Hospital admission rates for meningitis and septicaemia caused by Haemophilus influenzae, Neisseria meningitidis, and Streptococcus pneumoniae in children in England over five decades: a population-based observational study

Articles

Hospital admission rates for meningitis and septicaemia caused by Haemophilus influenzae, Neisseria meningitidis, and Streptococcus pneumoniae in children in England over five decades: a population-based observational study Natalie G Martin, Manish Sadarangani, Andrew J Pollard, Michael J Goldacre

Summary Background Infection with Haemophilus influenzae, Neisseria meningitidis, and Streptococcus pneumoniae causes substantial mortality and long-term morbidity in children. We know of no study to assess the long-term trends in hospital admission rates for meningitis and septicaemia caused by these pathogens in children in England. We aimed to do such a study using routinely reported data in England. Methods In this population-based observational study, we used datasets that include routinely collected administrative statistics for hospital care: the Hospital In-Patient Enquiry (data for England from 1968 to 1985), the Hospital Episode Statistics dataset (data for England from 1989 onwards), and the Oxford record linkage study (data for Oxfordshire and surrounding areas from 1963 to 2011). We analysed annual age-specific and age-standardised admission rates in children younger than 15 years with H influenzae, meningococcal and pneumococcal meningitis, and septicaemia. Findings We saw a reduction in hospital admission rates for childhood invasive bacterial disease after the introduction of conjugate vaccines against H influenzae, N meningitidis, and S pneumoniae in England. Annual incidence of H influenzae meningitis per 100 000 children decreased from 6·72 admissions (95% CI 6·18–7·26) in 1992 to 0·39 admissions (0·26–0·52) in 1994, after the introduction of routine H influenzae type b vaccination. We saw a small rise in admissions in the early 2000s, peaking at 1·24 admissions per 100 000 children (0·99–1·48) in 2003, which decreased to 0·28 per 100 000 children (0·17–0·39) by 2008 after the introduction of catch-up (2003) and routine (2006) booster programmes for young children. Meningococcal disease increased during the 1990s, reaching a peak in 1999, with 34·54 admissions (33·30–35·78) per 100 000 children. Hospital admissions decreased after the meningococcal serogroup C vaccine was introduced in 1999 and was 12·40 admissions (11·68–13·12) per 100 000 in 2011. Admissions for invasive pneumococcal disease increased from the 1990s reaching a peak in 2006 at 4·45 admissions for meningitis (95% CI 4·0–4·9) per 100 000 children and 2·81 admissions for septicaemia (2·45–3·17) per 100 000 children. A reduction in admissions occurred after the introduction of the pneumococcal conjugate vaccine in 2006: hospital admission rates in 2011 were 2·03 per 100 000 children for meningitis and 1·12 per 100 000 children for septicaemia.

Lancet Infect Dis 2014; 14: 397–405 Published Online March 14, 2014 http://dx.doi.org/10.1016/ S1473-3099(14)70027-1 See Comment page 363 Oxford Vaccine Group, Department of Paediatrics (N G Martin MBChB, M Sadarangani DPhil, Prof A J Pollard PhD), and Unit of Health-Care Epidemiology, Nuffield Department of Population Health (Prof M J Goldacre FFPH), University of Oxford, Oxford, UK; and NIHR Oxford Biomedical Research Centre, Oxford, UK (N G Martin, M Sadarangani, Prof A J Pollard) Correspondence to: Dr Natalie G Martin, Department of Paediatrics, University of Oxford, Old Road, Headington, Oxford OX3 7LE, UK natalie.martin@paediatrics. ox.ac.uk

Interpretation Vaccine-preventable invasive bacterial disease in children has decreased substantially in England in the past five decades, most notably with the advent of effective conjugate vaccines since the 1990s. Ongoing disease surveillance and continued development and implementation of vaccines against additional pneumococcal serotypes and serogroup B meningococcal disease are important. Funding None.

Introduction Haemophilus influenzae, Neisseria meningitidis, and Streptococcus pneumoniae can cause severe childhood disease and are associated with substantial mortality and long-term morbidity. Case-fatality rates in high-income countries for childhood bacterial meningitis caused by these pathogens are between 5% and 10%.1,2 The median risk of long-term disabling sequelae in survivors of bacterial meningitis is about 20%, with such sequelae most likely after pneumococcal meningitis.3,4 Unlike polysaccharide vaccines, conjugate vaccines are highly effective in children younger than 2 years and www.thelancet.com/infection Vol 14 May 2014

induce immunological memory responses, with protection boosted by subsequent doses.5,6 In the UK, a conjugate vaccine against H influenzae type b (Hib) was introduced into the routine infant vaccination schedule with a catchup programme in October, 1992.7–10 N meningitidis causes septicaemia and meningitis. A serogroup C meningococcal conjugate vaccine was introduced in September, 1999, as three infant doses along with a catch-up campaign to 19 years of age after an increase in meningococcal serogroup C disease, the second most common serogroup causing disease in the UK at the time.11,12 S pneumoniae causes a wide range of invasive disease including 397

Articles

meningitis, bacteraemia, pneumonia, and bone and joint infection. The seven-valent pneumococcal conjugate vaccine (PCV7) was introduced in September, 2006, at months 2, 4, and 13 after birth with a catch-up programme to 2 years of age, and was replaced by the 13-valent pneumococcal vaccine from April, 2010.13–15 Findings from microbiological surveillance studies in the UK have shown a substantial reduction in invasive bacterial disease caused by H influenzae, N meningitidis, and S pneumoniae after the introduction of effective conjugate vaccines.7,11,16 However, long-term trends in hospital admission rates have not been reported. Here, we assessed trends for meningitis and septicaemia caused by these pathogens using hospital data for England from the 1960s to 2011.

Methods Procedures In this population-based observational study, we used datasets that include routinely collected administrative statistics for hospital care. From 1968 to 1985, hospital admission statistics in England were collected in the Hospital In-Patient Enquiry (HIPE). HIPE was a 10% sample of every hospital admission in the English National Health Service (NHS). Health authorities were issued with “a detailed compendium of instructions to achieve consistency, and the submission of a complete and unbiased sample”.17 From 1989, English national data have been collected in the Hospital Episode Statistics (HES) dataset on a 100% basis covering all NHS day patients and inpatients. We combined HIPE and HES data, and scaled up the HIPE data by a factor of ten. For most of the years covered, English data were based on numbers of episodes of care rather than numbers of people receiving care. From 1999 the data were linkable, and, from then on, both episode-based and person-based rates are reported here. We linked data using encrypted values of the individuals’ unique NHS number and encrypted postcode, date of birth, and sex. In the linked data, people with more than one admission for each disease were counted only once for that disease. We also used data from the Oxford record linkage study (ORLS). The ORLS includes data for all NHS hospital admissions in the former Oxford NHS Regional Health Authority area from 1963 to 2011. Successive admissions Years

ICD disease classification and code Meningococcal infection

Pneumococcal meningitis

Pneumococcal septicaemia

Haemophilus meningitis

Haemophilus septicaemia

ICD revision 7

1963–67

057

340·1

053·2

340·0

Not coded

ICD revision 8

1968–78

036

320·1

038·2

320·0

Not coded

ICD revision 9

1979–94

036

320·1

038·2

320·0

Not coded

ICD revision 10

1994–2011

A39

G00·1

A40·3

G00·0

A41·3

Table 1: International Classification of Disease (ICD) Codes used to identify hospital admission rates from 1963 to 2011

398

for the same individuals were always linked in the ORLS, allowing the same patient to be counted only once through multiple admissions or inter-hospital transfers. We linked data using the individuals’ regional personal number until 1998, and, from 1999 onwards, the ORLS was taken as the regional subset of English national HES linked as previously described. The Oxford dataset was a 100% dataset from 1963, without sampling, and was continuous. English national hospital statistics were not collected in the second half of the 1980s. In the period covered by both, the regional ORLS and national HIPE datasets were collected independently of each other. Therefore, we assessed whether findings in one corroborate findings in the other. We present the ORLS data in the figures only, for annual trends, alongside data for all of England. Tables include only the all-England data. Ethical approval for a programme of work using the record linkage datasets was obtained from the Central and South Bristol Multi-Centre Research Ethics Committee (04/Q2006/176).

Statistical analysis We analysed annual age-specific and age-standardised admission rates in children younger than 15 years for meningitis and septicaemia caused by N meningitidis, S pneumoniae, and H influenzae. We used population denominators from the Office for National Statistics. We calculated admission rates and 95% CIs using the direct method of standardisation (in 5-year age groups, confined to people younger than 15 years) and the European standard population.18 We analysed annual age-standardised admission rates for S pneumoniae meningitis and septicaemia combined in adults aged 65 years or older from 1999 to 2011 to assess any effect of herd immunity driven by the infant PCV7 vaccine programme. We defined diseases according to International Classification of Disease (ICD) codes (table 1) and included cases if the diagnosis was recorded as the primary or one of the secondary diagnoses.

Role of the funding source The sponsor of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Results Annual age-standardised episodes of H influenzae (any type) meningitis per 100 000 children per year in England varied between 2·86 episodes and 6·72 episodes between 1968 and 1992, reaching a peak of 6·72 (95% CI 6·18–7·26) in 1992 (figure 1). After the introduction of routine Hib vaccination in 1992, there was a 94% decrease to 0·39 admissions per 100 000 children per year (0·26–0·52) within 2 years. We saw a small rise in www.thelancet.com/infection Vol 14 May 2014

Articles

www.thelancet.com/infection Vol 14 May 2014

A

National (episodes) National (people) Oxford (people)

9

Age-standardised admission rates per 100 000 children

8 1 7 6 5 4 3 2

2 3

1

19 63 19 65 19 67 19 69 19 71 19 73 19 75 19 77 19 79 19 81 19 83 19 85 19 87 19 89 19 91 19 93 19 95 19 97 19 99 20 01 20 03 20 05 20 07 20 09 20 11

0

B

National (episodes) National (people)

Age-standardised admission rates per 100 000 children

0·6 0·5 0·4 0·3 0·2 0·1 2

3

19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11

0

Year

Figure 1: Hospital-admission rates for Haemophilus influenzae meningitis (A) or septicaemia (B) in children younger than 15 years in England Arrow 1 is the 1992 introduction of the H influenzae type b (Hib) vaccine. Arrow 2 is the 2003 introduction of the catch-up Hib booster vaccine for children aged 6 months to 4 years. Arrow 3 is the 2006 introduction of routine 12-month Hib booster vaccine in the second year of life.

National (episodes) National (people) Oxford (people)

40 35 Age-standardised admission rates per 100 000 children

30 25 20 15 10 5 1

0

2

19 63 19 65 19 67 19 69 19 71 19 73 19 75 19 77 19 79 19 81 19 83 19 85 19 87 19 89 19 91 19 93 19 95 19 97 19 99 20 01 20 03 20 05 20 07 20 09 20 11

admissions in the early 2000s, peaking at 1·24 admissions per 100 000 children (0·99–1·48) in 2003, which decreased to 0·28 per 100 000 children (0·17–0·39) by 2008. Admission rates for H influenzae (any type) septicaemia remained low from 1995 to 2011 (figure 1). The all-England and the Oxford admission rates were much the same (figure 1). Before 1985, the episode-based admission rate for meningococcal disease in England was stable at about five admissions per 100 000 children per year, with a small rise during the mid-1970s (figure 2). We saw an increase in admission rates for N meningitidis infection during the 1990s, reaching a peak in 1999 at 34·54 admissions per 100 000 children (33·30–35·78). The rate of hospital admissions decreased after the meningococcal serogroup C vaccine was introduced and was 12·40 per 100 000 children (11·68–13·12) in 2011. Person-based admission rates per 100 000 children decreased by 66%, from 26·68 (25·59–27·77) in 1999 to 9·10 (8·48–9·71) in 2011. The comparison between episode-based and personbased rates (figure 2) shows the extent of multiplecounting of the same people as shown by record linkage. Annual person-based rates were 19–27% lower than episode-based rates. The pattern of admission rates in allEngland and Oxford were similar, albeit slightly lower in the Oxford dataset (figure 2). Hospital admission rates in England were higher than previously reported incidence from laboratory surveillance data in England and Wales (table 2).12 From 1968 to 1985, the admission rate per 100 000 children per year for pneumococcal meningitis in England ranged between 1·13 and 2·29 (figure 3). Annual rates increased throughout the 1990s and early 2000s, peaking in 2006 at 4·45 admission episodes per 100 000 children (95% CI 4·0–4·9). A reduction in admissions occurred after 2006, coinciding with the introduction of the PCV7 vaccine, and the overall hospital admission rate in 2011 was similar to the rate in the early 1990s at 2·03 episodes per 100 000 children (1·74–2·32). In analysis of linked data, the person-based rate per 100 000 children in 2006 was 2·67 (2·32–3·02) and in 2011 was 1·19 (0·97–1·41), less than half that in 2006. We noted a similar decrease in pneumococcal septicaemia admissions from 2·81 (2·45–3·17) admission episodes per 100 000 children in 2006 to 1·12 per 100 000 children (0·91–1·34) in 2011 (figure 3). We noted a small increase in admissions in 2009 for pneumococcal meningitis and in 2010 for pneumococcal septicaemia, decreasing again after the replacement of PCV7 with PCV13 in the routine infant immunisation schedule. Person-based hospital admission rates for pneumococcal meningitis in children younger than 5 years between 2000 and 2006 were 6·37 per 100 000 per year (6·03–6·72) compared with 3·18 per 100 000 per year (95% CIs not stated) in laboratory surveillance data.14 Similarly, between 2008 and 2010, person-based hospital admission rates in children younger than 5 years were 3·29 children per

Year

Figure 2: Hospital-admission rates for meningococcal infection in children younger than 15 years in England Arrow 1 is the 1999 introduction of the meningococcal serogroup C vaccine. Arrow 2 is the 2006 introduction of meningococcal serogroup C booster vaccine at age 12 months.

399

Articles

<1 year Hospital admissions

1–4 years Laboratory cases

Hospital admissions

5–9 years Laboratory cases

Hospital admissions

10–14 years Laboratory cases

Hospital admissions

Laboratory cases

2006

58·86

44·8

18·37

13·8

5·04

3

1·95

1

2007

58·19

44·5

19·76

15·3

3·98

3

1·91

1·1

2008

61·58

42·3

18·89

13·4

3·62

2·5

2·49

1·5

2009

46·69

31·6

16.00

9·3

4·26

2

2·19

1·2

2010

49·90

30·5

13·31

11·6

3·51

2·4

1·34

1·4

2011

41·38

··

14·95

··

3·78

··

1·70

··

Data are annual incidence of linked person-based hospital admissions in England and laboratory surveillance incidence in England and Wales (data from Ladhani and colleagues12). Hospital admission data are for calendar years 2006–11. Laboratory surveillance data are for epidemiological years 2006–07, 2007–08, 2007–08, 2008–09, and 2010–11.

Table 2: Comparison of rates of meningococcal infection per 100 0000 children younger than 15 years

A

Age-standardised admission rates per 100 000 children

6

National (episodes) National (people) Oxford (people)

5 4 3

2

1 1

0

2

although the Oxford data showed more fluctuation, which is probably because the data were from a smaller sample (figure 3). Hospital admission rates for pneumococcal meningitis or septicaemia combined in adults aged 65 years or older showed no consistent change between 1999 and 2011. The annual rate in 1999 was 4·79 people per 100 000, and in 2011 was 4·85 people. Annual rates per 100 000 fluctuated between 4·50 people and 5·46 people. The incidence of every disease was higher in male patients than in female patients: male patients comprised 56% of cases of meningococcal infection, 62% of cases of pneumococcal meningitis, 59% pneumococcal septicaemia, and 55% of H influenzae meningitis. The greatest burden of disease occurred in infants (table 3).

Age-standardised admission rates per 100 000 children

B 5·0

Discussion

4·5

Our findings show a striking reduction in childhood meningitis and septicaemia after the introduction of conjugate vaccines for H influenzae, N meningitidis, and S pneumoniae in England. The long period of preimmunisation surveillance provides a reliable baseline of admission rates against which to judge the postimmunisation decrease. These are the most complete data available about the population-based incidence of these diseases over the five decades analysed. Alternative sources are mortality statistics from death registrations and statutory infectious disease notifications. In the early years covered by our study, the statutory infectious disease notifications were incomplete for bacterial meningitis—eg, from 1969 to 1973 only half of meningococcal meningitis cases and less than a quarter of other bacterial meningitis cases were notified.19 The decrease in admission rates for these diseases after the introduction of conjugate vaccines occurred at a time when emergency hospital admissions have increased in children in England20 and when microbiological diagnostic methods have become more sensitive, most notably through the routine use of PCR testing. Our data show that excellent disease control was achieved after the introduction of the Hib vaccine programme from 1992 in England. We noted a 94%

4·0 3·5 3·0 2·5 2·0 1·5 1·0 0·5 1

2

19 63 19 65 19 67 19 69 19 71 19 73 19 75 19 77 19 79 19 81 19 83 19 85 19 87 19 89 19 91 19 93 19 95 19 97 19 99 20 01 20 03 20 05 20 07 20 09 20 11

0

Year

Figure 3: Hospital-admission rates for pneumococcal meningitis (A) and pneumococcal septicaemia (B) in children younger than 15 years in England Arrow 1 is the 2006 introduction of the PCV7 vaccine. Arrow 2 is the 2010 introduction of the PCV13 vaccine.

100 000 per year (2·93–3·66) compared with 1·44 per 100 000 laboratory cases per year (95% CIs not stated).14 Reductions in hospital admissions and laboratory cases of S pneumoniae infection in this age group were much the same between the two time periods.14 Admission rates in the all-England and Oxford datasets were similar, 400

www.thelancet.com/infection Vol 14 May 2014

Articles

reduction in H influenzae (any type) meningitis hospital admissions from 1992 to 1994. This decrease has also been shown in findings from laboratory surveillance studies, which indicated a 98% reduction in invasive Hib disease in children younger than 5 years by 1998.7,9 However, from 1999, laboratory surveillance data showed an increase in the incidence of Hib cases per 100 000 from 0·26 in 1998 to 1·80 in 2002, before falling to 0·27 in 2008.8 Analysis of the HES data has shown a similar trend in hospital admission rates for H influenzae meningitis, suggesting that most of these cases are likely to have been caused by Hib and that the trend for meningitis accurately shows the overall burden of Hib disease. The resurgence in 2002 was thought to be caused by several factors. Children aged 1–4 years who had been vaccinated as infants had a greater than expected reduction in Hib antibody concentrations after vaccination, leading to levels that were no longer sufficient for protection. Additionally, there was a decrease in herd immunity after the cessation of the catch-up campaign and a change in the combination Hib vaccine used (panel).8,21,22 A similar increase in cases of invasive Hib disease was also seen in the Netherlands in 2002.23 However, longitudinal hospital admissions data for Hib meningitis from Finland showed a pronounced decrease in admissions after the introduction of the Hib vaccine in 1986, but did not show disease resurgence in the early 2000s. Differences in trends between countries might have resulted from differences in vaccine schedules.24 Disease control in the UK was achieved with the use of an Hib catch-up booster vaccine in 2003 for children aged 6 months to 4 years, and then with the introduction of a routine 12 month booster from 2006.9,22 The increase in total Hib disease reported in laboratory surveillance data reached almost a third of the peak number of cases in the early 1990s, whereas the increase in H influenzae meningitis described in our hospital study were only about a sixth of the previous peak.8,10,25 This finding suggests that cases seen in the early 2000s might have been less severe than previous cases, which might be indicative of partial immunity from infant immunisation or a lower susceptibility of older children to meningitis during Hib infection compared with infants. We saw an increase in the number of hospital admissions for meningococcal disease during the midto-late 1990s, peaking in 1999. A similar increase was also seen in a longitudinal study in the Merseyside region of the UK.26 This increase was seen in analysis of microbiological surveillance data and was largely caused by the emergence of a hypervirulent serogroup C clone from the ST-11 clonal complex, which caused especially high incidences of disease in adolescents.11,27 Although our database did not include national data between 1985 and 1990, analysis of microbiological surveillance data showed an increase in serogroup B meningococcal disease between 1984 and 1990.28 Different serogroup B strains caused localised outbreaks during this time. www.thelancet.com/infection Vol 14 May 2014

Improved PCR-based diagnostic techniques could also have contributed to the increased number of laboratoryconfirmed cases identified in the late 1990s.11,27 This Admission episode rate (aged <1 year)

Admission episode rate (aged 1–14 years)

Age-standardised admission episode rate for children aged <15 years (95% CI)

Age-standardised admission linked people rate for children aged <15 years (95% CI)

Meningococcal infection 1968–70

31·09

3·91

5·89 (5·07–6·70)

1971–75

38·98

6·29

8·66 (7·87–9·46)

·· ··

1976–80

27·22

3·91

5·61 (4·91–6·30)

··

5·44 (4·73–6·14)

··

1981–85

27·49

3·71

1986–90

··

··

1991–96

68·87

11·51

15·68 (15·34–16·01)

1997–2001

129·76

22·03

29·87 (29·19–30·54)

23·35 (22·76–23·95)

2002–06

89·15

12·83

18·38 (17·96–18·79)

14·22 (13·85–14·58)

2007–11

70·34

9·60

14·02 (13·67–14·36)

10·37 (10·08–10·67)

··

·· ··

Pneumococcal meningitis 1968–70

10·09

1·53

2·15 (1·66–2·64)

1971–75

8·81

1

1·57 (1·23–1·91)

··

1976–80

9·84

0·98

1·62 (1·25–2·00)

··

1·59 (1·21–1·97)

··

1981–85

10·06

0·92

1986–90

··

··

1991–96

17·26

1·03

·· ·· 2·21 (2·09–2·34)

··

·· ··

1997–2001

27·02

1·17

3·05 (2·83–3·27)

2·25 (2·06–2·44)

2002–06

34·89

1·73

4·14 (3·94–4·34)

2·68 (2·52–2·83)

2007–11

19·66

1·23

2·57 (2·42–2·71)

1·44 (1·33–1·55)

Pneumococcal septicaemia 1968–70

0·81

0·19

0·23 (0·07–0·39)

··

1971–75

1·07

0·22

0·28 (0·14–0·43)

··

1976–80

2·62

0·20

0·38 (0·19–0·56)

··

1981–85

2·68

0·52

0·68 (0·43–0·93)

··

1986–90

··

··

1991–96

5·85

0·68

·· 1·05 (0·96–1·14)

·· ··

1997–2001

12·06

1·18

1·97 (1·80–2·15)

1·81 (1·64–1·97)

2002–06

12·71

1·48

2·30 (2·15–2·45)

1·97 (1·83–2·11)

2007–11

7·94

1·01

1·51 (1·40–1·63)

1·18 (1·08–1·28)

Haemophilus meningitis 1968–70

14·13

2·29

3·16 (2·56–3·75)

··

1971–75

16·02

2·40

3·39 (2·88–3·89)

··

1976–80

19·02

2·94

4·11 (3·50–4·72)

··

1981–85

26·49

3·35

5·04 (4·35–5·72)

··

1986–90

··

··

1991–96

15·76

1·62

··

1997–2001

2·67

0·30

0·47 (0·38–0·56)

0·38 (0·30–0·45)

2002–06

3·55

0·55

0·76 (0·68–0·85)

0·52 (0·45–0·59)

2007–11

2·19

0·11

0·26 (0·21–0·31)

0·17 (0·13–0·21)

2·65 (2·51–2·79)

·· ··

Haemophilus septicaemia 1968–90

··

··

1991–96

0·32

0·03

0·05 (0·03–0·07)

··

··

1997–2001

1·28

0·21

0·29 (0·22–0·36)

2002–06

1·18

0·29

0·35 (0·30–0·41)

0·29 (0·24–0·34)

2007–11

1·17

0·11

0·19 (0·15–0·23)

0·16 (0·13–0·20)

·· 0·26 (0·20–0·33)

Table 3: National hospital admission rates per 100 000 children

401

Articles

assumption is lent support by a finding from a study showing that more than 50% of confirmed cases of invasive meningococcal disease in England in 2009–10 were diagnosed by positive PCR with a negative culture.29 The reduction in admissions from 1999 occurred after the introduction of routine immunisation in England with a conjugate vaccine against serogroup C meningococcal disease, with national surveillance of laboratory-confirmed cases reporting a 97% reduction in meningococcal serogroup C disease at all ages from 1998–99 to 2006–07.11 During 1994–98, when the hypervirulent strain was circulating, the proportion of disease caused by serogroup C increased from 26% to 34%.27 The hospital admission data show an overall reduction in meningococcal infections (all serogroups) of 66% from 1999 to 2011. Analysis of the laboratory surveillance data shows that some of this decrease was not due to immunisation but to stochastic changes in disease cause by serogroup B meningococci, for which no vaccine has been used.11,12 Environmental factors—eg, the introduction of legislation banning smoking in public places in England in July, 2007—might also have contributed to the reduction in disease burden.30 Our data show higher incidences of meningococcal disease from hospital-discharge coding than microbiological surveillance data, a degree of which is expected because of the stricter criteria needed for laboratoryconfirmed cases (table 2).12 However, our linked personbased rates are about 30–40% higher than seen in laboratory data.12 Between 2006–07 and 2010–11, 87% of meningococcal disease (94% in infants) was caused by serogroup B in the UK, and serogroup C disease was well controlled (accounting for 2% of disease at all ages in the same period).12 The first broadly protective multicomponent vaccine against meningitis B (4CMenB; Novartis Vaccines and Diagnostics, MA, USA), comprised of non-serogroupspecific subcapsular proteins, has now been licensed in Europe and Australia.31,32 A decision on the use of the 4CMenB vaccine in the UK is expected in March, 2014. Our findings show an increase in hospital admissions for pneumococcal meningitis and septicaemia during the 1990s and early 2000s before the introduction of routine pneumococcal vaccination. Microbiological surveillance data also indicated an increase in invasive pneumococcal disease between 1997–98 and 2005–06, from 8·8 cases to 11·9 cases per 100 000 people at all ages, in addition to shifts in serotype distribution.15 Findings from another UK study reporting HES and Health Protection Agency data also showed an increase in pneumococcal meningitis incidence from 1998 to 2005 at all ages, with the greatest disease burden occurring in infants aged 2–11 months.33 A further study reported HES data for pneumococcal meningitis in adults from 1996 to 1999, and showed an increase in disease during this period.34 Analysis of our hospital admission data showed a reduction in pneumococcal meningitis and septicaemia 402

in children after the introduction of PCV7 into the routine schedule at months 2, 4, and 13 along with a catch-up campaign. The effect of vaccination on pneumococcal hospital admissions statistics is less pronounced than that for meningococcal and H influenzae disease because the data are for a disease caused by more than 90 serotypes, whereas the childhood vaccines used in the UK cover only seven or 13 of the commonest serotypes. Although the hospital admission rates are about double those reported in microbiological surveillance data,14 we saw a simlar reduction in pneumococcal meningitis reported in children younger than 5 years from 2000–06 to 2008–10. The admission data also show a halving in pneumococcal meningitis incidence in children younger than 15 years from 2006 to 2011, and a similar decrease in pneumococcal septicaemia. Consistent with our data, laboratory surveillance data in England and Wales indicated an overall reduction in invasive pneumococcal disease (of 34% at all age groups and 56% in children younger than 2 years from 2000–06 to 2009–2010).14,16 The overall reduction in pneumococcal meningitis seen in laboratory data from 2000–06 to 2008–10 was only statistically significant in children younger than 5 years (44% reduction) due to serotype replacement.14,35 We did not see any effect in hospital admissions for invasive pneumococcal disease from 1999 to 2011 in people aged 65 years or older, which might have been expected as a result of increased herd immunity driven by the infant vaccine programme from 2006, or as a result of direct protection induced through the pneumococcal polysaccharide vaccine programme (PPV23) introduced in this age group in 2003.36 By contrast with our data, findings from a study of laboratory surveillance data showed a 19% reduction in invasive pneumococcal disease from 2000–06 to 2009–10 in people aged 65 years or older after the introduction of the infant PCV7 vaccination programme in the UK.14 This reduction seems to have been caused by a reduction in PCV7 serotypes through herd immunity rather than any effect of the pneumococcal polysaccharide vaccination programme.14,36 Analysis of laboratory surveillance data indicated a greater reduction in invasive pneumococcal disease caused by PCV7 serotypes than the overall disease reduction because of serotype replacement. From 2000–06 to 2009–10, there was a 98% reduction in invasive pneumococcal disease caused by PCV7 serotypes in children younger than 2 years accompanied by a reduction in PCV7 serotypes of 75% or more in adults aged 65 years or older through herd immunity.14,16 After the introduction of the PCV13 vaccine into the UK’s routine schedule in 2010, microbiological surveillance data indicated that the number of cases of invasive pneumococcal disease due to additional serotypes in PCV13 that are not in PCV7 halved in children younger than 2 years in the first epidemiological year after its introduction.13 www.thelancet.com/infection Vol 14 May 2014

Articles

Similar trends in hospital admissions from ICD coding for H influenzae meningitis and invasive pneumococcal disease have been reported in the USA after the licensure of the Hib conjugate vaccine in 1987 and the PCV7 vaccine in 2000. After the licensure of the Hib conjugate vaccine, there was an average decrease in admissions in children younger than 5 years of 34% per year from 1988 to 1991.37 When comparing hospital admissions before (1998 to 2000) and after (2001 to 2003) PCV7 licensure, pneumococcal meningitis reduced statistically significantly by 67% at all ages.38 The reduction in pneumococcal bacteraemia (36% decrease) was only statistically significant in adults aged 65 years or older.38 Findings from another US study showed a reduction in pneumococcal meningitis when comparing hospital admissions from 1994–97 to 2001–04 of 66% in children younger than 2 years, 52% in children aged 2–4 years, 26% in adults aged 18–39 years, and 33% in adults aged 65 years or older.39 The datasets in our study show lower rates of hospital admissions in the 1960s and 1970s and a subsequent increase in incidence. Various contributory factors caused the fluctuations in incidence seen in our study, although the relative contribution of each is not well understood. Changes in diagnostics and coding are important causes of systematic changes in reporting, but other factors such as the emergence and disappearance of new invasive bacterial clones, variation in population immunity, increases in transmission of predisposing viral infections, reduction in smoking rates and household crowding, and other social changes are all important. Hospital admissions data show higher rates of meningitis and septicaemia than seen in microbiological surveillance data.12,14 This difference suggests that use of data for only microbiologically confirmed cases might underestimate disease incidence. Looking at trends, however, we saw similar profiles in analysis of both hospital admission and microbiological surveillance data. Our study had several limitations. Data collection, including clinical coding, in the ORLS was managed entirely by medically qualified epidemiologists from its inception until 1985, and disease rates were broadly similar between ORLS and HIPE in the period covered independently by both, suggesting both datasets are probably reliable. Admission rates for meningococcal disease were slightly lower in Oxford than in all of England. However, relative to England as a whole, Oxford is a prosperous and healthy region and meningococcal disease in childhood tends to be associated with socioeconomic disadvantage.40,41 Also, the dataset for Oxford was from a smaller population than that for the whole of England, leading to much fluctuation. There are no other regional datasets for comparison with Oxford in the years before HES. The quality and consistency of HES data are largely unknown. Findings from a www.thelancet.com/infection Vol 14 May 2014

Panel: Research in context Systematic review We searched PubMed up to Nov 18, 2013, for papers reporting paediatric invasive bacterial disease trends. We used the search terms “Haemophilus influenzae”, “meningococcal”, “Neisseria meningitidis”, and “pneumococcal” in combination with “meningitis”, “septicaemia”, “hospital admissions”, “hospital episode statistics”, and “epidemiology”. We also reviewed references from relevant articles that were not identified in the original search. We reviewed studies reporting disease incidence from microbiological surveillance data or hospital admission statistics in the UK and other countries with similar disease burden and vaccine programmes. To the best of our knowledge, our study reports the most complete populationbased dataset in the UK of disease incidence for paediatric meningitis and septicaemia caused by H influenzae, N meningitidis, and S pneumoniae between the 1960s and 2011. Interpretation Our findings show a reduction in paediatric meningitis and septicaemia in England after the introduction of conjugate vaccines for H influenzae, N meningitidis, and S pneumoniae. Disease incidences calculated from hospital admission data are higher than those calculated from microbiological surveillance data in the UK, although we saw similar post-vaccination disease trends in both datasets. Hospital admission rates for meningococcal disease in children younger than 15 years were about 30–40% higher, and for pneumococcal meningitis in children younger than 5 years were about double, the incidence rates calculated from laboratory surveillance data when direct comparisons were possible. Ongoing disease surveillance and continued development and implementation of vaccines against additional pneumococcal serotypes and serogroup B meningococcal disease will be needed to further improve disease control.

systematic review in the UK suggest that the accuracy of reporting might be improving.42 Changes in hospital coding practice could have affected our findings. However, the trends shown in these data are consistent with available microbiological surveillance data.14 Microbiological diagnostic techniques have improved during the study period, which might account for some increase in reporting. Nevertheless, despite these limitations, which would generally result in an increase in reported incidence, all of the disorders studied have substantially reduced since the introduction of relevant vaccination programmes. Declining lumbar puncture rates are a potential confounding factor that could have affected data precision. Although there have been changes in ICD coding because of revisions to the ICD, such discontinuity does not seem to have affected our findings: our data show no sudden changes in continuity at points of change of ICD. Our findings show that the incidence of vaccinepreventable paediatric invasive bacterial disease has reduced substantially in England with the advent of effective conjugate vaccines during the past 21 years. However, despite effective vaccination programmes, there were still many children younger than 15 years admitted to hospital with meningococcal infection, pneumococcal septicaemia, or meningitis, although few were admitted with Hib meningitis or septicaemia, showing the importance of ongoing vaccine development and deployment to ensure child health. Disease surveillance will continue to inform research into the 403

Articles

priorities for development of effective vaccines and policy about immunisation schedules to further improve disease control. Contributors All authors had full access to all of the data (including statistical reports and tables) in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis. MJG, NGM, AJP, and MS had the idea for and designed the study. MJG collected the data and designed the analysis. MJG, NGM, AJP, and MS contributed to data interpretation. NGM wrote the first draft of the paper. MJG, NGM, AJP, and MS contributed to the review and final approval of the report. Declaration of interests Oxford University has received grant funding from manufacturers of Hib, meningococcal, and pneumococcal vaccines including Novartis, Pfizer, Glaxosmithkline, Sanofi Pasteur MSD for clinical trials directed by AJP, who receives no personal payments from them. AJP is chair of the UK Department of Health’s Joint Committee on Vaccination and Immunisation. The views expressed in the article do not necessarily represent the views of the UK Department of Health. The other authors declare that they have no competing interests. Acknowledgments The Unit of Health-Care Epidemiology was funded by the English National Institute for Health Research to build the record linkage datasets (grant number RNC/035/002). The work of the Oxford Vaccine Group was also supported with funding from the National Institute of Health and Research Oxford Biomedical Research Centre. This study has no specific funding. The funders had no role in the design or analysis of the data and the views expressed in this paper do not necessarily reflect those of the funders. We thank David Yeates and Nick Hall, Unit of Health-Care Epidemiology, for their programming support. References 1 Saez-Llorens X, McCracken GH Jr. Bacterial meningitis in children. Lancet 2003; 361: 2139–48. 2 European Centre for Disease Prevention and Control. Annual epidemiological report: reporting on 2009 surveillance data and 2010 epidemic intelligence data. http://www.ecdc.europa.eu/en/ publications/Publications/1111_SUR_Annual_Epidemiological_ Report_on_Communicable_Diseases_in_Europe.pdf (accessed Sept 11, 2013). 3 Edmond K, Clark A, Korczak VS, Sanderson C, Griffiths UK, Rudan I. Global and regional risk of disabling sequelae from bacterial meningitis: a systematic review and meta-analysis. Lancet Infect Dis 2010; 10: 317–28. 4 Hudson LD, Viner RM, Christie D. Long-term sequelae of childhood bacterial meningitis. Curr Infect Dis Rep 2013; 15: 236–41. 5 Peltola H, Kayhty H, Sivonen A, Makela H. Haemophilus influenzae type b capsular polysaccharide vaccine in children: a double-blind field study of 100,000 vaccinees 3 months to 5 years of age in Finland. Pediatrics 1977; 60: 730–37. 6 Chapel H, Haeney M, Misbah S, Snowden N. Immune modulation. In: Essentials of clinical immunology. Oxford; Blackwell Publishing, 2006: 137–41. Heath PT, McVernon J. The UK Hib vaccine experience. 7 Arch Dis Child 2002; 86: 396–99. 8 Ladhani SN, Ramsay M, Slack MP. The impact of Haemophilus influenzae serotype B resurgence on the epidemiology of childhood invasive Haemophilus influenzae disease in England and Wales. Pediatr Infect Dis J 2011; 30: 893–95. 9 Lee YC, Kelly DF, Yu LM, et al. Haemophilus influenzae type b vaccine failure in children is associated with inadequate production of high-quality antibody. Clin Infect Dis 2008; 46: 186–92. 10 Ramsay ME, McVernon J, Andrews NJ, Heath PT, Slack MP. Estimating Haemophilus influenzae type b vaccine effectiveness in England and Wales by use of the screening method. J Infect Dis 2003; 188: 481–85. 11 Campbell H, Borrow R, Salisbury D, Miller E. Meningococcal C conjugate vaccine: the experience in England and Wales. Vaccine 2009; 27 (suppl 2): B20–29.

404

12

13

14

15

16

17

18 19 20

21

22

23

24

25

26

27

28

29

30

31

32

33

Ladhani SN, Flood JS, Ramsay ME, et al. Invasive meningococcal disease in England and Wales: implications for the introduction of new vaccines. Vaccine 2012; 30: 3710–16. Miller E, Andrews NJ, Waight PA, Slack MP, George RC. Effectiveness of the new serotypes in the 13-valent pneumococcal conjugate vaccine. Vaccine 2011; 29: 9127–31. Miller E, Andrews NJ, Waight PA, Slack MP, George RC. Herd immunity and serotype replacement 4 years after seven-valent pneumococcal conjugate vaccination in England and Wales: an observational cohort study. Lancet Infect Dis 2011; 11: 760–68. Trotter CL, Waight P, Andrews NJ, et al. Epidemiology of invasive pneumococcal disease in the pre-conjugate vaccine era: England and Wales, 1996-2006. J Infect 2010; 60: 200–08. Ladhani SN, Slack MP, Andrews NJ, Waight PA, Borrow R, Miller E. Invasive pneumococcal disease after routine pneumococcal conjugate vaccination in children, England and Wales. Emerg Infect Dis 2013; 19: 61–68. Department of Health and Office of Population Censuses and Surveys. Hospital In-patient enquiry. Series MB4 no 29. London: Her Majesty’s Stationery Office; 1989. Public Health England. Statistical tools. http://www.erpho.org.uk/ statistical_tools.aspx (accessed Sept 19, 2013). Goldacre MJ, Miller DL. Completeness of statutory notification for acute bacterial meningitis. BMJ 1976; 2: 501–03. Gill PJ, Goldacre MJ, Mant D, et al. Increase in emergency admissions to hospital for children aged under 15 in England, 1999–2010: national database analysis. Arch Dis Child 2013; 98: 328–34. McVernon J, Andrews N, Slack MP, Ramsay ME. Risk of vaccine failure after Haemophilus influenzae type b (Hib) combination vaccines with acellular pertussis. Lancet 2003; 361: 1521–23. Ladhani S, Slack MP, Heys M, White J, Ramsay ME. Fall in Haemophilus influenzae serotype b (Hib) disease following implementation of a booster campaign. Arch Dis Child 2008; 93: 665–69. Rijkers GT, Vermeer-de Bondt PE, Spanjaard L, Breukels MA, Sanders EA. Return of Haemophilus influenzae type b infections. Lancet 2003; 361: 1563–64. Peltola H, Salo E, Saxen H. Incidence of Haemophilus influenzae type b meningitis during 18 years of vaccine use: observational study using routine hospital data. BMJ 2005; 330: 18–89. Public Health England. Haemophilus influenzae epidemiological data: laboratory reports of Haemophilus influenzae type b infection by age group and quarter: England 1990–2012. http://www.hpa.org.uk/ Topics/InfectiousDiseases/InfectionsAZ/HaemophilusInfluenzaeTypeB/EpidemiologicalData/ (accessed July 25, 2013). Stanton MC, Taylor-Robinson D, Harris D, et al. Meningococcal disease in children in Merseyside, England: a 31 year descriptive study. PLoS One 2011; 6: e25957. Miller E, Salisbury D, Ramsay M. Planning, registration, and implementation of an immunisation campaign against meningococcal serogroup C disease in the UK: a success story. Vaccine 2001; 20 (suppl 1): S58–67. Jones DM, Kaczmarski EB. Meningococcal infections in England and Wales: report of the Meningococcal Reference Laboratory for 1990. CDR (Lond Engl Rev) 1991; 1: R76–78. Heinsbroek E, Ladhani S, Gray S, et al. Added value of PCR-testing for confirmation of invasive meningococcal disease in England. J Infect 2013; 67: 385–90. Norheim G, Sadarangani M, Omar O, et al. Association between population prevalence of smoking and incidence of meningococcal disease in Norway, Sweden, Denmark and the Netherlands between 1975 and 2009: a population-based time series analysis. BMJ Open 2014; 4: e003312. Martin NG, Snape MD. A multicomponent serogroup B meningococcal vaccine is licensed for use in Europe: what do we know, and what are we yet to learn? Expert Rev Vaccines 2013; 12: 837–58. O’Ryan M, Stoddard J, Toneatto D, Wassil J, Dull PM. A multicomponent meningococcal serogroup B vaccine (4CMenB): the clinical development program. Drugs 2014; 74: 15–30. Johnson AP, Waight P, Andrews N, Pebody R, George RC, Miller E. Morbidity and mortality of pneumococcal meningitis and serotypes of causative strains prior to introduction of the 7-valent conjugant pneumococcal vaccine in England. J Infect 2007; 55: 394–99.

www.thelancet.com/infection Vol 14 May 2014

Articles

34

35

36

37

38

Gjini A, Stuart JM, George RC, Nichols T, Heyderman RS. Capture-recapture analysis and pneumococcal meningitis estimates in England. Emerg Infect Dis 2004; 10: 87–93. Pichon B, Ladhani SN, Slack MP, et al. Changes in molecular epidemiology of Streptococcus pneumoniae causing meningitis following introduction of pneumococcal conjugate vaccination in England and Wales. J Clin Microbiol 2013; 51: 820–27. Andrews NJ, Waight PA, George RC, Slack MP, Miller E. Impact and effectiveness of 23-valent pneumococcal polysaccharide vaccine against invasive pneumococcal disease in the elderly in England and Wales. Vaccine 2012; 30: 6802–08. Schoendorf KC, Adams WG, Kiely JL, Wenger JD. National trends in Haemophilus influenzae meningitis mortality and hospitalization among children, 1980 through 1991. Pediatrics 1994; 93: 663–68. Shah SS, Ratner AJ. Trends in invasive pneumococcal diseaseassociated hospitalizations. Clin Infect Dis 2006; 42: e1–5.

www.thelancet.com/infection Vol 14 May 2014

39

40

41

42

Tsai CJ, Griffin MR, Nuorti JP, Grijalva CG. Changing epidemiology of pneumococcal meningitis after the introduction of pneumococcal conjugate vaccine in the United States. Clin Infect Dis 2008; 46: 1664–72. Williams CJ, Willocks LJ, Lake IR, Hunter PR. Geographic correlation between deprivation and risk of meningococcal disease: an ecological study. BMC Public Health 2004; 4: 30. Goldacre MJ, Wotton CJ, Maisonneuve JJ. Maternal and perinatal factors associated with subsequent meningococcal, Haemophilus or enteroviral meningitis in children: database study. Epidemiol Infect 2014; 142: 371–78. Burns EM, Rigby E, Mamidanna R, et al. Systematic review of discharge coding accuracy. J Public Health (Oxf) 2012; 34: 138–48.

405