5% of the patients with meningitis admitted to the PICU had complications during the acute phase, as was observed in other studies, such as the one published by Casado et al.11 Most children with IPD admitted to the PICU recovered and were free of sequelae. The most frequent sequelae were neurologic, followed by respiratory sequelae. Only 15.2% of patients with complicated pneumonia with pleural effusion or empyema required admission to the PICU, which demonstrates that these patients may be managed in a hospital without an intensive care unit under close monitoring.11,13

Although the most prevalent serotypes (especially 19A) are covered by the PCV13, 11.8% of the patients with complete PCV13 vaccination were admitted to the PICU compared to 18.5% of patients without complete vaccination (P=.23), a difference that was not statistically significant. Thus, in our study, complete vaccination with PCV13 did not exhibit a protective effect as concerned admission to the PICU. The only case of vaccine failure in a patient that required intensive care involved serotype 3, against which the PCV13 proved to be less effective compared to other serotypes, as found in previous studies.14–16 Asner et al.16 reported vaccine failure in 26% of patients that required admission to the PICU, with many of these cases involving serotype 3. We ought to mention 2 serotypes that are not included in the PCV13, 10A and 11A, which, while infrequent among our patients, are associated with a high proportion of PICU admission. In a prospective study conducted in Switzerland that included 117 patients with IPD, patients with meningitis, especially those with infection by non-vaccine serotypes, were most likely to require admission to the PICU.16

When it came to the serotypes involved in IPD, in our study serotype 3 was the one corresponding to the highest proportion of affected patients requiring admission to the PICU, which may be attributable to the characteristics of this serotype, which has capsule proteins resulting in greater virulence,17,18 and the lower efficacy of the PCV13 against this serotype, described in the previous literature.14,17,18 This was also the serotype most frequently associated with development of IPD and with vaccine failure in our study.

These results evince the need of ongoing surveillance of pneumococcal serotypes involved in IPD and their genotypes to have updated information on the aetiological agents of IPD in the local population at all times.19

There are limitations to our study. First of all, the sample only included 30.4% of the population under 5 years in Catalonia, as we only studied IPD in 3 of the hospitals in the region. Secondly, only 2 of the 3 participating hospitals had a PICU, and while these 2 units are the referral PICUs in Catalonia, there are other hospitals that manage patients with IPD. Third of all, there may have been biases associated with the documentation of clinical data in the health records, and some pneumococcal strains were not analysed, meaning the resulting information was not available for 100% of the cases. Lastly, data collection in the last year included in the study period ended in June, which means that the study did not include data corresponding to the bronchiolitis season of that year.

In our series, the presence of septic shock and/or neurologic involvement in the context of IPD was associated with admission to the PICU, and the proportion of patients admitted to the PICU was higher in the group of patients with underlying disease. The frequency of PICU admission increased during the annual bronchiolitis season. Admission to the PICU was associated with a longer length of stay in hospital, a higher prevalence of complications in the acute phase and a higher prevalence of sequelae, mainly neurologic. The most frequent serotypes involved in the group of patients admitted to the PICU were vaccine serotypes (19A, 14, 3 and 1). Admission to the PICU in the context of IPD is necessary in patients that develop septic shock or neurologic complications. In case of complicated pneumonia, it is possible to manage the patient in the paediatric ward under close monitoring.

National R&D& programme, Instituto de Salud Carlos III-Vice Directorate General of Evaluation and Promotion of Research in Health Care. Grant no. P111/02081, PI 11/2345; European Regional Development Fund (ERDF) and Agencia de Gestión de Ayudas Universitarias y de Investigación (AGAUR). Grant no. 2017 SGR 1342.

No conflicts of interest.