Pneumococcal disease is a major cause of morbidity and mortality in young children, both in developed and developing country settings. In September 2006, the United Kingdom introduced 7-valent pneumococcal conjugate vaccine (PCV7) into the national childhood immunization program at 2, 4 and 12–13 months of age, with a 12-month catch-up for all children up to 2 years of age. This program resulted in a rapid reduction of invasive pneumococcal disease (IPD) caused by PCV7 serotypes across all age groups through direct and indirect (herd) protection.1 By 2009/2010, PCV7-type IPD dropped by 98% in children <2 years of age and more than 75% in older age groups, with an overall decline in IPD incidence of 34%.1
In April 2010, a 13-valent pneumococcal conjugate vaccine (PCV13) replaced PCV7 in the childhood UK immunization program, without any catch-up.2 At that time, the extra 6 serotypes in PCV13 accounted for more than half of IPD cases in children <5 years of age. Compared with PCV7 serotypes the 6 extra PCV13 serotypes were more likely to cause lower respiratory tract infection (LRTI) in healthy children.3 At the same time, a higher prevalence of underlying comorbidity was observed among children with IPD due to non-PCV13 serotypes. We also observed a higher case fatality rate (CFR) among children with comorbidities and those who presented with meningitis compared with previously healthy children.3
The replacement of PCV7 with PCV13 had a major impact against 2 key replacing serotypes—7F and 19A—and, 4 years after implementation, IPD rates fell further by 32% compared with the pre-PCV13 baseline and 56% compared with the pre-PCV7 baseline.2 In children <5 years of age, IPD due to the extra PCV13 serotypes declined by 69% in 2013/2014, accounting for only 14% of cases in this age group. The decline in PCV13-type IPD, however, was associated with a small and steady increase in replacement disease, particularly due to serotypes 8, 15A, 15B/C, 22F, 23B and 24F.2
Since individual serotypes have different propensities for causing invasive disease, we aimed to describe the clinical presentation, comorbidity prevalence, serotype distribution and outcomes of IPD in children 3–59 months of age during the first 6 years after PCV13 introduction in England and Wales.
Public Health England (PHE) conducts enhanced national IPD surveillance in England and Wales.1 Briefly, National Health Service hospitals routinely report clinically significant infections electronically to PHE and submit invasive isolates to the PHE reference laboratory for confirmation and serotyping. Reported cases without submitted isolates are actively followed up with the reporting laboratories for sample submission. Laboratory reports are regularly reconciled with serotype data into a single dataset. Cases <5 years of age are routinely followed up by postal questionnaire with their general practitioners for further information about vaccination history, comorbidity status, clinical presentation and outcomes. For this study, laboratory-confirmed cases 3–59 months of age in the birth cohorts eligible for PCV13 (children born since April 01, 2009 and vaccinated from April 01, 2010) who were diagnosed with IPD until March 31, 2016 were included.
A case of IPD was defined as Streptococcus pneumoniae cultured from a normally sterile site, or, for culture-negative cases, detection of pneumococcal DNA in cerebrospinal, joint or pleural fluid. Meningitis was defined as S. pneumoniae identified in cerebrospinal fluid through culture/polymerase chain reaction or radiologic/clinical features of meningitis with S. pneumoniae isolated from blood culture. LRTI was defined as S. pneumoniae in empyema fluid or in blood with radiologic/clinical diagnosis of pneumonia. Septicemia was defined as S. pneumoniae cultured in blood with no distinctive clinical syndrome. Repeat samples from sterile sites within 30 days from the same individual were regarded as part of the same episode. Comorbidity was defined as presence of a high-risk condition as defined in the Green Book on Immunization.4 In children with >1 IPD episode, only the first episode was included in this analysis; repeat IPD episodes are described separately.
PCV7 serotypes included 4, 6B, 9V, 14, 18C, 19F and 23F. PCV13 serotypes included PCV7 serotypes plus 1, 3, 5, 6A, 7F and 19A. The additional 11 serotypes included in PPV23 were 2, 8, 9N, 10A, 11A, 12F, 15B/C, 17F, 20, 22F and 33F.
Data were exported from Microsoft Access to Stata version 13 (StataCorp, College Station, TX) for analysis and are mainly descriptive. IPD characteristics in PCV13-eligible children were compared with cases in PCV7-eligible children, before PCV13 introduction.3 Categorical variables were described as percentages and initially compared in bivariate analyses using χ2 test or 2-tailed Fisher exact test, as appropriate. Multinomial logistic regression was used to calculate relative risk ratios (RRRs) to assess the association between patient factors (age [months], sex [male/female], premature birth [yes/no] and comorbidity [present/absent]) and serotype groups (PCV13, additional PPV23, non-vaccine type [base]). The number of comorbidities was not included in the model as multiple comorbidities were rare in our cohort. All these factors (including serotype group) were then assessed in a second multinomial model with clinical presentation (septicemia, meningitis, LRTI [base], other) as the outcome. In these models, the RRRs were interpreted as follows: a RRR of 2 for sex (male vs. female) for the meningitis outcome would mean that the male:female ratio for meningitis was twice that seen for LRTI (after adjusting for other factors). To test for the overall relationship between each factor and the outcome, likelihood ratio tests were used. Multivariable logistic regression was used to investigate independent risk factors for death, including age, serotype group, premature birth, comorbidity status and clinical presentation with meningitis.
In England and Wales, after PCV7 replacement with PCV13 in the childhood immunization program in April 2010, the total number of IPD cases in children 3–59 months of age declined from 381 in 2010/2011 to 201 in 2012/2013, before increasing to 271 in 2015/2016 (Fig. 1). During this 6-year period, there were 1280 laboratory-confirmed IPD episodes in 1255 children eligible for PCV13, including 17 children with 2 episodes and 4 with 3 episodes (Fig. 1). Completed questionnaires were returned for all confirmed cases. Immunization rates were high, with 97.3% (1221/1255) having received ≥1 PCV13 dose before developing IPD. Overall, 84.4% (1059/1255) of the isolates were serotyped and PCV13 serotypes accounted for 22.5% (238/1059), the extra 11 PPV23 serotypes for 46.8% (496/1059) and non-vaccine serotypes (NVTs) for 30.7% (325/1059) of serotyped isolates. The cohorts eligible for PCV13 increased with time since PCV13 introduction, while those in the PCV7 cohort declined (Fig. 1). Differences in case characteristics between the PCV7 and PCV13 periods are summarized in Table 1.
Clinical presentation varied with age (Table 2); LRTI was the most common presentation (34.7%, 378/1089) followed by meningitis (30.5%, 332/1089) and septicemia (29.2%, 318/1089). Meningitis was most prevalent among 3- to 11-month olds (45.8%, 209/456) compared with 12- to 23-month olds (22.1%, 77/348; RRR compared with LRTI, 0.35, 95% CI: 0.23–0.52) and 24- to 59-month olds (16.1%, 46/285; RRR 0.22; 95% CI: 0.14–0.35). The reverse was true for LRTI: 115/456 (25.2%), 130/348 (37.4%) and 133/285 (46.7%), respectively.
IPD due to PCV13 serotypes declined with time since PCV13 introduction, while IPD due to the additional PPV23 and NVTs increased. The contribution of PCV13 serotypes to total IPD cases within the different childhood age-groups, however, remained relatively consistent (Table 2). More than half the PCV13-type IPD cases presented with LRTI (51.4%, 110/214) compared with 26.5% (115/434) for PPV23 and 29.4% (85/289) for NVT IPD cases.
Of the cases with complete serotype and clinical presentation data, the most common serotypes among children with LRTI cases were 3 (12.6%, 40/317), 19A (11.7%, 37/317), 12F (8.2%, 26/317), 33F (6.6%, 21/317) and 22F (6.3%, 20/317).
The most common serotypes responsible for meningitis included 22F (10.3% 30/292), 33F (10.3%, 30/292), 15B/C (9.9%, 29/292), 10A (9.9%, 29/292) and 12F (6.5%, 19/292). The most common serotypes causing septicemia cases were 15B/C (11.4%, 31/271), 12F (10.7%, 29/271), 23B (9.6%, 26/271), 22F (8.5%, 23/271) and 33F (8.1%, 22/271).
Children with PCV13-type IPD were least likely to present with septicemia (16.4%, 35/214) compared with PPV23 (32.5%, 141/434) and NVT IPD cases (33.9%, 95/280) (Table, Supplemental Digital Content 1, http://links.lww.com/INF/C880). This relationship persisted in the multinomial model (RRR 0.27, 95% CI: 0.17–0.45 compared with NVT IPD cases). Children with PCV13-type IPD were also least likely to present with meningitis (RRR 0.43, 95% CI: 0.27–0.70). Table 2 summarizes the clinical presentations by serotype group.
Overall, 259 children (20.6%) had 292 comorbidities, particularly immunosuppression (31.6%, 92/292), chronic heart disease (23.3%, 68/292) and chronic respiratory disease (17.8%, 52/292) (Table 2). Comorbidity prevalence varied between 12.0% and 29.6% per year and increased with age, being 14.2% (72/508) in 3- to 11-month olds, 16.5% (69/419) in 12- to 23-month olds and 36.0% (118/328) in 24- to 59-month olds. Congenital heart disease was more prevalent in 3- to 11-month olds (33.8%, 27/80), and 12- to 23-month olds (29.9%, 23/77), while immunosuppression was responsible for almost half the comorbidities reported among cases 24–59 months of age (46.7%, 63/135).
Children with comorbidity were more likely to present with septicemia compared with previously healthy children (95/217 [43.8%] vs. 223/872 [25.6%]; RRR 1.90, 95% CI: 1.26–2.87); no such associations were identified for other clinical presentations.
Children with comorbidities were less likely to develop PCV13-type IPD (14.7%, 35/238) compared with those with the additional 11 PPV23 (19.4%, 96/496; RRR 1.36, 95% CI: 0.87–2.13), or NVT (31.1%, 101/325; RRR 2.39, 95% CI: 1.52–3.76).
Overall, 170 children (14.1% 169/1198) had been born prematurely; this group was also more likely to have additional comorbidities (31.8%, 54/170) compared with those born at term (196/1028, 19.1%, P < 0.001). The main serotypes causing IPD in children with comorbidity over the 6 surveillance years included 23B (11.2%, 26/232), 24F (8.2%, 19/232) 15B/C (7.8%, 18/232) 22F (7.8%, 18/232) and 12F (6.9%, 16/232) (Fig. 2).
Sixty-four children died over the 6-year period (64/1255; CFR, 5.1%; 95% CI: 3.9%–6.5%); CFR after PCV13-type IPD was 6.7%, 5.2% with the extra PPV23 serotypes and 4.6% with NVT (Table 3). Meningitis cases had the highest CFR (9.6%), followed by septicemia (5.3%) and LRTI (3.7%). CFR was also higher for children with comorbidities (8.1%), compared with 4.3% for previously healthy children; this difference was most apparent among <2-year olds (Table 3).
Of those with comorbidities, although the number of cases were small, a higher CFR was observed for children with sickle cell disease (25.0%, 4/16), splenic dysfunction (20.0%, 2/10) and diabetes mellitus (20.0%, 1/5) compared with those with chronic respiratory (7.7%, 4/52), chronic heart (13.2%, 9/68), chronic renal (4.3%, 1/23) or chronic liver (11.1%, 1/9) diseases. Immunosuppression had the lowest CFR with one case (1/92, 1.1%) proving fatal. In children with comorbidities, CFR was higher in 3–11 months (12.5%, 9/72) and 12–23 months (13.0%, 9/69), but lower in 24–59 months (4.2%, 5/118).
In the univariate analysis, a fatal outcome was associated with male gender, underlying comorbidity and meningitis presentation (Table, Supplemental Digital Content 2, http://links.lww.com/INF/C881), but not with age at IPD onset or serotype group. In a multivariable logistic regression model, only comorbidity (aOR 2.41; 95% CI: 1.25–4.64) and meningitis (aOR 3.53; 95% CI: 1.62–7.7) were independently associated with death.
Repeat IPD Episodes
Twenty-one children had repeat IPD episodes (21/1255, 1.7%), including 17 (1.4%, 17/1255) with 2 episodes and 4 with 3 episodes (0.3% 5/1255). Of the 17 children with 2 repeat IPD episodes, 7 had the same serotype for both episodes (38, 17F, 23F, 24F, 22F, 23B and 10A). Of the 4 children with 3 episodes, the responsible serotypes were 15A/15A/38, 22F/22F/6D, 23B/unserotyped/23B and 17F/16F/15A. Nineteen children had an underlying comorbidity, including immunosuppression (n = 9, 47.4%), cochlear implant (n = 4, 21.1%), chronic heart (n = 2, 10.5%), chronic real (n = 2, 10.5%) and chronic liver (n = 1, 5.3%) disease, while one had multiple comorbidities (splenic dysfunction, liver and renal). Two cases were 3-dose PCV13 failures. None of the children with repeat IPD episodes died.
Cases in 2015/2016
In 2015/2016, when all children <5 years of age were eligible for PCV13, 38.0% (103/271) cases were 3–11 months, 25.1% (68/271) were 12–23 months and 36.9% (100/271) were 24–59 months. Comorbidity prevalence was 19.2% (52/271) (Table, Supplemental Digital Content 3, http://links.lww.com/INF/C882) and CFR was 4.1% (11/271); of the 255 cases with information about clinical presentation, 91 (35.7%) had LRTI, 74 (29.0%) had meningitis and 73 (28.6%) had septicemia. Figure 3 summarizes the clinical presentation for the 10 most common serotypes.
PCV13 serotypes were responsible for 10.8% (25/232) of cases with serotyped isolates, the additional 11 PPV23 serotypes for 60.8% (141/232) and NVT for 28.4% (66/232) cases. The most prevalent non-PCV13 serotypes were 12F (16.4%, 38/232), 10A (10.8%, 25/232), 23B (8.6%, 20/232), 33F (8.6%, 20/232), and 15B/C (7.3%, 17/232) and 8 (7.3%, 17/232).
Of the 25 children with PCV13-type IPD, 17 had received a full course of PCV13 (median age at IPD, 33 [interquartile range, 22–50] months); 3 received 2 doses (two 3–11 months, one 24–59 months of age) and 2 received one dose (3–11 months) of age; the remaining 3 (one 3–11 months, two 24–59 months) were unimmunized. The most prevalent serotypes associated with PCV13 failure were 19A (11/25, 44%) and 3 (8/25, 32%); vaccine failure due to other PCV13 serotypes was rare. Five children (20.0%) had comorbidity and 2 healthy children died of serotype 19A IPD.
We have previously reported large and significant declines in childhood IPD incidence following PCV7 introduction into the UK immunization program and described the characteristics of IPD in the PCV7-eligible children. In particular, we found that comorbidities were more prevalent in children with IPD due to non-PCV13 serotypes and were associated with higher case fatality.1–3 Following the introduction of PCV13 in April 2010, IPD cases in <5-year olds declined until 2012/2013, before increasing in subsequent years. The rapid decline in IPD due to the additional PCV13 serotypes has been offset by replacement disease with non-PCV13 serotypes, suggesting that the maximum benefit of the childhood PCV program may already have been achieved. By 2015/2016, 6 years after PCV13 replaced PCV7, the PCV13 serotypes were responsible for only 11% of IPD cases in our cohort. Most cases are now due to non-PCV13 serotypes and associated with higher comorbidity prevalence. The clinical presentation has also changed such that children are equally likely to present with septicemia, meningitis or LRTI. Reassuringly, vaccine failures, recurrent IPD and CFRs remain low, the latter being independently associated with underlying comorbidities and presentation with meningitis.
When comparing the 6 post-PCV13 years with the pre-PCV13 period, comorbidity prevalence among IPD cases increased from 14.9% to 20.6%.3 This had been predicted in our previous study because of serotype replacement disease; both the extra 11 PPV23 and NVTs were more likely to cause IPD in children with underlying comorbidities. In 2015/2016, 13.7% of children with IPD had also been born prematurely; this vulnerable group is known to be at increased risk for IPD.5 Notably, though, despite the dynamic shift in serotypes causing IPD, childhood CFR has remained low throughout the surveillance period. The CFR in children with comorbidity who developed IPD during the PCV13 period (8.1%) is similar to the 8.5% reported during the PCV7 period (Table 1).3
We also observed a shift in clinical presentations of IPD. The rise and fall of IPD due to PCV13 serotypes following PCV7 introduction and its subsequent replacement with PCV13, respectively, was associated with an increase followed by a decrease in LRTI presentations. This is in keeping with other studies that found non-PCV13 serotypes were more likely to cause septicemia in patients with underlying diseases. In a recent Swedish study using clinical data from 2096 adults and 192 children with IPD alongside 165 invasive and 550 carriage isolates from children, for example, the authors found a lower invasive potential for non-PCV13 compared with PCV13-type strains. Moreover, those infected with non-PCV13 strains were more likely to have underlying diseases, less likely to develop LRTI and, in adults, tended to have a higher CFR.6 In 2015/2016, although, overall, the children were equally likely to present with LRTI, septicemia or meningitis, we did observe an age effect, with infants more likely to present with meningitis and older children with LRTI. Other studies, albeit with fewer cases over a shorter post-PCV13 surveillance period have also observed a higher comorbidity prevalence among childhood IPD cases but changes in clinical presentations of IPD have been variable. In Spain, hospitalizations for pneumococcal pneumonia and meningitis among <5-year olds were significantly reduced during the first 3 years after PCV13 replaced PCV7,7 while an observational study from Massachusetts did not identify any significant changes in clinical presentation.8 In Calgary, Canada, the proportion of cases presenting with pneumonia increased after PCV7 introduction and then nearly halved after replacement with PCV13.9 The proportion of cases with empyema and meningitis presentations, however, increased over the same period, as did hospitalization rates for IPD, raising concerns about more severe disease due to the replacing serotypes. Interestingly, recent studies from Israel found that PCV impact may vary for the different IPD presentations.10–12 In the most recent study, for example, meningitis rates declined by 24% only compared with 57% for bacteremic pneumonia and 70% for other presentations. Notably, too, bacteremic pneumonia rates did not decline after PCV7 introduction and fell only after replacement with PCV13. The authors also found differential increases in clinical presentations following IPD due to non-PCV13 serotypes after PCV13 introduction.12
We have consistently observed higher case fatality associated with meningitis presentations across the age groups.3 Other recent studies have been unable to identify any differences in the clinical characteristics of children with pneumococcal meningitis even though cases due to PCV13 serotypes declined rapidly after PCV13 introduction and those due to non-PCV13 serotypes increased.13 It is likely that, once meningitis occurs, the course, severity and outcomes are more dependent on host factors rather than the infecting pneumococcal serotype.
The comorbidities associated with an increased risk of IPD are well described.14 The serotypes causing IPD in children with comorbidity were different from those identified in healthy children; these serotypes usually have low invasive potential, but are often associated with higher CFR because of the child’s vulnerable state. In our cohort, children with splenic dysfunction, particularly those with sickle cell disease, were more likely to die of IPD. This group continues to have a significantly higher risk of IPD despite current preventative measures such as early detection through universal screening, early penicillin prophylaxis and high immunization uptake.15 On the other hand, although immunosuppression (including malignancy) was the most prevalent comorbidity in children with IPD, especially among 24- to 59-month olds, only one child died. Consequently, CFR in children 24–59 months of age with and without comorbidity was remarkably similar (Table 3). This reassuring finding is likely to be a result of rapid access to medical care for immunosuppressed children presenting with fever and strict protocol adherence with low threshold for initiating antimicrobial therapy by clinicians.16
The strength of this study lies in the consistently high case ascertainment and extensive national follow-up of all confirmed cases during the past decade, alongside a national reference laboratory to serotype nearly all invasive isolates across England and Wales. This has allowed us to monitor not only serotype-specific trends over time, but also changes in disease characteristics and outcomes in young children, who are most at risk for IPD. Like all large-scale surveillance programs, this study has some limitations. Not all laboratories, for example, routinely submit clinical isolates for serotyping to the national reference laboratory and some information in the questionnaires remained incomplete despite multiple follow-up attempts; these limitations, however, were generally overcome by the large number of cases included in our analysis. Another limitation of this study is the lack of long-term complications among survivors of pneumococcal meningitis, especially because this pathogen has the worst prognosis among encapsulated bacteria causing meningitis in children.17
In England and Wales, most cases are now due to non-PCV13 serotypes, which are more likely to cause septicemia, especially in children with comorbidities. The lack of any serotype predominance in replacement disease highlights the need for a universal vaccine against the pneumococcus. Additional interventions are also required to protect children with specific comorbidities and those presenting with meningitis from fatal outcomes.
We thank (1) Sarah Collins for maintaining the pneumococcal surveillance database in the Immunisation Department at PHE; (2) Melissa Kephalas and Rashmi Malkani for the follow-up of confirmed cases; (3) Pauline Waight who managed the pneumococcal surveillance before 2016; (4) Catrin Moore for reporting IPD cases in southern England serotyped by the John Radcliffe Hospital Oxford (Oxford, United Kingdom) laboratory; (5) the staff at hospital laboratories across England and Wales who referred isolates for serotyping and provided additional information on request and (6) general practitioners of patients with IPD for completing the surveillance questionnaires.
1. Miller E, Andrews NJ, Waight PA, et alHerd immunity and serotype
replacement 4 years after seven-valent pneumococcal conjugate vaccination in England and Wales: an observational cohort study. Lancet Infect Dis. 2011;11:760–768.
2. Waight PA, Andrews NJ, Ladhani SN, et alEffect of the 13-valent pneumococcal conjugate vaccine on invasive pneumococcal disease in England and Wales 4 years after its introduction: an observational cohort study. Lancet Infect Dis. 2015;15:535–543.
3. Ladhani SN, Slack MP, Andrews NJ, et alInvasive pneumococcal disease after routine pneumococcal conjugate vaccination in children, England and Wales. Emerg Infect Dis. 2013;19:61–68.
4. Salisbury D, Ramsay MImmunisation against infectious disease [Public Health England web site]. December 17, 2013. Available at: https://www.gov.uk/government/collections/immunisation-against-infectious-disease-the-green-book
. Accessed June 21 2017.
5. Rückinger S, van der Linden M, von Kries REffect of heptavalent pneumococcal conjugate vaccination on invasive pneumococcal disease in preterm born infants. BMC Infect Dis. 2010;10:12.
6. Browall S, Backhaus E, Naucler P, et alClinical manifestations of invasive pneumococcal disease by vaccine and non-vaccine types. Eur Respir J. 2014;44:1646–1657.
7. Georgalis L, Mozalevskis A, Martínez de Aragón MV, et alChanges in the pneumococcal disease-related hospitalisations in Spain after the replacement of 7-valent by 13-valent conjugate vaccine. Eur J Clin Microbiol Infect Dis. 2017;36:575–583.
8. Iroh Tam PY, Madoff LC, Coombes B, et alInvasive pneumococcal disease after implementation of 13-valent conjugate vaccine. Pediatrics. 2014;134:210–217.
9. Ricketson L J, Conradi N G, Vanderkooi O G, et alChanges in the nature and severity of invasive pneumococcal disease in children, before and after the 7-valent and 13-valent pneumococcal conjugate vaccine programs in Calgary, Canada. Pediatr Infect Dis J. 2017 [Epub ahead of print].
10. Ben-Shimol S, Greenberg D, Hazan G, et alIsraeli Bacteremia and Meningitis
Active Surveillance Group. Differential impact of pneumococcal conjugate vaccines on bacteremic pneumonia versus other invasive pneumococcal disease. Pediatr Infect Dis J. 2015;34:409–416.
11. Ben-Shimol S, Greenberg D, Givon-Lavi N, et alIsrael Bacteremia and Meningitis
Active Surveillance Group. Impact of PCV7/PCV13
introduction on invasive pneumococcal disease (IPD) in young children: comparison between meningitis
IPD. Vaccine. 2016;34:4543–4550.
12. Ben-Shimol S, Greenberg CL, Grisaru-Soen G, et alComparative incidence dynamics and serotypes of meningitis
, bacteremic pneumonia and other-IPD in young children in the PCV era: insights from Israeli surveillance studies. Vaccine. 2017.
13. Olarte L, Barson WJ, Barson RM, et alImpact of the 13-valent pneumococcal conjugate vaccine on pneumococcal meningitis
in US children. Clin Infect Dis. 2015;61:767–775.
14. van Hoek AJ, Sheppard CL, Andrews NJ, et alPneumococcal carriage in children and adults two years after introduction of the thirteen valent pneumococcal conjugate vaccine in England. Vaccine. 2014;32:4349–4355.
15. Oligbu G, Collins S, Streetly A, et alP08 Risk of invasive pneumococcal disease in children with sickle cell disease in England: National observational cohort study, 2010 – 2015. Arch Dis Childhood. 2017;102(suppl 1):A4.
16. NICE. Neutropenic sepsis: prevention and management in people with cancer [National Institute for Health and Care Excellence web site]. September 2012. Available at: https://www.nice.org.uk/guidance/cg151
. Accessed March 13, 2017.
17. Jit MThe risk of sequelae due to pneumococcal meningitis
in high-income countries: a systematic review and meta-analysis. J Infect. 2010;61:114–124.