Background
Streptococcus pneumoniae is a Gram-positive bacterium and a commensal of the human nasopharynx [
1]. Failure of natural immunity to
S. pneumoniae leads to pneumococcal infection and in some cases to invasive pneumococcal disease (IPD) [
1‐
3]. The most frequent manifestation of pneumococcal disease, however, is pneumococcal pneumonia where the pneumococcus may be responsible for up to 60 % of cases of community-acquired pneumonia (CAP) [
4]. Hospitalised CAP carries a mortality rate of 5–15 % rising to more than 30 % in patients admitted to the intensive care unit [
5,
6]. The highest rates of pneumococcal disease are observed in infants, the elderly, patients with chronic respiratory disease and in patients with immune compromise [
7‐
10]. This is despite the availability of effective antimicrobial treatments against
S. pneumoniae, emphasising the importance of preventing pneumonia wherever possible [
11,
12]. The impact of pneumococcal disease in the UK is substantial with approximately 6000 cases of IPD reported annually and 192,281 hospital admissions for pneumonia in 2013/14 in England of which up to 50 % may be pneumococcal [
13,
14]. The cost to the UK National Health Service is estimated at more than £1 billion [
15].
Pneumococcal disease is, at least partially, a vaccine preventable disease. The 23-valent pneumococcal polysaccharide vaccine (PPSV) has been recommended in the UK for patients at high risk of pneumococcal disease since 2003, including adults over the age of 65 years. A systematic review and meta-analysis of the data supporting PPSV show that it protects against IPD in adults in high income countries (OR 0.20 95 % CI 0.10–0.39,
n = 27886), but limitations include uncertainty over its protection against IPD in patients with chronic illnesses (OR 1.56 95 % CI 0.35–0.694
n = 3230), protection against non-invasive pneumococcal CAP and its duration of protection [
16].
The 13-valent pneumococcal conjugate vaccine (PCV13, Prevenar-13) has been evaluated for the prevention of vaccine type IPD in children and in elderly subjects [
17,
18]. The recent CAPITA trial conducted in the Netherlands demonstrated the efficacy of PCV13 for the prevention, in those aged ≥65, of vaccine type pneumococcal CAP and also non-invasive CAP caused by vaccine serotypes [
18]. An analysis based on the frequency of IPD and CAP caused by
S. pneumoniae in the Netherlands concluded that PCV13 was cost-effective [
19].
PCV13 is not currently part of the UK adult vaccination programme, neither for elderly patients aged >65 years nor for specific high risk groups. Determining whether PCV13 would be cost-effective in the UK requires accurate information on the burden of vaccine preventable pneumococcal disease in the UK.
We conducted a systematic review to determine the incidence and burden of vaccine preventable pneumococcal disease in the adult UK population.
Methods
This manuscript reports a systematic review of observational studies and was conducted and is reported according to the MOOSE (meta-analysis and systematic review of observational studies in epidemiology) guidelines [
20]. The review protocol was registered on PROSPERO (CRD42015025043).
Search strategy
A librarian searched electronic databases from 1990 until September 2015 for relevant studies using PUBMED and EMBASE. A combination of text words and controlled vocabulary terms related to the subject of interest (pneumococcal disease) and possible outcome measures was used to develop a sensitive search strategy. Terms entered were (Streptococc* [tiab] OR pneumococc* [tiab]) AND (Serotype [Title/Abstract] OR serogroup [Title/Abstract) AND (incidence OR frequency OR prevalence OR distribution). Further searches were conducted for specific data on risk groups and UK regions as described in the relevant sections below. No language restrictions were applied to the search. The search was supplemented by reviews of reference lists, bibliographies and the investigators files where appropriate.
Inclusion criteria
The review included observational cohort studies (including prospective, retrospective, registry and surveillance designs) reporting any of the following study outcomes; 1) Original data reporting of the incidence of vaccine type and non-vaccine type pneumococcal disease; 2) Inclusion or enrolment of patients in the United Kingdom; 3) Sufficient data to generate or infer incidence of disease in the general population or specific risk groups.
For data extraction, articles were independently reviewed by two investigators. Non relevant studies were excluded based on title and abstract review alone.
Outcomes
The primary outcome was the incidence of vaccine type pneumococcal disease in the adult UK population, expressed as an incidence per 100,000 population. Secondary measures included the proportion of pneumococcal disease caused by PCV13 vaccine serotypes, other vaccine serotypes and the proportion of cases of CAP caused by vaccine serotypes was recorded. In addition the proportion of cases of vaccine type pneumococcal pneumonia in risk groups expressed as incidence per 100,000 population where possible. Risk groups include patients with splenectomy, chronic respiratory disease, chronic heart disease, chronic kidney disease and diabetes. Immunosuppression included any disorder leading to significant immune suppression (whether inherited or acquired) including HIV and iatrogenic immune suppression (incl. chemotherapy and corticosteroids).
Vaccine type pneumococcal disease
Vaccine type (VT) pneumococcal disease was defined as being caused by one of the following serotypes: 4, 6B, 9 V, 14, 18C, 19 F, 23 F, 1, 3, 5, 6A, 7 F and 19A (PCV13 VT pneumococcal disease) OR pneumococcal disease caused by one of the following serotypes: 4, 6B, 9 V, 14, 18C, 19 F, 23 F, 1, 3, 5, 7 F, 19A, 2, 8, 9 N, 10A, 11A, 12 F, 15B, 17 F, 20, 22 F and 33 F. (PPSV VT pneumococcal disease).
Results were stratified according to the period of the study, following PPSV introduction (2003), following PCV7 introduction (2007) and following PCV13 introduction (2010).
Discussion
Our systematic review identifies a high burden of pneumococcal disease in adults in the UK, while also revealing substantial ongoing changes in the epidemiology of pneumococcal disease. The most recent data from 2013/14 shows an incidence of 6.85 per 100,000 population across all age groups for IPD, and an incidence of 20.58 per 100,000 population in those aged >65 years [
21]. The corresponding incidences for PCV13 serotype IPD were 1.4 per 100,000 and 3.72 per 100,000. The most recent available data for CAP including non-invasive disease showed an incidence of 20.6 per 100,000 for hospitalised adults with pneumococcal CAP and 8.6 per 100,000 population for hospitalised PCV13 serotype CAP [
21,
29]. We have estimated that if these most recent estimates are applicable to UK population as a whole, there would be at least 10,000 cases of hospitalised pneumococcal CAP, with 4000 caused by PCV13 serotypes and more than 900 cases of PCV13 serotype IPD. The estimates of burden for non-invasive disease are likely to be an underestimate due to the absence of data from CAP managed in the community.
These data have limitations as discussed below, but suggest that pneumococcal disease and PCV13 vaccine type pneumococcal CAP continue to have a significant burden in adults, even after the introduction of PCV13 in children.
There is, however, an ongoing trend of reduced incidence of PCV13 serotype IPD and CAP in the UK, demonstrated both in the study of Waight et al., who demonstrated a 69 % incidence reduction between 2008 and 2014, and in the study by Rodrigo et al. who demonstrated a reduction of 48 % in PCV7 CAP, and a 13 % reduction in the additional serotypes contained in PCV13 [
21,
29].
Pneumococcal vaccination is now a core part of public health policy [
3,
45]. The pneumococcal polysaccharide vaccination covers 23 common serotypes and has been available since 1983. It was introduced into the routine vaccination schedule in England and Wales in 2003/4 for patients aged 80 and over, followed by patients aged 75 and over in 2004/5 and all patients aged 65 and over in 2005/6. The 7-valent conjugate vaccine was introduced in children in 2007, followed by the introduction of the 13-valent conjugate vaccine for children in 2010 [
3].
S. pneumoniae is capable of causing IPD and non-invasive pneumococcal pneumonia [
46]. We identified a large body of evidence on the incidence of IPD in the UK thanks to ongoing surveillance programmes in England, Wales and Scotland.
Patients aged >65 years have the highest incidence of pneumococcal disease and non-invasive pneumococcal pneumonia and are therefore the primary target of vaccination programmes [
18]. Although high uptake of paediatric conjugate vaccines has led to a reduction in cases of adult IPD through herd protection, there remains a substantial burden of IPD and pneumococcal pneumonia in adults [
47,
48].
The CAPITA trial was a randomized double blind placebo controlled trial conducted in the Netherlands which demonstrated efficacy in the reduction of vaccine type pneumococcal CAP in those adults >65 years receiving the 13-valent conjugate vaccine [
18]. In the per-protocol population, vaccine efficacy for the prevention of a first episode of vaccine-type CAP was 46 %, and protection persisted for at least 4 years. There are, however, important differences between the UK and the Netherlands. The Netherlands has no PPSV programme, while uptake of the PPSV vaccination in the UK is among the best in Europe [
45]. In addition, as noted above, the PCV13 vaccine has been used in the UK since 2010 for children, while the CAPITA trial results were obtained in a population where PCV7 was introduced for newborns in 2006 and replaced by PCV10 in 2011 [
49]. Therefore our data allows a degree of comparison between the epidemiology of pneumococcal disease in the UK and the Netherlands. In 2008 in the Netherlands, 68.4 % of IPD episodes in patients aged 65 years or older were caused by PCV13 serotypes, compared to 42.3 % in 2013 [
49]. The corresponding figures from Waight et al. for England and Wales were 44.1 and 20.3 % [
21]. Rodrigo et al. reported that 40.6 % of cases of pneumococcal CAP were PCV13 VT in 2012/13 [
29].
The impact of pneumococcal vaccination programmes in children has been accompanied by a concern about serotype replacement and the potential implication of this on public health [
50‐
52]. Waight et al. reported a significant increase in non-vaccine serotype IPD following the introduction of PCV13 (IRR 1.25 95 % CI 1.17–1.35 comparing 2008–10 to 2013/14) [
21]. Data reported, only in abstract form to date, from the BSAC surveillance project provides further evidence that this is occurring [
53]. Using data from up to 40 clinical laboratories in the UK and Ireland, they showed for bacteraemic pneumococcal disease, 76 % of cases were PCV13 serotypes prior to PCV7 introduction (data from Jan 04- Dec 06) falling to 21 % (Jan-Dec 14). The corresponding data for respiratory isolates was 59 % falling to 17 % [
53]. The most frequent serotypes for both bacteraemia and lower respiratory tract infection in 2004–6 were covered by PCV13, but in 2014 were not, with serotypes 8, 22 F and 12 F being most frequent in bacteraemia and 15A and 23B most frequent in LRTI [
53]. They report associated significant increases in penicillin, tetracycline and multidrug resistance which appears to be mostly due to expansion of serotype 15A [
53]. Antimicrobial resistance is an increasing problem in CAP generally and these data emphasise the importance of ongoing surveillance and consideration of the indirect impact of childhood vaccination [
50‐
52,
54].
This study identified important gaps in the literature regarding the burden of pneumococcal disease in the UK. There was limited data outside of surveillance of IPD. Data for non-invasive pneumococcal pneumonia was limited to hospitalised patients in a single UK city, in Nottingham, England and a series of respiratory tract isolates forming part of the British Society for Antimicrobial Chemotherapy surveillance studies [
31,
36,
53] which would include patients with CAP, but would also include respiratory/sputum isolates from patients with other respiratory tract infections such as exacerbations of COPD [
55], bronchiectasis (where
S. pneumoniae can be a coloniser) [
56] and upper airway commensals. There are important differences both in terms of serotypes responsible for invasive vs non-invasive disease, in disease outcomes and in vaccine effectiveness against non-invasive disease vs invasive disease. Inclusion of additional data on non-invasive CAP in the UK would be valuable. Our available data suggests that there are up to 10 cases of non-invasive pneumococcal CAP for each case of IPD and so studies restricted to IPD are likely to greatly underestimate the burden of pneumococcal disease.
None of the studies identified in the systematic review were conducted in the community and so the burden of pneumococcal disease in the community is unknown. Multiple studies suggest that
S. pneumoniae remains the most common cause of CAP in outpatients, where the majority of CAP is managed [
57‐
60]. Although the mortality rate of CAP in the community is low at less than 1 %, the impact is significant at a population level in terms of lost days of work, reduced productivity and long term complications [
60,
61]. More data on the burden of disease in the community is needed, as hospital based studies may underestimate the true impact.
At the other end of the severity spectrum, we identified little data on the pneumococcal vaccine serotypes associated with the most severe pneumonia causing intensive care unit admission. Although Van Hoek et al. reported data on the mortality attributable to different serotypes using record linkage, similar data for ICU admission was not available [
39]. Mortality and severe pneumonia requiring ICU admission are not necessarily synonymous as the majority of pneumonia related deaths occur outside the ICU with up to 50 % occurring due to co-morbidities rather than directly due to pneumonia [
61‐
63].
The incidence of CAP, and pneumococcal disease in particular is greatly increased in at risk populations such as patients with COPD, chronic heart failure and immunosuppression [
42]. COPD patients have a greatly increased frequency of pneumonia at 22.4 per 1000 person years due to a combination of reduced local immunity, co-morbidities and the impact of immunosuppression through inhaled corticosteroids [
64,
65]. Additional risk factors identified through analysis of UK health records include diabetes, chronic heart disease, chronic renal disease, asplenia, chronic liver disease, sickle cell disease, HIV, previous stroke, rheumatoid arthritis, Parkinson’s disease, multiple sclerosis, dementia, osteoporosis and malignancy [
66]. IPD data suggests a greatly increased risk of IPD in patients with one or more risk factors, particularly chronic liver disease, COPD and immunosuppression with evidence of “risk stacking” with the incidences greatest in patients with more risk factors, or a combination of increasing age and high risk co-morbidities [
39]. Mortality from IPD is also greatly increased in patients from high risk groups, ranging from a 20 % increase in risk of death among elderly patients with COPD, to an estimated 1000 % increased risk of death among patients aged 16–64 years with chronic liver disease, compared to patients of the same age without liver disease [
39].
The UK population is ageing, and therefore the burden of CAP and pneumococcal disease in general is expected to increase even with the impact of childhood and adult vaccine programmes.
There are important regional variations in the incidence of CAP in the UK. Millet et al. demonstrated a significantly lower incidence of LRTI and CAP in London and the South East of England compared to the North, Yorkshire and the Midlands in both men and women [
67]. This may be partly explained by differences in socioeconomic deprivation which is a major risk factor for CAP. Given the large differences observed in this study between regions, estimates of risk averaged across the whole of the UK should be interpreted in the local context. [
67]
Competing interests
JDC reports grant support from AstraZeneca, Bayer Healthcare, Aradigm corporation and Pfizer Ltd. outside the submitted work. JC and HM are full time employees of Pfizer Ltd, UK. AD and MW declare that they have no competing interests.
Authors’ contributions
JDC conceived the study, supervised the literature review, reviewed papers/abstract and wrote the manuscript. JC contributed to study design, literature review and wrote the manuscript. AD supervised the literature review, reviewed papers and abstracts and wrote the manuscript. MW contributed to study design and wrote the manuscript. HM conceived the study, contributed to study design and wrote the manuscript. All authors read and approved the final manuscript.