- Email: [email protected]
The current burden of pneumococcal disease in England and Wales
Journal of Infection (2006) 52, 37–48
www.elsevierhealth.com/journals/jinf
The current burden of pneumococcal disease in England and Wales A. Melegaroa,*, W.J. Edmundsa, R. Pebodyb, E. Millerb, R. Georgec a
Modelling and Economics Unit, Centre for Infections, 61 Colindale Avenue, Health Protection Agency, London NW9 5EQ, UK b Immunisation Department, Centre for Infections, Health Protection Agency, London NW9 5EQ, UK c Respiratory and Systemic Infection Laboratory, Centre for Infections, Health Protection Agency, London NW9 5EQ, UK Accepted 1 February 2005 Available online 17 March 2005
KEYWORDS Pneumococcus; Pnc disease; IPD; Pneumonia; Otitis media; England and Wales; AOM; CAP
Summary Objective: To evaluate the potential impact of various pneumococcal conjugate vaccination strategies, it is critical to ascertain the pre-vaccination epidemiology and to have a detailed evaluation of the current burden of pneumococcal disease. Method: A variety of national data sources and GP sentinel surveillance systems were used to estimate the incidence, number of hospital admissions, deaths, and GP consultations due to pneumococcal disease in England and Wales. Clinical outcomes included pneumococcal meningitis, bacteraemia, pneumonia and otitis media. A statistical model was used to attribute GP consultation recorded as pneumonia and acute otitis media to specific aetiological causes when these were not recorded. Results: The burden of pneumococcal disease is considerable, with incidence rates of both invasive and non-invasive disease peaking in children (!5 years) and in the elderly (75C years). Around 5800 hospitalisations specifically mentioning Streptococcus pneumoniae are estimated to occur annually in England and Wales, almost 40 000 for lobar pneumonia and over 15 000 for otitis media. There may be an additional 70 000 GP consultations for pneumococcal related community acquire pneumonia and over 630 000 for otitis media. A significant proportion of hospitalisations and GP consultations for pneumococcal disease occur among highrisk groups, with over 80% of hospital admissions reporting more than one diagnosis. Q 2005 The British Infection Society. Published by Elsevier Ltd. All rights reserved.
Introduction
* Corresponding author. Tel.: C44 20 8327 6048. E-mail address: [email protected] (A. Melegaro).
Streptococcus pneumoniae is a bacterial pathogen normally residing in the nasopharynx which causes a wide range of invasive and non-invasive diseases, the most important of which are: meningitis,
0163-4453/$30.00 Q 2005 The British Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jinf.2005.02.008
38 septicaemia, pneumonia and otitis media. The burden of invasive pneumococcal (Pnc) disease (IPD) is high worldwide, with reported incidence rates ranging from 23.2 per 100 000 in the U.S.A.1 (166 in !2 years of age) to around 10 per 100 000 in European countries.2 In developing countries extremely high incidence rates of IPD have been reported among infants (139–224 per 100 000 in !2 years of age)3 as well as among indigenous adult populations (190 per 100 000 in 65C years of age)4 Less serious non-invasive pneumococcal disease (non-IPD) represents most of the pneumococcal conditions and is much more difficult to estimate. Microbiological confirmation of S. pneumoniae infection is both difficult and often not performed; hence the aetiological cause remains unknown. Estimates from previous studies suggest that 15–43% of community-acquired pneumonia (CAP) cases are attributable to pneumococcal infection;5 around 30–35% of acute otitis media (AOM) has been attributed to Pnc.6,7 A pneumococcal conjugate vaccine (PCV) has been proved to be safe and effective against the most serious form of pneumococcal infection8–11 and also moderately effective against AOM12,13 and pneumonia.14,15 Widespread vaccination has been introduced in the U.S.A.16 whereas in the U.K. the vaccine is currently recommended only for children less than 5 years of age with specific high-risk conditions.17 Moreover, since, July 2003, a pneumococcal immunisation program for the healthy elderly with a 23-valent polysaccharide vaccine has also been introduced in England (www.dh.gov.uk) in addition to the previous policy of vaccinating high-risk individuals over 2 years of age. The aim of this work is to estimate the amount of pneumococcal disease that is present in England and Wales in order to provide baseline information for the assessment of the potential benefits that may derive from vaccination.
Methods Data sources Laboratory reports The enhanced surveillance of pneumococcal disease set up jointly between the Communicable Disease Surveillance Centre’s (CDSC) national laboratory reporting scheme and the Respiratory and Systemic Infection Laboratory (RSIL) at the Specialist and Reference Microbiology Division of the Health Protection Agency is the main data source for the ascertainment of the burden of IPD in
A. Melegaro et al. England and Wales.18,19 The system was set up in 1996 in order to improve the estimate of the burden of IPD throughout England and Wales and to gain further information on serotype distribution and disease incidence in different age groups. From this national surveillance system, cases of laboratory confirmed IPD (pneumococcal bacteraemia and meningitis) reported from laboratories around England and Wales from January 1996 to December 2000 were extracted. The extracted data included age, sex, earliest specimen date, serogroups and serotype (when available), antimicrobial susceptibility information (penicillin and erythromycin), region and method of confirmation. Hospital episode statistics (HES) Hospital episode statistics (HES—department of health) (http://www.dh.gov.uk) is a computerised hospital discharge database that covers all National Health Service Hospitals in England. It contains information on individual episodes of illness, together with patient details (age, date of birth, postcode, sex), clinical conditions, number of days spent in the hospital and admissions to intensive care unit (ICU). For each record, seven diagnostic fields are available, in which the condition(s) of each patient are specified using the tenth revision of the International Classification of Disease coding system (ICD-10). All hospital admissions that occurred over the period April 1995–March 2000 which included an occurrence of one of the pneumococcal related ICD-10 codes (Table 1) in Table 1 International classification of diseases (ICD) codes for pneumococcal related disease Definition
ICD-9 code
ICD-10 code
Pneumococcal meningitis Pneumococcal septicaemia Pneumococcal pneumonia Lobar pneumonia, organism unspecified S. pneumoniae as the cause of the disease Non-suppurative otitis media Suppurative and unspecified otitis media Otitis media in diseases classified elsewhere
3201
G001
0382
A403
481
J13X
481
J181
410
B953
3810–3814
H650–H659
3820–3829
H660–H669
n.a.
H670–H678
n.a., no equivalent ICD-9 code was available.
Burden of pneumococcal disease any of the seven diagnostic fields were extracted from the database and cleaned for possible duplicates. This included strictly IPD codes (pneumococcal meningitis and septicaemia), pneumococcal pneumonia code (under which both IPD and non-IPD can be recorded) as well as non-invasive codes such as otitis media and lobar or unspecified pneumonia. Admissions due to lobar pneumonia, organism unspecified (J181), were extracted as pneumococcus is considered one of the major cause of lobar pneumonia.20 GP consultations The weekly number of consultations to general practitioners for a diagnosis of either pneumonia/pneumonitis or AOM and population at risk was obtained form the Royal College of General Practitioners (RCGP) weekly returns system.21 The bi-weekly number of pneumonia and AOM was calculated for the period 1996–2000 and used as outcome variable in a multivariate regression analysis (see later). Additional information was available from the morbidity survey of general practices fourth edition (MSGP4), which is a 1-year prospective survey of around 500 000 patients attending GP practices undertaken in 1991/1992.22 This database contains information on people attending the practices in the survey, such as socio-economic status, details on family composition, individual characteristics (i.e. smoking status) and the history of the patient’s health status during the year of the survey (recorded using the ICD-9 coding system).
Data analysis The total number of pneumococcal confirmed cases identified by laboratory reports and HES for the period 1995–2000 was derived and comparisons were made between the data sources. The seasonal patterns were investigated from both laboratory reports and hospitalisations data. Multiple linear regression analysis was used, following the technique developed by Ryan and colleagues,23 to ascertain the underlying aetiologies of CAP and AOM GP consultations and to estimate the proportion attributable to S. pneumoniae infection. The observed seasonality in agents that could potentially cause pneumonia and AOM was compared to the number of consultations reported for the two clinical outcomes over the same time period. The bi-weekly number of consultations for the period 1996–2000 was used as the dependent variable of the regression. The independent variables were the bi-weekly number
39 of national laboratory reports for the following agents associated with CAP and AOM: S. pneumoniae, Haemophilus influenzae, adenovirus, influenza A (Flu A), influenza B (Flu B), Klebsiella pneumoniae, Legionella pneumophila, Moraxella catarrhalis. Mycoplasma pneumoniae, parainfluenza, Pneumocistis carinii, rhinovirus, RSV, Chlamydia pneumoniae and Chlamydia psittaci. A backward stepwise regression was performed using STATA 6.0. This technique is based on a variable selection procedure in which all variables are entered into the equation and then sequentially removed if non-significant (P-valueO0.05). The importance of two-way interaction terms for the explanation of the dependent variables was also investigated, assessing the significance (P!0.01) of each interaction term between the organisms in the final model. The validity of the model was assessed (R2) and the impact of changes in the model specification was investigated. Age-specific incidence rates were calculated from laboratory confirmed IPD cases (pneumococcal bacteraemia, meningitis and other conditions where pneumococcus was isolated from a normally sterile site) extracting the number of invasive isolates reported to the national enhanced surveillance system for the period 1996–2000 and using Office for National Statistics (ONS) population estimates for the appropriate years and age groups. Similarly, hospitalisation rates were derived for any pneumococcal confirmed hospitalisations, for lobar pneumonia and for otitis media (organism unspecified) using ONS population estimates for England. The serotype distribution of IPD cases was derived for each year from 1996 to 2000 examining the isolate characteristics. The proportions of invasive infections that were caused by the serotypes contained in the different formulations of pneumococcal conjugate and polysaccharide vaccines were estimated for different age groups. Case-fatality rates (CFR) from 1995/1996 to 1999/2000 were estimated by dividing the number of deaths among hospitalised patients by the number of pneumococcal-related admissions (HES). Due to the presence of other serious comorbidities in many patients (especially the elderly), it was not possible to ascertain with confidence whether the patient had died because of the pneumococcal infection or, simply, with a pneumococcal infection. For this reason we estimated CFR for patients reporting a pneumococcal code in the first diagnostic code, in the first three, or in any of the seven diagnostic fields. Assuming that the first diagnosis is more likely to report the underlying cause of the hospitalisation, we assumed that CFR calculated on the first diagnosis only were
40 more likely to be related to S. pneumoniae infection. From ONS, the number of pneumococcal related deaths (i.e. pneumococcal code as the underlying cause of death) for the same period were also derived and comparisons were made with HES figures. Average age-specific episode and consultation rates for CAP and AOM were estimated using MSGP4 data and compared to the rates provided by RCGP. Moreover, from MSGP4 data, episode rates of both CAP and AOM due to pneumococcal infection were also derived (using the proportions estimated with the multivariate model) for high and low-risk individuals, stratifying the patients according to whether they had consulted their GP over the 1year observation period for one or more of the following chronic medical conditions: diabetes mellitus, chronic renal, hepatic, or pulmonary disease, alcoholism, or neoplastic disease, chronic immunosuppression. These represent the high-risk group for which the polysaccharide pneumococcal vaccine has been in the past17 and is currently recommended (http://www.dh.gov.uk).
Results The number of laboratory reports of S. pneumoniae infection for each year since 1990 was derived and showed an increase in the annual figure in 1996, when the enhanced surveillance system was implemented (data not shown). Since, then, an annual average of around 4828 IPD cases had been reported (range: 4692–5159). Of these, an average of 343 (range: 314–394) were identified as cases of pneumococcal meningitis (pneumococci isolated from the CSF) whereas the remaining isolates were obtained from bacteraemic patients with pneumococcus isolated from the blood or other sterile body sites. A marked seasonal pattern in the presentation of IPD with preponderance in the winter months was observed from both weekly laboratory reports and hospital admissions due to confirmed pneumococcal infection. Reported IPD cases reach a low during August and peak in December and January when reporting rates are 3–5 times higher. Fig. 1 shows a comparison between laboratory weekly reports of IPD and hospitalisations due to all pneumococcal confirmed infection and also admissions with a diagnosis of lobar pneumonia, organism unspecified. The almost overlapping weekly pattern observed when considering all confirmed pneumococcus patients suggests that hospital episodes with an ICD-10 code mentioning. S. pneumoniae are an
A. Melegaro et al. accurate reflection of IPD incidence as reported in the enhanced surveillance system. Moreover, a similar seasonal pattern also characterises lobar pneumonia admissions though the numbers are much higher. A multivariate regression analysis was performed to assess the proportion of GP consultations for CAP and AOM that are likely to be pneumococcal in origin. Three models were estimated for both CAP and AOM. We found no blologically plausible statistically significant interaction terms. Tables 2 and 3 provide estimates of the proportion of CAP and AOM attributable to each of the different agents. The variables laboratory reports for S. pneumoniae, adenovirus, M. pneumoniae, RSV and Flu A stayed in the model for unspecified pneumonia cases (model CAP1). The adjusted R2 indicates that 67% of the variation in the weekly number of unspecified cases was explained by the model. Using this model, it was estimated that 31% of unspecified pneumonia could be attributed to S. pneumoniae, 38% to adenovirus, 17% to M. pneumoniae, 6% to RSV and 8% to Flu A. A slightly more adequate fit (R2Z68%) was achieved by model CAP2, where the variable Adenovirus was dropped from the model and ‘other organism’ was added to consider all the other possible infectious and noninfectious causes of pneumonia that were not included in the analysis. Using CAP2, the proportion of unspecified pneumonia attributed to S. pneumoniae was 26%, 14% to M. pneumoniae, 8% to RSV, 11% to H. influenzae, 10% to Flu A and 31% to ‘other organism’. H. influenzae was kept in the model, though the P-value was less significant, but still within the acceptable range (P-valueZ0.019). The sensitivity of model results to the variable H. influenzae was assessed in CAP3. Using this model, it was estimated that the proportion of pneumonia cases attributable to S. pneumoniae was 29%, 12% to M. pneumoniae, 6% to RSV, 8% to Flu A and 44% to ‘other organism’. Similarly to the previous models, the latter explained 67% for the data variablilty. For the unspecified acute otitis media episodes, the variables S. pneumoniae, Flu B, M. pneumoniae, RSV, rhinovirus and adenovirus stayed in the model (model AOM1), explaining 68% of the data variability. From this model, S. pneumoniae was estimated to cause 23% of AOM episodes, 4% were due to Flu B, 12% to M. pneumoniae, 8% to RSV, 18% to rhinovirus and 36% to adenovirus. Dropping adenovirus from the model reduced the fit to 63% and 58% when the variable ‘other organism’ was, respectively, included and excluded from the model specification. In summary, from the six models just described, S. pneumoniae was estimated to cause from 26%
Burden of pneumococcal disease
41
Figure 1 Weekly reports of pneumococcal disease (IPD and lobar pneumonia) from the national enhanced surveillance system and from hospital admissions (HES). Pnc confirmed hospitilisations include the following ICD-10 codes: G001, A403, J13, B953 reported in any of the seven diagnostic fields.
[95% CI: 10–37%] to 31% [95% CI: 21–39%] of pneumonia consultations, whereas from 23% [95% CI: 14–29%] to 42% [95% CI: 36–47%] of AOM consultations. Figs. 2 and 3 show the observed and estimated seasonal patterns and the contribution of each pathogen. Although not much variation was found in the proportion of CAP attributable to pneumococcal infection in the three models, CAP2 was the one that showed the best fit (highest adjusted R-square) and, consequently, was chosen as our final model. The proportion of GP consultations for CAP attributable, to S. pneumoniae was thus assumed to be 26% in subsequent calculations.
Similarly, the proportion of AOM attributable to pneumococcus was assumed to be 23% according to model AOM1, this being the model with the best fit. Age-specific incidence rates of pneumococcal meningitis and other invasive pneumococcal disease were calculated using the overall number of isolates reported over the period 1996–2000 and the annual population figures for England and Wales, produced by the Office of National Statistics (ONS). Incidence rates are reported in Table 3 and show, as reported in previous studies,24,25 IPD rates peaking in infants and among the elderly. Incidence rates for pneumococcal meningitis are highest in infants
Table 2 Estimation of the proportion of unspecified CAP and AOM attributable to specific pathogens and sensitivity results to model specification Pathogen
Model CAP1 (%)a
Model CAP2— final model (%)
Model CAP3 (%)
Model AOM1— final model (%)a
Model AOM2 (%)
Model AOM3 (%)a
S. pneumoniae M. pneumoniae RSV H. influenzae Influenza A Influenza B Adenovirus Rhinovirus Other organisms Adjusted R2 square
31 17 6 – 8 – 38 – – 67
26 14 8 11 10 – – – 31 68
29 12 6 – 8 – – – 44 67
23 12 8 – – 4 36 18 – 68
27 10 7 – – 4 – 25 27 63
42 15 5 – – 3 – 35 – 58
a
Regression through the origin (no-intercept model).
Other IPD
140.5 (128.4–153.4) 255.8 (249.8–261.9) 131.7 (129.7–133.8) 39.7 (38.8–40.7) 17.2 (16.6–17.9) 19 (18.8–19.3) 45.8 (45.3–46.4) 146.5 (144.9–148.2) 404.7 (401.7–407.6) 73.7 (73.4–74.1) 50.9 (43.7–58.9) 82.8 (79.4–86.3) 16.3 (15.6–17.1) 3.7 (3.4–4) 2 (1.8–2.3) 3.5 (3.4–3.6) 8.6 (8.4–8.9) 23.1 (22.4–23.7) 44.8 (43.8–45.8) 10.8 (10.7–11) 75.3 (65.2–86.6) 38.6 (36.4–41) 11.6 (11–12.2) 2 (1.8–2.2) 1 (0.9–1.2) 3.1 (2.9–3.2) 7.3 (7–7.5) 20.2 (19.6–20.8) 44.3 (43.4–45.2) 9.2 (9.1–9.3) 59.7 (50.8–69.8) 23.4 (21.7–25.2) 9.9 (9.4–10.4) 1.8 (1.6–2) 0.8 (0.7–1) 2.8 (2.7–2.9) 6.7 (6.5–6.9) 19.3 (18.7–19.9) 43.6 (42.7–44.5) 8.5 (8.4–8.7) 15.6 (11.2–21.1) 15.3 (13.9–16.8) 1.7 (1.5–1.9) 0.2 (0.1–0.3) 0.2 (0.1–0.3) 0.2 (0.2–0.3) 0.6 (0.5–0.7) 0.9 (0.8–1) 0.7 (0.6–0.8) 0.7 (0.6–0.7) !1 month 1–11 months 1–4 years 5–9 years 10–14 years 15–44 years 45–64 years 65–74 years 75C years Total
Lobar pneumonia, organism unspecified Pnc confirmed infection Pnc meningitis
All IPD
Hospitalisations (HES) Laboratory reports (CDSC/RSIL) Age group
Average annual incidence rate (and 95% CI) per 100 000 population of pneumococcal related disease Table 3
30.2 (16.9–49.9) 345 (329.7–361) 406.8 (398.8–414.8) 236.2 (231–241.6) 52.1 (49.6–54.6) 13.2 (12.7–13.7) 16 (15.3–16.7) 15.9 (14.7–17.1) 13.5 (12.4–14.8) 54.2 (53.6–54.8)
A. Melegaro et al. Otitis media, organism unspecified
42
(15 per 100 000 population, 95% CI: 11–21) and remain low in all other age groups. No significant change was observed over the time period. As part of the enhanced surveillance of pneumococcal disease, the number of invasive isolates with serotype information has increased since 1996, rising from 1696 (36%) in 1996 to 2351 (50%) in 2000. An average of 31 different serotypes were detected annually, with the 15 most prevalent types counting for 93% of the overall number. The derived proportions of IPD in England and Wales that are caused by the serotypes contained in the available or potentially available formulations of vaccines (7, 9, 11, 23-valent) are given in Table 4 assuming cross protection within serogroups occurs (i.e. vaccine containing serotype 6B also confers protection against serotype 6A). Eighty-two percent of the 1815 IPD cases reported in children 0–5 years of age over the period 1996–2000 were caused by serotypes contained in the 7-valent pneumococcal conjugate vaccine. Information on serotype distribution and antimicrobial susceptibilities amongst the IPD isolates will be reported elsewhere. Overall, HES data show a similar incidence of admission due to confirmed pneumococcal infection (10.8 vs. 9.2) (Table 3). Otitis media hospitalisation rates are also shown for the different age groups and the incidence rate peaks in the 1–4 years of age, although the incidence remains high in all age groups for children less than 10 years old. Among all hospitalised patients, 81% were recorded with more than one diagnostic code, 54% with more than two, and 38% had more than three diagnoses. Co-morbidities were clearly directly associated with the age of the patients, being much higher for older age groups (Fig. 4). S. pneumoniae was present in the first diagnosis in 72% of the patients, and the remainder was distributed among neoplasm cases (11%), disease of the circulatory system (23%), disease of the respiratory system (20%), and various other conditions (46%). In Fig. 5 the age-specific proportions of hospitalisations that reported a non-pneumococcal diagnosis in the first diagnostic fields are shown. In the second and further diagnoses the proportion of pneumococcus related codes were of course lower, being 15% in the second diagnosis, and 7% in the third (co-morbidities, respectively, 60 and 46%, the remaining being blank fields). Age-specific episode and consultation rates for both CAP and AOM derived from MSGP4 data22 were more than twice as high as those routinely produced by the RCGP system21 (GP consultation rates for CAP: 679 vs. 266 per 100 000 population per year). Higher episode rates were found in patients with
Burden of pneumococcal disease
43
Figure 2 Comparison of observed bi-weekly number of pneumonia GP episodes with estimated number derived from final best fitting model (Table 2—model CAP2: % CAP due to S. pneumoniae is 26%).
high-risk conditions. In Fig. 6 pneumococcal pneumonia and AOM age-specific episode rates per 100 000 populations per year, calculated using MSGP4 data, are reported for high and low risk individuals. For otitis media, higher rates are observed in children 10 years old, and rates almost twice as high are reported for children !5 years of age when they also have other high-risk conditions.
A similar pattern is observed for pneumonia episodes although, in this case, higher rates are reported at a much later age (60C years). The total number of deaths for the period April 1995–2000 was derived from both ONS and HES data (Table 5), extracting those records for which one of the pneumococcal codes was either specified as the underlying cause of death (ONS) or recorded as
Figure 3 Comparison of observed bi-weekly number of otitis media GP episodes with estimated number derived from final best fitting model (Table 3—model AOMI: % of AOM due to S. pneumoniae is 23%).
44
A. Melegaro et al.
Table 4 Number and percent of IPD caused by the 7, 9, 11 or 23 serotypes contained in the formulations of the available vaccines over the period 1996–2000 7-valenta
Age group
!1 month 1–11 months 1–4 years 5–9 years 10–14 years 15–44 years 45–64 years 65–74 years 75C years Unknown Total a b c
9-valentb
11-valentc
23-valent
Cases
%
Cases
%
Cases
%
Cases
%
59 587 845 88 29 601 884 960 2008 250 6311
47 81 88 53 43 47 55 62 68 62 64
81 606 892 126 44 863 1026 1043 2134 277 7092
64 84 93 76 66 67 64 68 72 69 72
96 647 908 137 47 977 1177 1204 2364 311 7868
76 89 94 83 70 76 74 78 80 77 80
117 708 944 160 66 1249 1544 1495 2855 386 9524
93 98 98 96 99 98 97 97 97 96 97
Serotypes of the 7-valent vaccine: 4, 6B, 9V, 14, 18C, 19F, 23F. Nine-valent: 7-valent serotypes C1, 5. Eleven-valent; 9-valent C3, 7F.
primary diagnoses of the patient (HES). Deaths reporting a pneumococcal related ICD-10 code in any of the seven diagnostic fields were also extracted from HES. The number of ONS registered deaths from invasive pneumococcal diseases (Pnc meningitis and septicaemia) ranged between 72 and 117 per year in the period 1995–2000. The number of ONS registered deaths from lobar (pneumococcal) pneumonia was in the range of 1807–2064. These figures represent around 70% of the total number of deaths reported in HES for patients admitted with a pneumococcal meningitis or septicaemia code in the first diagnostic field though only around 40% of cases that died in the hospital and for which the pneumococcal related diagnosis was extracted from any of the seven diagnostic fields. The lower number of deaths reported in ONS
Figure 4 2000).
may represent cases that die with a pneumococcal related condition but for which the underlying cause of death (as reported in the death certificate) was non-pneumococcal specific. In Fig. 7 age-specific case fatality ratios derived from HES are shown for invasive and non-invasive pneumococcal conditions and show a marked rise in 15C years of age. Case-fatality rates of around 40% are reached among elderly hospitalised patients reporting strictly invasive pnuemococcal codes (Pnc meningitis and septicaemia) as their first diagnosis. Case-fatality ratios were calculated also for hospitalised patients reporting a pneumococcal code in the first three or in any diagnostic fields and little variation was shown in the proportion of individuals who died during the hospital admission (all ages: 13%; in 65C: 21–28%).
Proportion of hospital admissions with 1, 2–4 and 5C diagnoses reported during a hospital stay (HES 1995–
Burden of pneumococcal disease
45
Figure 5 Proportion of hospitalisations with a non-pneumococcal diagnosis reported in the first diagnostic field (HES 1995–2000).
The burden of pneumococcal disease in England and Wales was quantified using hospitalisation rates and GP consultation rates just provided (Table 6). The number of consultations to the general practitioner for CAP and AOM attributable to the pneumococcus was estimated using the results of the multivariate regression analysis and the average consultation rate derived from MSGP4 and RCGP data. Otitis media represents a major component of the overall burden of disease with an estimated 630 470 consultations to the general practitioner per year. Of these, around 60% occurs in children less than 10 years of age. Similarly, 64% of otitis media
Figure 6
hospitalisation occurs in children as well as around 40% of pneumococcal meningitis cases. The burden of pneumococcal disease in the elderly is represented mostly by pneumococcal pneumonia and septicaemia. Forty-nine and 58% of hospitalisations for, respectively, pneumococcal confirmed disease and lobar pneumonia occurs in 65C year of age.
Discussion In the light of the current discussions on whether or not to introduce widespread vaccination of infants
Pneumococcal pneumonia and otitis media episode rates by age in high and low risk individuals (MSGP4 data).
46 Table 5 Year
95/96 96/97 97/98 98/99 99/00
A. Melegaro et al. Number of deaths from pneumococcal disease ONSa
HES—first diagnosisb
HES—any diagnosisb
Lobar (Pnc) pneumonia
IPD
Lobar (Pnc) pneumonia
IPD
Lobar (Pnc) pneumonia
IPD
1907 2064 1807 1991 2059
94 117 99 104 72
2468 2841 2738 3471 3696
133 140 132 114 127
4553 5286 5323 6357 6853
219 268 223 229 234
a The ICD-9 coding system was used for ONS records until 2000 data (lobar, pneumococcal pneumonia codeZ481, IPDZ0382 (Pnc septicaemia)C3201 (Pnc meningitis). b The ICD-10 coding system was used in HES since 1995 (lobar, pneumococcal pneumonia codeZJ181 (lobar pneumonia organism unspecified)CJ13 (pneumococcal pneumonia), IPDZA403 (Pnc septicaemis)CG001 (Pnc meningitis).
with the pneumococcal conjugate vaccine, baseline information on the actual burden of disease is essential in order to perform a reliable economic evaluation and to provide baseline information against which future control programs can be measured. Although several national data sources are available to gain insights into the burden of pneumococcal disease in the U.K., uncertainties are nevertheless present due, mainly, to the fact that the aetiological agent for clinical conditions such as pneumonia and otitis media is not determined for the majority of the cases. A number of studies have attempted to estimate the burden of Pnc disease, by applying the study results of more sensitive alternative laboratory diagnostic methods (e.g. PCR, antigen detection) to unspecified syndromic surveillance data,26 or by reviewing individual hospital admissions to estimate the proportion likely to be attributable to S. pneumoniae infection.20 In this study, we quantified the amount of pneumococcal disease, invasive and non-invasive,
that is present in England and Wales, using both national surveillance database and multivariate regression techniques based on the underlying seasonality of pneumococcal disease reports to estimate the proportion of GP consultations for both pneumonia and AOM that is attributable to Pnc infection. The strength of this modelling technique has been shown in the past when similar questions were addressed for rotavirus23 and hospital admissions for RSV.27 As was shown in previous studies18,28–30 the estimated burden of pneumococcal disease in England and Wales is considerable, with incidence rates of both IPD and non-IPD peaking in children (!5 years) and in the elderly (75C years). Morbidity appears to be related to the presence of high-risk conditions, with over 80% of hospitalised patients presenting with more than one diagnosis. In particular, neoplasm and diseases of the respiratory and circulatory systems are common co-morbid conditions associated with invasive Pnc disease.
Figure 7 Case-fatality rate by age of Pnc meningitis and septicaemia (G001 and A403), Pnc pneumonia (J13) and lobar pneumonia, organism unspecified (J181)—HES (primary diagnosis).
Burden of pneumococcal disease
47
Table 6 Burden of pneumococcal related disease in England and Wales (ONS England and Wales 2002 population estimates) Age group
!1 years 1–4 years 5–9 years 10–14 years 15–44 years 45–64 years 65–74 years 75C years Total
Hospitalisationsa
GP consultationsb
Deathsc
Pnc confirmed
Lobar pneumonia
OM
Pnc pneumonia
Pnc OM
Pnc confirmed
Lobar pneumonia
488 392 119 70 772 1096 1021 1808 5766
1500 3164 1285 592 4171 5826 6486 16 324 39 349
1055 4881 4007 1180 2074 1492 488 304 15 565
933 2665 1568 879 12 916 14 520 10 676 26 152 70 309
53 311 223 065 98 732 25 294 142 599 60 438 17 762 9268 630 470
7 5 1 1 32 143 119 421 730
5 4 2 4 41 303 762 3392 4513
a
ICD-10 codes—Pnc confirmed: G001, A403, J13, B953; lobar pneumonia: J181; OM: H65–H67. Average MSGP4 and RCGP rates (from the multivariate regression analysis: 26 and 23% of, respectively, pneumonia and OM consultations are attributable to S. pneumoniae). c Age-specific CFR shown in Fig. 7 (HES primary diagnosis only) is applied to the number of hospitalisations shown in this table. b
Similarly, non-invasive disease is more common in those with underlying illnesses, with CAP and AOM occurring almost twice as frequently in high-risk individuals than in the non-high-risk groups. A polysaccharide and a conjugate vaccine, nevertheless, are available and currently recommended in England and Wales for high-risk individuals and more recently, polysaccharide vaccine for all those aged 65 years or over. In this latter group, the vaccine appears to be effective in reducing morbidity due to IPD, though discussions are still ongoing on the protection it confers among certain at-risk categories which appear to be highly vulnerable to pneumococcal disease.31 This paper demonstrates a large burden due to pneumococcal disease in both the elderly and high-risk categories. Ongoing surveillance will be critical to ascertain the impact of these programs. In addition, the 7-valent pneumococcal conjugate vaccine also seems to reduce S. pneumoniae colonisation and have an impact on the overall transmission of the organism in the population, potentially providing protection to unvaccinated populations (i.e. herd immunity).10,11,32 However, the potential threat of serotype replacement as a consequence of vaccination, as seems to have been observed in the U.S.A.9 and in Finland,12 will require prospective monitoring for any changes in the serotype distribution of pneumococcal disease patients and, ideally asymptomatic carriers. These factors will have to be combined to disease burden data to assist decision-makers in the most appropriate intervention strategy with the pneumococcal conjugate vaccine in countries considering introduction. Moreover, results from cost-effectiveness analyses showed that indirect effects such as herd
immunity and serotype replacement may have a crucial impact on the economic acceptability of alternative programs.33,34 The serotype diversity of S. pneumoniae and its many clinical manifestations make it hard to answer questions related to the short as well as long-term effects that vaccination may have at the population level. Here, we have attempted to give some baseline information on the epidemiology of both invasive and non-invasive pneumococcal condition and on the potential impact of the available pneumococcal vaccines. Ongoing surveillance is in place to assess any impact on disease epidemiology of the current high-risk strategy in England and Wales and to assist decisions on future vaccination strategy.
Acknowledgements We would like to thank Dr Douglas Fleming of the Royal College of General Practitioners Scheme for providing consultation rates for CAP and AOM and the DH for hospitalisation records. We also want to thank Pauline Kaye and Usha Gungabissoon for data extraction. The MSGP4 are Crown Copyright and are reproduced with permission. The study was funded by a grant from the EU (QLG4-CT-2000-00640).
References 1. Robinson K, Baugham W, Rothrock G, Barret NL, Pass M, Lexau C, et al. Epidemiology of invasive Streptococcus
48
2.
3.
4.
5. 6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
A. Melegaro et al. pneumoniae infections in the United States, 1995–1998. Opportunities for prevention in the conjugate vaccine era. JAMA 2001;285(13):1729–35. Cartwright K. Pneumococcal disease in Western Europe: burden of disease, antibiotic resistance and management. Eur J Pediatr 2002;161:188–95. Usen S, Adegbola R, Mulholland K, Jaffar S, Hilton S, Oparaugo A, et al. Epidemiology of invasive pneumococcal disease in the Western Region, The Gambia. Pediatr Infect Dis J 1998;17(1):23–8. Watt JP, O’Brien KL, Benin AL, Whitney CG, Robinson K, Parkinson AJ, et al. Invasive pneumococcal disease among Navajo adults, 1989–1998. Clin Infect Dis 2004;38(4): 496–501. Brown PD, Lerner SA. Community-acquired pneumonia. Lancet 1998;352:1295–302. Pelton SI. Acute otitis media in the era of effective pneumococcal conjugate vaccine: will new pathogens emerge? Vaccine 2001;19:S96–S9. Jacobs MR, Dagan R, Appelbaum PC, Burh DJ. Prevalence of antimicrobial-resistant pathogens in middle ear fluid: multinational study of 917 children with acute otitis media. Antimicrob Agents Chemother 1998;42(3):589–95. Black SB, Shinefield HR, Fireman B, Lewis E, Ray P, Hansen JR, et al. Efficacy, safety and immunogenicity of heptavalent pneumococcal conjugate vaccine in children. Pediatr Infect Dis J 2000;19:187–95. Kaplan SL, Mason Jr EO, Wald ER, Schutze GE, Bradley JS, Tan TQ, et al. Decrease of invasive pneumococcal infections in children among 8 children’s hospitals in the United States after the introduction of the 7-valent pneumococcal conjugate vaccine. Pediatrics 2004;113(3):443–9. Black S, Shinefield H, Baxter R, Austrian R, Bracken L, Hansen J, et al. Postlicensure surveillance for pneumococcal invasive disease after use of heptavalent pneumococcal conjugate vaccine in Northern California Kaiser Permanente. Pediatr Infect Dis J 2004;23(6):485–9. Whitney CG, Farley MM, Hadler J, Harrison LH, Bennett NM, Lynfield R, et al. Decline in invasive pneumococcal disease after the introduction of protein-polysaccharide conjugate vaccine. N Engl J Med 2003;348(18):1737–46. Eskola J, Kilpi T, Palmu A, Jokinen J, Haapakoski J, Herva E, et al. Efficacy of a pneumococcal conjugate vaccine against acute otitis media. N Engl J Med 2001;344(6):403–9. Fireman B, Black SB, Shinefield HR, Lee J, Lewis E, Ray P. Impact of the pneumococcal conjugate vaccine on otitis media. Pediatr Infect Dis J 2003;22(1):10–16. Black SB, Shinefield HR, Ling S, Hansen J, Fireman B, Spring D, et al. Effectiveness of heptavalent pneumococcal conjugate vaccine in children younger than five years of age for prevention of pneumonia. Pediatr Infect Dis J 2002; 21(9):810–5. Poehling KA, Lafleur BJ, Szilagyi PG, Edwards KM, Mitchel E, Barth R, et al. Population-based impact of pneumococcal conjugate vaccine in young children. Pediatrics 2004; 114(3):755–61. CDC. Prevention of pneumococcal disease: recommendations of the advisory committee on immunisation practices (ACIP). MMWR 1997;46(RR-08):1–24.
17. [www.dh.gov.uk] Salisbury D, Begg N. Immunisation against infectious disease. London: HMSO; 1996. 18. Miller E, Waight P, Efstratiou A, Brisson M, Johnson A, George R. Epidemiology of invasive and other pneumococcal disease in children in England and Wales 1996–1998. Acta Paediatr Suppl 2000;435:11–16. 19. George R, Melegaro A. Invasive pneumococcal infection England and Wales, 1999. CDR Weekly 2001;11(21):6–12. 20. Djuretic T, Ryan M, Miller E, Fairley CK, Goldblatt D. Hospital admission in children due to pneumococcal pneumonia in England. J Infect 1998;37:54–8. 21. Fleming DM. Weekly returns service of the Royal College of General Practitioners. Commun Dis Public Health 1999;2(2): 96–100. 22. McCormick A, Fleming D, Charlton J. Morbidity statistics from general practice, Fourth National Survey 1991–1992. London: Office of Population Census and Survey (Series MB5:3), HMSO: 1995. 23. Ryan M, Ramsay M, Brown D, Gay N, Farrington C, Wall P. Hospital admissions attributable to rotavirus infection in England and Wales. J Infect Dis 1996;174(Suppl 1):S12–S18. 24. Smith M, Stuart J, Andrews N, Telfer Brunton W, Cartwright K. Invasive pneumococcal infection in South and West England. Epidemiol Infect 1998;120:117–23. 25. Laurichesse H, Grimaud O, Waight P, Johnson AP, George R, Miller E. Pneumococcal bacteremia and meningitis in England and Wales, 1993 to 1995. Commun Dis Public Health 1998;1:22–7. 26. McIntosh EDG, Booy R. Invasive pneumococcal disease in England and Wales: what is the true burden and what is the potential for prevention using 7 valent pneumococcal conjugate vaccine? Arch Dis Child 2002;86:403–6. 27. Muller-Pebody B, Edmunds W, Zambon M, Gay N, Crowcroft N. Contribution of RSV to bronchiolitis and pneumonia-associated hospitalizations in English children, April 1995–March 1998. Epidemiol Infect 2002;129:99–106. 28. Smith MD, Stuart J, Andrews NJ, Telfer Brunton WA, Cartwright KA. Invasive pneumococcal infection in South and West England. Epidemiol Infect 1998;120:117–23. 29. Sleeman K, Knox K, George R, Miller E, Waight P, Griffiths D, et al. Invasive pneumococcal disease in England and Wales: vaccination implications. J Infect Dis 2001;183:239–46. 30. Kyaw MH, Clarke S, Jones IG, Campbell H. Incidence of invasive pneumococcal disease in Scotland, 1988–99. Epidemiol Infect 2002;128:139–47. 31. Melegaro A, Edmunds WJ. The 23-valent pneumococcal polysaccharide vaccine/Part I—efficacy of PPV in the elderly: a comparison of meta-analyses. Eur J Epidemiol 2004;19(4): 353–63. 32. Givon-Lavi N, Fraser D, Dagan R. Vaccination of day-care center attendees reduces carriage of Streptococcus pneumoniae among their younger siblings. Pediatr Infect Dis J 2003;22(6):524–32. 33. Melegaro A, Edmunds WJ. Cost effectiveness analysis of pneumococcal conjugate vaccination in England and Wales. Vaccine 2004;22(31–33):4203–14. 34. McIntosh ED. Cost-effectiveness studies of pneumococcal conjugate vaccines. Expert Rev Vaccines 2004;3(4):433–42.
www.elsevierhealth.com/journals/jinf
The current burden of pneumococcal disease in England and Wales A. Melegaroa,*, W.J. Edmundsa, R. Pebodyb, E. Millerb, R. Georgec a
Modelling and Economics Unit, Centre for Infections, 61 Colindale Avenue, Health Protection Agency, London NW9 5EQ, UK b Immunisation Department, Centre for Infections, Health Protection Agency, London NW9 5EQ, UK c Respiratory and Systemic Infection Laboratory, Centre for Infections, Health Protection Agency, London NW9 5EQ, UK Accepted 1 February 2005 Available online 17 March 2005
KEYWORDS Pneumococcus; Pnc disease; IPD; Pneumonia; Otitis media; England and Wales; AOM; CAP
Summary Objective: To evaluate the potential impact of various pneumococcal conjugate vaccination strategies, it is critical to ascertain the pre-vaccination epidemiology and to have a detailed evaluation of the current burden of pneumococcal disease. Method: A variety of national data sources and GP sentinel surveillance systems were used to estimate the incidence, number of hospital admissions, deaths, and GP consultations due to pneumococcal disease in England and Wales. Clinical outcomes included pneumococcal meningitis, bacteraemia, pneumonia and otitis media. A statistical model was used to attribute GP consultation recorded as pneumonia and acute otitis media to specific aetiological causes when these were not recorded. Results: The burden of pneumococcal disease is considerable, with incidence rates of both invasive and non-invasive disease peaking in children (!5 years) and in the elderly (75C years). Around 5800 hospitalisations specifically mentioning Streptococcus pneumoniae are estimated to occur annually in England and Wales, almost 40 000 for lobar pneumonia and over 15 000 for otitis media. There may be an additional 70 000 GP consultations for pneumococcal related community acquire pneumonia and over 630 000 for otitis media. A significant proportion of hospitalisations and GP consultations for pneumococcal disease occur among highrisk groups, with over 80% of hospital admissions reporting more than one diagnosis. Q 2005 The British Infection Society. Published by Elsevier Ltd. All rights reserved.
Introduction
* Corresponding author. Tel.: C44 20 8327 6048. E-mail address: [email protected] (A. Melegaro).
Streptococcus pneumoniae is a bacterial pathogen normally residing in the nasopharynx which causes a wide range of invasive and non-invasive diseases, the most important of which are: meningitis,
0163-4453/$30.00 Q 2005 The British Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jinf.2005.02.008
38 septicaemia, pneumonia and otitis media. The burden of invasive pneumococcal (Pnc) disease (IPD) is high worldwide, with reported incidence rates ranging from 23.2 per 100 000 in the U.S.A.1 (166 in !2 years of age) to around 10 per 100 000 in European countries.2 In developing countries extremely high incidence rates of IPD have been reported among infants (139–224 per 100 000 in !2 years of age)3 as well as among indigenous adult populations (190 per 100 000 in 65C years of age)4 Less serious non-invasive pneumococcal disease (non-IPD) represents most of the pneumococcal conditions and is much more difficult to estimate. Microbiological confirmation of S. pneumoniae infection is both difficult and often not performed; hence the aetiological cause remains unknown. Estimates from previous studies suggest that 15–43% of community-acquired pneumonia (CAP) cases are attributable to pneumococcal infection;5 around 30–35% of acute otitis media (AOM) has been attributed to Pnc.6,7 A pneumococcal conjugate vaccine (PCV) has been proved to be safe and effective against the most serious form of pneumococcal infection8–11 and also moderately effective against AOM12,13 and pneumonia.14,15 Widespread vaccination has been introduced in the U.S.A.16 whereas in the U.K. the vaccine is currently recommended only for children less than 5 years of age with specific high-risk conditions.17 Moreover, since, July 2003, a pneumococcal immunisation program for the healthy elderly with a 23-valent polysaccharide vaccine has also been introduced in England (www.dh.gov.uk) in addition to the previous policy of vaccinating high-risk individuals over 2 years of age. The aim of this work is to estimate the amount of pneumococcal disease that is present in England and Wales in order to provide baseline information for the assessment of the potential benefits that may derive from vaccination.
Methods Data sources Laboratory reports The enhanced surveillance of pneumococcal disease set up jointly between the Communicable Disease Surveillance Centre’s (CDSC) national laboratory reporting scheme and the Respiratory and Systemic Infection Laboratory (RSIL) at the Specialist and Reference Microbiology Division of the Health Protection Agency is the main data source for the ascertainment of the burden of IPD in
A. Melegaro et al. England and Wales.18,19 The system was set up in 1996 in order to improve the estimate of the burden of IPD throughout England and Wales and to gain further information on serotype distribution and disease incidence in different age groups. From this national surveillance system, cases of laboratory confirmed IPD (pneumococcal bacteraemia and meningitis) reported from laboratories around England and Wales from January 1996 to December 2000 were extracted. The extracted data included age, sex, earliest specimen date, serogroups and serotype (when available), antimicrobial susceptibility information (penicillin and erythromycin), region and method of confirmation. Hospital episode statistics (HES) Hospital episode statistics (HES—department of health) (http://www.dh.gov.uk) is a computerised hospital discharge database that covers all National Health Service Hospitals in England. It contains information on individual episodes of illness, together with patient details (age, date of birth, postcode, sex), clinical conditions, number of days spent in the hospital and admissions to intensive care unit (ICU). For each record, seven diagnostic fields are available, in which the condition(s) of each patient are specified using the tenth revision of the International Classification of Disease coding system (ICD-10). All hospital admissions that occurred over the period April 1995–March 2000 which included an occurrence of one of the pneumococcal related ICD-10 codes (Table 1) in Table 1 International classification of diseases (ICD) codes for pneumococcal related disease Definition
ICD-9 code
ICD-10 code
Pneumococcal meningitis Pneumococcal septicaemia Pneumococcal pneumonia Lobar pneumonia, organism unspecified S. pneumoniae as the cause of the disease Non-suppurative otitis media Suppurative and unspecified otitis media Otitis media in diseases classified elsewhere
3201
G001
0382
A403
481
J13X
481
J181
410
B953
3810–3814
H650–H659
3820–3829
H660–H669
n.a.
H670–H678
n.a., no equivalent ICD-9 code was available.
Burden of pneumococcal disease any of the seven diagnostic fields were extracted from the database and cleaned for possible duplicates. This included strictly IPD codes (pneumococcal meningitis and septicaemia), pneumococcal pneumonia code (under which both IPD and non-IPD can be recorded) as well as non-invasive codes such as otitis media and lobar or unspecified pneumonia. Admissions due to lobar pneumonia, organism unspecified (J181), were extracted as pneumococcus is considered one of the major cause of lobar pneumonia.20 GP consultations The weekly number of consultations to general practitioners for a diagnosis of either pneumonia/pneumonitis or AOM and population at risk was obtained form the Royal College of General Practitioners (RCGP) weekly returns system.21 The bi-weekly number of pneumonia and AOM was calculated for the period 1996–2000 and used as outcome variable in a multivariate regression analysis (see later). Additional information was available from the morbidity survey of general practices fourth edition (MSGP4), which is a 1-year prospective survey of around 500 000 patients attending GP practices undertaken in 1991/1992.22 This database contains information on people attending the practices in the survey, such as socio-economic status, details on family composition, individual characteristics (i.e. smoking status) and the history of the patient’s health status during the year of the survey (recorded using the ICD-9 coding system).
Data analysis The total number of pneumococcal confirmed cases identified by laboratory reports and HES for the period 1995–2000 was derived and comparisons were made between the data sources. The seasonal patterns were investigated from both laboratory reports and hospitalisations data. Multiple linear regression analysis was used, following the technique developed by Ryan and colleagues,23 to ascertain the underlying aetiologies of CAP and AOM GP consultations and to estimate the proportion attributable to S. pneumoniae infection. The observed seasonality in agents that could potentially cause pneumonia and AOM was compared to the number of consultations reported for the two clinical outcomes over the same time period. The bi-weekly number of consultations for the period 1996–2000 was used as the dependent variable of the regression. The independent variables were the bi-weekly number
39 of national laboratory reports for the following agents associated with CAP and AOM: S. pneumoniae, Haemophilus influenzae, adenovirus, influenza A (Flu A), influenza B (Flu B), Klebsiella pneumoniae, Legionella pneumophila, Moraxella catarrhalis. Mycoplasma pneumoniae, parainfluenza, Pneumocistis carinii, rhinovirus, RSV, Chlamydia pneumoniae and Chlamydia psittaci. A backward stepwise regression was performed using STATA 6.0. This technique is based on a variable selection procedure in which all variables are entered into the equation and then sequentially removed if non-significant (P-valueO0.05). The importance of two-way interaction terms for the explanation of the dependent variables was also investigated, assessing the significance (P!0.01) of each interaction term between the organisms in the final model. The validity of the model was assessed (R2) and the impact of changes in the model specification was investigated. Age-specific incidence rates were calculated from laboratory confirmed IPD cases (pneumococcal bacteraemia, meningitis and other conditions where pneumococcus was isolated from a normally sterile site) extracting the number of invasive isolates reported to the national enhanced surveillance system for the period 1996–2000 and using Office for National Statistics (ONS) population estimates for the appropriate years and age groups. Similarly, hospitalisation rates were derived for any pneumococcal confirmed hospitalisations, for lobar pneumonia and for otitis media (organism unspecified) using ONS population estimates for England. The serotype distribution of IPD cases was derived for each year from 1996 to 2000 examining the isolate characteristics. The proportions of invasive infections that were caused by the serotypes contained in the different formulations of pneumococcal conjugate and polysaccharide vaccines were estimated for different age groups. Case-fatality rates (CFR) from 1995/1996 to 1999/2000 were estimated by dividing the number of deaths among hospitalised patients by the number of pneumococcal-related admissions (HES). Due to the presence of other serious comorbidities in many patients (especially the elderly), it was not possible to ascertain with confidence whether the patient had died because of the pneumococcal infection or, simply, with a pneumococcal infection. For this reason we estimated CFR for patients reporting a pneumococcal code in the first diagnostic code, in the first three, or in any of the seven diagnostic fields. Assuming that the first diagnosis is more likely to report the underlying cause of the hospitalisation, we assumed that CFR calculated on the first diagnosis only were
40 more likely to be related to S. pneumoniae infection. From ONS, the number of pneumococcal related deaths (i.e. pneumococcal code as the underlying cause of death) for the same period were also derived and comparisons were made with HES figures. Average age-specific episode and consultation rates for CAP and AOM were estimated using MSGP4 data and compared to the rates provided by RCGP. Moreover, from MSGP4 data, episode rates of both CAP and AOM due to pneumococcal infection were also derived (using the proportions estimated with the multivariate model) for high and low-risk individuals, stratifying the patients according to whether they had consulted their GP over the 1year observation period for one or more of the following chronic medical conditions: diabetes mellitus, chronic renal, hepatic, or pulmonary disease, alcoholism, or neoplastic disease, chronic immunosuppression. These represent the high-risk group for which the polysaccharide pneumococcal vaccine has been in the past17 and is currently recommended (http://www.dh.gov.uk).
Results The number of laboratory reports of S. pneumoniae infection for each year since 1990 was derived and showed an increase in the annual figure in 1996, when the enhanced surveillance system was implemented (data not shown). Since, then, an annual average of around 4828 IPD cases had been reported (range: 4692–5159). Of these, an average of 343 (range: 314–394) were identified as cases of pneumococcal meningitis (pneumococci isolated from the CSF) whereas the remaining isolates were obtained from bacteraemic patients with pneumococcus isolated from the blood or other sterile body sites. A marked seasonal pattern in the presentation of IPD with preponderance in the winter months was observed from both weekly laboratory reports and hospital admissions due to confirmed pneumococcal infection. Reported IPD cases reach a low during August and peak in December and January when reporting rates are 3–5 times higher. Fig. 1 shows a comparison between laboratory weekly reports of IPD and hospitalisations due to all pneumococcal confirmed infection and also admissions with a diagnosis of lobar pneumonia, organism unspecified. The almost overlapping weekly pattern observed when considering all confirmed pneumococcus patients suggests that hospital episodes with an ICD-10 code mentioning. S. pneumoniae are an
A. Melegaro et al. accurate reflection of IPD incidence as reported in the enhanced surveillance system. Moreover, a similar seasonal pattern also characterises lobar pneumonia admissions though the numbers are much higher. A multivariate regression analysis was performed to assess the proportion of GP consultations for CAP and AOM that are likely to be pneumococcal in origin. Three models were estimated for both CAP and AOM. We found no blologically plausible statistically significant interaction terms. Tables 2 and 3 provide estimates of the proportion of CAP and AOM attributable to each of the different agents. The variables laboratory reports for S. pneumoniae, adenovirus, M. pneumoniae, RSV and Flu A stayed in the model for unspecified pneumonia cases (model CAP1). The adjusted R2 indicates that 67% of the variation in the weekly number of unspecified cases was explained by the model. Using this model, it was estimated that 31% of unspecified pneumonia could be attributed to S. pneumoniae, 38% to adenovirus, 17% to M. pneumoniae, 6% to RSV and 8% to Flu A. A slightly more adequate fit (R2Z68%) was achieved by model CAP2, where the variable Adenovirus was dropped from the model and ‘other organism’ was added to consider all the other possible infectious and noninfectious causes of pneumonia that were not included in the analysis. Using CAP2, the proportion of unspecified pneumonia attributed to S. pneumoniae was 26%, 14% to M. pneumoniae, 8% to RSV, 11% to H. influenzae, 10% to Flu A and 31% to ‘other organism’. H. influenzae was kept in the model, though the P-value was less significant, but still within the acceptable range (P-valueZ0.019). The sensitivity of model results to the variable H. influenzae was assessed in CAP3. Using this model, it was estimated that the proportion of pneumonia cases attributable to S. pneumoniae was 29%, 12% to M. pneumoniae, 6% to RSV, 8% to Flu A and 44% to ‘other organism’. Similarly to the previous models, the latter explained 67% for the data variablilty. For the unspecified acute otitis media episodes, the variables S. pneumoniae, Flu B, M. pneumoniae, RSV, rhinovirus and adenovirus stayed in the model (model AOM1), explaining 68% of the data variability. From this model, S. pneumoniae was estimated to cause 23% of AOM episodes, 4% were due to Flu B, 12% to M. pneumoniae, 8% to RSV, 18% to rhinovirus and 36% to adenovirus. Dropping adenovirus from the model reduced the fit to 63% and 58% when the variable ‘other organism’ was, respectively, included and excluded from the model specification. In summary, from the six models just described, S. pneumoniae was estimated to cause from 26%
Burden of pneumococcal disease
41
Figure 1 Weekly reports of pneumococcal disease (IPD and lobar pneumonia) from the national enhanced surveillance system and from hospital admissions (HES). Pnc confirmed hospitilisations include the following ICD-10 codes: G001, A403, J13, B953 reported in any of the seven diagnostic fields.
[95% CI: 10–37%] to 31% [95% CI: 21–39%] of pneumonia consultations, whereas from 23% [95% CI: 14–29%] to 42% [95% CI: 36–47%] of AOM consultations. Figs. 2 and 3 show the observed and estimated seasonal patterns and the contribution of each pathogen. Although not much variation was found in the proportion of CAP attributable to pneumococcal infection in the three models, CAP2 was the one that showed the best fit (highest adjusted R-square) and, consequently, was chosen as our final model. The proportion of GP consultations for CAP attributable, to S. pneumoniae was thus assumed to be 26% in subsequent calculations.
Similarly, the proportion of AOM attributable to pneumococcus was assumed to be 23% according to model AOM1, this being the model with the best fit. Age-specific incidence rates of pneumococcal meningitis and other invasive pneumococcal disease were calculated using the overall number of isolates reported over the period 1996–2000 and the annual population figures for England and Wales, produced by the Office of National Statistics (ONS). Incidence rates are reported in Table 3 and show, as reported in previous studies,24,25 IPD rates peaking in infants and among the elderly. Incidence rates for pneumococcal meningitis are highest in infants
Table 2 Estimation of the proportion of unspecified CAP and AOM attributable to specific pathogens and sensitivity results to model specification Pathogen
Model CAP1 (%)a
Model CAP2— final model (%)
Model CAP3 (%)
Model AOM1— final model (%)a
Model AOM2 (%)
Model AOM3 (%)a
S. pneumoniae M. pneumoniae RSV H. influenzae Influenza A Influenza B Adenovirus Rhinovirus Other organisms Adjusted R2 square
31 17 6 – 8 – 38 – – 67
26 14 8 11 10 – – – 31 68
29 12 6 – 8 – – – 44 67
23 12 8 – – 4 36 18 – 68
27 10 7 – – 4 – 25 27 63
42 15 5 – – 3 – 35 – 58
a
Regression through the origin (no-intercept model).
Other IPD
140.5 (128.4–153.4) 255.8 (249.8–261.9) 131.7 (129.7–133.8) 39.7 (38.8–40.7) 17.2 (16.6–17.9) 19 (18.8–19.3) 45.8 (45.3–46.4) 146.5 (144.9–148.2) 404.7 (401.7–407.6) 73.7 (73.4–74.1) 50.9 (43.7–58.9) 82.8 (79.4–86.3) 16.3 (15.6–17.1) 3.7 (3.4–4) 2 (1.8–2.3) 3.5 (3.4–3.6) 8.6 (8.4–8.9) 23.1 (22.4–23.7) 44.8 (43.8–45.8) 10.8 (10.7–11) 75.3 (65.2–86.6) 38.6 (36.4–41) 11.6 (11–12.2) 2 (1.8–2.2) 1 (0.9–1.2) 3.1 (2.9–3.2) 7.3 (7–7.5) 20.2 (19.6–20.8) 44.3 (43.4–45.2) 9.2 (9.1–9.3) 59.7 (50.8–69.8) 23.4 (21.7–25.2) 9.9 (9.4–10.4) 1.8 (1.6–2) 0.8 (0.7–1) 2.8 (2.7–2.9) 6.7 (6.5–6.9) 19.3 (18.7–19.9) 43.6 (42.7–44.5) 8.5 (8.4–8.7) 15.6 (11.2–21.1) 15.3 (13.9–16.8) 1.7 (1.5–1.9) 0.2 (0.1–0.3) 0.2 (0.1–0.3) 0.2 (0.2–0.3) 0.6 (0.5–0.7) 0.9 (0.8–1) 0.7 (0.6–0.8) 0.7 (0.6–0.7) !1 month 1–11 months 1–4 years 5–9 years 10–14 years 15–44 years 45–64 years 65–74 years 75C years Total
Lobar pneumonia, organism unspecified Pnc confirmed infection Pnc meningitis
All IPD
Hospitalisations (HES) Laboratory reports (CDSC/RSIL) Age group
Average annual incidence rate (and 95% CI) per 100 000 population of pneumococcal related disease Table 3
30.2 (16.9–49.9) 345 (329.7–361) 406.8 (398.8–414.8) 236.2 (231–241.6) 52.1 (49.6–54.6) 13.2 (12.7–13.7) 16 (15.3–16.7) 15.9 (14.7–17.1) 13.5 (12.4–14.8) 54.2 (53.6–54.8)
A. Melegaro et al. Otitis media, organism unspecified
42
(15 per 100 000 population, 95% CI: 11–21) and remain low in all other age groups. No significant change was observed over the time period. As part of the enhanced surveillance of pneumococcal disease, the number of invasive isolates with serotype information has increased since 1996, rising from 1696 (36%) in 1996 to 2351 (50%) in 2000. An average of 31 different serotypes were detected annually, with the 15 most prevalent types counting for 93% of the overall number. The derived proportions of IPD in England and Wales that are caused by the serotypes contained in the available or potentially available formulations of vaccines (7, 9, 11, 23-valent) are given in Table 4 assuming cross protection within serogroups occurs (i.e. vaccine containing serotype 6B also confers protection against serotype 6A). Eighty-two percent of the 1815 IPD cases reported in children 0–5 years of age over the period 1996–2000 were caused by serotypes contained in the 7-valent pneumococcal conjugate vaccine. Information on serotype distribution and antimicrobial susceptibilities amongst the IPD isolates will be reported elsewhere. Overall, HES data show a similar incidence of admission due to confirmed pneumococcal infection (10.8 vs. 9.2) (Table 3). Otitis media hospitalisation rates are also shown for the different age groups and the incidence rate peaks in the 1–4 years of age, although the incidence remains high in all age groups for children less than 10 years old. Among all hospitalised patients, 81% were recorded with more than one diagnostic code, 54% with more than two, and 38% had more than three diagnoses. Co-morbidities were clearly directly associated with the age of the patients, being much higher for older age groups (Fig. 4). S. pneumoniae was present in the first diagnosis in 72% of the patients, and the remainder was distributed among neoplasm cases (11%), disease of the circulatory system (23%), disease of the respiratory system (20%), and various other conditions (46%). In Fig. 5 the age-specific proportions of hospitalisations that reported a non-pneumococcal diagnosis in the first diagnostic fields are shown. In the second and further diagnoses the proportion of pneumococcus related codes were of course lower, being 15% in the second diagnosis, and 7% in the third (co-morbidities, respectively, 60 and 46%, the remaining being blank fields). Age-specific episode and consultation rates for both CAP and AOM derived from MSGP4 data22 were more than twice as high as those routinely produced by the RCGP system21 (GP consultation rates for CAP: 679 vs. 266 per 100 000 population per year). Higher episode rates were found in patients with
Burden of pneumococcal disease
43
Figure 2 Comparison of observed bi-weekly number of pneumonia GP episodes with estimated number derived from final best fitting model (Table 2—model CAP2: % CAP due to S. pneumoniae is 26%).
high-risk conditions. In Fig. 6 pneumococcal pneumonia and AOM age-specific episode rates per 100 000 populations per year, calculated using MSGP4 data, are reported for high and low risk individuals. For otitis media, higher rates are observed in children 10 years old, and rates almost twice as high are reported for children !5 years of age when they also have other high-risk conditions.
A similar pattern is observed for pneumonia episodes although, in this case, higher rates are reported at a much later age (60C years). The total number of deaths for the period April 1995–2000 was derived from both ONS and HES data (Table 5), extracting those records for which one of the pneumococcal codes was either specified as the underlying cause of death (ONS) or recorded as
Figure 3 Comparison of observed bi-weekly number of otitis media GP episodes with estimated number derived from final best fitting model (Table 3—model AOMI: % of AOM due to S. pneumoniae is 23%).
44
A. Melegaro et al.
Table 4 Number and percent of IPD caused by the 7, 9, 11 or 23 serotypes contained in the formulations of the available vaccines over the period 1996–2000 7-valenta
Age group
!1 month 1–11 months 1–4 years 5–9 years 10–14 years 15–44 years 45–64 years 65–74 years 75C years Unknown Total a b c
9-valentb
11-valentc
23-valent
Cases
%
Cases
%
Cases
%
Cases
%
59 587 845 88 29 601 884 960 2008 250 6311
47 81 88 53 43 47 55 62 68 62 64
81 606 892 126 44 863 1026 1043 2134 277 7092
64 84 93 76 66 67 64 68 72 69 72
96 647 908 137 47 977 1177 1204 2364 311 7868
76 89 94 83 70 76 74 78 80 77 80
117 708 944 160 66 1249 1544 1495 2855 386 9524
93 98 98 96 99 98 97 97 97 96 97
Serotypes of the 7-valent vaccine: 4, 6B, 9V, 14, 18C, 19F, 23F. Nine-valent: 7-valent serotypes C1, 5. Eleven-valent; 9-valent C3, 7F.
primary diagnoses of the patient (HES). Deaths reporting a pneumococcal related ICD-10 code in any of the seven diagnostic fields were also extracted from HES. The number of ONS registered deaths from invasive pneumococcal diseases (Pnc meningitis and septicaemia) ranged between 72 and 117 per year in the period 1995–2000. The number of ONS registered deaths from lobar (pneumococcal) pneumonia was in the range of 1807–2064. These figures represent around 70% of the total number of deaths reported in HES for patients admitted with a pneumococcal meningitis or septicaemia code in the first diagnostic field though only around 40% of cases that died in the hospital and for which the pneumococcal related diagnosis was extracted from any of the seven diagnostic fields. The lower number of deaths reported in ONS
Figure 4 2000).
may represent cases that die with a pneumococcal related condition but for which the underlying cause of death (as reported in the death certificate) was non-pneumococcal specific. In Fig. 7 age-specific case fatality ratios derived from HES are shown for invasive and non-invasive pneumococcal conditions and show a marked rise in 15C years of age. Case-fatality rates of around 40% are reached among elderly hospitalised patients reporting strictly invasive pnuemococcal codes (Pnc meningitis and septicaemia) as their first diagnosis. Case-fatality ratios were calculated also for hospitalised patients reporting a pneumococcal code in the first three or in any diagnostic fields and little variation was shown in the proportion of individuals who died during the hospital admission (all ages: 13%; in 65C: 21–28%).
Proportion of hospital admissions with 1, 2–4 and 5C diagnoses reported during a hospital stay (HES 1995–
Burden of pneumococcal disease
45
Figure 5 Proportion of hospitalisations with a non-pneumococcal diagnosis reported in the first diagnostic field (HES 1995–2000).
The burden of pneumococcal disease in England and Wales was quantified using hospitalisation rates and GP consultation rates just provided (Table 6). The number of consultations to the general practitioner for CAP and AOM attributable to the pneumococcus was estimated using the results of the multivariate regression analysis and the average consultation rate derived from MSGP4 and RCGP data. Otitis media represents a major component of the overall burden of disease with an estimated 630 470 consultations to the general practitioner per year. Of these, around 60% occurs in children less than 10 years of age. Similarly, 64% of otitis media
Figure 6
hospitalisation occurs in children as well as around 40% of pneumococcal meningitis cases. The burden of pneumococcal disease in the elderly is represented mostly by pneumococcal pneumonia and septicaemia. Forty-nine and 58% of hospitalisations for, respectively, pneumococcal confirmed disease and lobar pneumonia occurs in 65C year of age.
Discussion In the light of the current discussions on whether or not to introduce widespread vaccination of infants
Pneumococcal pneumonia and otitis media episode rates by age in high and low risk individuals (MSGP4 data).
46 Table 5 Year
95/96 96/97 97/98 98/99 99/00
A. Melegaro et al. Number of deaths from pneumococcal disease ONSa
HES—first diagnosisb
HES—any diagnosisb
Lobar (Pnc) pneumonia
IPD
Lobar (Pnc) pneumonia
IPD
Lobar (Pnc) pneumonia
IPD
1907 2064 1807 1991 2059
94 117 99 104 72
2468 2841 2738 3471 3696
133 140 132 114 127
4553 5286 5323 6357 6853
219 268 223 229 234
a The ICD-9 coding system was used for ONS records until 2000 data (lobar, pneumococcal pneumonia codeZ481, IPDZ0382 (Pnc septicaemia)C3201 (Pnc meningitis). b The ICD-10 coding system was used in HES since 1995 (lobar, pneumococcal pneumonia codeZJ181 (lobar pneumonia organism unspecified)CJ13 (pneumococcal pneumonia), IPDZA403 (Pnc septicaemis)CG001 (Pnc meningitis).
with the pneumococcal conjugate vaccine, baseline information on the actual burden of disease is essential in order to perform a reliable economic evaluation and to provide baseline information against which future control programs can be measured. Although several national data sources are available to gain insights into the burden of pneumococcal disease in the U.K., uncertainties are nevertheless present due, mainly, to the fact that the aetiological agent for clinical conditions such as pneumonia and otitis media is not determined for the majority of the cases. A number of studies have attempted to estimate the burden of Pnc disease, by applying the study results of more sensitive alternative laboratory diagnostic methods (e.g. PCR, antigen detection) to unspecified syndromic surveillance data,26 or by reviewing individual hospital admissions to estimate the proportion likely to be attributable to S. pneumoniae infection.20 In this study, we quantified the amount of pneumococcal disease, invasive and non-invasive,
that is present in England and Wales, using both national surveillance database and multivariate regression techniques based on the underlying seasonality of pneumococcal disease reports to estimate the proportion of GP consultations for both pneumonia and AOM that is attributable to Pnc infection. The strength of this modelling technique has been shown in the past when similar questions were addressed for rotavirus23 and hospital admissions for RSV.27 As was shown in previous studies18,28–30 the estimated burden of pneumococcal disease in England and Wales is considerable, with incidence rates of both IPD and non-IPD peaking in children (!5 years) and in the elderly (75C years). Morbidity appears to be related to the presence of high-risk conditions, with over 80% of hospitalised patients presenting with more than one diagnosis. In particular, neoplasm and diseases of the respiratory and circulatory systems are common co-morbid conditions associated with invasive Pnc disease.
Figure 7 Case-fatality rate by age of Pnc meningitis and septicaemia (G001 and A403), Pnc pneumonia (J13) and lobar pneumonia, organism unspecified (J181)—HES (primary diagnosis).
Burden of pneumococcal disease
47
Table 6 Burden of pneumococcal related disease in England and Wales (ONS England and Wales 2002 population estimates) Age group
!1 years 1–4 years 5–9 years 10–14 years 15–44 years 45–64 years 65–74 years 75C years Total
Hospitalisationsa
GP consultationsb
Deathsc
Pnc confirmed
Lobar pneumonia
OM
Pnc pneumonia
Pnc OM
Pnc confirmed
Lobar pneumonia
488 392 119 70 772 1096 1021 1808 5766
1500 3164 1285 592 4171 5826 6486 16 324 39 349
1055 4881 4007 1180 2074 1492 488 304 15 565
933 2665 1568 879 12 916 14 520 10 676 26 152 70 309
53 311 223 065 98 732 25 294 142 599 60 438 17 762 9268 630 470
7 5 1 1 32 143 119 421 730
5 4 2 4 41 303 762 3392 4513
a
ICD-10 codes—Pnc confirmed: G001, A403, J13, B953; lobar pneumonia: J181; OM: H65–H67. Average MSGP4 and RCGP rates (from the multivariate regression analysis: 26 and 23% of, respectively, pneumonia and OM consultations are attributable to S. pneumoniae). c Age-specific CFR shown in Fig. 7 (HES primary diagnosis only) is applied to the number of hospitalisations shown in this table. b
Similarly, non-invasive disease is more common in those with underlying illnesses, with CAP and AOM occurring almost twice as frequently in high-risk individuals than in the non-high-risk groups. A polysaccharide and a conjugate vaccine, nevertheless, are available and currently recommended in England and Wales for high-risk individuals and more recently, polysaccharide vaccine for all those aged 65 years or over. In this latter group, the vaccine appears to be effective in reducing morbidity due to IPD, though discussions are still ongoing on the protection it confers among certain at-risk categories which appear to be highly vulnerable to pneumococcal disease.31 This paper demonstrates a large burden due to pneumococcal disease in both the elderly and high-risk categories. Ongoing surveillance will be critical to ascertain the impact of these programs. In addition, the 7-valent pneumococcal conjugate vaccine also seems to reduce S. pneumoniae colonisation and have an impact on the overall transmission of the organism in the population, potentially providing protection to unvaccinated populations (i.e. herd immunity).10,11,32 However, the potential threat of serotype replacement as a consequence of vaccination, as seems to have been observed in the U.S.A.9 and in Finland,12 will require prospective monitoring for any changes in the serotype distribution of pneumococcal disease patients and, ideally asymptomatic carriers. These factors will have to be combined to disease burden data to assist decision-makers in the most appropriate intervention strategy with the pneumococcal conjugate vaccine in countries considering introduction. Moreover, results from cost-effectiveness analyses showed that indirect effects such as herd
immunity and serotype replacement may have a crucial impact on the economic acceptability of alternative programs.33,34 The serotype diversity of S. pneumoniae and its many clinical manifestations make it hard to answer questions related to the short as well as long-term effects that vaccination may have at the population level. Here, we have attempted to give some baseline information on the epidemiology of both invasive and non-invasive pneumococcal condition and on the potential impact of the available pneumococcal vaccines. Ongoing surveillance is in place to assess any impact on disease epidemiology of the current high-risk strategy in England and Wales and to assist decisions on future vaccination strategy.
Acknowledgements We would like to thank Dr Douglas Fleming of the Royal College of General Practitioners Scheme for providing consultation rates for CAP and AOM and the DH for hospitalisation records. We also want to thank Pauline Kaye and Usha Gungabissoon for data extraction. The MSGP4 are Crown Copyright and are reproduced with permission. The study was funded by a grant from the EU (QLG4-CT-2000-00640).
References 1. Robinson K, Baugham W, Rothrock G, Barret NL, Pass M, Lexau C, et al. Epidemiology of invasive Streptococcus
48
2.
3.
4.
5. 6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
A. Melegaro et al. pneumoniae infections in the United States, 1995–1998. Opportunities for prevention in the conjugate vaccine era. JAMA 2001;285(13):1729–35. Cartwright K. Pneumococcal disease in Western Europe: burden of disease, antibiotic resistance and management. Eur J Pediatr 2002;161:188–95. Usen S, Adegbola R, Mulholland K, Jaffar S, Hilton S, Oparaugo A, et al. Epidemiology of invasive pneumococcal disease in the Western Region, The Gambia. Pediatr Infect Dis J 1998;17(1):23–8. Watt JP, O’Brien KL, Benin AL, Whitney CG, Robinson K, Parkinson AJ, et al. Invasive pneumococcal disease among Navajo adults, 1989–1998. Clin Infect Dis 2004;38(4): 496–501. Brown PD, Lerner SA. Community-acquired pneumonia. Lancet 1998;352:1295–302. Pelton SI. Acute otitis media in the era of effective pneumococcal conjugate vaccine: will new pathogens emerge? Vaccine 2001;19:S96–S9. Jacobs MR, Dagan R, Appelbaum PC, Burh DJ. Prevalence of antimicrobial-resistant pathogens in middle ear fluid: multinational study of 917 children with acute otitis media. Antimicrob Agents Chemother 1998;42(3):589–95. Black SB, Shinefield HR, Fireman B, Lewis E, Ray P, Hansen JR, et al. Efficacy, safety and immunogenicity of heptavalent pneumococcal conjugate vaccine in children. Pediatr Infect Dis J 2000;19:187–95. Kaplan SL, Mason Jr EO, Wald ER, Schutze GE, Bradley JS, Tan TQ, et al. Decrease of invasive pneumococcal infections in children among 8 children’s hospitals in the United States after the introduction of the 7-valent pneumococcal conjugate vaccine. Pediatrics 2004;113(3):443–9. Black S, Shinefield H, Baxter R, Austrian R, Bracken L, Hansen J, et al. Postlicensure surveillance for pneumococcal invasive disease after use of heptavalent pneumococcal conjugate vaccine in Northern California Kaiser Permanente. Pediatr Infect Dis J 2004;23(6):485–9. Whitney CG, Farley MM, Hadler J, Harrison LH, Bennett NM, Lynfield R, et al. Decline in invasive pneumococcal disease after the introduction of protein-polysaccharide conjugate vaccine. N Engl J Med 2003;348(18):1737–46. Eskola J, Kilpi T, Palmu A, Jokinen J, Haapakoski J, Herva E, et al. Efficacy of a pneumococcal conjugate vaccine against acute otitis media. N Engl J Med 2001;344(6):403–9. Fireman B, Black SB, Shinefield HR, Lee J, Lewis E, Ray P. Impact of the pneumococcal conjugate vaccine on otitis media. Pediatr Infect Dis J 2003;22(1):10–16. Black SB, Shinefield HR, Ling S, Hansen J, Fireman B, Spring D, et al. Effectiveness of heptavalent pneumococcal conjugate vaccine in children younger than five years of age for prevention of pneumonia. Pediatr Infect Dis J 2002; 21(9):810–5. Poehling KA, Lafleur BJ, Szilagyi PG, Edwards KM, Mitchel E, Barth R, et al. Population-based impact of pneumococcal conjugate vaccine in young children. Pediatrics 2004; 114(3):755–61. CDC. Prevention of pneumococcal disease: recommendations of the advisory committee on immunisation practices (ACIP). MMWR 1997;46(RR-08):1–24.
17. [www.dh.gov.uk] Salisbury D, Begg N. Immunisation against infectious disease. London: HMSO; 1996. 18. Miller E, Waight P, Efstratiou A, Brisson M, Johnson A, George R. Epidemiology of invasive and other pneumococcal disease in children in England and Wales 1996–1998. Acta Paediatr Suppl 2000;435:11–16. 19. George R, Melegaro A. Invasive pneumococcal infection England and Wales, 1999. CDR Weekly 2001;11(21):6–12. 20. Djuretic T, Ryan M, Miller E, Fairley CK, Goldblatt D. Hospital admission in children due to pneumococcal pneumonia in England. J Infect 1998;37:54–8. 21. Fleming DM. Weekly returns service of the Royal College of General Practitioners. Commun Dis Public Health 1999;2(2): 96–100. 22. McCormick A, Fleming D, Charlton J. Morbidity statistics from general practice, Fourth National Survey 1991–1992. London: Office of Population Census and Survey (Series MB5:3), HMSO: 1995. 23. Ryan M, Ramsay M, Brown D, Gay N, Farrington C, Wall P. Hospital admissions attributable to rotavirus infection in England and Wales. J Infect Dis 1996;174(Suppl 1):S12–S18. 24. Smith M, Stuart J, Andrews N, Telfer Brunton W, Cartwright K. Invasive pneumococcal infection in South and West England. Epidemiol Infect 1998;120:117–23. 25. Laurichesse H, Grimaud O, Waight P, Johnson AP, George R, Miller E. Pneumococcal bacteremia and meningitis in England and Wales, 1993 to 1995. Commun Dis Public Health 1998;1:22–7. 26. McIntosh EDG, Booy R. Invasive pneumococcal disease in England and Wales: what is the true burden and what is the potential for prevention using 7 valent pneumococcal conjugate vaccine? Arch Dis Child 2002;86:403–6. 27. Muller-Pebody B, Edmunds W, Zambon M, Gay N, Crowcroft N. Contribution of RSV to bronchiolitis and pneumonia-associated hospitalizations in English children, April 1995–March 1998. Epidemiol Infect 2002;129:99–106. 28. Smith MD, Stuart J, Andrews NJ, Telfer Brunton WA, Cartwright KA. Invasive pneumococcal infection in South and West England. Epidemiol Infect 1998;120:117–23. 29. Sleeman K, Knox K, George R, Miller E, Waight P, Griffiths D, et al. Invasive pneumococcal disease in England and Wales: vaccination implications. J Infect Dis 2001;183:239–46. 30. Kyaw MH, Clarke S, Jones IG, Campbell H. Incidence of invasive pneumococcal disease in Scotland, 1988–99. Epidemiol Infect 2002;128:139–47. 31. Melegaro A, Edmunds WJ. The 23-valent pneumococcal polysaccharide vaccine/Part I—efficacy of PPV in the elderly: a comparison of meta-analyses. Eur J Epidemiol 2004;19(4): 353–63. 32. Givon-Lavi N, Fraser D, Dagan R. Vaccination of day-care center attendees reduces carriage of Streptococcus pneumoniae among their younger siblings. Pediatr Infect Dis J 2003;22(6):524–32. 33. Melegaro A, Edmunds WJ. Cost effectiveness analysis of pneumococcal conjugate vaccination in England and Wales. Vaccine 2004;22(31–33):4203–14. 34. McIntosh ED. Cost-effectiveness studies of pneumococcal conjugate vaccines. Expert Rev Vaccines 2004;3(4):433–42.